Narrow Vein Mining Proceedings

Conference 2008

-
14 15 October 2008
BaIIarat, Victoria

The Australasian Institute of Mining and Metallurgy

Publication Series No 9/2008

Edited by Dr Simon Dominy

Published by

THE AUSTRALASIAN INSTITUTE OF MINING AND METALLURGY

Level 3, 15 - 31 Pelham Street, Carlton Victoria 3053 Australia

i
 0 The Australasian Institute of Mining and Metallurgy 2008

All papers published i n this volume were refereed prior to publication.

The Institute is not responsible as a body for the facts and opinions advanced i n any of its publications.

ISBN 978 1 920806 89 7

Desktop published by:

Kristy Pocock, Kylie McShane and Olivia Tet Fong for
The Australasion Institute of Mining and Metallurgy

Printed by:
BPA Digital
11 Evans Street
Burwood VIC 3125

ii
 Contents
Keynote Address
Beyond Airleg Mining – Narrow Vein Mining in the 21st Century P L McCarthy 3

Case Studies
Narrow Vein Mining at Charters Towers, Queensland, by C A J Towsey 9
Longhole Open Stoping

Return to Hillgrove P Ganza and C Dell 13

The Challenger Gold Mine P Androvic, P Bamford and 19

M Sandy

Exploration and Resource Geology

Continuity Risk in Narrow Reef Gold Deposits – Implications for S C Dominy and I M Platten 31
Evaluation and Exploitation

Resource Estimation and Grade Assignment – A Comparison M D Goodz, J Rea and 51

Between Historical Production and Current Maxwell Mining P Jackson
Validation Case Study at Morning Star Gold Mine, Woods Point

The Ballarat East Goldfield – New Insights on an Old Model D J Osborne 59

Significance of Geological Control on Assay Data in the Narrow, D Potter, C Sheriff and 71
High-Grade Gold Quartz Veins at the Sand Queen Mine, Comet P Collins
Vale, Western Australia

Resource Modelling and Mine Design at Obuasi Mine, Ghana H Eybers, J Visser and 79
C K Boafo

Metallurgy
Gold Particle Characteristics in Narrow Vein Deposits – S C Dominy, Y Xie and 91
Implications for Evaluation and Metallurgy I M Platten

The Python – An Underground Processing Plant for Narrow Vein T R Hughes and N J Grigg 105
Mining
 New Applications/Research and Development

Mechanised Ore Removal from the Floor of Narrow, Flat Stopes S L Tombs, A Buckingham and 111
R M Hancock

Recovery of Broken Ore Stocks Left on the Mine Floor and Other K Biegaj 115
Applications of the Underground Mobile Supersucker

Thermal Fragmentation – An Innovative Process for Mining D Brisebois 123

Narrow High-Grade Precious Metal Veins

Monorail Technology – A Rapid and Cost-Effective Method of E Chanda and B Besa 129
Decline Development

Rock Mechanics/Mining Geotechnics

Augusta Gold-Antimony Mine M McCarthy, A Fowler and 145

S Marshall

Strategies for Minimising and Predicting Dilution in Narrow Vein P C Stewart and R Trueman 153
Mines – The Narrow Vein Dilution Method

The Use of Geotechnical Instrumentation to Optimise an A R Penney, P B Hills and 165

Engineered Mine Design at Beaconsfield Gold Mine, Tasmania R J Walton

Competing Factors in Support Selection for the West Zone of the C Scott, A R Penney and 173
Beaconsfield Gold Mine, Tasmania P Fuller

Stope Design and Mining Methods

Design Factors Leading to Prevention of Bridged Stopes at C Miles and G Lind 181
Jundee Operations

Drift and Fill – The High Value, High Recovery Mining System L Dawson, M Yumlu and 187
M English

Resue Firing and Dilution Control in Narrow Vein Mining M Tuck 195

The Development and Implementation of a Fully Remote Stoping P B Hills, J Mills, A R Penney 199
Method at Beaconsfield Gold Mine, Tasmania and S Arthur
Wallaby – The Horizontally Challenged Opportunity N S Rauert 207

Victorian Gold Case Studies

Recent Experiences at the Kangaroo Flat Mine, Bendigo R McLean and M Hernan 217

Historical Challenges, Modern Solutions at Ballarat East K Williams and M Sykes 229

Tarnagulla Gold Mine P L McCarthy and L Faulkner 237

Reviving the Inglewood Goldfield in Central Victoria – Exploration J Cahill 249

and Operations
Beyond Airleg Mining – Narrow Vein Mining in the 21st Century
P L McCarthy1

ABSTRACT
Higher commodity prices and environmental considerations will make
selective narrow-vein mining more attractive in future. Unfortunately,
hand-held airleg (jackleg) mining is a mature technology that has safety
and productivity limitations. Despite decades of attempted innovation, no
obvious successor to airleg mining has emerged. A review of some of the
possible technologies is followed by a speculative description of the mine
of the future, intended to stimulate thinking and discussion during the
conference.

INTRODUCTION
Underground mining will become more important in coming
decades as environmental and social concerns make surface
mining less attractive. With an increased focus on carbon
emissions it will be difficult to justify a surface mining option
that breaks, lifts and moves vast quantities of useless overburden
if there is a viable underground alternative. Despite surface
subsidence, caving methods are generally much less obtrusive FIG 1 - Miner drilling underground using airleg drill in a shaft,
than surface mining, while other underground methods can be Kalgoorlie, 1951 (Western Australian Government collection).
managed to have minimal impact. Increasing product prices will
make high-grade, narrow vein mines particularly attractive.
Changes in underground mining methods are driven by
technology. The progressive introduction of bar-mounted
rockdrills, electric locomotives, rail boggers, rope scrapers, airleg
drills, longhole drills, bulk explosives, LHDs, hydraulic jumbos
and low-profile diesel trucks over the course of a century brought
a series of changes to development and stoping techniques. The
changes to methods were made when the equipment became
available, and operations personnel often made innovations that
were not envisaged by the equipment developers.
The most recent innovation was the widespread introduction of
sublevel retreat benching, made possible by the availability of
remote and tele-remote controlled LHDs. In an early application
of this method to narrow vein mining at William Ford Decline,
stopes 0.5 to 1.2 m wide were mined over 8 to 19 m stope heights
and 200 m strike lengths (Mills and Will, 1993). However, where
one or both walls are weak, longhole drilling and blasting is
likely to give unacceptable dilution in veins narrower than 2 m.
FIG 2 - An airleg miner today (Placer Dome Inc).
The situation is even more difficult in flat-dipping veins where
broken ore may not rill to the extraction drive. Often, airleg
mining is still the only viable mining method. the load; the load is compact and easy to grasp; it is held close to
the trunk and not carried frequently or for long distances, all of
AIRLEG MINING which can be a problem in mining.
Airleg mining often involves using a rope scraper, which has
Airleg (or jackleg) mining has been practiced for more than its own hazards. After blasting, for example, the miner must hang
60 years without significant change, as illustrated in Figures 1 a tail block from a pin in the new face before any broken rock
and 2. Using a small, hand-held rockdrill enables close control of can be moved and thus before the back can be properly barred
stope walls but requires good geological skills from the mining down. Rockfall injuries are a significant issue for airleg mining
team and good liaison with geologists. Safe airleg mining (McCarthy and Faulkner, 2008).
requires highly skilled operators (Roberts, 2008). Even then, it
The author was a member of the WMC Limited Elimination of
can be argued that the method is intrinsically hazardous. The
Fatalities Taskforce and formed the view that the intrinsic nature
popular Atlas Copco Panther rockdrill weighs 27 kg with a
of airleg mining hazards was one of the reasons that drove WMC
further 16 - 19 kg for the pusher leg, taking it well into the risk
to sell its Kambalda nickel mines. Today, larger mining
range of 16 to 55 kg for safe manual handling (Commission for
companies and contractors avoid the method.
Occupational Safety and Health, 2000). Risk factors are
minimised only if no bending or twisting is required to pick up In March 2008 Barminco’s Stuart Tonkin was reported as
saying that six Western Australian operations using the
company’s contracting services had airleg miners doing slot
1. FAusIMM(CP), Managing Director, AMC Consultants, Level 19, rising or stoping. Narrow higher grade nickel mines such as the
114 William Street, Melbourne Vic 3000. Mincor operations around Kambalda supported the use of
Email: pmccarthy@amcconsultants.com.au airleggers to enable selective mining:

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 3

P L McCARTHY
We have a large skill base in this field, however,
mine owners are steering away from the use of
this discipline for stoping and rising, Tonkin
said. We are not currently training airleg miners
(Roberts, 2008).
Steve Coughlan, Managing Director of another major
Australian-based underground mining contractor, Byrnecut
Mining, said he didn’t ‘see a big future for those skills in our
business’ (ibid).
The few Australian mines using airleg as their primary stoping
method include Otter Juan (Mincor), Blair (Australian Mines)
and Bullen (Norseman Gold). Other mines or projects using
some airleg for stoping or development include Redross
(Mincor), Cassilis (Mutiny Gold), Inglewood (Greater Bendigo
Mines), Walhalla (Goldstar), Radio Hill (Fox Resources), Hill
End (Hill End Gold), Higginsville (Avoca Resources), Daisy
Milano (Silverlake Resources) and Costerfield (Cambrian)
Although airleg mining has declined in popularity in Australia
it remains an important mining method around the world.
INNOVATIONS IN NARROW VEIN MINING
A narrow vein mining research project by Natural Resources
Canada between 1999 and 2003 produced a useful database of
mines but failed to identify any promising technologies (Natural
Resources Canada, 2003). With the exception of the use of
Styrofoam blocks to form a slot raise in fill at Barrick’s Bousquet
mine, all of the innovations and mining methods identified could
have come from the 1970s.
The following is a brief discussion of technologies that have
been proposed elsewhere and may have a role in narrow vein
mining in the 21st Century.
Thermal rock fragmentation in narrow vein extraction is
proposed in US Patent 6913320, with the burner moving at a
controlled rate of travel between the sidewalls of the vein to
cause the ore comprised in the vein to spall into fragments
(Gauthier, 2007). However, one can imagine some serious issues
with introducing such a device into the underground environment.
An alternative reef mining technology which drew
considerable attention at South Africa’s Electra Mining ’94
exhibition was the Reefeater wire rope cutting system developed
by Anglo American Technical Development Services. The
Reefeater used diamond cutters on a wire rope to cut slots above
and below the reef, which was then broken with a small
explosive charge. A total of 18 installations were planned by mid
1995, and there was a small trial at Central Norseman gold, but
this idea seems to have faded away.
There have been multiple attempts to introduce rock-breaking
cartridges, or rock splitting, into South African narrow vein gold
and platinum longwall mines. A parallel is drawn with the
changes in the underground coal mining industry following the
elimination of blasting using explosives at the coal face, which
eventually led to the introduction of automation and advanced
technology. However, this was achieved by continuous cutting
systems. One product is NoneX™, a non-detonating chemical
compound enclosed in a cartridge, which reacts very quickly
when ignited to produce high volumes of harmless gas, mainly
consisting of nitrogen, carbon dioxide and steam (Nonex™,
2004). Trials of several similar products seem not to have led to a
commercially viable mining system.
A more direct approach to continuous mining would be to
develop a roadheader or similar machine to cut hard, abrasive
rock. One promising technology is CSIRO’s SMART*CUT, an
advanced, very strong and highly wear-resistant, super-material
cutting head system, which CSIRO developed for mining and
excavation of soft to very hard rock. SMART stands for Super
Material Abrasive Resistant Tools, which use diamond-
composite cutting elements to improve the efficiency of all rock-
4 Ballarat, Vic, 14 - 15 October 2008
cutting operations. SMART*CUT, with its small and compact
cutting head system, is designed to cut any hard rock at a
minimum specific energy, and also to operate in difficult,
selective and narrow vein mines under any tight and hazardous
environmental conditions (CSIRO, 2008).
The oscillating disc cutter is an interesting technology and is
logically an efficient way to cut rock. It is was heavily promoted
some eight years ago at the prototype stage and is reportedly
being commercialised by Joy Mining Machinery, South Africa
and Odyssey Technology Pty Ltd (SME), but recent information
is hard to come by (CRCMining, 2005).
An impact ripper developed in South Africa was trialled in
1995 at Telfer Gold Mine, The impact ripper, powered by
high-pressure water, travels along a longwall face and breaks the
reef down onto a face conveyor. Telfer did not have the
advantage of very high rock stresses that, at depth in Kloof gold
mine, caused the reef to break readily under the pick’s action
with moderate pick wear.
Commissioned at Broken Hill in mid 1992, the 260 t Robbins
MM130 Mobile Miner was used for development, although it
had originally been envisaged as a cut-and-fill stoping machine.
Manufactured by Perry Engineering Pty Ltd in South Australia to
a Robbins Co design, the prototype machine had a 4 m diameter
rotating cutter head, which was set vertically and oscillated from
side to side, so potentially suitable for narrow veins. It operated
successfully in hard, abrasive metasediments with compressive
strengths which range from 70 MPa to over 300 MPa, but at a
higher cost than the drill and blast alternative.
A new range of high-frequency hydraulic rockdrills is giving
substantially increased penetration rates, with an ability to tune
the hammer frequency to the rock properties. The typical impact
rate of a hammer is about 60 Hz with a current maximum of
about 100 Hz. If the percussion rate could be doubled at the same
energy per blow, double penetration rate could be expected. The
result will not only be a shorter development cycle. Engineers in
operations will find entirely new ways to do development and to
design stopes using these machines. If drilling is quicker and
cheaper then more holes can be drilled, with better distribution of
explosive.
Stoping at AngloGold Ashanti’s TauTona mine has been
100 per cent electric-powered since January of this year. The
drills are Hilti 240 V, 2.2 kW water-cooled, with a water leg for
thrust. The drills have improved face advance and productivity,
with lower noise levels and vibration. This implementation
follows many years of development to resolve issues relating to
the performance of the drill, including the drill string, the drilling
bits, and electrics (Copans, 2008).
What if an automated airleg-sized hydraulic or electric drill (or
two) was mounted on a carrier like a tracked Bobcat? Weight
could be kept low by using the airleg principle to transfer thrust
to the ground. Perhaps previous attempts at small jumbos have
tried to scale down existing jumbo designs, rather than starting
with a clean sheet of paper. Perhaps ‘airleg’ mining has a future
after all.
THE MINE OF THE FUTURE
Let us visit the mine of the future, which is an underground
mine. In this age, open pit mining has been banished to remote
corners of the least developed countries. Most minerals now
come from small, efficient underground mines located wherever
the ore deposits are found.
This mine appears, on the surface, to be a small manufacturing
plant. It is located in the industrial suburb of a provincial city. It
is a zero-discharge site, so contaminated groundwater and
mineral waste are treated and retained on the site, or sold as
useful products including potable water.
Narrow Vein Mining Conference
 BEYOND AIRLEG MINING – NARROW VEIN MINING IN THE 21ST CENTURY
If we pass through the landscaped portal into the service
decline, we notice that it is relatively small. All mine equipment
has been compactly designed to travel or be transported down its
sealed surface into the production areas. Efficient diesel engines
are still used, but the mine manager expects clean-burning
hydrogen engines to be available within a year or two.
There is no truck traffic in the decline. Ore and waste are
processed underground and only the valuable concentrate and a
saleable sand product are pumped to surface. The plant can crush
waste for roadbase if needed.
Pilot headings are mined 2.5 m high and 2.5 m wide, which is
sufficient for ore delineation, bulk sampling and transport of
diamond drill rigs to their drill cuddies. If this development is
later found to be poorly placed, it has little effect on the mine
plan. If the pilot heading is required for production haulage it is
stripped to full size, for a final cost no greater than for full-face
development. The pilot heading allows preplanning of ground
support and often the prior installation of ventilation raises.
The pilot heading concept became widely used when a suitable
mucking machine was developed. The muck from a 2 m round is
loaded out in a single pass, giving a four-hour mining cycle and
an advance rate of 80 m per week. Hand-held hydraulic drills are
carried in on a power-pack buggy. In the stopes, the emphasis is
on mining contained metal. ‘Artificial noses’ sniff sulfide content
on every exposed face and the ore outline is marked up at the
chosen cut-off grade. Cut-and-fill mining gives samplers access
to the ore and enables separate blasting and stowing of low-grade
or waste. Cemented tailings fill is used to provide a working
floor.
The proponents of mass-mining methods with large equipment
took several years to acknowledge the economics of this process.
They had grown up in an era when ‘tonnes was king’ and
measured their success by the size of the openings they created.
Their approach was inefficient and environmentally destructive
and led to great public opposition to mining because of the large
waste dumps and tailings dams which were required.
The workforce here is multi-skilled. All have diplomas in the
theory and practice of mechanical maintenance and equipment
operation. They are rotated regularly through underground
production, the treatment plant and the planning office. One in
ten holds a university degree and others are encouraged to
continue their studies. There are more employees than in past
years, consistent with the company’s policy of maximising
productive employment.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
The communication system allows all employees instant
contact, but it is rarely used. Work is preplanned and runs to a
comfortable schedule. Ground support designs are agreed in
advance based on examination of the pilot headings and the use
of remote-sensing equipment. Sufficient stopes are available to
give a flexible schedule.
The mine manager would like to spend more time with us, but
he is playing golf this afternoon.
REFERENCES
Commission for Occupational Safety and Health, 2000. Code of Practice:
Manual Handling, 18 December (Government of Western Australia).
Copans, G, 2008. During Eskom’s power curb, we didn’t lose production.
We just turned on our new electric rockdrills – TauTona gold mine
[online]. Creamer Media’s Mining Weekly Online. Available from:
<http://www.miningweekly.com/article.php?a_id=129968>
[Accessed 4 April 2008].
CRCMining, 2005. Annual Report 2004 - 2005.
CSIRO, 2008. SMART*CUT makes hard rock mining faster and more
efficient [online]. CSIRO Exploration and Mining. Available from:
<http://www.csiro.au/solutions/ps14h.html>.
Gauthier, B, 2007. Thermal fragmentation: A new technology for mining
narrow high-grade precious metal vein ore bodies [online].
InvestorIdeas.com. Available from: <http://www.investorideas.com/
Articles/041207a_page1.asp> [Accessed: 14 December 2007].
McCarthy, P L and Faulkner L, 2008. Tarnagulla gold mine, in
Proceedings Narrow Vein Mining Conference, pp 237-248 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Mills, G and Will, D, 1993. Coolgardie Gold NL – William Ford decline:
Narrow vein sublevel up-hole mechanised stoping and long hole raise
blasting practices, in Proceedings Narrow Vein Mining Seminar,
pp 71-78 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Natural Resources Canada, 2003. Narrow vein mining research project:
Project results [online]. Natural Resources Canada. Available from:
<http://mmsd1.mms.nrcan.gc.ca/canmet/database_choice-e.asp>.
Nonex™, 2004. Contents [online]. Available from: <http://www.nonex.
co.za/3_3coninuous_mining.htm>.
Roberts, R, 2008. The beginning of the end [online]. High Grade.
Available from: <http://www.highgrade.net> [Accessed: 24 March
2008].
5
Narrow Vein Mining at Charters Towers, Queensland, by
Longhole Open Stoping
C A J Towsey1
ABSTRACT
Citigold Corporation operates the Warrior gold mine at Charters Towers,
northern Queensland, mining narrow quartz veins (0.1 to 2 m wide) in
granodiorite by long hole open stoping on levels 15 m apart. Access is by
decline and cross-cuts with on-ore development drives.
The mineralisation comprises Palaeozoic mesothermal quartz veins
containing gold, pyrite, sphalerite and galena, hosted by the Ordovician
age Towers Hill Granite. The gold grain size is variable from 1 to 500 µm
with 1 m face samples up to 800 ppm Au, and average run-of-mine
grades of 10 to 14 ppm. Despite the spectacular gold tenor, little visible
gold is present.
Gold grades are highly variable presenting challenges in ore reserve
estimation. Tonnages are discounted by 70 per cent to allow for the low
payabilities (30 to 50 per cent) of the veins. Estimation techniques allow
high confidence in overall global grade estimates but poor reliability on a
local scale (25 m). Kriging ranges as low as 5 - 8 m means that obtaining
a Proved Ore Reserve by diamond drilling is not economically feasible,
and broad-scale mine planning decisions are made at the Inferred Mineral
Resource stage and mine design work at the Probable Ore Reserve stage.
Proved Ore Reserves can only be estimated after underground
development has proceeded through the mineralised body with levels at
15 m vertical spacing sampled at 2 - 3 m intervals along the level.
However, reconciliations between actual gold produced and earlier
resource estimates produce workable results.
A further three mines are scheduled to be developed underneath the
city of Charters Towers.
INTRODUCTION
Charters Towers is located 1000 km north of Brisbane and about
130 km south-west of Townsville in northern Queensland, and the
gold field produced 6.6 million ounces of gold from six million
tons of rock between 1871 and 1920. Citigold Corporation
Limited has re-opened the Charters Towers gold field after lying
dormant for over 80 years. Over A$120 million has been invested
to date in exploration, drilling and infrastructure, with a 340 000
tonne per year processing plant and two million tonne capacity
tailings dam constructed. The Charters Towers Gold Project is
scheduled to produce up to 300 000 ounces of gold per year for
30 years. The project plans to mine 6.8 million ounces over the
life of the project, from within the ten million ounce Inferred
Mineral Resource (23 million tonnes @ 14 g/t Au). Four main
production areas will be utilised, three within the main central
Charters Towers area (Brilliant, Day Dawn and Sunburst), and
the Warrior area approximately 4 km to the south of the city. All
ore will be treated at the existing gold processing plant 6 km
south of the city.
The gold-bearing reefs at Charters Towers are typically
0.3 - 1.5 m thick, comprising hydrothermal quartz veins in granite
and granodiorite host rocks. There are some 80 major reefs in and
around Charters Towers city, of which 22 are included to date in
Citigold’s resource estimate. The main east-west reef systems are
the Brilliant, the Day Dawn, the Mexican, the Queen and the
Sunburst, extending over a strike length of 5 km and cut by NNW
trending cross-veins. There is a second E-W system 800 m to
the south comprising the Golden Sunrise, Mary and Clark’s
1. FAusIMM(CP), Chief Operating Officer, Citigold Corporation
Limited, Level 15, IBM Centre, 348 Edward Street, Brisbane Qld
4000. Email: ctowsey@citigold.com
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
Moonstone line of reefs, and a third system 500 m further south,
comprising the Ruby and Gladstone line of reefs. They are found
in extensive sheet-like alteration zones (lodes). The most
productive gold-bearing lodes (the Day Dawn, Brilliant and
Queen) dip to the north beneath the city of Charters Towers. The
majority of the ore mined in the past is concentrated within a set
of fractures over 5 km long east-west, and 500 m to 1600 m down
dip in a north-south direction. The mineralised lodes lie in two
predominant directions dipping at moderate to shallow angles to
the north (main production), and the cross-veins, which dip to the
ENE. The E-W and NNW trends seen at the regional scale are
repeated at local scale on the Citigold tenements. The veins are
hydrothermal quartz-gold systems with a gangue of pyrite, galena,
sphalerite, carbonate, chlorite and clays. The reefs occur within
sericitic hydrothermal alteration, historically known as
‘formation’.
While the reefs are typically 0.3 to 1.5 m wide, they range
locally up to 6 m thick, and in isolated cases up to 15 m.
Blatchford (1953) suggested an average width of less than 0.9 m
over most of the field, and this was confirmed by Citigold’s
modelling of stoped volumes. The ore shoots occur with a
periodicity typically in the order of 120 - 300 m within the lodes,
and extend from 200 - 700 m in the down plunge direction, and
are 70 - 300 m wide. The deepest drilling to date by BHP-
Homestake in 1980 - 1984 and Citigold in 2002 - 2003 has
demonstrated that the gold mineralised lodes persist to at least
1300 m vertically and remain open at depth.
Warrior gold mine
The underground Warrior mine is the first of four mines to be
developed. It is accessed by a 1 km decline from the floor of the
Washington open pit, which was mined in 1998 - 2000. Gold
production commenced from Warrior in November 2006. Narrow
vein mining at Warrior mine is by long-hole open stoping,
initially on 10 m sublevels then later expanded to 15 m sublevels
as confidence increased in the accuracy of blasthole drilling and
the dip remained predictable. As tonnage increases, dilution will
be reduced after blasting by photometric ore sorting to remove
waste at a throughput rate of up to 100 tonnes per hour. The dip
of 45 to 48 degrees is steep enough for the ore to run under
gravity and a clean separation of ore on the footwall is achieved
without mechanical assistance. Operations have increased to a
24/7 operation from an initial six day week and two ten-hour
shifts. The Warrior gold mine is scheduled to produce gold at the
rate of 100 000 oz per year at an average run-of-mine grade of
10 g/t Au.
Citigold’s underground mining breakeven grade is 3.5 g/t at
A$900 per ounce.
BENEFITS OF MECHANISED NARROW VEIN
MINING
Mechanised narrow vein mining offers advantages in:
• low mining costs resulting from low tonnages,
• safety if workers are excluded from stopes, and
• accurate grade control.
9
C A J TOWSEY
Low mining costs
The main benefits of mechanised narrow vein mining are the
breaking, handling, transport and milling of low tonnages. The
cost per tonne for underground mining is relatively fixed for any
one mining method. Where gold ore is high grade, the high dollar
value per tonne ensures high profit margins of 20 to 50 per cent of
revenue. The selection of an appropriate mining method is
dependent on safety considerations, ground support techniques,
cost of ground support, mining cost per tonne, orebody
dimensions and orientation, ground stress regime, intact rock
strength and the decision on whether or not to leave stopes open or
backfilled.
At Warrior, the mining method selected is longhole open
stoping. Citigold’s Gold production plan of September 2002
originally envisaged computer-controlled stope drilling by breast
mining on 20 m × 20 m panels. Subsequent analysis of costs and
recent developments in low-profile mining equipment indicated
that longhole open stoping was feasible.
The Warrior orebody dips at 48 degrees to the north and strikes
east-west, and has an average width of 2.4 m. Internal declines at
a gradient of one in seven are developed usually in the footwall,
with cross-cuts from the declines accessing the orebody,
preferably in the centre of the economic sections. Strike drives
are driven along the level with face sampling every cut to verify
block model grades. Longhole open stoping at Warrior mine uses
15 m sublevels, drilling 38 mm diameter up-holes with an Atlas
Copco Simba rig within 4 m × 3.5 m shanty-back drill drives,
breaking through into the floor of the level above. The
breakthrough enables a check to be made on the accuracy of the
drilling. The reef is planned to enter the hanging wall of the drive
at about 2.5 m up the sidewall and exit the drive close to the floor
on the footwall, depending on the reef dip and thickness.
Keeping the drill sublevels to 15 m intervals minimises the
chance of the orebody deviating substantially between levels, and
minimises dilution. Stoping commences with the opening of an
expansion slot blasted up a drilled rise, and then a series of blast
rings are fired to slice ore into the slot. The stope face retreats
back to the central cross-cut, with remote mucking in the open
stope.
Safety
The longhole process has safety advantages, as workers are not
required to work inside open stopes. Conventional narrow vein
mining has used air-leg mining in underhand or overhand shrink
stoping, or cut and fill. Resue mining, where the ore and waste
are drilled and fired separately, is also used for narrow vein
mining. All these conventional methods involve miners working
inside the stope, requiring additional ground support. Working on
the broken ore in shrink stoping or on rock fill in cut-and-fill
mining creates unstable floor conditions that may cause trip and
fall injuries. The initial expansion slot often requires rise mining,
with miners working under recently fired ground, and requiring
significant manual handling to construct and remove stages and
drill platforms, and barring down in a confined space. With
long-holing, the expansion slot can be blasted by using vertical
crater retreat in the initial blast rings, with timed firings
downhole and having the cavity retreat upwards from the lower
breakthrough level.
Grade control
Citigold’s initial first pass drill spacing is 100 m apart
horizontally and 50 m apart vertically on a staggered or diamond
pattern. This spacing is designed to jag one or two high-grade
intersections in pods that are assumed to be 120 m along strike
by 50 to 80 m down dip. Once a high-grade intersection is
encountered, follow-up diamond drilling is undertaken on 25 m
centres to define the mining block. Citigold owns and operates
10 Ballarat, Vic, 14 - 15 October 2008
five electrohydraulic Atlas Copco rigs, three U8 rigs and two U6
rigs, which are set up to operate either underground or from
surface. Narrow vein mining normally allows the orebody to be
driven along, creating long exposures that can be bulk sampled
or channel sampled for accurate grade control along the
extraction levels. Grade control at Warrior uses face channel
samples every cut along the drill and extraction levels, with a
minimum of five 5 kg samples taken across the vein exposure in
the face. Sludge sampling is occasionally undertaken in the blast
holes, or in holes driven into the hanging or footwalls if splits in
the vein are suspected.
CHALLENGES
Narrow vein mining also has some inherent challenges, primarily:
• multiple working faces to achieve tonnage output,
• unplanned overbreak and dilution from hanging wall failures,
• planned dilution from the requirement for the stope to be
wider than the orebody, and
• mucking in narrow confines in flat-dipping stopes.
Multiple working faces to achieve tonnage output
While narrow veins are usually high grade, sufficient tonnage has
to be mined to ensure profitable operations. At Warrior, the
average width is 2.4 m. An average density of 2.7 tonnes per
cubic metre was used in conservative ore reserve estimates, but
the actual density is usually higher, around 3.0 to 3.9 at head
grades of 15 to 20 g/t Au due to the association of sphalerite and
galena with high grades, delivering higher tonnages from smaller
volumes than lower grade ore. There is a roughly linear
relationship between grade and density. The key to achieving
productivity is to have multiple working faces, which in turn
requires adequate drilling ahead of operations so mine planners
are able to schedule stoping operations and particularly the
decline advance to guarantee continuity of ore supply.
Planned and unplanned dilution
Dilution control is critical to the profitability of all narrow vein
mines. Dilution will occur from planned dilution, where the
stope shape is controlled by the longhole pattern but the orebody
will pinch and swell within the stope outline. Unplanned dilution
may result from unscheduled ground movements such as hanging
wall slabbing or failure.
In conventional operations, this dilution has to be transported
to surface and milled, increasing haulage and processing costs
and diluting the mill head grade.
Dilution is planned to be controlled at Warrior and Sunburst
after blasting by photometric ore sorting, using 2 × 50 t/h sorters
manufactured by UltraSort at Hornsby in Sydney. Test work
completed to date shows that photometric sorting successfully
removes the expected 30 per cent dilution with minimal gold
losses and minimal misidentification of waste as ore. Run-of-
mine ore will be crushed to -50 mm, with the -7 mm fines going
direct to the treatment plant, and the +7 mm material sorted.
Sorting at Warrior will be above-ground on the floor of the
Washington pit, but at Sunburst the crushing, screening and
sorting will be conducted underground to minimise waste
transport costs, and also to minimise environmental impacts of
noise, dust and night lighting within the Charters Towers city
limits. Waste will be placed back into stope voids as fill.
Mucking challenges in flat-dipping stopes
Mucking of ore in flat-dipping narrow veins, where the dip is too
steep for trackless equipment and too shallow for the ore to run
under gravity, has conventionally been done using scrapers or
Narrow Vein Mining Conference
 NARROW VEIN MINING AT CHARTERS TOWERS, QUEENSLAND, BY LONGHOLE OPEN STOPING
slushers. These are triangular steel shoes winched up and down
the stope by air or electric winches, and require a pulley system
to be installed into the upper wall and the winch installed below.
The winch operator is at risk from snapped cables recoiling in his
direction, and loose rocks that may roll down slope.
At Warrior, the dip is about 48° and ore runs under gravity.
Where the orebody rolls over and flattens to the point where the
ore will no longer run, it will be washed from stopes by
high-pressure low-volume water jets for trucking by trackless
load-haul-dump units. The jets will be mounted on a Hiab style
hydraulic arm fitted to a Toyota Landcruiser. This will allow the
arm to articulate inside the stope for up to 3 m from a vehicle
running on a rock-bolted and meshed level, without the need for
workers to enter the stope.
RESOURCE AND RESERVE ESTIMATION
Resources
Resources are estimated by validating the geological model from
drilling and constructing wireframe orebody models in SURPAC.
The mineral resource drilling database used in the 2005 estimate
included 147 053 m of drilling from 1809 drill holes, of
which 44 259 m is diamond-core (mainly HQ [63.5 mm] and
NQ [47.6 mm] diameter) in 322 holes, 94 694 m is reverse
circulation (RC) percussion drilling in 1240 holes and 8100 m
of other non-core drilling (mainly open-hole percussion) in
247 holes. The holes intersected down-dip and along-strike
extensions of known structures. From this database, there are
1559 significant drill intersections for which a mineralised body
code could be identified, in 644 drill holes. A significant drill
intersection is one assaying 0.1 g/t Au and/or over 100 ppm lead,
or sometimes an intersection with identifiable quartz vein or
formation material indicating that the vein had been intersected
even if assay values were low. Over 97 per cent of the
intersections are above 0.1 m.g/t Au and over 80 per cent of
intersections exceed 1 m.g/t Au. There are 22 significant drill
intersections deeper than 1000 m, of which 19 are deeper than
1100 m and ten deeper than 1200 m (maximum gold grade
20.54 g/t). The deepest significant intersection is 1300.1 m
(2.3 m.g/t Au). This is positive proof that economic gold grades
persist along strike and down dip from the previously mined
areas down to at least 1200 m vertical depth.
Paragraph 19 of the JORC Code 2004 defines a ‘Mineral
Resource’ as a concentration or occurrence of material of
intrinsic economic interest in or on the earth’s crust in such form,
quality and quantity that there are reasonable prospects for
eventual economic extraction. The location, quantity, grade,
geological characteristics and continuity of a Mineral Resource
are known, estimated or interpreted from specific geological
evidence and knowledge. Mineral Resources are subdivided, in
order of increasing geological confidence, into Inferred,
Indicated and Measured categories.
The JORC Code dwells on continuity of grade, as well as
geological continuity of the host structure or body. One
significant issue is that the economic gold grades at Charters
Towers are known to be not continuous, based on some 180 km
of previous drives and the extraction of 6.6 million ounces of
gold from six million tonnes of rock. Gold occurs as small grains
mostly less than 200 microns although a nugget effect can be
detected in statistics as cluster of large numbers of small grains.
There is very little visible gold, even at grades of over 800 g/t
Au. There are few coherent shoots within the structures, rather
patches of mineralisation separated by areas of subeconomic
grades. Therefore the concept of grade continuity is difficult to
prove at Charters Towers. What can be proven from drilling is
that the host structure (lode or formation) persists between drill
holes and a good geometrical fit can be obtained to match a
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
cloud of 3D intersections up to the planar structures defined from
previous workings (inclined shafts or ‘underlies’ in the plane of
the lode, and drives and old stopes in the plane of the lode). This
establishes geological continuity but not grade continuity.
What can also be proven is that while economic gold grades
may not be continuous, the mineralisation process is. Very few of
the holes drilled resulted in undetectable gold – only some 14 per
cent of holes drilled had no detectable mineralisation. The
majority contained gold with associated base metal signatures in
zinc and lead. Grades and widths vary wildly along strike and
down dip, and there is no statistical correlation between grade
and width, or between grade, width or depth. Therefore
determining a valid statistical or scientific basis for drill spacings
to represent various categories of resource is not possible. Using
past experience with some other orebody somewhere else in the
world may be comforting but has no scientific basis when
applied to a deposit that is not identical.
On this basis, drill spacings in some Indicated categories may
be as wide as 50 m to 80 m where there was confidence that drill
intersections could be unequivocally linked to known lodes
extrapolated from and interpolated between old workings.
Inferred resources may be interpolated for up to 300 m or 400 m
from old workings to an outlying drill hole, but any extrapolation
away from a boundary hole was limited to a maximum of 100 m
in the Inferred category. Very early kriging on limited drill hole
data in 1994 recorded variography ranges as low as 5 - 8 m.
Recent variography work by Rankin (2008) determined a range
of 50 - 80 m for the Warrior orebody, which could imply a drill
spacing of twice the range (100 - 160 m) could be used for the
Indicated category. However, it was decided it would be prudent
to restrict spacings only half this range, with drill spacings of
50 - 80 m as a maximum.
Grades are essentially a polygonal estimate, based on
averaging drill grades above cut-off within areas blocked out
from the wire-frame model. Acknowledging that polygonal-type
estimates tend to overestimate tonnage, a mining factor,
payability, was introduced into the resource stage to cut back the
tonnage based on the payability of the previously mined areas.
Old surveyed stope plans indicated that around 30 per cent of the
area defined by level driving, raising and winzing was actually
stoped, and a similar payability adopted in the modern estimates.
Grades at the resource stage were left uncut, as log probability
curves indicated single populations that were log normal and
relatively unskewed, and therefore there was no statistical basis
for arbitrarily cutting high grades. If a top cut cannot be justified
or validated based on statistical studies, then there is no
justification for an arbitrary cut based on ‘common practice’ or
‘previous experience’.
A second mining factor, dilution, was also introduced into the
resource category, where ore intersections less than 1 m true
width were diluted out to a minimum of 1 m, as it was regarded
as improbable that widths less than 1 m could be successfully
mined. At the Reserve stage, planned dilution was introduced by
assuming a minimum stope width of 1.4 m was the likely
achievable limit.
Reserves
The project has no Measured Mineral Resources or Proved Ore
Reserves at this time. The lack of continuity of grade means that
an area can only be proven to be mineralised after levels have
been driven along the lode. A 25 m drill spacing is used for
Indicated Mineral Resources to convert to Probable Ore Reserves
ahead of mining but even at this spacing there is potential for
uneconomic zones between holes. Drilling at closer intervals is
seen as uneconomic when a decision to mine has been made at
the Inferred or outer Indicated category and grade control will be
by face sampling during mining.
11
C A J TOWSEY
As stoping commences within one to two months of a level
being developed, it is a pointless exercise to estimate Proved
Reserves when the area will be mined out within a month or two.
Reconciliations between original estimates and 12 months of
production at Warrior in the 2007 - 2008 financial year indicate
that run of mine grades were higher than originally estimated,
confirming the prudent and conservative nature of the grade
estimates undertaken. Kriged estimates on 25 m drill centres gave
a grade of 7 g/t with run of mine production grades averaging
9 g/t. There is evidence that diamond drilling in this deposit tends
to underestimate grades, as the spectacular face grades of up to
800 g/t samples underground were not seen in the diamond core
assays where highest grades tended to be 20 - 40 g/t.
CONCLUSION
Narrow vein mining of high-grade veins has advantages in low
tonnage breaking, handling and milling costs balanced by high
ounces per tonne, resulting in relatively low mining costs per
ounce. There are safety advantages in not having workers inside
open stopes.
Dilution control and mucking of flat-dipping stopes are
significant challenges, combined with a requirement for multiple
working faces to obtain sufficient tonnage to feed the mill.
Gold grades are highly variable presenting challenges in ore
reserve estimation. Tonnages are discounted by 70 per cent to
allow for the low payabilities (30 to 50 per cent) of the veins.
Estimation techniques allow high confidence in overall global
grade estimates but poor reliability on a local scale (25 m).
Kriging ranges as low as 5 - 8 m means that obtaining a Proved
Ore Reserve by diamond drilling is not economically feasible, and
broad-scale mine planning decisions are made at the Inferred
Mineral Resource stage and mine design work at the Probable Ore
Reserve stage. Proved Ore Reserves can only be estimated after
underground development has proceeded through the mineralised
body with levels at 15 m vertical spacing sampled at 2 - 3 m
intervals along the level. However, reconciliations between actual
gold produced and earlier resource estimates produce workable
results.
12 Ballarat, Vic, 14 - 15 October 2008
REFERENCES
Blatchford, A, 1953. Charters Towers goldfield, in Geology of Australian
Ore Deposits (ed: A B Edwards), pp 796-806 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Dominy, S C, Johansen, G F and Annels, A E, 2001. Bulk sampling as a
tool for the grade evaluation of gold-quartz reefs, Transactions of the
Institute of Mining and Metallurgy, Applied Earth Science,
110(3):176-191.
Kreuzer, O, Dominy, S C, Raine, M, Platten, I, Hodgkinson, I H and
Morrison, R J, 2002. Ore controls and grade distribution in
mesothermal veins of the Charters Towers goldfield, North
Queensland, Australia, Australian Institute of Geoscientists Bulletin,
No 36.
Laing, W P, 1991. Ore controls and predicted extensions at Charters
Towers, report to Mt Leyshon Gold Mines Ltd.
Morrison, R J, Storey, N J M and Box, S, 2001. Charters Towers gold
mines – The importance of historical data, in Mine Geologists’
Workshop Abstracts Volume, pp 26-27 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Morrison, R J, Storey, N J M and Towsey, C A J, 2004. Management of
geological risks associated with quartz reef gold deposits, Charters
Towers, Queensland, in Mine and Resource Geology: EGRU-NQ
AusIMM Symposium (ed: S C Dominy) EGRU Contribution 62,
148 p (James Cook University of North Queensland: Townsville).
Rankin, R A, 2008. Charters Towers Warrior East Project resources,
report for Charters Towers Gold Pty Ltd, GeoRes Project No
GR0718.
Reid, J H, 1917. The Charters Towers goldfield, in Geological Survey of
Queensland Publication No 256, 236 p (Geological Survey of
Queensland).
Towsey, C A J, 2003. The Charters Towers gold mine project, in Northern
Engineering Conference – NEC 2003, Townsville (Institute of
Engineers Australia).
Towsey, C A J, 2005. Case study: Specialised mining methods in narrow
vein mining at Warrior gold mine, Charters Towers, Queensland, in
Third Annual Underground Mining 2005 (IIR Conferences).
Towsey, C A J, Morrison, R J, Foord, G E and Storey, N J M, 2002. The
Charters Towers gold project gold production plan, Charters Towers
Gold Mines Ltd, Brisbane, Qld [online]. Available from: <http://
www.citigold.com>.
Narrow Vein Mining Conference
Return to Hillgrove
P Ganza1 and C Dell2
ABSTRACT
Hillgrove, situated on the New England plateau of New South Wales, is
an area rich in history and resources. Key resources are predominantly
antimony, gold and scheelite and the area surrounding the historical field
is environmentally sensitive and topographically challenging.
Since acquiring the Hillgrove leases, Straits Hillgrove Gold has
undertaken a significant exploration program on these historical mining
areas with the goal to return the field to production. This exploration
program has enabled Straits to significantly increase the resource base
and has subsequently enabled the planned redevelopment of Hillgrove to
be realised. In October 2006, installation of mining infrastructure and the
initial stripping of existing development had commenced.
To successfully redevelop Hillgrove, Straits has had to adopt a highly
mechanised, trackless, diesel mining operation which will see the
production rate commence at an initial rate of 250 kt/a of ore from a
narrow vein environment. This ore will be processed on site to produce
antimony metal, gold bullion and a scheelite concentrate.
INTRODUCTION
The Hillgrove Mineral Field is situated in north-eastern New
South Wales, approximately 25 km east of Armidale The field
has been mined since the 1870s and produced over 720 000 oz
of gold with current resources in the order of 1.3 Moz gold
equivalent. In excess of 200 individual occurrences have been
identified to date. Recorded gold production of 720 000 oz along
with current resources in the order of 1.3 million ounces of gold
equivalent highlight the significance of this district. Historical
antimony production is estimated to be in excess of 50 000 t.
Significant tungsten of over 2000 tonnes in the form of scheelite
has also been produced from the field. Figures 1a and 1b depict
the Bakers Creek and Sunlight mines in the early 1900s.
1. MAusIMM, Mining Manager, Straits Hillgrove Gold, 51 Bracken
Street, Hillgrove NSW 2350. Email: pganza@hillgrove.straits.com.au
2. Project Geologist, Straits Hillgrove Gold, 51 Bracken Street,
Hillgrove NSW 2350. Email: cdell@hillgrove.straits.com.au
A B
FIG 1 - (A) Bakers Creek mine, (B) Sunlight mine.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
REGIONAL GEOLOGY
The Hillgrove Mineral Field is focused in an area of
approximately 9 km × 6 km in a north-west striking belt between
the Chandler and Hillgrove faults (Figure 2). The field is situated
on the edge of the New England Plateau and is dissected
by a gorge system up to 500 m in depth (Figure 2). Mineral
occurrences of the Hillgrove Au-Sb-W district are hosted in late
Palaeozoic polydeformed, hornfelsed metasediments and
Permo-Carboniferous granitoids of the New England Orogen.
Nearly all of the mineralised structures of the Hillgrove region
lie between two major north-east striking regional structures, the
Hillgrove and Chandler Faults (Figure 2). These structures are
largely ductile and mylonitic in character, cutting the granitoids
and metasedimentary rocks. Metamorphic grade changes across
the Chandler Fault imply significant asymmetric uplift along the
northern side of the fault, occurring between 266 - 256 Ma.
GEOLOGY AND STRUCTURE
The mineralisation at Hillgrove is hosted predominantly in a
central zone of biotite-grade metamorphosed sedimentary rocks
of the late Palaeozoic Girrakool Beds (originally shale, siltstone,
argillite, greywacke). The meta sediments are bounded to the
north by biotite monzogranite of the ~300 Ma Hillgrove
Adamellite, and to the south by granodioritic-dioritic rocks of the
early Permian Bakers Creek Diorite Complex. Although the
majority of deposits are hosted by metasediments several
significant lodes are developed in both the Hillgrove Adamellite
and Bakers Creek Diorite.
Syn-mineralisation lamprophyre dykes dated at 247 - 255 Ma
are both cut by mineralisation as well as having intruded
mineralised structures (Ashley et al, 1994; Boyle and Hill, 1988).
The majority of dykes are closely related spatially to mineralised
structures and are up to a few metres wide, rarely exceeding
three metres. Geochemically, the lamprophyres are related to the
high-K I-type granitoids of the Permo-Triassic Moonbi Plutonic
Suite.
13
P GANZA and C DELL
FIG 2 - Hillgrove location and regional geology map.
Mineralisation is developed in veins, vein breccias, sheeted
veins, network stockworks and as alteration sulfide haloes to
the main structures. The majority of mineralised veining has a
north-west oriented strike, with dips commonly 70° to vertical
and vary in widths of up to 20 m in places. Mineralised
structures commonly pinch and swell, according to the presence
of local dilatational sites leading to large variation in widths of
mineralised veins and breccias (from <1 cm to several metres).
Ore shoots are typically subvertical with internal movement
indicators suggesting a component of left lateral movement
associated with the latest stibnite–quartz–gold vein breccia
stage.
Paragenetic studies have previously indicated that the earliest
mineralising event was a scheelite-bearing phase of quartz
veining. Subsequent phases of arsenopyrite–pyrite–quartz–
carbonate veining were accompanied by gold and minor base
metal sulfides. Alteration is typically sericite–ankerite–quartz.
Overprinting stibnite–quartz veining with gold-electrum,
aurostibite and arsenopyrite form an important subsequent phase.
The sulfidic haloes about the mineralised structures vary from
being narrow and tight, to up to 20 m wide. In these zones, it has
been shown that gold grades (with little or no accompanying Sb)
can be significant, with gold hosted ‘invisibly’ in the sulfides.
Arsenopyrite is the main host to invisible gold, although a
smaller proportion is also hosted in pyrite that tends to be of
arsenical composition (Ashley, Creagh and Ryan, 2000). The
alteration-mineralisation process has led to addition of Au, Sb,
As, S, CO2, K and Rb, with depletion of Na and Sr. Fluid
inclusion studies have indicated homogenisation temperatures in
gangue quartz in the range 100 - 250°C and that fluids were of
low salinity.
Veining can be inferred from historical records to extend for
vertical depths of over 1 km. Metal zonation within the field is
variable between the lode systems but on the basis of past
production can generally be inferred to be from Au-As at depth
to Sb-Au-As at shallower levels, with minor scheelite occurring
throughout the production interval.
RESOURCE AND RESERVE
The current resource and reserve, as at June 30 2008, is detailed
in Table 1.
14 Ballarat, Vic, 14 - 15 October 2008
TABLE 1
Hillgrove resource.
Category Tonnes (kt) Sb (%) Au (g/t)
Resource 4676 1.9 4.7
Reserve 2210 2.5 4.0
THE REDEVELOPMENT PROCESS
An extensive exploration program led to the decision to
redevelop Hillgrove with the focus on the Syndicate orebody
located on the Metz side of the gorge. The redevelopment plan
identified immediate projects that required attention due to
potential impacts on the project. These included:
• development of the Bakers Creek road, to provide access
from Hillgrove to the Syndicate mine by means of a road
through the gorge rather than using public roads around the
plateau;
• planning and installation of mine services to immediately
satisfy the exploration program and aid in the
recommencement of mining; and
• metallurgical test work and plant design.
Running along side these projects, plans for the
recommencement of mining operations were formulated by
on-site mining and geology personnel and implemented to see
the recommencement of mining operations at Hillgrove in
October 2006.
MINING
The approach at Hillgrove has been to move from a labour
intensive, low productivity but highly selective shrink stoping
operation to a highly mechanised, high productivity sublevel
longhole open stoping operation ensuring a feed of 250 kt per
annum to the plant.
Shrink stoping was previously developed on 50 m sublevels
with timber lagging and chutes (Figures 3a and 3b), with ore
trucked to surface tipples within the gorge by rail. The ore was
then loaded into conventional highway trucks and hauled to the
ROM pad.
The redevelopment plan identified sublevel longhole open
stoping as the safest, productive and flexible method for the site.
Initial stoping has commenced on 16.5 m sublevels (due to
pre-existing infrastructure) with plans to expand the sublevel
interval as confidence in stoping increases. The average ore
width is 2.0 m with the grade varying across the ore face with
respect to the antimony, gold and tungsten location within the
shear zone. Ground conditions vary across the ore with respect to
the degree of alteration. The ore is steeply dipping with an
average dip of 80 degrees to the west with strike lengths of
450 m. Stopes are planned to be 25 m in length with a 5 m rib
separating stopes along strike. Crown pillars are located every
100 m vertical and attempts to recover these crowns will occur
after stoping has been completed and the stope has been
backfilled. Figure 4 depicts the planned stope layout.
A top down sequence retreating to central cross-cuts is being
practiced.
Stope drilling will be with 64 mm blastholes with a dice five
pattern being the predominant pattern. Ring spacings of 1.3 m
are currently practiced with plans to expand with increasing
confidence. Trials of 64 and 76 mm ‘zipper’ patterns are
currently being trialled in narrower (1 m and less) areas.
The main decline is being developed at 5.0 m wide × 5.2 m
high (arched profile) with ore drives developed at 3.5 m wide ×
4.5 m high (shanty profile).
All vertical development has been undertaken by Alimak and
longhole rising techniques.
Narrow Vein Mining Conference
 RETURN TO HILLGROVE

A B

FIG 3 - (A) G81 stope, (B) Sy77 stope.

FIG 4 - Long section of syndicate.

OWNER OPERATOR VERSUS CONTRACTOR
Straits Hillgrove embarked on the tendering process
simultaneously with the development of owner operator costs
utilising new and second-hand equipment. All contractors were
informed of Straits intention to develop and compare submitted
contractor costs with the owner operator costs derived by an
external consultant. The contract period tendered for was a two
year period with the opportunity to extend for an additional two
years. Results of the comparison are depicted in Figure 5.
The tendering process identified a number of areas that
represented high uncertainty and risk with the redevelopment of
the Syndicate mine. The risks included:
• personnel,
• equipment,
• access and material movement,
• power – initial start up and long term, and
• ventilation – initial start up and long term.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 15

P GANZA and C DELL

90 Power was limited to 415 volts delivered underground and

74.8
80 therefore plans were required to introduce an 11 kV feed direct
Total Cost ($M)

62.4 69.8
70 59.7 to underground using a pre-existing shaft. The initial
60 46.5 42.9
50 development was undertaken on generator power supply until
40 access to this 11 kV power could be realised. The gorge itself
30 represents the biggest hindrance to power supply, with a
20
10 maximum deliverable power to the Syndicate mine set at
0 1.25 MVA due to aerial conductor size running across the gorge
(approximately 2 km). Investigations are underway to identify

p
3

4
2
1

ui

ui
or

or
or
or

Eq

Eq
other means of increasing the delivered power to Syndicate and
ct

ct
ct
ct

tra

tra
tra
tra

ew

d
an
other future prospects within the gorge.
on

on
on
on

-N

-h
C

C
C
C

nd
p
Ventilation was limited to a small volume circuit relying on
rO

-2
ne

p
rO hand-held rises between pre-existing levels. This was adequate
w
O

ne

for shrink stoping previously practiced and subsequent

w
O

exploration programs but it could not support a move to diesel,

trackless mining. Initial network designs and simulations were
FIG 5 - Contractor and owner operator cost comparison.
developed to upgrade the system to an efficient low pressure
system to minimise power draw and maintain adequate flow to
The comparison of contractor and owner operator costs the underground workings. To avoid leakage issues posed by the
highlighted significant financial advantages by going owner existing workings and multiple break-throughs from old stopes, a
operator but the greatest advantage was in the owner operator’s new ventilation decline was developed to provide the primary
ability to address changing situations presented by the lower ventilation return airway. The current primary ventilation circuit
three identified risks above. The personnel and equipment risks has seen an increase in flow by 500 per cent with dedicated fresh
were perceived to be faced equally by contractor and owner and return airways.
operator scenarios.
EQUIPMENT
PERSONNEL
The recent increase in mining activity across Australia has led to
The goal was to employ people who wanted to move to a the procurement of underground equipment experiencing ever
residential location within close distance to the coast and have a increasing lead times. With this in mind during the planning
roster that was amenable to promoting a good family oriented phase, Straits Hillgrove Gold opted to commence mining with a
lifestyle. To do this, Straits had to employ a core group of second hand fleet of equipment in reasonable condition, with the
experienced operators and embark on a rigorous training aim of rebuilding these items as required and introduce new
campaign on local personnel interested in working at Hillgrove. equipment when and where available. The core mining fleet is
The core group included management, supervisors, jumbo detailed in Table 2.
operators and truck drivers. ‘Green’ personnel were employed
and trained to fill various positions throughout the mining crews. TABLE 2
The initial roster was five days a week, Monday to Friday, with Hillgrove core mining equipment list.
two crews operating ten hours per shift for 50 hours per week.
This roster was in place for approximately 15 months with mixed Item Quantity Purchased
success due to fatigue and productivity issues. A review of this Atlas Copco H282 Boomer 3 Second-hand – two rebuilt
roster was implemented which resulted in the change to a four
days a week roster, Monday to Thursday, with the same two Atlas Copco 1257 Simba 1 New
crews working 12 hours per shift for 48 hours per week and a Elphinstone R1700 series II 2 Second-hand – one rebuilt
longer break between the change from dayshift to nightshift. The
Elphinstone R1300 series II 1 Second-hand – rebuilt
change was prompted by ever present fatigue and traffic
management issues and an anticipated improvement in Caterpillar R1300G 1 New
productivity, due to better communication between crews and Caterpillar IT28G 2 Second-hand
ability for the mining cycle not to be ‘stop–start’ every work day.
Caterpillar 120H 1 Second-hand
In the three days off, the mine is closed, except for maintenance
and exploration activities. Volvo FM12 haul truck 2 Second-hand
The change in roster has seen an immediate improvement in Tatra Jamal haul truck 2 New (due late Sept 2008)
average development rates of 270 m per month to 350 m per
month. No stoping activities were performed during the initial
roster. Non-core equipment items include two Toro 151D boggers,
Since the recommencement of mining operations the turn over one Tamrock 105 diesel hydraulic jumbo and numerous items of
in mining personnel including professional staff has averaged rail and hand-held equipment for rehabilitation and exploration
15 per cent annually and of this 15 per cent, at least 95 per cent programs.
have not been residents of the local area. The selection of mining equipment has been based on safety,
dilution, production rates, availability and the proposed mining
method. The selection rationale of the main items of equipment
INFRASTRUCTURE can be summarised as follows.
The existing infrastructure was less than suitable for the
recommencement of mining operations. The mines of the Jumbo
Hillgrove area were primarily developed for rail access only with
the Syndicate mine being the only exception where a decline was The jumbo had to have the smallest possible carrier, smallest
developed to access the lowest two levels of the orebody. This powerpack size, longest available drilling length whilst still
decline was developed at 4.0 m wide × 3.0 m high (average) and being able to install ground support items (2.1 m, 2.4 m and
presented immediate access issues for the move to trackless 3.0 m splitsets and mesh) in the small ore drives. The Atlas
mining. Copco H282 boomer with BMH2343 feeds best suited this

16 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 RETURN TO HILLGROVE

application. The selection of this jumbo was a compromise TABLE 3

between acceptable dilution and productivity and has ultimately Hillgrove truck costs.
affected the bogger selection. The site has standardised on the 282.
Truck t/h $/t $/t.km $/h
Bogger Mack Titan 575 27 5.45 1.82 146
The size of the ore drives and expected productivities drove Volvo FM12 37 1.87 0.62 72
the bogger selection to the 6 tonne class and ultimately the Elphinstone AD40 II† 37 7.57 2.52 282
availability of boggers in this class dictated the selection of the
Caterpillar R1300. The R1300 boggers are to operate in the ore † Reduced tub capacity to 30 tonnes.
drives and utilise stockpiles and orepasses to deliver ore to the
R1700 boggers for loading trucks.
truck available and consequently purchased for utilisation on site.
Production drill Recent trials have identified the Tatra Jamal (Figure 6d), as the
most compatible to the local conditions. Two Tatra Jamal trucks
With the selection of development jumbos, the selection of a have been ordered for delivery to site mid to late September 2008.
suitable production drill was constrained to the same brand. The
logical choice with site specifications was the Simba 1257 which The haul from the Syndicate mine to Hillgrove is unique in that
would satisfy all production, cablebolt, service hole and rise the surface component accounts for 85 per cent of the haul with
drilling. positive and negative gradients in both directions of travel. Surface
haul is approximately 5.5 km with the underground accounting for
Haul trucks approximately 1.0 km at this stage. Figure 7 depicts the typical
terrain encountered in the materials handling process.
The selection of suitable haul trucks has been an arduous task with
numerous trials performed on site involving different makes of
haul trucks. The trials focused on safety, speed and productivity.
Trials were conducted within the gorge and trial vehicles ranged
from specific underground vehicles to tractor-trailer combinations.
Considerations for all trials were braking (with respect to
gradeability), fire suppression, catalytic converter, ergonomics,
fuel burn rates, tyre usage/availability. All trucks utilised
underground have AFFF, catalytic converter and adequate braking
for the application within the gorge and underground. Actual
costs, including fuel, operator and maintenance, realised to date
for site based trucks are detailed in Table 3.
During the recommencement of mining operations, haulage
started with a Mack Titan 575 (Figure 6a) with the addition of
an Elphinstone AD40 Series II (Figure 6b), for short tram waste
haulage. As the trials continued, two second-hand Volvo FM12
trucks (Figure 6c), were identified as the best immediate haul FIG 7 - Typical haul within the gorge.

A B

C D

FIG 6 - (A) Mack Titan 575, (B) Elphinstone AD40, (C) Volvo FM12, (D) Tatra Jamal.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 17

P GANZA and C DELL
FUTURE DIRECTION
The topographical challenges at Hillgrove present significant
logistical problems for personnel and material movement. During
the course of 2008, initial scoping studies are planned to
investigate potential haulage methods which will supplement the
lengthy truck haulage currently practiced. Truck haulage is only
a short- to medium-term solution for haulage but there will
always be a requirement for trucks on-site to deliver ore to an
alternative haulage system. Alternatives include shaft hoisting,
inclined rail down the side of the gorge, inclined conveyor or
aerial skipway. The local topography also impacts the
exploration potential of the site due to access difficulties and the
usual complexities with drilling out a narrow vein system. As
Straits accesses additional workings, the flow on exploration
effect will be enormous as exploration has been limited to
approximately ten of the 200 plus known lodes on site.
The narrow vein nature of orebodies at Hillgrove also lend
themselves to trialling alternative stoping methods. Initial
discussions are underway with respect to Alimak stoping. The
incorporation of paste fill is being assessed for application within
the present stoping areas and filling of existing voids on site.
Paste fill will enable the site to realise increased extraction from
planned stopes, commence remnant mining and also move to a
continuous mining system.
18 Ballarat, Vic, 14 - 15 October 2008
CONCLUSION
Straits Hillgrove Gold is taking significant steps to ensure
Hillgrove becomes a major long-term producer of antimony, gold
and tungsten within the Australian mining industry. The current
mining and exploration is focused on three main mining areas
which have enabled the Hillgrove operation to recommence. As
these are developed, exploration will focus on near mine and
regional exploration targets to ensure the longevity of the
operation.
REFERENCES
Ashley, P M, Cook, N D J, Hill, R L and Kent, A J R, 1994. Shoshonitic
lamprophyre dykes and their relation to mesothermal Au-Sb veins at
Hillgrove, New South Wales, Australia, Lithos, 32:249-272.
Ashley, P M, Creagh, C J and Ryan, C G, 2000. Invisible gold in ore and
mineral concentrates from the Hillgrove gold-antimony deposits,
NSW, Mineralium Deposita, 35:285-301.
Boyle, G O and Hill, R L 1988. The Hillgrove antimony and gold field, in
New England Orogen Tectonics and Metallogenesis (ed: J D
Kleeman) (University of New England: Armidale).
Narrow Vein Mining Conference
The Challenger Gold Mine
P Androvic1, P Bamford2 and M Sandy3
ABSTRACT
The Challenger Gold deposit was discovered in 1995 by Dominion
Mining Ltd in a remote western area of the Gawler Craton in South
Australia. Mine development commenced in 2002 with initial production
from an open pit supplemented by ore from a very small adjacent
auxiliary pit. Development of the underground mine commenced in 2004.
The orebody plunge of just less than 30 degrees requires careful
attention to mine design. Production from the M1 orebody, with less than
800 ounces per vertical metre (ovm) and very limited ‘surge’ capacity,
requires careful scheduling and implementation to ensure continuity of
ore feed for milling.
Initial production results from underground stopes included more
dilution than anticipated. The narrow, high-grade zones proved to have
much more complex geometry than early interpretations had indicated.
Greatly improved understanding of the mineralisation controls resulted
in important changes being made to the underground mine design and
method. Coupled with careful control of development and drilling and
blasting practices, this resulted in an immediate reduction in dilution and
high recoveries of the resource.
The benefits of these changes flowed through in the 2005/06 financial
year, resulting in a 150 per cent increase in gold production to 108 080 oz
and a sharp reduction in cash operating costs to A$280/oz. This was
followed in 2006/07 with production of 108 91 oz of gold at an
operating cost of $309/oz and in 2007/08 with 109 326 oz of gold
at $367/oz.
Production to June 2008 has been mainly from the M1 orebody, with
the lowest stoping to date on the 580 m Level (615 m below surface)
above development mining on the 520 Level. The rate of vertical advance
of mining this orebody has averaged 150 vertical metres per year.
The increased current combined JORC compliant resource and
reserve position of over one million ounces, now includes the M2 and
M3 orebodies with a current total average gold inventory up to 1400 -
1500 ovm. Exploration has intersected other shoots and all shoots are
open at depth.
The adjacent M2 and M3 orebodies are being developed for stoping.
The narrow width of these zones from 1 m up to 4 m will require
continued focus on ore zone definition, accurate development and careful
drill and blast practice. In addition, to minimise cost increases in the
current period of rapidly escalating labour, equipment, consumables and
energy costs, high productivities will need to be maintained to allow the
established rapid mining advance to continue at low unit costs.
LOCATION
The Challenger Gold Mine is owned and operated by Dominion
Gold Operations, a wholly owned entity of Dominion Mining
Ltd. Challenger is located 750 km NW of Adelaide in the Gawler
Craton (Figure 1) with the nearest community of Coober Pedy,
300 km by road to the NE.
The mining lease is situated within the Commonwealth
Hill/Mobella Pastoral Lease as well as within the Woomera
Prohibitive Area (WPA).
1. MAusIMM, Manager, Mining and Geology, Dominion Gold
Operations, PO Box 453, Torrensville SA 5041.
Email: pandrovic@dgo.com.au
2. FAusIMM, General Manager, Operations, Dominion Mining
Limited, PO Box 465, West Perth WA 6872.
Email: pbamford@dml.com.au
3. FAusIMM, Principal Geotechnical Engineer, AMC Consultants, 9
Havelock Street, West Perth WA 6005.
Email: msandy@amcconsultants.com.au
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
INTRODUCTION
The Challenger Gold deposit was discovered in 1995 by
Dominion Mining Ltd (Dominion) in a remote western area of the
Gawler Craton in South Australia. Mine development commenced
in 2002 with initial production from an open pit supplemented by
ore from a very small adjacent auxiliary pit (SEZ). Development
of the underground mine commenced in 2004.
Greatly improved understanding of the mineralisation
controls resulted in important changes being made to the
underground mine design and method. Coupled with careful
control of development and drilling and blasting practices, this
resulted in an immediate reduction in dilution and increased
recovery of resource.
The benefits of these changes flowed through in the 2005/06
financial year, resulting in a 150 per cent increase in gold
production to 108 080 oz and a sharp reduction in cash operating
costs to A$280/oz. This was followed in 2006/07 with 108 191 oz
of gold produced at an operating cost of $309/oz and in 2007/08
with 109 326 oz of gold at $367/oz.
Production to June 2008 has been mainly from the M1
orebody, with the lowest stoping to date on the 580 m Level
(615 m below surface) above development mining the 520 Level.
The adjacent M2 and M3 orebodies are also being developed for
stoping in the near future. The narrow width of these zones will
require continued focus on ore zone definition, accurate
development and careful drill and blast practice.
The rate of vertical advance of mining this orebody has
averaged 150 vertical metres per year.
DISCOVERY
The Challenger Gold Deposit was discovered in 1995 by
Dominion. Dominion had been exploring in the Gawler Craton
since 1991 and was one of the first companies to use the
pre-PIRSA funded exploration initiative data from the early 1990s.
Using a calcrete sampling method to generate subsoil anomalies
pioneered by Dominion, a number of interesting anomalies were
discovered in the region (outlined in Bonwick, 1997).
Initially using a 1.6 km × 1.6 km grid a number of significant
anomalies were generated and subjected to infill sampling.
Infilling of the regional sampling grids occurred by
progressively decreasing the size of the patterns down to 400 m ×
400 m then 100 m × 50 m, eventually homing in on the Challenger
gold and arsenic anomaly, (Figure 2). An initial RAB drilling
program targeting the calcrete anomaly delivered an intersection
of 28 m @ 5.8 g/t, which launched the Challenger project.
PROJECT DEVELOPMENT
Following a period of frenetic drilling in 1997 and 1998 that
defined the original resource, the project was the subject of a
50/50 Joint Venture with Resolute Ltd who were managers of the
project. A prefeasibility was conducted but little development took
place. Dominion took full control of the Challenger project again
in December 2000 and completed a feasibility study in late 2001.
Construction of a conventional CIP plant was started with
refurbishment of the Mt Monger ball mill from January 2002.
Civils and earthworks began on site in April 2002 and were
completed as planned within a budget of $18.3 million inclusive
of all prestrip mining, owners costs up to plant commissioning,
and the construction of village, borefield and treatment plant
facilities.
19
P ANDROVIC, P BAMFORD and M SANDY
FIG 1 - Challenger location.
FIG 2 - Infill calcrete sampling to 100 x 100 m.
Prestrip mining had commenced in April 2002 to uncover the
orebody and ensure sufficient ore was available to commission
the treatment plant in October 2002.
OPEN PIT MINING
Mining began in May 2002 and the open pit produced 120 000 oz
from an initial reserve estimate of 110 000 oz including a small
satellite pit, the SEZ. The open pit was mined using a PC1600
20 Ballarat, Vic, 14 - 15 October 2008
excavator and a fleet of Komatsu 50 t haul trucks plus ancillary
plant. The weathering profile allowed the pit to be excavated
‘free dig’ down to 33 m below surface before fresh rock required
blasting. The massive nature of the Challenger gneiss, poor
natural fracturing and indistinct geology in the pit resulted in
difficult interpretation of ore zones.
The 380 m by 320 m open pit was grade controlled on 18 m
panels using an RC drill rig and benches were fired on 5 m
heights and mined on 2.5 - 3.0 m flitches. Ore was classified into
four categories, ‘LG’ (1.0 - 1.7 g/t), ‘Blue’ (1.7 - 2.5 g/t), ‘Green’
(2.5 - 5.0 g/t) and ‘Red’ (5.0 g/t+) and placed on the ROM pad
into discrete stockpiles per bench that enabled an accurate
reconciliation tracking process as well as the development of
some large low grade stockpiles.
Blast hole logging and mapping of bench faces did not provide
significant geological control. The true structural geometry of the
ore was still not apparent as no ore was exposed for mapping in
the pit walls until the lower sections of the pit. While there was
an understanding that the ore was folded, the nature and
complexity was still not apparent with the ore mined in robust
blocks that did not necessarily highlight any structural continuity.
The open pit was mined to a depth of 135 m and completed as
scheduled in April 2004.
REGIONAL GEOLOGY
Challenger occurs within the Mulgathing Complex with the
earliest recognised event being the Sleafordian Orogeny (2950 -
3150 Ma?) followed by the Kimban Orogeny between 1745 and
1845 Ma.
The area is characterised by Archean to mid Proterozoic, high-
grade metamorphism (granulite) gneissic basement. The
original granulite facies metamorphism then retrograded to
amphibolite facies recrystallisation around 1650 - 1540 Ma
(Tomkins, 2002).
Narrow Vein Mining Conference

Lamprophyre sills cross cut all structures and have an age of
c.1650 Ma. Dolerite Dykes have been imprecisely dated at
between 750 and 1425 Ma.
CHALLENGER GEOLOGY
The saprolitic clays extended to 40 - 50 m depths in the ore
zones with the base of oxidation tending to be deeper in relation
to the ore.
There are five main rock types:
1. Massive, granulite facies gneiss. Medium grained, psammo-
pelitic in origin with dominant plagioclase, garnet and
lesser biotite and orthpyroxene.
2. Medium to coarse-grained massive pegmatitic granulite
gneiss.
3. Garnet rich plus biotite foliated feldspathic gneiss.
4. Massive lamprophyre dykes, massive with porphyritic
phlogopite and clinopyroxenes.
5. Thin mafic dolerite dykes.
The mafic dykes are the youngest lithological unit and the
contact orientation is the same across the gneiss and
lamprophyre. The lamprophyre sill has been observed in the
mapping to be truncated by the mafic dolerite and the
lamprophyre is significantly younger than the gneiss.
The main mineralogy in the ore zones is arsenopyrite/
loellingite, pyrite, pyrrhotite, Au-Bi telluride and Au. Biotite
appears as a retrograde product of pyroxene. A pervasive
alteration found in late fracture selvedges and on lamprophyre
dyke margins. This is characterised by sericite, carbonate,
chlorite, serpentine and leucoxene.
There are three main leucosome/vein styles:
1. quartz dominant veins, which may be remnant pre-
metamorphic mineralised veins;
2. polysilicate veins, which are dominant in the main ore
zones and are well mineralised; and
3. pegmatitic veins which are late stage veins with cross-
cutting relationships and are unmineralised.
High-grade gold appears associated with coarse grain quartz
veins with arsenopyrite and lesser feldspar and telluride. Garnet
and biotite also occur with free gold and are associated with
sulfides and/or Au-bismuth telluride. The first two leucosome/vein
types also contain the coarse visible gold. This high nugget-effect
coarse gold is responsible for the variability in grade
reconciliation and to date the life of mine reconciliation is ~8 per
cent more than the reserve estimates.
Current thoughts on the genesis of the ore shoots are that they
are hosted within migmatites that have undergone partial melting.
It is possible that this melting ‘fertility’ may reflect a precursor
hydrothermal alteration event, (McFarlane, Mavrogenes and
Tomkins, 2007). The gold mineralisation is structurally controlled
through emplacement of the partial melt into relatively low
strain positions. Monazite geochronology has outlined a 40 Ma
period (2460 - 2420 Ma) of repeated high temperature events
(McFarlane, 2006).
STRUCTURE
The Challenger orebody is defined by a laterally extensive shear
zone with the shoots plunging ~30° to 030°- 035° (AMG) or
055°- 060° mine grid. The ore shoots are defined by leucosome
veins which show characteristic ptygmatic folding. This small
scale folding is parasitic to the larger scale asymmetric folding
recognised in orientated drill core. Early work by Jon Standing
(Standing, 2003), had hypothesised that these shoots are defined
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
THE CHALLENGER GOLD MINE
by flexures of drag folds in the foliation in response to regional
folding. These inferred fold axes are interpreted to have the same
plunges as the ore shoots.
M1 veins have a highly folded nature and the M1 shoot is an
antiformal fold that is characterised by a thick package of
upright, tightly folded quartz and leucosome veins that have fold
axes plunging parallel with the ore shoot, 30° → 060°.
There is evidence (Standing, 2003) of extreme transposition and
extensional rodding of leucosome bands within the psammo pelitic
gneiss. The long axis of the rodded leucosome quartz is parallel
with the plunge of the drag folds and the overall M1 plunge.
With folding being a prepeak metamorphic structural feature
as with the gold mineralisation, quartz veins should be deformed
in a similar manner to the leucosome. An inference can be drawn
that the Challenger shoots were once a continuous ore zone or
two parallel ore zones striking NW-SE, prior to extreme
WNW-ESE shortening and extension directed shallowly towards
the NE (Figure 3).
FIG 3 - Cartoon showing the progressive shortening of psammitic
beds (So) to produce tight to isoclinal folds that ultimately are
transposed by the axial planar foliation (S1) and local shear zones
along attenuated limbs (S2). The psammitic bed could also
represent leucosome melt bands and quartz veins and is
independent of scale (from Standing, 2003).
Detailed mapping is necessary to lead to a better
understanding of the fold structures. The scale and geometry of
the folding was and is still critical to the understanding and
predictability of the ore shoots.
There are three main fold types (Standing, 2006):
1. F1 – isoclinal folds;
2. F2 – upright to inclined, tight to close folds; and
3. F3 – reclined open folds.
The F2 event is the dominant fold structure which controls the
morphology of the ore shoots. Understanding fold vergences is
important where fold closures are obscured. The refolding by F2
folds of F1 folds often accounts for the sudden disappearance of
ore zones with no obvious continuity. The wrapping around of
these F1 folds into F2 geometry dominates the Challenger ore
shoots. F2 appears more complex due to its larger extent and
difficulty in pinning any reference location when cross cutting
through an asymmetrical long limb position.
21
P ANDROVIC, P BAMFORD and M SANDY

UNDERGROUND GEOLOGY
Mapping of the shoots is critical in maintaining and understanding
the dynamic nature of the folded shoots (Figure 4). Changes in
plunge dip and orientation can create apparent lengthening of ore
drives. Changes in wavelength and frequency also affect the
positioning of ore drives. Geological control is important in
ensuring that each successive level is positioned appropriately and
that changes in orientation are noted so that level infrastructure
can be adjusted. Development drives following survey control are
a rarity. The ability to map and interpret the geology is crucial to
the ongoing development of the Challenger orebody. Mapping is
part of the mining cycle and most ore drives are under ‘geological
control’ which necessitates that a good relationship be developed
with the mining contractor.
Figure 5 highlights the distribution of the ore shoots at
Challenger that have been mined. The M1 shoot is a more
symmetrical and open fold system with M2 a combination of
asymmetrical and symmetrical dominant fold zones with the
M3 dominated by long limb asymmetrical folding.
Figure 5 also highlights the target rich environment which will
be investigated in due course. An exploration drive has been
started on the 580 level to delineate the M3 zone at depth. The
M3 shoot has been mined underground in the upper levels
adjacent to the open pit and after M2 shoot, will be the next ore
shoot brought into the mining cycle.
A second 250 m long exploration drive has been designed to
target the M1 footwall zone, M2 Footwall and allow for drilling
positions onto Challenger West. This drive will start on the
800 level at the end of the M2 ore drive. An exploration drive is
currently being developed on the 580mRL along the M3 structure.
By understanding the local structural geometry of the shoots
and development of target positions, the Challenger operations are
most likely to develop additional reserves and extend mine life.
RESOURCE
The Challenger Resource is advanced by directional surface
diamond drilling. Section lines are spaced every 100 m with
average hole lengths reaching 1.3 km on the deepest sections to

FIG 4 - Typical faces within M1.

FIG 5 - Challenger mine scale ore geometry and targets.

22 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 THE CHALLENGER GOLD MINE

date. Each parent hole has four to five daughter holes branching
off and these vary from 250 - 300 m in length each. The
philosophy is to target the interpreted hinge position of the M1
while also intercepting the M3 and M2 before intersecting the
M1 shoot.
Using a long directional drilled ‘navi’ bend to lift the parent
hole ~20°, allows for the five daughter holes to be sequential
drilled back up the bend of the drill hole. This allows for a spread
of holes to cover any variation in the shoots plunge azimuth or
dip, (Figure 6). The ultimate target window is 15 - 20 m and
holes are adjusted mid stream to account for natural hole
deviation and the pierce point is adjusted up or down dip to suit.
The nuggetty high-grade orebody has also proven to possess
both a consistent structural geometry and robust grade continuity
vertically. A block modelling approach to estimate resource tonnes
and grade was used through the open pit and first few years of the
underground. The block model interpolations underestimated
tonnes, over estimate and under estimate grade and generally
under estimated contained metal. A subsequent block modelling
approach using ‘unfolding techniques’ and trialed in parallel
during late 2004 through to 2006 resulted in a closer correlation
but a similar under representation.
The poor ability of the ID2 block modelling interpolation
techniques was demonstrated as not being suitable for resource
estimation because of the nugget effect in addition to the local
variability of the geometry. To this end, the Challenger resource
methodology is based on a ‘generic’ approach which takes into
account both historic reconciled data from the underground
mining using a 180 g/t Au top cut and continuity of orebody
geometry as interpreted from both drilling and underground
development. This approach has proven to be a more accurate
basis for estimation than traditional block modeling. Block
modeling is used in other shoots where appropriate. The resource
to reserve conversion is based upon the geological outlines and
stope designs that incorporate planned dilution to allow for mining
and geotechnical considerations.
An updated Resource statement, summarised in Table 1, was
released to the ASX in July 2008 and highlights the success of
the ongoing Challenger Deeps drilling program.
GROUND CONDITIONS AND ROCK MECHANICS
Conditions in the upper levels were generally ‘extremely good’
as described using the Norwegian Geotechnical Institute’s (NGI)
‘Q’ system (Barton, Lien and Lunde, 1974), with typical Q
values of >100 in the massive garnet gneiss that hosts the
mineralisation (Figure 7, Sandy, 2005).
As part of an ongoing, formal rock mechanics program at the
project, rock property testwork has been undertaken for the major
rock types present, using the facilities at the Western Australian
School of Mines (Scott, Machuca and Villaescusa, 2006). The
results are consistent with data from other sites in similar

TABLE 1
July 2008 updated Resource and Reserve summary, ASX release July 2008.

Challenger Gold Project – 30 June 2008 Resource and Reserve summary

Resource Reserve
t g/t oz t g/t oz
Measured 456 100 9.1 132 810 Proven 522 300 7.3 122 040
Indicated 2 387 100 8.8 678 150 Probable 2 551 100 7.4 605 000
Inferred 1 133 900 9.6 348 870
Total 3 977 100 9.1 1 159 830 Total 3 073 400 7.4 727 040

FIG 6 - Schematic view of hole 07CDDH0078 and daughters.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 23

P ANDROVIC, P BAMFORD and M SANDY
FIG 7 - Good development practices in extremely good ground
(Q>100) resulting in a high ‘half barrel’ count.
lithologies, with an average uniaxial compressive strength of
183 MPa in the granulite gneiss which hosts the gold
mineralisation.
Stress measurements have been conducted at several sites using
the overcoring method and CSIRO Hollow Inclusion (HI) cells
(Litterbach, 2006, 2008). Additional tests have been conducted
with the Acoustic Emission (AE) technique (Villaescusa and
Machuca, 2008). The results are somewhat different, as shown in
Figure 8 which is a plot of stress magnitudes against depth. To
date, very little stress-related damage has been observed in the
mine other than localised spalling of pillars near the lower stope
abutment. Back-analysis of observed damage using numerical
modelling indicates that the currently-observed behaviour is
consistent with the HI stress field, but work in this area is ongoing.
FIG 8 - Stress measurements at Challenger plotted against depth.
24 Ballarat, Vic, 14 - 15 October 2008
Ground conditions continue to be very good in much of the
development in the lower levels, although there has been an
increased occurrence of shallow (generally) east-dipping
lamprophyre-parallel structures/contacts. These can be persistent
over large distances (>100 m) and usually contain very weak
infillings. Increased jointing and associated poorer conditions
sometimes occur near these structures, requiring systematic
application of meshing as part of the development support.
UNDERGROUND MINING – INITIAL
PRODUCTION
An underground feasibility study was approved by Dominion to
proceed in September 2003 and the portal was cut in February
2004 when the production rate slowed whilst mining the base of
the open pit. Open pit mining was completed in April 2004.
It was during this initial stage that one level was thoroughly
drilled and mapped and the ‘true’ nature of the M1 began to
‘unfold’. The narrow, high-grade zones proved to have much
more complex geometry than early interpretations had indicated.
Although the new geometry was complex, when projected back
up into the pit and overlain with the open pit mining blocks the
correlation with high grade was almost exact (Figure 9).
FIG 9 - Ore blocks defined during open pit mining and comparison
to actual M folded geometry.
With increased confidence in the geometry of the M1, a
number of mining methods were examined initially to deal with
the low plunge angle of the orebody at close to but less than 30°.
Three mining methods had been costed within the underground
feasibility study:
1. inclined room and pillar using a jumbo to mine all ore;
2. sublevel open stoping from a waste horizon below the base
of the orebody; and
3. benching, retreating back to the hinge or fold zone of
the shoots.
Uphole benching was adopted as the most suitable and
profitable method (Figure 10). In addition a cut and fill method
was trialled on the 1055 level. Trial stopes on the 1055, 1040 and
1020 levels revealed that the geological definition of upright limbs
and mining out fold hinges discretely was not only possible but
achievable and highly profitable. A bulk mining approach was
shown to be uneconomic. To incorporate interstitial waste as
Narrow Vein Mining Conference

FIG 10 - Mining method used at Challenger.
planned dilution between all the fold limbs reduced the grade of
the stopes significantly. Below these levels using tight geological
geometry constraints mining has achieved low dilution and close
positive reconciliation.
MINE DESIGN
As understanding of the orebody developed through a program of
careful mapping, sampling and sludge drilling, stope designs
were modified to follow the main M1 orebody geometry more
accurately. Through careful implementation of these designs,
minimal overbreak was achieved.
In exploiting the improved knowledge of the orebody
geometry, it is important that design stope shapes are kept as
simple as possible, avoiding sharp changes in profile and
‘re-entrants’. Even in a very good quality rock mass, unduly
complex stope shapes tend to break back to more simple
outlines. The associated fall-off often occurs as large slabs
which can disrupt production prevent access to broken ore or
trap equipment.
Ground conditions assessed from drill core and the exposures
in the Challenger open pit indicated that stope voids could be left
open using the uphole bench retreat method, with an allowance
for local pillars to ensure hangingwall spans were kept below a
nominal design value (hydraulic radius not exceeding 10 m).
Waste rock from development below is tipped into mined out
stopes to minimise waste haulage costs and in some areas
provides a little passive support.
In practice, much larger stope spans have been achieved
than the original design studies suggested would be possible,
(Figure 11). The initial M1 stope extended down-plunge
(east-west) for a total of more than 500 m, with a maximum
north-south span of about 40 m.
This is considered to be partly because the arched shape of the
stopes is naturally stable and the effective span is thus smaller.
The Stability Graph Method on which the nominal stope spans
prior to mine development were based assumes planar stope
surfaces and does not cater for arched stope shapes.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
THE CHALLENGER GOLD MINE
FIG 11 - M1 stope showing clean extraction of the OD1 and OD2
limbs with the central waste pillar left in situ.
STOPING
The consistent orebody geometry has allowed the mine to follow
a highly repetitive mine design, where each level is very similar
to the previous one. The focus is to consistently achieve the
development schedule fundamental to achieving the planned
production rate. At a vertical advance rate in excess of 150 m per
annum, which has been maintained for over three years, the mine
is at the forefront of Australian practice.
The current cycle of mining at Challenger requires ongoing
development diamond drilling for each 20 m level to ‘capture’
the local scale position of the M1 geometry. This allows for
adjustments to the development design to approach the M1 shoot
across the hinge area. Following geological mapping of the cross
cut ore, drives then develop into and mine the main limbs.
Once development has been completed, grade control drilling
using ‘sludge’ openhole percussion is used on 5 m sections to
delineate the cross sectional shape (Figure 12). Each grade
control section is designed to highlight the M1 geometry and
any variations that may occur. The wavelength and amplitude of
the M1 folds can vary down plunge and parasitic folds can also
25
P ANDROVIC, P BAMFORD and M SANDY
FIG 12 - Sludge drilling stope definition.
develop and retreat. The M1 geometry is very dynamic and
mapping of backs along with each face is critical in not only the
initial development of the drives but also in the final
stoping design.
The dense drilling is required to ensure that the geometry is
tightly constrained. Holes are designed based upon the
extrapolated geology model and hinge zones, limbs, parasitic
folds are preferentially targeted. Each section will be designed
based upon the attitude of the fold geometry in that position.
The sludge drilling uses the production drilling rig, a Tamrock
720 Solo with samples taken each half rod, 0.75 m and flushed
down a clear hose into a carousel containing prenumbered sample
bags. A mine technician ensures quality control is maintained and
that each interval is flushed and a sieved sample is collected in a
chip tray for logging by the geologist. Sample sizes vary between
5 - 7 kg and the entire sample is then delivered to the onsite Lab,
dried, split in an automatic crusher and splitter unit allowing a
sample of 500 - 700 g to be assayed in the PAL 1000. The PAL
1000 is an aggressive pulverising and leaching method of assaying
with a turn around of five hours possible for urgent samples.
The scheduling of the sludge drilling is based upon half the
drilling rate of production drilling to allow for the slower
completion of programs. Programs are scheduled on completion
of the ore development drives and they may take over a week to
complete the stope delineation. Turn around from stope modeling
by geologists and mining engineers is within four to five days.
Geological modeling incorporates face sampling and mapping,
backs mapping, sludge logging data as well as the sludge grade
control results. Production rings are ready for mark up once the
final mining shape and ring design has been reviewed by both
geology and mining a second time.
The main challenge presented to date in stoping has been
associated with a lamprophyre sill. This caused significant
hanging wall instability around the 640 mL, at a depth of about
550 m below surface. Stope surveys clearly show the influence of
the overlying sill, with overbreak profiles exactly matching the
interpreted position of the upper contact (Figure 13).
A series of local pillars was designed to arrest the overbreak.
The pillar designs were confirmed using numerical modelling to
ensure that they were appropriately sized. Stoping has been
26 Ballarat, Vic, 14 - 15 October 2008
successfully re-established below the pillars, although the
potential for further instability if more intrusives are encountered
is recognised. In that case, additional pillars will be designed as
required, based on the performance of the current pillars.
DEVELOPMENT PRACTICES AND
GROUND SUPPORT
At Challenger, in many areas the combination of very good to
extremely good, generally massive rockmass conditions and
excellent development drilling and blasting practices has resulted
in conditions where surface support is not generally required. In
the widely-used NGI ‘Q’ system, the conditions described above
will plot in the area of the NGI support selection chart that is
traditionally referred to as the ‘no support’ area.
In practice, light pattern bolting is used as a minimum standard
at Challenger where the conditions do not require surface
support, to ensure that there is always an allowance of bolts to
deal with local, isolated structures. A rigorous program of
inspection and check-scaling is in place to ensure that any scats
that do develop are identified and removed. Mesh is installed by
default when:
• Less than six ‘half-barrels’ are visible in the backs, or less
than two half-barrels are visible in either wall, per cut.
• Unusual scaling effort is required to get back to sound rock.
• A dyke or sill is present in the backs or face (these are
reviewed and monitored for at least five cuts after their initial
exposure as stress conditions may still be affected by
subsequent development at three drive diameters back from
the face).
• A fault or other persistent structure is present in the face or
backs. Similarly these require monitoring for several cuts
after exposure.
• Conditions are ‘blocky’ (due to the presence of several
closely spaced joints from two or more joint orientations).
• The operator has any concerns.
A key factor in obtaining the extremely good development
conditions seen in Figure 7 has been the application of industry-
Narrow Vein Mining Conference

FIG 13 - Stope surveys confirmed that overbreak was strongly controlled by structure.
best drilling and blasting practice, including the use of
prepackaged explosives in the perimeter holes in the backs and
sidewalls. After an initial period of using these explosives, a
change was introduced to lower costs using a technique referred
to as ‘snow loading’ with full strength ANFO. Snow loading is a
technique used in perimeter holes of a development cut, to
reduce impact on the surrounding ground by partially filing the
holes with ANFO rather than tightly packed.
After some experience with the new technique, a study to
formally assess blasting results was undertaken by Kieran Rich,
an undergraduate student at Curtin University. A simple count of
‘half barrels’ and partial half barrels were used as a measure of
the effectiveness of perimeter blast damage minimisation. The
results from these locations are shown in Figure 14.
FIG 14 - Average number of half barrels per cut (after Rich, 2007).
The 1080, 640 600 and 570 level results reflect the use of
prepackaged explosives, whilst those from the 780 and 740 levels
relate to the use of ANFO and ‘snow’ charging.
The results confirmed the observation that development wall
damage appeared to have increased since the use of ANFO and
snow charging was introduced. Rich also undertook a simple
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
THE CHALLENGER GOLD MINE
economic analysis to compare the savings obtained by using the
cheaper explosives against the potential costs if universal
meshing had to be introduced to effectively manage the more
damaged ground. There was a compelling case to revert to the
original blasting practices, and this has since occurred. A
subsequent resurvey of half barrels in development has
confirmed that the initial high quality result in terms of visible
half barrels has been largely regained.
Wider development spans are assessed on a case by case basis
for cable bolting requirements. Situations in which cables are
specified include intersections where an adverse structure may be
present above the backs, and wide ore drives, where cables may be
required to ensure brow stability during stope retreat. Where
possible, stope blasts are designed to fire through ‘difficult’ ground
in a single blast, to avoid having to re-establish stope brows in
adverse conditions such as close to a lamprophyre intrusion.
SUMMARY
From an initial reserve in 2001 of 110 000 oz of contained gold,
the Challenger gold mine produced 467 408 oz up to 30 June
2008. Challenger has been producing a consistent 108 000+ oz
per year since the start of underground stoping in mid-2005
producing 325 600 oz from 1.18 million tonnes at an average
grade of 9.1 g/t in the three financial years to 30 June 2008 at an
average operating cost of $319/oz.
Recent exploration drilling has shown the main mineralised
structure extends well below the current base of planned mining.
An ongoing stress measurement program, coupled with careful
observation of stress-related damage, is being used with numerical
modelling to assess the likely mining environment at depth.
In the medium term, investigations are being undertaken into
ventilation requirements at depth, including geotechnical
investigations for a possible raise bored ventilation shaft.
Planning continues for the extraction of some of the other
resources at Challenger, including the M2 and M3 series, and
the M1 ‘Shadow’ zone, all of which have a similar general
plunge to the main M1 series and can be exploited from
existing infrastructure.
27
P ANDROVIC, P BAMFORD and M SANDY
Extremely good ground conditions and a low to moderate
stress environment have provided generally excellent mining
conditions at Challenger thus far, Industry-best development
drilling and blasting practices are employed to ensure minimal
wall damage. A strong focus on quality and achieving
development schedules has allowed mining to proceed at an
average vertical advance of 150 metres per annum, which is at
the forefront of industry practice.
Strict discrimination utilising all available geological data
including diamond drilling, grade control drilling, face and backs
mapping all contribute to tightly constraining the dynamic
geometry of the Challenger Ore Shoots. With the often rapid
changes in amplitude, wavelength, parasitic folding and plunge
dip and azimuth variations, a geological focus is paramount.
Understanding the folded structural geology, coupled with
high-grade gold drilling intersections away from the lodes
currently being mined, has led to the development of multiple
targets of other similar lodes within reach of the existing
underground infrastructure.
Combined with tight blast design and control it is possible to
mine out nearly all the ore taking only additional planned
dilution to maintain stability. The ore is constrained entirely
within the folded geometry and if this geometry is correctly
delineated then nearly all the metal will be recovered. This is
demonstrated by positive reconciliations with gold produced
from the treatment plant.
REFERENCES
Barton, N, Lien, R and Lunde, J, 1974. Engineering classification of rock
masses for the design of tunnel support, Rock Mechanics, 6:189-236.
Bonwick, C, 1997. Discovery of the Challenger gold deposit – Implications
for future exploration on the Gawler Craton: Case histories of
discovery, in Proceedings New Generation Gold Mines Conference,
Perth, pp 7.1-7.16.
28 Ballarat, Vic, 14 - 15 October 2008
Litterbach, N, 2006. Challenger gold mine rock stress measurement 780
stockpile, Mining Measurement Services Pty Ltd, internal report
Dominion Gold Operations Pty Ltd.
Litterbach, N, 2008. Challenger gold mine rock stress measurement
jumbuck decline 580 external stockpile, Mining Measurement
Services Pty Ltd, DGO internal report.
McFarlane, C R M, 2006. Palaeoproterozoic evolution of the Challenger
Au deposit, South Australia, from monazite geochronology,
J Metamorphic Geol, 24:75-87.
McFarlane, C R M, Mavrogenes, J A and Tomkins, A G, 2007.
Recognising hydrothermal alteration through a granulite facies
metamorphic overprint at the Challenger Au deposit, South
Australia, Chemical Geology, 243:64-89.
Rich, K, 2007. Challenger half barrels comparison, DGO internal report.
Sandy, M P, 2005. Challenger SE zone geotechnical review, AMC
Consultants internal report to Dominion Gold Operations.
Sandy, M P, 2006. Challenger Gold Mine – Geotechnical Review, January
2006, AMC Consultants, DGO internal report.
Scott, B, Machuca, L and Villaescusa, E, 2006. Report on intact rock
properties, Western Australian School of Mines, DGO internal report.
Standing, J, 2003. Update on the nature of the M1 shoot and general
geology of the Challenger pit, Challenger gold mine, South
Australia, DGO internal report.
Standing, J, 2006. Ongoing structural investigations of the M2 Shoot,
Challenger gold mine, South Australia, DGO internal report.
Tomkins, A G, 2002. Evolution of the granulite-hosted Challenger gold
deposit, South Australia: Implications for ore genesis, PhD thesis
(unpublished), Australian National University, Canberra.
Villaescusa, E and Machuca, L, 2008. Stress measurements from oriented
core using the acoustic emission method, Western Australian School
of Mines, DGO internal report.
Narrow Vein Mining Conference
Continuity Risk in Narrow Reef Gold Deposits – Implications for
Evaluation and Exploitation
S C Dominy1,2 and I M Platten3
ABSTRACT
Narrow reef gold deposits are generally less than 4 m in width and
relatively complex geological phenomena. They often display variable
geometry, attitude, continuity and internal architecture. Economic grades
are mostly restricted to specific structural domain(s) and localised within
discrete ore shoots that have a nuggety internal grade distribution. On
account of the often extended history of geological events, the specific
factors localising gold-rich domains are likely to be unique to a particular
mine or even reef. These features lead to a high resource risk and require
careful management during both evaluation and mining. The gross form,
location and global continuity of gold reefs can generally be determined
from surface diamond drilling leading to the definition of Inferred
Mineral Resources. It is likely that for the definition of Indicated and
Measured Mineral Resources underground development and/or close-
spaced diamond drilling will be required. Understanding geometry and
continuity permits the resource to be more effectively managed, lowering
risk/uncertainty. This in turn will allow for well-informed financial
and operational decisions. Most importantly, project expectations and
milestones can be set at achievable levels.
INTRODUCTION
The major challenge during narrow gold reef resource evaluation
and exploitation is determining and managing their inherently
high risk profile. Risk relates to issues of grade and geological
continuity, and in particular variable vein geometry and internal
architecture, erratic grade distribution and generally a high to
extreme nugget effect (Dominy et al, 1997, 1999, 2000a, 2003,
2004).
If the key risks can be understood, then an evaluation program
can be designed to manage them through appropriate strategies.
Most importantly, project expectations and milestones can be set
at realistic levels. The role of geological interpretation is pivotal
in understanding continuity risk and thus managing wider
resource risk.
Many narrow reefs are characterised by coarse gold and a high
nugget effect, where representative sampling and evaluation will
be challenging. The difficulty of evaluating narrow high-nugget
effect systems is well-documented (Dominy et al, 1997, 1999,
2000b, 2001, 2003; Dominy, Stephenson and Annels, 2003).
A key part of any risk control strategy will include the
application of structural geological principals through mapping,
core logging, data interpretation and geological modelling.
Structural geological studies of reef systems are often absent
during evaluation and exploitation programs or where present,
lack rigour or focus. This is despite the fact that the JORC Code
and other reporting codes require geology and hence continuity
to be considered (Sinclair and Vallée, 1994; Dominy and Phelps,
1996; Dominy et al, 1999; Dominy, Platten and Raine, 2003;
1. FAusIMM(CP), Executive Consultant and General Manager (UK),
Snowden Mining Industry Consultants Limited, Abbey House,
Wellington Way, Brooklands Business Park, Weybridge Surrey KT13
0TT, England. Email: sdominy@snowdengroup.com
2. Adjunct Principal Research Fellow, School of Science and
Engineering, University of Ballarat, University Drive, Mount Helen
Vic 3353.
3. Principal Geologist, Snowden Mining Industry Consultants Limited,
Abbey House, Wellington Way, Brooklands Business Park,
Weybridge Surrey KT13 0TT, England.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
Platten and Dominy, 2003; Dominy, 2004a; 2005). The definition
of continuity models can form the basis for understanding
continuity in the context of evaluation (Dominy, Platten and
Raine, 2003; Dominy, 2005).
Geological studies are shown to have a positive impact on
evaluation and mining by lowering uncertainty and risk (Peters,
1993; Cox et al, 1995; Cuffley et al, 1998; McCuaig, Vann and
Seymour, 2000; Baxter and Yates, 2001; Nugus et al, 2003;
Dominy, Platten and Raine, 2003; Tripp and Vearncombe, 2004;
Monteiro, Fyfe and Chemale, 2004; Snowden, 2006b, 2008a).
Many studies are academic and, however, whilst scientifically
rigorous, often lack the focus required to aid evaluation and
exploitation. In particular, they fail to investigate the relationship
between geological and grade continuity.
In some projects the re-interpretation of historical data using
modern 3D modelling techniques has played a major role in
understanding reef geology and grade distribution/controls
(Dominy et al, 2004; Morrison, Storey and Towsey, 2004; Fraser,
Bartlett and Quigley, 2004; Dominy, in press).
Resource evaluation depends upon geological models that
provide a sound, confident expectation that the defined volume is
mineralised throughout. Investigations must yield information as
to the nature of reef geometry, internal architecture, geological
continuity and controls on gold localisation and distribution.
REEF GEOLOGY
Narrow reefs are generally less than 4 m in width and may
comprise one individual vein, or be made up of numerous
individual veins forming a larger ‘lode’ or ‘reef’ structure
(Figure 1). They contain both barren and productive segments
with gold grade varying laterally, vertically and across the body
(eg Platten and Dominy, 2003). Reef formation generally has a
complex and extended history, related to the emplacement of
relatively barren and gold-rich vein elements at different stages.
Although it is a comparatively easy matter to establish the
geological continuity of the global structure during diamond
drilling, the tracing of individual gold-rich veins or domains
(eg local geological continuity) and determining their extent
within the reef is more difficult. This is accentuated when the
gold distribution in the vein is erratic (high-nugget effect), as it
may not be present in cores where the gold-carrying structure is
intersected in a locally barren area.
Based on various field-based studies, there is evidence to show
that different gold grade and particle sizing characteristics are
often related to specific structural (continuity) domains. For
example, studies at Ballarat (Australia) have shown that
fine-gold particle populations are associated with low-grade
(<3 g/t Au) steep west dipping veins, whereas coarse gold is
principally related to flat- to east-dipping veins (>3 g/t Au) that
extend away from the larger west-dipping structures (Dunn,
1929; Olsen, 2005; Osborne, 2008). Similar studies in the
Dolgellau gold-belt (United Kingdom) have shown that high-
grade (>120 g/t Au) coarse gold mineralisation is restricted to
discrete veins within a larger low-grade (<5 g/t Au) mother reef
(Platten and Dominy, 2003). At Bendigo (Australia), Dominy,
Platten and Raine (2003) noted that on the local scale, the
gold-rich veins (>100 g/t Au) are generally narrow and
discontinuous containing clusters of visible gold. These veins
form part of a larger lower grade reef structure (<2 g/t Au).
31
S C DOMINY and I M PLATTEN

grade at various scales. Grade risk is related to grade information

that should be based on quality sampling and assaying of the reef
from either drilling and/or underground development. A number
of continuity risks are associated with the resource evaluation of
gold-quartz reefs (Table 1).

TABLE 1
Key continuity criteria that impact on resource risk for narrow
gold reefs.
Uncertainty criteria Principal resource Comment
risk contribution
Gross geological Tonnage risk Gross continuity
continuity controls the overall
economic mineralised
zone volume.
Reef geometry: dip, Tonnage risk Local variations in reef
strike, width, etc geometry will lead to
local tonnage
uncertainty. Prediction
of effects will depend
upon scale and data
density.
Grade distribution Grade risk Gross localisation of
control: small-scale gold within shoots
geological continuity controls grade risk. The
and ore shoot controls scale of the shoots will
determine ease of
definition.

Resource risk includes additional factors such as mineralogy,

FIG 1 - Examples of narrow gold reef structures. (A) Main Vein,
440 west level in the Nalunaq gold mine, Southern Greenland.
High-grade (>20 g/t Au) laminated quartz vein showing relatively
continuous wallrock layers (extending >1 m) and discontinuous
layers traceable for a few centimetres. Field of view 2 m (source:
Snowden, 2006b). (B) Cononish Vein, exploration adit, Cononish
gold mine, Scotland, UK. Massive low-grade (1 - 3 g/t Au) white
quartz vein bearing a high-grade (>20 g/t Au) shear zone with
sulfides. Field of view 3 m (source: Snowden, 2008a).
There are three levels of complexity involved in localisation of
gold in reef structures. Firstly, the internal distribution in the
immediate host structure may be controlled by textural
development of the major reef filling minerals to give an
irregular distribution of gold. Secondly, the primary shape and
dimensions of the host reef are controlled mostly by slip
directions and initial fracture shape. Fracture pattern and
deformation style may be controlled in part by the behaviour of
the rest of the reef and its host rock. Finally, the variable
relationships of the gold-rich vein(s) to whole reef, particularly
the effect of later deformation, minor igneous intrusions and
veins on its position and continuity relative to the gross reef
envelope.
CONTINUITY RISK
There are two significant risks related to resource evaluation: risk
on tonnage (‘geological risk’) and risk on grade (‘grade risk’).
Both geological and grade continuity (and variability) have an
obvious impact on both.
Grade risk is usually greater than geological risk, though the
effect of the latter should not be understated. Geological risk is
related to the identification of economic volumes from both
geological and grade data (ie drilling and/or underground
development), and must consider continuity of both geology and
sampling and assaying quality, survey data, data density, bulk
density, estimation methods, etc which are out of the scope of
this discourse (Dominy, Noppé and Annels, 2004).
The application of matrices to communicate aspects of
resource risk is common (Annels, 1996; McCuaig, Vann and
Seymour, 2000; Annels and Dominy, 2002; Dominy, Noppé and
Annels, 2004). This contribution presents a simple six score
classification for continuity risk based on the criteria given
in Table 1. The system uses numerical scores which are defined
in Table 2. Examples of the classification are given in the case
studies.
The more continuous and uniform the mineralisation, the
fewer the number of data points (eg drill holes and samples)
needed for a given level of confidence and acceptable risk.
Continuity risk can be directly linked to resource classifications
using international reporting codes (eg JORC, 2004). The JORC
Code defines the confirmation of grade and geological continuity
for Measured Mineral Resources, through to assumed geological
(and/or grade) continuity for Inferred Mineral Resources
(Table 3). The JORC Code and other reporting codes are
important risk communication tools.
CASE STUDIES
This contribution presents two case studies representing project
evaluation (Clogau mine, North Wales, United Kingdom) and
production (Nalunaq mine, Southern Greenland). Each presents
how geological studies have led to a better understanding of the
nature of continuity risk and show how it is being used to
manage resource/reserve risk.
Clogau gold mine
Introduction
The Clogau gold mine is situated within the Dolgellau gold-belt
of North Wales (Figure 2). With a recorded total production of
over 100 000 oz Au, Clogau is Britain’s largest producing mine.

32 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 CONTINUITY RISK IN NARROW REEF GOLD DEPOSITS – IMPLICATIONS FOR EVALUATION AND EXPLOITATION

TABLE 2
Definition of the continuity risk score and its practical meaning. Note that precision figures are given for indicative purposes only and do not relate
to any specific resource classification or estimation methodology. Precision will depend upon geology, nugget effect and data density, etc.

Risk score Implied meaning Practical meaning for resource

1-2 Little or no perceived risk Such a score is unlikely for a narrow vein gold reef unless it is characterised by a low nugget effect and
(Low risk) (low uncertainty) simple geology and geometry. Indicated precision on tonnes and grade less than ±10%.
3-4 Some risk that could lead to Indicates a moderate risk profile related to grade and/or geological variability. Drilling alone may be
(Moderate risk) material error in the resource able to define Ore Reserves, though underground development may be required. Indicated precision
model on grade and tonnes in the range ±10% to ±20%.
5-6 Strong risk that could lead to Indicates a high level of grade and/or geological variability. Likely to result in uncertainly in both
(High risk) material error in the resource grade and tonnes. Possible that Ore Reserves will be difficult to define in large quantities and will be
model (high uncertainty) driven by underground development and a ‘blocking out’ approach. Indicated precision on grade and
tonnes in the range ±20% to ±40%, or possibly higher.

TABLE 3
JORC (2004) Code requirements for geological and grade continuity in the definition of Inferred, Indicated and Measured Mineral
Resources.
Inferred Mineral Resource Indicated Mineral Resource Measured Mineral Resource
It is inferred (the resource) from geological The locations (drill hole intersections, etc) are too The locations (drill hole intersections, etc) are
evidence and assumed but not verified geological widely or inappropriately spaced to confirm spaced closely enough to confirm geological and
and/or grade continuity geological and/or grade continuity but are spaced grade continuity
closely enough for continuity to be assumed

FIG 2 - Geological summary and location map of the Dolgellau gold belt (from Platten and Dominy, 1999).

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 33

S C DOMINY and I M PLATTEN
Production was first recorded around 1840 and the main mines of
the belt (Clogau, Gwynfynydd and Cefn Coch mines) have been
worked sporadically since then (Hall, 1988). Clogau operated
during 1982 to 1987 and 1992 to 1998, and was again explored
during 2000 and 2006 though no mining took place.
A Mines Royal Gold Licence (exploration) has recently been
acquired (September 2008) for a 120 km2 area of the gold-belt,
including the Clogau mine by Gold Mines of Wales Limited
(‘GMWL’), a subsidiary of the privately owned Victorian Gold
Limited. GMWL have commenced a thorough revaluation of the
gold-belt, including the Clogau mine. There are no mineral
resources defined at Clogau.
A unique feature of the Dolgellau gold-belt is that the gold
produced has traditionally been used to produce Welsh Gold
jewellery, which can be sold at a premium above the gold price.
The product and its premium are related to the notion of ‘Royal
Gold’ and the use of Welsh Gold by the British Royal Family. In
Great Britain, the rights to all precious metals are vested in The
Crown.
Dolgellau gold-belt mineralisation shows many broad
similarities to other ‘slate-belt’ hosted deposits such as those in
Australia (Central Victorian Goldfields: Cox et al, 1995;
Cuffley et al, 1998; Turnbull and McDermott, 1998; Shaubs and
Wilson, 2002; Phillips and Hughes, 2003), Canada (Nova Scotia
and Meguma: Keppie, Boyle and Haynes, 1986) and China
(Guizhou: Lu et al, 2005).
This case study is based upon a recent review of the gold-belt
(Snowden, 2008b) and extensive previous work by the authors
(Dominy, 1996; Platten and Dominy, 1999; Dominy and
Platten, 1999; Platten and Dominy, 2003).
Mine area geology
The Dolgellau gold-belt occurs within the northern part of the
Welsh Lower Palaeozoic basin (Allen and Jackson, 1985).
Subsidence and filling of this basin took place in three stages,
represented by the Dyfed, Gwynedd and Powys Supergroups.
Basin evolution ended with an Acadian (late Caledonian)
tectonic event, which resulted in folding and development of the
regional cleavage.
The Dyfed Supergroup begins with a volcanic sequence of
which the Bryn-Teg Formation, the inferred source of gold
mineralisation, is a part. Thick, coarse turbidite sands, the
Harlech Grits Group and the following siltstone-mudstone-
dominated Mawddach Group constitute the main, 4 km thick,
part of the Dyfed Supergroup (Figure 3). These host the quartz
reefs with the main exploitable gold accumulations associated
with the carbonaceous mudstones of the Clogau Formation (90 m
FIG 3 - Stratigraphic sequence of rocks in the Dolgellau area (from Howells, 2007). The basal Maentwrog Formation and entire
Clogau Formation are key exploration targets for gold (box shown).
34 Ballarat, Vic, 14 - 15 October 2008
to 105 m thick) at the top of the Harlech Group (Figure 3).
The Clogau shales consist of dark-grey or black-banded
carbonaceous mudstones and silty mudstone, with rare quartoze
silty laminae and even less common beds of fine sandstone. The
deposition of the Dyfed Supergoup closed with emplacement of
the andesitic Rhobell Fawr Volcanic Centre.
The region contains Doleritic sills, locally called
‘greenstones’, attributed to the Rhobell Fawr volcanic event.
They are principally conformable sheet-like bodies, which pinch
and swell and sometimes split and rejoin. Their thickness can
reach a maximum of about 25 m and can be traced on surface for
up to 2 km. At Clogau, nine main sills are identified as four
separate groups. In addition, dykes of intermediate composition
are also found (known as Clogau Stone dykes). These dykes
follow the gold veins, occurring within or at the side of the vein
network.
The Clogau mine is hosted principally within the Clogau
Formation and contains greenstone sills, which increase its
thickness to approximately 120 m (Figure 3). The Clogau
Formation at the mine dips about 30° SE, and is cut by a 70° SE
cleavage (Figure 2).
Mineralisation
The Dolgellau gold reefs form an east-northeast to
west-southwest trending array of generally steeply dipping veins
(Figure 2). The veins are internally complex and are composed
of swarms of narrow (1 mm to 300 mm) subparallel veins
separated by country rock sheets giving a total width of up
to 6 m. The reef forming gold-quartz veins are relatively
discontinuous on a scale of tens of metres or less. The main
gold-bearing reef at Clogau can be traced for about 3 km,
striking approximately northeast-southwest dipping steeply
between 60° and 90° SE (Figure 4).
The reef system pinches and swells, changes strike and splits
into a number of smaller offshoots and branches (Figures 5 and 6).
Reef widths are variable, ranging from over 6 m to a few
centimetres, and in places reducing to small irregular quartz
blobs or veinlets.
Veining is complex and characterised by vein networks formed
by successive fracture re-activation. Three quartz vein events are
noted (Figure 7): early non-auriferous quartz-carbonate veins
(EQV); main gold-quartz veins (GQV); and late barren quartz
veins (LQV). Their mineralogy is characterised by specific
paragenesis: early pyrrhotite-pyrite-arsenopyrite-cobaltite (EQV);
chalcopyrite-pyrrhotite (GQV): and telluride-gold-galena (GQV/
high-grade pocket domains). The LQV comprise bucky and fibre
textured white quartz with minor pyrite.
Narrow Vein Mining Conference
 CONTINUITY RISK IN NARROW REEF GOLD DEPOSITS: IMPLICATIONS FOR EVALUATION AND EXPLOITATION

FIG 4 - Schematic map of Clogau mine local geology. Gross trend of reefs (lodes) northeast-southwest.

FIG 5 - Clogau Main Reef geological map from a section of the Tyn-y-Cornel Level at the No 1 Shaft. The reef shows pinch and swell structure
and splits into the upper John Hughes Reef. Gross reef trend northeast-southwest. Key: dark grey – quartz reef; light grey – Clogau Stone dyke;
hatching – Clogau Formation; stipple – greenstone. Scale: strike length of drive 70 m (source: mapping by S C Dominy, 1996).

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 35

S C DOMINY and I M PLATTEN

FIG 6 - (A) Clogau main reef geological map from a section of the Cornel level close to No 2 shaft. The reef shows pinch and swell structure
and splits (east facing) into a minor branch. Beyond the No 2 shaft this branch ultimately becomes the major South or Paraffin reef. Scale:
strike length of drive 45 m. (B) Detail of the split section of the Main-South reef with local occurrence of high-gold grades in sublevel 4 m
below Cornel level. Gross reef trend northeast-southwest. Key: dark grey – quartz reef; light grey – Clogau Stone dyke; hatching – Clogau
Formation; stipple – greenstone. Scale: strike length of drive 15 m (source: mapping by Dominy, 1996).

The majority of the gold occurs as free particles with very
little gold intergrown with sulfides. In some sections of the mine,
gold is seen to occur in close spatial association with
concentrations of bismuth sulfides (eg bismuthinite) and
tellurides (eg tellurobismuthite and tetradymite), and sometimes
galena. Where chalcopyrite, pyrite and pyrrhotite concentrations
are high, they show a negative relationship to gold occurrence.
Gold generally has a fineness of >750 and is seen as particles
up to 5 mm in size. Historically larger masses of gold on the tens
of centimetres scale have been reported (Hall, 1988). Within
high-grade pockets over 75 per cent of the gold reports to the
+300 µm fraction and over 85 per cent to the +150 µm fraction,
and less than five per cent reports to the -50 µm fraction (Dominy
et al, 2000a). Approximately 85 per cent of the gold is generally
gravity recoverable.
The wider quartz reef zones comprise a series of thin and
laterally discontinuous GQV. Inclined, cross-cutting barren LQV
is abundant in the GQV. The LQV are lenticular in three
dimensions, tapering and vanishing up and down dip and across
their strike. LQV dimensions vary considerably, veins ranging
from 8 m to 10 mm in length and from 1 mm to 1 m thick.
The north-northwest – south-southeast dimension is essentially

36 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 CONTINUITY RISK IN NARROW REEF GOLD DEPOSITS: IMPLICATIONS FOR EVALUATION AND EXPLOITATION
FIG 7 - Clogau Main Reef in the Williams Stope, Cornel Level.
(A) Laminated gold veins (GQV) in a stope pillar cut by late barren
quartz veins (LQV), width of view 1.75 m. (B) Laminated gold veins
(GQV) in the stope backs effectively replaced by late barren quartz
veins (LQV), width of view 1 m (source: Dominy, 1996).
controlled by the host GQV width. Very large veins occur in
thick massive sections of the GQV or penetrate greenstone
wallrocks. The smallest are found where the GQV is composed
of well-spaced very narrow (<10 mm) veins.
Mineralisation at Clogau is related to specific events which
controlled the distribution of gold and its associated minerals. For
example, the John Hughes Reef trends about 30° to the gross trend
of the Main Reef, but is parallel to some local segments in the
Main Reef (Figure 5). The Main Reef gold was associated with
tellurides, whereas in the John Hughes Reef gold was related to
galena and very minor tellurides. This mineralogical difference
must indicate that the Main Reef was inactive at the time of John
Hughes Reef activity. This is supported by field evidence based on
the fact that the widest part of the John Hughes Reef (1 m) is
parallel to the thinnest sparsely-mineralised part of the Main Reef.
Beyond this 50 m long section, the John Hughes Reef thins to the
east and west, where the Main Reef regains its width of between
0.5 m and 2 m. Time relations between the John Hughes and Main
Reefs are unknown.
Figure 8 summarises the geology of an idealised reef structure
and includes normal movement across the fault zone and the
intrusion of Clogau Stone dykes.
Regional structural history
Recent studies at Clogau and elsewhere have determined the
general sequence of geological events within the Dolgellau gold-
belt (Dominy and Platten, 1999; Platten and Dominy, 1999):
• Deposition of Cambrian host sediments.
• Emplacement of greenstone sills.
• Normal faulting on east-northeast to west-southwest trends.
• Ribbon vein emplacement along fault planes forming EQV
and GQV.
• Emplacement of north-east to south-west dyke swarm.
• Fracture of GQV and dykes with initiation of formation of
LQV. East-dipping LQV veins first, west-dipping veins
second. Cataclastic internal deformation of the GQV. Onset
of main metamorphic stage.
• Continued deformation, producing cleavage and deforming
all pre-existing rocks. Metamorphic recrystallisation generates
strain-free quartz in the GQV.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
FIG 8 - Schematic showing an idealised section through a Clogau
gold reef structure and relationship between the various geological
elements.
• Post-cleavage, post LQV events such as the calcite-marcasite
veins and the regional Trawsfynydd Fault Zone.
Ore shoot/mineralisation controls
Introduction
Two key controls to gold mineralisation and ore shoot formation
have been identified at Clogau; lithogeochemical and structural.
The lithogeochemical control relates to the interaction between
gold-bearing fluids and the Clogau Formation shales. The
structural controls combine the effects of vein-hosting fracture
refraction by greenstone sills and vein splitting/branching
(Dominy, Phelps and Camm, 1996).
Lithogeochemical control of the Clogau Formation
Gold mineralisation is exclusively contained within the quartz
reefs where they are hosted within dark graphitic shales. The
main host is the Clogau Formation, but some dark shales occur
within the uppermost Gamlan and basal Maentwrog Formations.
This gives a total prospective thickness of approximately 210 m
(Figure 3). This control has been shown to be related to the
interaction between the carbonaceous shales and gold-bearing
fluids. A fluid-inclusion study on GQV material indicates the
presence of methane, where reduction of the hydrothermal fluids
by methane is considered to be the main gold-precipitation
mechanism (Shepherd, Bottrell and Miller, 1991; Bottrell et al,
1988; Bottrell and Spiro, 1988). The methane is considered to
have been generated by the reaction of water with the graphite in
the Clogau Formation.
Greenstone sills
A critical control to ore shoot localisation at Clogau is related to
the mechanical effects of the greenstone sills on the reef – a
37
S C DOMINY and I M PLATTEN

feature common across the Dolgellau gold-belt (Hall, 1988;
Dominy, Phelps and Camm, 1996). The sills were intruded into
the Clogau Formation before reef formation. They are competent
intrusive rocks up to about 25 m in thickness and fracture
differently to the enclosing shales. The sills are offset across the
plane of the reef with an apparent down-throw to the southeast
(Figure 8). The sill/reef intersection pitches at a low to moderate
angle south-west on the vein. The reef-hosting fractures are
refracted strongly when they reach a sill. In the shales, the reef
maintains an approximate dip of 70° SE, whereas in the sill
refraction zone has a dip of closer to 50° NW. Ore shoots
pitching gently south-west are spatially associated with sill
intersections (Figures 9 and 10). They lie entirely below the sill
intersection on the north-west side of the structure, but extend
across the truncated surface of the sill intersection on the
south-east wall (the gross hanging wall). This may indicate flow
localisation related to local widening and closure of fractures as
a result of competency contrast between greenstone and shale as
well as refraction. Details are not known at present.
The presence of gold mineralisation below greenstone sills is
apparent in the John Hughes Lode (Figure 10b). However, an
important conclusion from recent mapping is that this control
was regarded by earlier miners as the most important, and has
guided mining of the shallow plunging ore shoots. The old stopes
pitch at about the same angle as the bedding against the plane of
the reef, and occur where the greenstone sills can be projected
onto the workings. It appears that as well as at John Hughes,
most of the early workings at Clogau were related to
intersections of the reefs with the greenstone sills (Figure 10a).
Some stopes are also seen at several localities above the sills.
These indicate that sill control is more complex and requires
further investigation.
Reef splits/branches
Significant splits (or branches) along the reefs have long been
known to carry gold mineralisation (Figures 5 and 6). Splits
provide steep-pitching conduits for ascending mineralising

FIG 9 - Schematic long section showing typical position of the ore shoot zone below greenstone bodies at Clogau.

FIG 10 - (A) Main and (B) John Hughes Reefs at Clogau. The ore shoots are located at the base of the greenstone sills cut by the reef
structures. High-grade pockets of gold are located within the shoots. (A) Strike length 275 m and (B) strike length 110 m.

38 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 CONTINUITY RISK IN NARROW REEF GOLD DEPOSITS: IMPLICATIONS FOR EVALUATION AND EXPLOITATION
fluids, resulting in steep-pitching ore shoots along the axes of the
splits. Due to the variable nature of GQV dip and the pitch
direction of the split axis at different localities, north-east and
south-west facing splits with different pitch axes exhibit different
styles of mineralisation.
The geometry of mineralised splits is more closely established
by recent mining in the Llechfraith Level. Activities during
1984 to 1987 followed the Llechfraith-No 1 shoot down-dip
(Figure 11). The shoot plunges 65° SW, is approximately 10 m
along strike and can be traced 40 m down-dip from surface to the
No 4 level below. The shoot is related to a split in the reef with
high-grades within the hinge zone. Two further parallel shoots
were established, but were less important.
The reef splits are not uniformly mineralised. Historical
records suggest that in east facing splits the mineralisation was in
the minor vein away from the split, while in west facing splits it
was in the main vein near to the split (hinge) zone. Recent
studies have shown that the situation is not simple. Open space
development is important for split-based mineralisation to occur
and two favourable sites are indicated: west facing split with a
west pitching junction and east facing split with an east pitching
junction. Other configurations result in good open space
development on the main reef rather than on the split branch
which are generally poorly mineralised. Reef splits can be
complex and may change plunge with depth or close and
disappear. In addition, their proximity to north-south faults
appears to enhance gold potential. Further work is required to
understand the nature of reef splits.
Gold grade distribution in ore shoots
High-grade gold pockets are located within ore shoots of 10 m to
25 m along-strike and >50 m down plunge. Sill/reef intersection
shoots plunge between 20° and 45° SW and reef split shoots
plunge between 45° and 80° SW. Historical records showed that
the most productive ore shoots contained gold-rich pockets in
approximately five per cent to ten per cent of the total shoot
volume. Three gold domains are recognised in the Clogau reefs:
1. barren: reef outside the ore shoot <0.5 g/t Au,
2. low/medium-grade: ore shoot – background grades >0.5
- 60 g/t Au, and
3. high/extreme-grades: ore shoot – pocket grades >200 g/t Au.
FIG 11 - Long section (northeast-southwest) of the Main Reef Llechfraith section and position of three ore shoots related to reef branching.
Particularly high-grade gold mineralisation was located between the No 3 and No 4 sublevels on the No 1 shoot. As branch/split based
shoots, they have a steeper plunge (70° SE) than is usual with the greenstone-hosted shoots (30° to 40° SE).
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
Gold is extremely erratic with high-grades (100s oz/t Au)
within tens of millimetres of very low-grades (<0.1 g/t Au). In
general, the extreme high-grade areas are relatively continuous
over very small areas. Hall (1988) reports a volumetrically small,
high-grade GQV (0.23 m by 1.8 m by 1.4 m) as part of a bigger
structure at Clogau mined in 1867. This GQV yielded 16 kg of
gold from about two tonnes of rock – a grade of ~260 oz/t Au.
Similar pockets are also recorded at the Gwynfynydd mine
(Platten and Dominy, 2003).
The average grade of ore processed at Clogau was historically
17 g/t Au, though this is misleading given that it reflects different
mining strategies (ie highly selective versus non-selective).
Historical and recent mining indicates that selective mining of
shoots to maximise gold pocket extraction is likely to yield
grades around 6 oz/t Au to 10 oz/t Au. Ore shoots generally
posses gold-rich pockets bearing bonanza grades of between
10s to 100s oz/t Au. The ore shoots have a moderate continuity
(tens of metres down plunge and <10 m along strike) once
identified, but the bonanza pockets are less easy to determine
without mining as they tend to be on a scale of <5 m.
Reef modifying features
Post-ore faulting resulted in normal movement downwards to the
southeast across the GQV. This resulted in portions of the GQV
being downthrown by up to 10 m effectively diluting grades
within ore shoots.
Open folding of the Clogau Formation has had little effect on
the localisation of ore deposition; however, disharmonic folding
has in some areas caused additional irregularities in the reefs,
with consequent mining complications.
Where the LQV are emplaced into areas of pre-existing GQV,
they frequently had a deleterious effect, and locally truncating,
dislocating and sometimes even folding the GQV.
Implications for evaluation and exploitation
Clogau continuity risk profile
Clogau mineralisation is high risk due to an extreme nugget
effect which is related to the localised nature of the ore shoot-
hosted gold pockets (Table 4). Resolution of grade continuity is
difficult due the relatively small-scale of the ore shoots and
pockets. There is thus a need for geologically controlled
underground development.
39
S C DOMINY and I M PLATTEN

TABLE 4
Clogau continuity risk review and associated project expectations.

Uncertainty criteria Risk score Comment

Gross geological continuity
Reef geometry: dip, strike, width, etc
Grade distribution control: ore shoot
controls
Grade distribution: local spatial
effects and small-scale continuity
issues within the ore zone/ore shoots –
‘nugget effect’
Post-reef effects of faulting, dykes, etc
Total risk profile
Project expectations
4 The Clogau Main Reef is relatively continuous, but shows some local variability with pinchouts.
Gross continuity is generally not material to ore shoot definition. Wide-spaced drilling at 100 m
will resolve gross reef and greenstone continuity.
5 (6) Locally reefs can show strong variability that requires underground development or close-spaced
drilling (~10 m by 10 m) to resolve. Locally material to evaluation.
6 Ore shoots are controlled by various structural features (eg vein splits and greenstones) which
show generally good down-plunge continuity. Requires underground development or
close-spaced drilling (~10 m) to resolve.
6 Ore shoots have a very high nugget effect, with gold grade very difficult to predict on a shoot
scale. Requires underground development or very close-spaced drilling (~5 m) to resolve
effectively. Provides the greatest risk at Clogau.
4 (5) Cross faulting and LQV can have a material effect on the continuity of ore shoots. Requires
underground development or very close-spaced drilling (~5 m) to resolve.
6 The overall continuity risk is considered to be high. This is principally driven by the very
high-nugget effect and pockety nature of the very high grades.
• Small system with total gold potential around 150 000 oz to 200 000 oz
• Boutique operation, highly selective stoping to produce a very high-grade
low-tonnage ore
• Continuity risk leads to high resource risk
• Only very small tonnages of resources likely to be defined (Inferred Mineral
Resource category only)

Evaluation grade could be stated as a ‘grade range’ based on different

The primary targets are quartz reefs within the Clogau
Formation. The region contains a number of phases of quartz
veining, many of which are non-auriferous. Platten and Dominy
(1999) and Dominy and Platten (1999) present the case for the
auriferous reefs being precleavage, which provides a primary
discriminator for gold potential.
Evaluation targeting will be based on the identification of
greenstone sills from surface and underground mapping. Recent
mapping has resulted in the identification of a number of targets
with good potential close to the existing workings. Definition
requires underground diamond drilling to determine local
structure and stratigraphy in more detail, and to focus further
drilling and/or underground development. Previous operators
have undertaken little diamond drilling, and the authors believe
that strong definition of structure and stratigraphy will be critical
to project development.
Drilling will rarely resolve actual ore shoots effectively unless
the drill spacing is prohibitively small (potentially <5 m). A
series of 50 m sections drilled from existing workings will
augment surface mapping and the south-east-trending Tyn-y-
Cornel cross-cut to resolve local structure and stratigraphy.
Resource evaluation in the traditional sense in Clogau will be
difficult, as is typical of all mines of the Dolgellau gold-belt.
Identification of ore shoots, hence resolution of small-scale
geological and grade continuity requires underground
development. This development can be targeted based on
drilling. Once an ore shoot is identified by development it is
likely to be possible to extrapolate its extent based on mapping
and drilling. The challenge is to define grade, based on the
highly pockety nature of the gold. Historical records suggest that
as little as five per cent to ten per cent of an ore shoot will be
very high-grade material.
Given the high geological uncertainty and risk, it is likely that
only an Inferred Mineral Resource (JORC, 2004) could be
defined even after development. An identified ore shoot can be
considered to be a high ‘probability’ zone for which a
gold-bearing tonnage can be estimated. Further evaluation of
historical data may allow a payability factor(s) to be defined and
probabilistic outcomes (eg low-, medium- and high-grade cases).
The reporting of grades within ranges has previously been
undertaken to provide a better measure of possible resource
outcomes (Dominy, 2004b, 2006).
Local grade may be defined by processing small parcels of
ore (eg 2 t to 5 t bulk samples), though these may either over or
underestimate grade depending proximity to gold pockets. All
available information should be combined to estimate the
proportion of high-grade pocket ore versus lower grade material,
and likely grade of each. Such an approach is based on a strong
knowledge of local geological and grade continuity. In addition,
it may be possible to gain further confidence in the ore shoot
based on the application of suitable proxies for gold (eg quartz
texture, geochemistry and mineralogy: Dominy and Johansen,
2004).
Exploitation/mining
Reef geometry and grade distribution control the nature of
exploitation. A selective mining operation will develop through
the ore shoot and focus on extracting high-grade pockets.
However, to some extent mining of the entire ore shoot zone is
more likely to ensure pockets are located. High-grade material
can be selectively extracted and milled during a focused ‘bulk’
mine operation. Mining of all reef material, inside and out of ore
shoots, would not be economic. Any mining operation at Clogau
is likely to extract between 1500 t and 3000 t per annum
maximum, and process 200 t to 400 t of high-grade ore.
The use of gold for a premium jewellery product allows for a
low cut-off grade and viability of relatively lower grade material.
Large mineral resources are not required for project viability due
to the small-scale and highly selective nature of the operation.
A number of geological features result in the modification of
the GQV and can hence lead to complexities during mining.
Normal faulting results in portions of the reef being downthrown
by up to 10 m to the south-east, effectively dislocating the shoot
and diluting grade locally. Folding of the Clogau Formation
appears to have minimal effect on gold localisation, but has in
some areas caused additional irregularities in the reef with

40 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 CONTINUITY RISK IN NARROW REEF GOLD DEPOSITS: IMPLICATIONS FOR EVALUATION AND EXPLOITATION

consequent mining complications. Where the LQV are emplaced
into areas of pre-existing GQV, they have a deleterious effect,
locally diluting, truncating, dislocating and sometimes even
folding the GQV. A selective extraction process such as reef
open stoping or some form of modified shrink stoping will
provide the flexibility to deal with these geological complexities
(Dominy and Phelps, 1996: Dominy et al, 1998).
Nalunaq gold mine
Introduction
The Nalunaq gold mine is situated in southernmost Greenland,
some 86 km north-west of Kap Farvel (Figure 12). It lies on the
south face of Nalunaq Mountain in Kirkspirdalen between the
Sarqa and Tasermiut Fjords. The nearest settlement is Nanortalik,
30 km to the south-west. The mine is operated by Nalunaq Gold
Mine A/S, a 100 per cent owned subsidiary of Crew Gold
Corporation.
between 1993 and 1996. Drilling from surface is difficult due to
the steep face of Nalunaq Mountain and the cover of glacial
deposits on lower ground. Surface samples were taken in 1997,
followed by underground development in 1998 and further
surface drilling. Extensive underground exploration development
and bulk sampling began in 2000, with production commencing
in 2004 (Dominy and Petersen, 2005).
The underground mine extracts approximately 180 000 t per
annum yielding 80 000 oz to 90 000 oz gold. Operations are
based on an Indicated Mineral Resource of 535 000 t (18 g/t Au)
and an Inferred Mineral Resource of 1.5 Mt (17 g/t Au) for a
total of 1.1 Moz gold reported at a zero g/t Au cut-off (Crew,
2007).
This case study stems from desk and site-based work
undertaken by the authors during 2005 and 2006 (Dominy and
Petersen, 2005; Dominy, 2006; Dominy et al, 2006; Snowden,
2005a, 2005b, 2006a, 2006b).
General geology
The deposit is situated within the early Proterozoic (1850 Ma
to 1750 Ma) Ketilidian Orogenic Belt in South Greenland
(Figure 12). It occurs within the Psammite Zone close to the NW
boundary against the Julianehab Batholith Zone (Chadwick and
Garde, 1996). The Julianehab Batholith Zone is considered to be
the plutonic root of an island arc accreted on to South Greenland
and the Psammite Zone represents an intra arc basin filled with
detritus from the island arc (Chadwick and Garde, 1996). Late
tectonic granitoids and post-tectonic rapakivi granites cut the
Psammite Zone metasediments.
The Nalunaq site lies in a metavolcanic thrust sheet resting on
gently dipping meta-arkoses of the highest tectonic unit in the
Psammite Zone (Windley, 1991). The rocks show amphibolite
facies metamorphism and limited deformation. The metavolcanics
are underlain and intruded by a late tectonic leucocratic granitoid
pluton which is associated with a network of aplite sheets in the
metavolcanics.

FIG 12 - Location and regional geology of Nalunaq (simplified
from Chadwick and Garde, 1996).
The deposit was discovered in 1992 by regional geochemical
surveys (Gowen et al, 1993; Stendal et al, 1995) and was
followed by a program of diamond drilling and surface mapping
Gross form of the deposit
The gross vein structure (Gowen et al, 1993; Kaltoft, Schlatter
and Kludt, 2000; Dominy et al, 2006) is a sheet with
northeast-southwest strike and gross dip 35° SE (range 20° to
55° SE) (Figure 13). The total vein thickness ranges from sub
1 cm up to 3 m, with a mean of about 0.7 m. The vein can be

FIG 13 - Mine plan showing the 300 m, 350 m, 400 m and 450 m initial exploration drives and subsequent development at approximately
10 m vertical intervals. Illustrating the relatively simple gross structure of the vein within the active mine area. The Target Block and South
Block are separated by the Pegmatite Fault, which interrupts gross continuity with an apparent down throw to the north-west. Mine grid north
is to 315° true North (source: Crew Gold Corporation).

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 41

S C DOMINY and I M PLATTEN

traced in outcrop for 1.4 km along the north side of Nalunaq

Mountain and down the west side of Kirkspirdalen. Drill holes
show the concealed vein extends a further 400 m east.
Exploration drives and drill holes prove vein continuity for at
least 600 m along strike in the mine area.
Gross geological continuity is good within the explored
ground. Only one late fault, the Pegmatite Fault, has sufficient
throw to affect mine layout (Figure 13).
The vein is emplaced in fine amphibolites and medium to
coarse amphibolites interpreted as lavas and sills respectively.
Minor sediments and sulfide-rich layers are intercalated in the
sequence. The gross vein dips at a slightly lower angle than the
host succession and cuts across the lavas and sills. Local
segments run parallel to layering, following the minor sediment
layers.
Based on surface and underground geological mapping and
drill data, it has been indicated that the high-grade sections of the
vein appeared to be hosted within medium-grained metadolerite FIG 14 - Nalunaq gold mine, 440 west level. Quartz vein with
predominantly massive quartz shows rapid variation in thickness.
sills or were located proximal to the metadolerite/metapillow
Moderately inclined aplite sheets with dilation vector at high angle
basalt contact. The sills appear to intersect the veins along a line
to steeper sections (seen near floor). These are converging with
that plunges 30° ENE, and to define a linear trend of plunging
the vein up dip from the drive. Vein will be extended at intersection
ore shoots.
(source: Snowden, 2006b).

Primary vein structure

The vein (Figures 1 and 14 to 17) commonly occurs as an array of
quartz sheets (0.01 m to 1 m thick) that can be in direct contact
with each other, separated by discontinuous, thin wall rock screens
or separated by wide (0.1 m to 2.0 m) screens of wall rock.
Mapping scales for the levels and raises do not permit the thinner
wall rock screens to be shown. Locally quartz veins diverge by
more than 2 m (Kaltoft, Schlatter and Kludt, 2000). A single
quartz sheet may occur or the vein material may be locally absent.
Veins show local tapered terminations that may be primary.
The pattern changes laterally as screens and component veins
change thickness along the vein (Figure 15). Veins may diverge
as screens thicken towards terminations. In packages of veins the
narrow veins may be less continuous than the large veins.
En-echelon patterns are also seen.
Wall rocks show local alteration adjacent to the vein, usually
within 1 m (Figure 16). Amphibolites are replaced by pyroxene-
bearing, equigranular textured, skarn materials. Foliated biotite-
rich rocks may represent metasediments.
Vein mineralogy
The vein mineralogy is simple (Kaltoft, Schlatter and Kludt,
2000; Grammatikopoulos et al, 2004), being mostly granular
coarse quartz. Trace to minor amounts of lollingite, arsenopyrite,
pyrite, pyrrhotite, chalcopyrite and bismuth sulfosalts occur in
some vein elements. Native gold occurs as fine (<100 µm) and
coarse (>100 µm) particles. In run-of-mine ore most gold
particles, about 50 per cent are >100 µm in size (Kaltoft, Schlatter
and Kludt, 2000; Grammatikopoulos et al, 2004; Dominy and
Petersen, 2005). Significant amounts of visible gold occur and
locally form very high-grade clusters (Dominy and Platten,
2007). There is a significant coarse gold problem, which gives
rise to sampling issues (Dominy and Petersen, 2005; Dominy et
al, 2006; Dominy, Xie and Platten, 2008). Visible gold is
commonly restricted to one or two individual veins at any one
site, the rest appearing barren.

FIG 15 - Vein section in raise, showing gross planar form but discontinuous and complex in detail. Note the change in total thickness (quartz
plus wallrock screens) in dip and strike direction the matching 2.5 m to 3 m gap in the vein array with a small en-echelon offset pitching
gently south-east on the vein stacked arrays of narrow veins with intervening wallrock screens splitting of veins in the up dip part (source:
redrawn from Crew, 2002).

42 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 CONTINUITY RISK IN NARROW REEF GOLD DEPOSITS: IMPLICATIONS FOR EVALUATION AND EXPLOITATION

FIG 16 - Vein section in raise, showing the general planar form of a simple, continuous vein segment. Local gaps and offsets occur along the raise
and differences in detail occur between the opposite walls of the raise. The vein thickness is mostly between 0.1 m and 0.2 m, locally 0.4 m,
significantly less than minimum stope width of 1.2 m. The relationship between vein and alteration envelope is shown. Note that internal
boundaries within the quartz veins are not resolved at the scale of mapping (source: redrawn from Crew, 2002).

FIG 17 - Vein (0.2 m to 0.5 m thick) undulation seen in raise walls. The axis (AL, AR) pitches gently south-west and the up dip vein
termination pitches gently south-east on the plane of the vein. The outline of a model 15 m by 1.2 m stope is superimposed and the bar
shows width of more practical 1.4 m stope. The separation of the enveloping surfaces to the structure is >1.4 m (source: redrawn from
Crew, 2002).

Early deformation of the vein
Three principal deformation types affect the Nalunaq Main Vein
comprising vein undulations, boudinage and folding.
The vein exposure in raises shows undulations at a range of
scales from <1 m to 30 m (Figures 17, 19 and 21). These have
open cross profiles in the raises and are the main cause of the
gross dip variation in the vein. Some have a gentle pitch in the
plane of the vein. The time of development is not well
constrained although the local presence of rigorously planar
aplites indicates the undulations predate at least part of the aplite
network. The larger undulations have sufficient amplitude to be
relevant to stope design.
The vein material has been subject to extension leading to
local necking and pinch and swell patterns, but also to complete
disruption by boudinage (Figure 18 and 20). Boudin structures
are cut by some aplite veins. Wall rock screens within the vein
also show boudinage at some sites. Layers in the immediate wall
rock commonly show boudin development, with necks filled with
quartzo-feldspathic material (quartz-feldspar blebs in Crew,

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 43

S C DOMINY and I M PLATTEN

FIG 20 - Nalunaq gold mine, strike section on 440 west level

looking south-east. Boudinaged block of quartz vein with blunt
ends cutting across internal layers. Wallrock smoothly deflected
around the ends. Field of view 1.5 m (source: Snowden, 2006b).

FIG 18 - Nalunaq gold mine, 530 east level. Photograph of boudin

neck (0.2 m gap) in quartz vein in coarse amphibolite host. Altered
wall rock is dragged into the neck and a seam of coarse (5 mm to
20 mm) pyroxene crystals occur along centre line of neck. Planar
aplite veins cross the deformation structures associated with the
boudinaged vein (arrowed). A channel sample site here (blue line)
lies across the boudin neck (source: Snowden, 2006b).

2002). This material locally invades boudin necks in the quartz

vein (Figure 20).
There are no proved large scale folds affecting the vein, but
minor folds at 1 cm to 100 cm scale are sometimes seen. The
narrowest quartz veins, particularly those oriented slightly
oblique to the main structure and isolated in wall rock,
occasionally show small folds with traces of axial plane foliation
in host rock. Narrow screens within the vein locally show small
folds. Planar aplite sheets were seen to cut one clearly folded
quartz vein. FIG 21 - Nalunaq gold mine, 350 level, looking south-east.
Photograph of undulation in 0.3 m vein and comparison with later,
planar, 0.4 m aplite sheet in drive. Vein dip at crest is 20°,
Aplite sheet network steepening to 50° in the limb. Note also the boudin neck in the
Aplite (used to describe various leucocratic granites with fine to quartz (Bn) and patch and veins of coarse quartz and feldspar (Qf)
coarse texture) sheets are seen at surface to form several sets in associated with host rock and vein rupture, structures that are
the vein host rocks. The most gently dipping set are traceable as seen in the vein but not the aplite sheet (source: Snowden, 2006b).
offshoots of the adjacent leucocratic biotite granodiorite pluton.
Underground two sets are conspicuous: a gently dipping set are commonly <1 m thick. Dilation on some is normal to sheet
intersecting vein at low angles and steep set running slightly walls but some show evidence of oblique dilation. Dilution and
oblique to the strike of vein (Figures 14, 18, 19 and 21). These small vein offsets result from aplite sheet emplacement

FIG 19 - Combination of structures affecting gross vein envelope in a raise. Two vein structures show varying separation (0.3 m minimum to
1 m maximum wall rock screen), undulation and some component of vertical displacement across aplite. Aplites show cumulative width of
0.9 m along the line of the raise, seven per cent of the 13 m segment illustrated. Rectangle shows upper 11 m of best fit position for 15 m
by 1.2 m best fit stope outline on the raise data (source: redrawn from Crew, 2002).

44 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 CONTINUITY RISK IN NARROW REEF GOLD DEPOSITS: IMPLICATIONS FOR EVALUATION AND EXPLOITATION
(Figure 19). Channel sample sites can lie within aplite sheets. At
surface the thicker sheets are seen to be continuous for hundreds
of metres.
Faults
The vein and wall rocks are cut by a small number of planar
faults with widely varying dip and strike. These faults are
discordant to the vein and the ductile deformation structures.
They are usually occupied by sheets of granitic or aplitic
material. These igneous materials do not show evidence of
faulting and where fractured the patterns are unrelated to the
fault orientation. The faults are inferred to predate or be
synchronous with aplite sheet emplacement.
The Pegmatite Fault controls mine layout, separating the
Target and Southern Blocks (Figure 13). It has an apparent
down-throw to the northwest, but the true movement vector
is unknown. The other faults have small (1 m to 10 m)
displacements and affect local design of drive and stopes.
Late joints
Planar joint sets cut the veins and aplite sheets. Conspicuous sets
of steep joints run northwest-southeast, 90° to the vein strike, and
a set of southwest-northeast joints dip southeast, subparallel to
the vein. These produce negligible offsets of the host structures.
The joints have mineralised fills and include examples with
limonitic stains recording recent weathering. These joints control
breakage in some parts of drive backs and can influence wallrock
breakage and dilution in the stopes.
Grade/resource estimate
Sampling and estimation practice at Nalunaq
At Nalunaq the complexity and presence of coarse gold was
recognised at an early stage and resulted in the use of bulk
samples and underground development during the feasibility
studies (Lind, Kludt and Ballou, 2001; Dominy and Petersen,
2005). Detailed mapping during development provided a record
of the vein complexity (Crew, 2002).
Dominy and Petersen (2005) describe the sampling protocols
developed to control the effects of coarse gold. Bulk samples
(480 samples of 50 t to 70 t) were taken from development
material in 2000 and 2001. Routine face and drive wall sampling
uses a composite of three channel samples across the vein
(channels spaced 1 m and cut to include a few centimetres of
external wall rock, minimum sample length 0.5 m; details are
given in Dominy and Petersen, 2005).
Controls on the geological nugget effect
The observation of coarse gold, visible clusters and the local
truncation of individual early gold bearing veins is a direct
warning that gold distribution is likely to be nuggety.
Boudin necks and primary vein terminations form distinct
short gaps in the vein where gold mineralisation is usually absent
(Figures 14 to 21). Aplite sheet intersections also produce short
‘gold absent’ sites. Systematic sampling in the plane of the vein
will pick some of these ‘gold absent’ sites. These produce
extreme low grades that are adjacent to full thickness vein
material and thus contribute to the geostatistical nugget effect
(effectively ‘negative nuggets’). The relative contributions of
large grains, or clusters of coarse gold, and the ‘gold absent’ sites
to the observed nugget effect are not known.
Exploitation
The Nalunaq deposit is exploited in longhole stopes developed
from sublevels driven at approximately 11 m vertical intervals
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
along the vein. The minimum design stope width is 1.2 m, but in
practice is commonly 1.4 m.
The thickness of all individual veins (<1 m) and many vein
assemblages is less that any practical minimum stoping width.
Open folds or undulations, offsets across aplite sheets and local
rapid lateral divergence of individual veins result in the gross
envelope containing the vein to be thicker than any local
measurement of vein thickness (Figures 15, 17 and 19). Stopes
are however designed to minimise dilution and thus take rock
only from the vein thickness, subject to a minimum practicable
working width (1.2 m ideal or 1.4 m practical minimum).
Prior to stoping the position of the vein system is known
precisely on the walls of the development sublevels, thus locating
the top and bottom of the stope panel. Vein position and
thickness between the sublevels is unknown in detail for most of
the intended stope. For commonly used longhole stoping, this
defines a stope cross-section with rectangular form and little or
no change in thickness between the upper and lower ends. This
assumes a simple planar form to the vein with limited variation
in thickness. Any departure from this assumption potentially
leads the unplanned dilution and possible failure to actually
recover vein material (Figures 17 and 19). The latter is
particularly serious where the vein is extremely narrow (<0.3 m)
and of very high grade. The absence, presence and position of
individual high grades or extreme low grades are not known
between the sublevels.
Exploration raises (Figures 15 to 17 and 19) and production
raises will however provide local information on vein position
and character between the sublevels. These will allow some
assessment of the risk of using narrow stopes based on the
observed vein characters in the levels. Clearly there is little risk
for stopes adjacent to Figure 16. In Figure 15 there is a risk of
leaving some part of thick vein assemblage in the stope if the
thick section is not seen in sublevel walls. The envelope around
the vein assemblage is however still simple. Figure 17 illustrates
a situation where there is severe risk of leaving a long section of
the ore material in the stope wherever the sublevels occur. In
Figure 17 the model stope outline is drawn for the best recovery
and any other position would leave significant amounts of either
the upper or lower leaf of the vein in place in the stope. In this
case the vein position is controlled by three geological features:
variation in separation of the two leaves, the undulation structure
and the offsets associated with aplite sheet emplacement. These
are unrelated structures and are likely to have different tracks
through any adjacent stope.
Successful projection of structures away from the raise
requires knowledge of the pitch of the structure in the vein
together with its likely continuity. Thicker aplites (Figures 19
and 21) are continuous and regular for relatively long distances
but continuity of individual undulations is not clearly established.
Nalunaq continuity risk profile
The Nalunaq mineralisation defines as moderate-high risk, due
to a high nugget effect of gold grade within the ore shoots
(Table 5). Resolution of local geological continuity and grade for
Ore Reserve definition requires underground development
(blocking out).
It is possible to make a more formalised, but not quantitative,
assessment of the geological risk in stope design (Snowden,
2006b; Dominy et al, 2006). Table 6 summarises vein
complexities and Table 7 and Figure 22 illustrate possible risk
categories for Nalunaq. Various potential stope outlines are
shown in Figure 22 to illustrate the variation in dilution and
recovery risk and the risk for different outlines is summarised in
Tables 6 and 7.
Geological risk features cannot be entirely eliminated due to
cost and the balance between full recovery and increased
45
S C DOMINY and I M PLATTEN

TABLE 5
Nalunaq continuity risk review and associated project expectations.

Uncertainty criteria Risk score Comment

Gross geological continuity
Reef geometry: dip, strike, width, etc
Grade distribution control: ore shoot controls
Grade distribution: local spatial effects and
small-scale continuity issues within the ore
zone/ore shoots – ‘nugget effect’
Post-reef effects of faulting, dykes, etc
Total risk profile
Project characteristics based on risk profile
3 The Nalunaq Main Vein is relatively continuous. Gross continuity is generally not material
to ore shoot definition. Wide-spaced drilling at 100 m to 150 m spacing will resolve gross
reef continuity.
(4) 5 Locally reefs can show strong variability that requires underground development or
close-spaced drilling (~10 m) to resolve. Locally material to evaluation and mining.
3 Ore shoots are controlled by structural and lithological features and show generally good
down-plunge continuity. Can be defined by drilling alone (~30 m).
(4) 5 Ore shoots have a high nugget effect (~70%), with gold grade moderately difficult to
predict locally. Requires underground development to resolve effectively and define Ore
Reserves.
5 Critical vein modifying features are aplite dykes, shearing/boudinage and minor cross
faulting, that has local material effect on the continuity of ore shoots. Requires
underground development to resolve or close spaced drilling (~10 m).
4 (5) The overall continuity risk is considered to be moderate-high. This is driven by the
high-nugget effect and local geological complexities and need for development to resolve
continuity issues.
• Moderate size gold system with total inventory of potentially 1.5 Moz Au
• Small high-grade operation producing around 80 000 to 90 000 oz Au per annum
• Definition of Inferred and Indicated Mineral Resources. Measured not defined even after development
due to nugget effect
• Some 75% of resource base in the Inferred category
• Only one to two years ahead defined as ore reserves based on development

TABLE 6
Summary of vein complexities at Nalunaq and implications for mining (after Snowden, 2006b).

Type of vein complexity Feature Scale Implication

Splitting up or uniting Vein breaks up or veins
converge
Dislocation Aplites and/or faults
cross-cutting the vein
Width variation in vein or Boudinage or primary
vein zone nature of the vein
Attitude Change in dip over short
range
>5 m High variation in total width of veined zone with likelihood of
sections of the vein passing out of the planned stope limits.
Gaps <1 m, spaced at Relocation of vein and dilution by barren aplitic material.
2 m to 10s m Likelihood of vein passing out of the planned stope limits.
Gaps up to 5 m, spaced High variation in vein width with barren gaps leading to dilution.
4 m to 10s m Likelihood of wider sections of the vein passing out of the planned
stope limits.
>5 m Change in vein attitude that may result in sections of the vein
passing out of the planned stope limits.

TABLE 7 CONCLUSIONS
Geological risk classification for stope design at Nalunaq It is well-known that gold-quartz reefs with a high-nugget effect
(after Snowden, 2006b). generally possess a high resource risk. A key component of
resource risk is ‘continuity’ risk, related to uncertainty of reef
Risk Geological features Implications for stope design
continuity at different scales and grade continuity within ore
Low Simple planar vein Vein unlikely to extend shoots. An evaluation program must collect the geological
beyond the limits defined information required, and will include diamond drill core logging
from development. Recovery and surface and/or underground mapping. It should then be
likely to be in excess of 90%. possible to produce a continuity risk profile, from which realistic
Moderate Vein shows some Some portions of the vein expectations of resource definition and problems during mining
variability in width, may extend beyond the limits can be understood. For early stage programs (eg with drilling
attitude and/or continuity defined from development. data only), this may be difficult and will rely on the experience
Recovery likely to be between of the geologist(s) concerned.
60 % and 90%.
The Clogau reef is a very high-risk style of mineralisation due
High Complex vein structure Vein very likely to extend to the small size of the ore shoots and highly-localised
with highly variable width, beyond the limits defined unpredictable gold pockets. On the macro-scale, the occurrence
attitude and/or continuity from development. Recovery
of ore shoots is controlled by the reef structure crossing
likely to be less than 60%.
greenstone sills and/or splits/branches. This localisation is due to
geochemical and structural factors which in combination created
the access for gold-bearing fluids and a geochemical
dilution. Risk factors can be identified and can guide response to environment for gold deposition. Such a deposit has a high
unacceptable shortfalls in grade and ounces. In addition, it resource risk and requires strong geological control during
should be possible to schedule blending of high-risk and low-risk evaluation and mining. The high risk nature of the deposit only
stopes to maintain extraction rates and optimise head grade. permits the definition of small Inferred Mineral Resources even

46 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 CONTINUITY RISK IN NARROW REEF GOLD DEPOSITS: IMPLICATIONS FOR EVALUATION AND EXPLOITATION
FIG 22 - Schematic risk assessment models (A, B, C1 and C2) for
stopes from raise mapping data. Refer also to Table 6
(from Snowden, 2006b).
after drilling and underground development. Potential economic
viability is only gained through the use of gold for a premium
jewellery product.
At Nalunaq the resource risk is medium-high and related to a
high nugget effect and continuity issues with the Main Vein.
Undulations in the vein profile are locally large enough relative
to planned stope widths to result in vein material being left
in situ. The deleterious effects begin when the undulations have a
relief of half the stope width, where the largest example may
leave as much as a 10 m length of vein in the stope. Steps in vein
elevation across aplite sheets and small faults may compound the
effects of the undulations. If the undulations can be shown to be
predominantly linear features, then their effects within the stopes
may be partly predictable from outcrop in drives and raises.
Boudinage, aplite sheet intersections and extensional faults are
the main structures that disrupt the gross continuity of the Main
Vein. These are linear features and produce ribbon like gaps or
zones of reduced total vein thickness, with gap widths from
<1 m up to 5 m.
The role of continuity risk definition lies in the wider
management of project and resource risk. If continuity issues can
be determined as early on in a project as possible, it will permit
the technical team to define realistic expectations and milestones.
The level of investment required to fully resolve continuity is
often high, and will involve underground development supported
by close-spaced diamond drilling and bulk sampling/trial mining
(Dominy and Petersen, 2005; Dominy, Platten and Xie, 2008).
However, in previously mined goldfields historical data usually
provides a key source of information. A strong geological
understanding must form the base of all mineral resource
estimates, particularly for complex narrow gold reefs
(Dominy et al, 2003; Dominy, Noppé and Annels, 2004).
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
ACKNOWLEDGEMENTS
This contribution results from on-going work by the authors on
the evaluation of high-nugget effect gold reefs. The following
organisations have supported studies in Dolgellau: Clogau Gold
Mines PLC, Welsh Gold PLC, Elenith Mining Company
Limited, Clogau St David’s Gold Mines Limited, Union Mining
Company Limited and Victorian Gold Limited, and in
Greenland: Crew Gold Corporation and Nalunaq Gold Mine
A/S. Discussions with George Hall, David Pelham and Roland
Phelps on Dolgellau, and Jon Steen Petersen, Ole Dahl, Mogens
Lind and Kurt Christensen on Nalunaq are acknowledged. In
addition, thanks are due to Andy Ross (Snowden Group) and
The AusIMM Conference reviewers, Brian Cuffley and Keith
Whitehouse, for their helpful comments on the manuscript. The
opinions expressed in this paper are those of the authors, and
not necessarily those of Snowden or the named operating
companies.
REFERENCES
Allen, P M and Jackson, A A, 1985. The Geology of the Country Around
Harlech, 111 p (British Geological Survey: Keyworth).
Annels, A E, 1996. Ore Reserves: Errors and classification, Transactions
of the Institutions of Mining and Metallurgy, Applied Earth Science,
105:A150-A156.
Annels, A E and Dominy, S C, 2002. The development of a Resource
Reliability Rating (RRR) system for mineral resource estimation, in
Proceedings Value Tracking Symposium 2002, pp 27-32 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Baxter, J L and Yates, M G, 2001. Estimation of reserves and resources in
shear zone hosted gold deposits, in Mineral Resource and Ore
Reserve Estimation – The AusIMM Guide to Good Practice,
pp 125-134 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Bottrell, S H, Shepherd, T J, Yardley, B W D and Dubessy, J, 1988.
A fluid inclusion model for the genesis of the ores of the Dolgellau
gold-belt, North Wales, J Geol Soc Lond, 145:139-145.
Bottrell, S H and Spiro, B, 1988. A stable isotope study of black shale
hosted mineralisation in the Dolgellau gold-belt, North Wales,
J Geol Soc Lond, 145:941-949.
Chadwick, B and Garde, A A, 1996. Palaeoproterozoic oblique plate
convergence in South Greenland: A reappraisal of the Ketilidian
Orogen, in Precambrian Crustal Evolution of the North Atlantic
Region, Geological Society of London special publication,
112:179-196.
Cox, S F, Sun, S-S, Etheridge, M A, Wall, V J and Potter, T F, 1995.
Structural and geochemical controls on the development of
turbidite-hosted gold quartz vein deposits, Wattle Gully mine,
Central Victoria, Australia, Econ Geol, 90:1722-1746.
Crew, 2002. Underground development in the Nalunaq concession, South
Greenland, unpublished report, Crew Development Corporation and
Nuna Minerals A/S, 110 p (author: L Porritt).
Crew, 2007. Updated mineral resource and reserve for Nalunaq gold mine,
press release dated 20 April 2007, Crew Gold Corporation, 3 p.
Cuffley, B W, Krokowski deVickerod, J, Evans, T J and Fraser, R, 1998.
A new structural model for fault-hosted gold mineralisation: An
example from the Nick O’Time oreshoot, Poverty reef, Tarnagulla, in
Second GPIC Conference on Developments in Victorian Geology
and Mineralisation, Bulletin 24:53-63 (Australian Institute of
Geoscientists: Perth).
Dominy, S C, 1996. Mineralisation and oreshoot controls at the Clogau
gold mine, Dolgellau, United Kingdom, unpublished report for
Clogau St David’s Gold Mines Limited, 18 p.
Dominy, S C, 2004a. Comments on classification and reporting of
mineral reporting for high-nugget effect gold vein deposits, Explor
Min Geol, 11:119-124.
Dominy, S C, 2004b. Reporting resource grade uncertainty in high-nugget
effect lode-gold deposits, in Proceedings EGRU Mining and
Resource Geology Symposium 2004, EGRU contribution No 62,
pp 46-60 (James Cook University: Townsville).
47
S C DOMINY and I M PLATTEN
Dominy, S C, 2005. Mineral resource reporting in mesothermal gold
veins: The critical role of structural (continuity) modelling, in
Proceedings Seminar on the Estimation and Reporting of Resources
and JORC: The Role of Structural Geology, Bulletin No 42, pp 17-27
(Australasian Institute Geoscientists: Perth).
Dominy, S C, 2006. Inferred mineral resources and gold veins –
Discussion and case studies, in Proceedings Sixth International
Mining Geology Conference 2006, pp 71-83 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Dominy, S C, in press. Role of historical data in the evaluation of old
goldfields, J Welsh Mines Soc, 1.
Dominy, S C, Annels, A E, Camm, G S, Cuffley, B W and Hodkinson, I
P, 1999. Resource evaluation of narrow gold-bearing veins: Problems
and methods of grade estimation, Transactions of the Institutions of
Mining and Metallurgy, Mining Technology, 108:A52-A70.
Dominy, S C, Annels, A E, Camm, G S, Wheeler, P D and Barr, S P,
1997. Geology in the resource and reserve estimation of narrow vein
deposits, Explor Min Geol, 6:317-333.
Dominy, S C, Annels, A E, Platten, I M and Raine, M D, 2003. A review
of problems and challenges in the resource estimation of high-nugget
effect lode-gold deposits, in Proceedings Fifth International Mining
Geology Conference 2003, pp 279-298 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Dominy, S C, Cuffley, B W, Johansen, G F, Annels, A E and Platten, I M,
2000b. Resource evaluation of nuggety slate-hosted gold-quartz
reefs, in Proceedings Fourth International Mining Geology
Conference 2000, pp 169-185 (The Australasian Institute of Mining
and Metallurgy: Melbourne).
Dominy, S C, Cuffley, B W, McCarthy, P L, Thomas, O K and Hill, R L,
2004. The challenges of evaluating and exploiting gold-quartz reefs,
in the Central Victorian goldfield, Australia, in Proceedings New
Leaders’ Conference 2004, pp 121-141 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Dominy, S C and Johansen, G F, 2004. Reducing grade uncertainty in
high-nugget effect gold veins – Application of geological and
geochemical proxies, in Proceedings PACRIM Congress 2004,
pp 291-302 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Dominy, S C, Johansen, G F, Annels, A E and Cuffley, B W, 2000a.
General considerations of sampling and assaying in a coarse gold
environment, Transactions of the Institutions of Mining and
Metallurgy, Applied Earth Science, 109:B145-B167.
Dominy, S C, Johansen, G F, Cuffley, B W, Platten, I M and Annels, A E,
2001. Estimation and reporting of mineral resources for coarse
gold-bearing veins, Explor Min Geol, 9:13-42.
Dominy, S C, Noppé, M A and Annels, A E, 2004. Errors and uncertainty
in mineral resource and ore reserve estimation: The importance of
getting it right, Explor Min Geol, 11:77-98.
Dominy, S C and Petersen, J S, 2005. Sampling coarse gold-bearing
mineralisation – Developing effective protocols and a case study
from Southern Greenland, in Proceedings Second World Conference
on Sampling and Blending 2005, pp 151-165 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Dominy, S C and Phelps, R F G, 1996. Gwynfynydd’s gold veins:
Geology plays an important role in the mining of complex narrow
veins, Mining Mag, 175(6):325-333.
Dominy, S C, Phelps, R F G and Camm, G S, 1996. Geological controls
on gold grade distribution in Chidlaw link zone, Gwynfynydd mine,
Dolgellau, North Wales, Transactions of the Institutions of Mining
and Metallurgy, Applied Earth Science, 105:B151-B158.
Dominy, S C, Phelps, R F G, Sangster, C J S and Camm, G S, 1998.
Narrow vein stoping practice – A review, Transactions of the
Institutions of Mining and Metallurgy, Mining Technology,
107:A122-A136.
Dominy, S C and Platten, I M, 1999. Criteria for the recognition of a
pre-cleavage age for gold-bearing quartz veins in weakly cleaved
terrains, Transactions of the Institutions of Mining and Metallurgy,
Applied Earth Science, 108:B73-B76.
Dominy, S C and Platten, I M, 2007. Gold particle clustering – A new
consideration in sampling applications, Transactions of the
Institutions of Mining and Metallurgy, Applied Earth Science,
116:B130-B142.
Dominy, S C, Platten, I M and Raine, M D, 2003. Grade and geological
continuity in high-nugget effect gold-quartz reefs: Implications for
48 Ballarat, Vic, 14 - 15 October 2008
resource estimation and reporting, Transactions of the Institutions of
Mining and Metallurgy, Applied Earth Science, 112:B239-B259.
Dominy, S C, Platten, I M and Xie, Y, 2008. Bulk sampling of complex
gold deposits: Material characterisation and program design and
management, in Proceedings Sampling 2008 Conference, pp 41-57
(The Australasian Institute of Mining and Metallurgy: Melbourne).
Dominy, S C, Sides, E J, Platten, I M and Dahl, O, 2006. Estimation and
exploitation in an underground narrow vein gold operation – Nalunaq
mine, Greenland, in Proceedings Sixth International Mining Geology
Conference 2006, pp 29-44 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Dominy, S C, Stephenson, P R and Annels, A E, 2003. Classification and
reporting of mineral resources for high-nugget effect gold vein
deposits, Explor Min Geol, 10:215-233.
Dominy, S C, Xie, Y and Platten, I M, 2008. Gold particle characteristics
in narrow vein deposits: Implications for evaluation and metallurgy,
in Proceedings Narrow Vein Mining Conference 2008, pp 91-104
(The Australasian Institute of Mining and Metallurgy: Melbourne).
Dunn, E J, 1929. Geology of Gold (South Africa, Australia and New
Zealand), 303 p (Charles Griffin & Company Limited: London).
Fraser, R, Bartlett, J K and Quigley, P, 2003. The role of historical
research in the development of the New Bendigo project, Central
Victoria, in Proceedings Fifth International Mining Geology
Conference 2003, pp 119-128 (The Australasian Institute of Mining
and Metallurgy: Melbourne).
Gowen, J, Christiansen, O, Grahl-Madsen, L, Petersen, J S and Robyn,
T L, 1993. Discovery of the Nalunaq gold deposit, Kirkspirdalen,
Southwest Greenland, Int Geol Rev, 35:1001-1008.
Grammatikopoulos, T A, Porritt, L, Petersen, J S and Christensen, K,
2004. Mineralogical characterisation and process mineralogy of
gold-bearing rocks from Nalunaq gold deposit, Greenland,
Transactions of the Institutions of Mining and Metallurgy, Applied
Earth Science, 113:B197-B203.
Hall, G W, 1988. The Gold Mines of Merioneth, p 99 (Griffin
Publications: Kington).
Howells, M F, 2007. British Regional Geology – Wales, p 230 (British
Geological Survey: Keyworth).
Joint Ore Reserves Committee (JORC), 2004. Australasian Code for
Reporting of Mineral Resources and Ore Reserves, Joint Committee
of The Australasian Institute of Mining and Metallurgy, Australian
Institute of Geoscientists and Minerals Council of Australia.
Kaltoft, K, Schlatter, D M and Kludt, L, 2000. Geology and genesis of
Nalunaq Palaeoproterozoic shear zone-hosted gold deposit, South
Greenland, Transactions of the Institutions of Mining and
Metallurgy, Applied Earth Science, 109:B23-B33.
Keppie, J D, Boyle, R W and Haynes, S R, 1986. Turbidite-Hosted Gold
Deposits, Geological Association of Canada, special paper 32, p 186.
Lind, M, Kludt, L and Ballou, B, 2001. The Nalunaq gold prospect, South
Greenland: Test mining for feasibility studies, Bull Geol Surv
Greenland, 189:70-75.
Lu, H, Wang, Z, Chen, W, Wu, X, Hu, R and Keita, M, 2005.
Turbidite-hosted gold deposits of Southeast Guizhou, China: Their
regional setting, mineralising styles and genetic constraints, in
Proceedings Eighth Biennial SGA Meeting: Mineral Deposit Research
– Meeting the Global Challenge, pp 545-548 (Springer Verlag: Berlin).
McCuaig, T C, Vann, J and Seymour, C, 2000. Dynamic links between
geology and the mining process, in Proceedings Fourth International
Mining Geology Conference 2000, pp 169-185 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Monteiro, R N, Fyfe, W S and Chemale, F, 2004. The impact of the
linkage between grade distribution and petrofabric on the
understanding of structurally controlled mineral deposits: Ouro Fino
gold mine, Brazil, J Struct Geol, 26:1195-1214.
Morrison, R J, Storey, N J M and Towsey, C A J, 2004. Management of
geological risks associated with quartz reef gold deposits, Charters
Towers, Queensland, in Proceedings Mine and Resource Geology
Symposium 2004, EGRU contribution No 62, pp 87-106 (James Cook
University: Townsville).
Nugus, M J, Blenkinsop, T G, Dominy, S C and Robson, S, 2003.
Enigmatic kinematics resolved in the Taurus shear zone, Golden Pig
mine, Southern Cross, Western Australia – Resource implications, in
Proceedings Fifth International Mining Geology Conference 2003,
pp 171-179 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Narrow Vein Mining Conference
 CONTINUITY RISK IN NARROW REEF GOLD DEPOSITS: IMPLICATIONS FOR EVALUATION AND EXPLOITATION
Olsen, S, 2005. Personal communication.
Osborne, D J, 2008. The Ballarat East goldfield: New insights on an old
model, in Proceedings Narrow Vein Mining Conference 2008,
pp 59-70 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Peters, S G, 1993. Nomenclature, concepts and classification of oreshoots
in vein deposits, Ore Geol Rev, 8:3-22.
Phillips, G N and Hughes, M J, 2003. Gold – Historical wealth, future
potential, in Geology of Victoria, Geological Society of Australia,
special publication 23, pp 377-433.
Platten, I M and Dominy, S C, 1999. A re-evaluation of quartz vein
history in the Dolgellau gold-belt, North Wales, United Kingdom,
Geol J, 34:369-391.
Platten, I M and Dominy, S C, 2003. The occurrence of high-grade gold
pockets in quartz reefs at the Gwynfynydd mine, Wales, United
Kingdom: A geological explanation of the nugget effect, Explor Min
Geol, 10:249-272.
Shaubs, P M and Wilson, C J L, 2002. The relative roles of folding and
faulting in controlling gold mineralization along the Deborah
anticline, Bendigo, Victoria, Australia, Econ Geol, 97:351-370.
Shepherd, T J, Bottrell, S H and Miller, M F, 1991. Fluid inclusion
volatiles as an exploration guide to black shale-hosted gold deposits,
Dolgellau gold-belt, North Wales, United Kingdom, J Geochem
Explor, 42:5-24.
Sinclair, A J and Vallée, M, 1994. Reviewing continuity: An essential
element of quality control for deposit and reserve estimation, Explor
Min Geol, 2:95-108.
Snowden, 2005a. Independent mineral resource estimate: Nalunaq gold
mine, Greenland, unpublished report by Snowden Mining Industry
Consultants Limited for Nalunaq Gold Mine A/S, project No 5122
(author: S C Dominy), 76 p.
Snowden, 2005b. Nalunaq gold mine optimisation study, unpublished
joint report by Snowden Mining Industry Consultants Limited and
AMC Consultants Limited for Crew Gold Corporation, project No
5335 (authors: S C Dominy and C J S Sangster), 70 p.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
Snowden, 2006a. Independent mineral resource estimate: Nalunaq gold
mine, Greenland, unpublished report by Snowden Mining Industry
Consultants Limited for Nalunaq Gold Mine A/S, project No L102
(author: S C Dominy), 123 p.
Snowden, 2006b. Aspects of structural geology at the Nalunaq gold mine,
Greenland: Implications for mining and resource estimation,
unpublished report by Snowden Mining Industry Consultants
Limited for Crew Gold Corporation, project No L102 (authors: S C
Dominy and I M Platten), 66 p.
Snowden, 2008a. Geological review and mineral resource estimate:
Cononish gold project, Scotland, United Kingdom, unpublished
report by Snowden Mining Industry Consultants Limited for
Scotgold Resources Limited, project No L136 (authors: S C Dominy
and I M Platten) 147 p.
Snowden, 2008b. Review of gold mineralisation and potential in the
Dolgellau gold-belt, North Wales, United Kingdom, unpublished
report by Snowden Mining Industry Consultants Limited for
Victorian Gold Limited, project No L132 (author: S C Dominy) 28 p.
Stendal, H, Grahl-Madsen, L, Olsen, H K, Schonwandt, H K and
Thomassen, B, 1995. Gold exploration in the early Proterozoic
Ketilidian orogen, South Greenland, Explor Min Geol, 4:307-315.
Tripp, G I and Vearncombe, J R, 2004. Fault/fracture density and
mineralisation: A contouring method for targeting in gold exploration,
J Struct Geol, 26:1087-1108.
Turnbull, D G and McDermott, G J, 1998. Deborah line of reef gold
deposits, Bendigo, in Geology of Australian and Papua New Guinea
Mineral Deposits, pp 521-526 (The Australasian Institute of Mining
and Metallurgy: Melbourne).
Windley, B F, 1991. Early Proterozoic collision tectonics and rapakivi
granites as intrusions in an extensional thrust thickened crust: The
Ketilidian Orogen, South Greenland, Tectonophysics, 195:1-10.
49
Resource Estimation and Grade Assignment – A Comparison
Between Historical Production and Current Maxwell Mining
Validation Case Study at Morning Star Gold Mine, Woods Point
M D Goodz1, J Rea2 and P Jackson3

ABSTRACT
The Morning Star Gold Mine has operated over a 101-year period,
producing gold from narrow quartz veins and sulfide stockworks hosted
mainly within a dioritic-gabbroic dyke. The deposit has been developed
over a vertical depth of 980 m with drilling extending the vertical depth to
greater than 1200 m. Currently there is a total Inferred Resource of
4.6 million tonnes at a grade of 6 g/t for 910 000 ounces of contained gold
resources, which comprises an underground project and a surface project.
Significant variation exists between the grade of drill core intersections
and that observed in face sampling and subsequent production
reconciliation (250 - 650 per cent). Studies were carried out on ten major
production reefs, where the grades of drill intercepts were compared with
face samples collected during development and mining. The implications
of this study have been used as justification to assign production or
stope/development grades where applicable to resources that are direct
extensions of the historical development headings. Grade assignment and
its impact on sensitivity for gold resources are detailed and show a 40 per
cent variation in the gold resource, depending upon whether a production
or calculated stope grade is used in grade assignment for the major reef
deposits.
Recent diamond drilling has intersected approximately 40 new gold
bearing quartz reefs and sulfide zones that contain all the key
mineralogical and structural indicators, and traditional large variations in
gold grades similar to the historical major production zones. The
implications of grade factoring and grade assignment are the greatest
challenges confronting the resource estimation team in defining grade for
these new zones.
Grade assignment is seen as a method of balancing known historical
production observations with the irregular grade distribution inferred
from drill core assay results. Ongoing mine development and bulk
sampling works are planned to validate the grade assignment process.
The first stage of this process is the trial mining of the newly discovered
Maxwell Reef (MR). The stope drawpoints were completed in August
2008 and the MRS1 stope is scheduled for completion in September. The
stope reconciliation results will be presented at the October 2008
conference.
INTRODUCTION
The Morning Star Gold Mine is located within the Woods
Point-Walhalla goldfield 500 m west of the town of Woods Point,
approximately 110 km northeast of Melbourne, Australia
(Figure 1). The Woods Point-Walhalla goldfield produced greater
than five million ounces of gold from hard rock underground
production between 1861 and 1992. Production at Morning Star
was 883 000 ounces (Goodz, 1994; Goodz et al, 1997;
VanderBerg et al, 2006). The mine is owned and operated by
Morning Star Gold NL (MCO) who have tenure to an 11 km
corridor of the Ross Creek Fault Zone (RCFZ) with an overall
production history of greater than 1.8 million ounces.
In the RCFZ mineralised zone, more than 1000 dyke mines
have been identified as hosting gold deposits, the largest having
1. MAusIMM, Technical Director, Morning Star Gold NL, PO Box 10488,
Kalgoorlie WA 6433. Email: morningstargold@goodzgmc.com.au
2. Managing Consultant, Gemcom Australia Pty Ltd, PO Box 7495,
Cloisters Square WA 6850. Email: jrea@gemcomsoftware.com
3. Chief Geologist, Morning Star Gold NL, Private Bag 2000,
Mansfield Vic 3724. Email: mcogeologist@bigpond.com
FIG 1 - Project location.
produced more than 1.2 million ounces, and several mines more
than 100 000 ounces. The gold is associated with narrow
quartz-carbonate-sulfide veins commonly 0.1 - 1.5 m thick and
oriented along various structural planes ranging from 15° to 60°
in dip. This requires an assortment of stoping and support
options. The veins strike between 30° and 70° oblique to the
strike of the dyke (trending 310° to 325°) and therefore have a
strike length of 100 to 700 m (Figure 2). MCO has commenced
advanced exploration on four of these deposits, including the
resource definition drilling underground at the Morning Star
mine (currently refurbished and working on the No 10 Level
~305 m below surface). MCO has recently purchased and
commissioned an owner-operator fleet of both underground and
surface diamond drilling rigs.
The current database exceeds 508 drill holes (450 diamond),
20 km of underground drive development (7 km currently
accessible) with extensive stope survey and sampling detail.
Significant 3D-modelling has been completed to determine
structural controls on grade distribution, including defining
models for exploration drilling and development. Vein mapping,
thickness and grade contouring have been collated on the main
production reefs with reference to a 3D-modelled average plane
orientation. This work has evolved from a series of contour
diagrams referred to as Conolly contour diagrams (Schwartz,
1986). We have used modified Conolly-style grade-thickness
modelling to understand relationships between high-grade shoot
orientation, frequency and position (Figure 3).
Diamond drill testing of these models has resulted in the
successful discovery of three new reef zones and substantial
extensions to historical production zones. Underground
development is in progress to commence a trial mining
feasibility study on the Maxwell Reef.
MCO has carried out pre-feasibility metallurgical test work to
design a processing facility to recover up to 93 per cent
free-milling gravity gold and in excess of 99 per cent with a
standard leach circuit. Trial processing has commenced on the
first bulk sampling parcel to design a wholly gravitational option
for a processing circuit (Ammtec, 2007).

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 51

M D GOODZ, J REA and P JACKSON

FIG 2 - Long section of Morning Star mine (isometric view) showing reef orientations and dyke morphology.

FIG 3 - Conolly diagram showing grade contours in gram-metres. HGZs represent significant tonnages of Achilles reef group remaining
intact and assaying >40 g/t/m (scale is 50 m grid).

Data analysis, followed up with underground drilling and
development, has provided confidence to establish methods for
resource identification and quantification, leading to the release
of a new mineral resource inventory statement. These methods
are currently being expanded to incorporate a regional model.
GEOLOGICAL OVERVIEW
The Morning Star Gold dyke-hosted deposits form a unique
geological opportunity whereby there is a series of quartz vein
deposits at multiple orientations. The Morning Star mine has
been most likened to the Norseman Gold deposits (production
>5 million ounces). Both of these deposit types have similar
geological, mineralogical and structural features, including the
multiple vein orientations, intersecting conjugate vein sets that
are predominately flat-lying reverse fault systems and sinistral
displacements. Gold Mines of Australia (GMA) concurrently
managed both the Morning Star and Norseman mines from 1933
to 1959. During this period, many of Australia’s famous
geologists and surveyors, utilised learnings from both mines to
develop successful exploration models, sampling and mining
methods (Campbell, 1990; Mason, 1994).

52 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 RESOURCE ESTIMATION AND GRADE ASSIGNMENT

FIG 4 - Examples of Morning Star gold mine gold-bearing quartz reefs exposed in mine headings and drill core: (A) laminated vein (Burns
Floor 1965L), photo height 60 cm; (B) breccia vein (Dickensons Floor 4L), photo height 40 cm; (C) stockwork vein (HW Whitelaws Floor 7L),
photo height 110 cm; (D) laminated four-vein set (Tills Floor 9L), photo 2 m high × 5 m deep; (E) conjugate vein set (Shamrock Floor 4L),
photo height 140 cm; (F) 5 m-thick (true width) laminated four-vein set and multiple stockwork veinlets (Whitelaws Floor 10L), NQ-core 1 m
trays; (G) visible gold clots in sulfidic vein laminate and styolitic selvages (Campbells Floor 7L), photo height 10 cm; (H) laminated vein
(Age of Progress 1L) crossing breccia and stockwork veins (AAA Floor), photo 2 m high × 3 m deep.

More than 90 per cent of all stoping is from quartz-carbonate-
sulfide veins and sulfidic stockwork within reverse fault systems
that cross the entire width of the dyke and show thrust offsets on
the dyke contact of 20 m up-dip. Examples of the vein types
exposed in the mine workings are given in Figure 4. These veins
are referred to as ‘Ladder Veins’ due to their interconnecting
stepped orientation that rise from one contact to the other.
Individual veins have up-dip component of between 80 m and
150 m, depending on their dip angle (Figure 5). Less than ten per
cent of the gold production came from the quartz veins after they
entered the more ductile sedimentary country rocks into which
the dyke had intruded.
The gold-bearing veins are associated with a carbonate-
sulfide-sericite wallrock alteration assemblage, which is
predominated by ankerite-arsenopyrite-fuchsite closest to the
vein wall (and is bright green). Pyrite becomes the prominent
sulfide when more than 30 cm from the vein boundary. The
major production reefs are generally a collective of two to three
quartz veins with an alteration envelope of 1.5 - 2.0 m; smaller
veins lose the hydrothermal envelope within 0.3 m. Traditionally
the veins were mined at a minimum working width to permit
access by the miner (1.2 - 2.4 m depending upon the dip angle
and the extent of the alteration envelope into the footwall of the
structure). Maintenance of the hanging wall ensured stope
stability. Many stopes dating back to the 1860s are still open.
Detailed surveys of the stopes have been carried out to accurately
constrain the void model for the mine (Figure 5).
Current drilling at Morning Star has confirmed that the host
dyke geological unit is open-ended in both strike and depth and
has been extended in width in the upper GAP Zone drilling.
Gold-bearing vein and sulfide deposits have been intersected on
spacings of less than 10 m, in portions of the upper GAP Zone
(Figure 2). Studies are underway on the relationship between
dyke morphology and vein frequency/orientation.
RESOURCE HISTORY
The Morning Star mine has had a production history of
883 000 ounces of gold from 1.14 million tonnes of ore for an
average head grade 26.5 g/t Au. Significant production records
document drilling, development, stoping and mill processing
reports, and have allowed for detailed analysis of grade
interpretation. These results have highlighted the impact of grade
factoring and their relevance in the assignment of grade ranges
and key mineral indicator proxies.
Review of the extensive historical database has provided
accurate and reproducible data relating to drilling, surveying,
development, stoping and production records. A detailed study of

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 53

M D GOODZ, J REA and P JACKSON

approximately 50 per cent of the available archival records was

FIG 5 - Cross-section showing vein spacing and dyke morphology
through shaft section (grid scale is 200 m).
carried out in 2008. Table 1 details the results of a preliminary
review based on 7570 face samples and 438 drill hole
intersection assays on ten major reef deposits in the Morning Star
Underground Project. Grade factoring has not been applied in
this resource estimate; however the implications of this study
have been used as justification to assign production or
stope/development grades where applicable to resources that are
direct extensions of the historical development headings. Grade
factoring would have a significant impact on the newly
discovered and undeveloped reef deposits at the Morning Star
mine and provide for a future additional opportunity to extend
the project’s mineral resource inventory.
More study is required to examine variations in both sample
type, size, assay methods, relation to geological and structural
models and possible nugget effect. The current study was limited
in time and personnel resources, and focussed on the ten major
production reef areas initially because of the volume and detail
on sample data. Future recommended work includes statistical
analysis and block modelling.
GMA identified a number of gold-bearing zones grading
between 0.5 and 4.0 g/t that were neglected while there was
ample 24+ g/t Au ore being extracted. Table 1 identifies that
more than 81 per cent of all the drill hole intersections of the
historic production reefs were less than 4 g/t Au. Understanding
the variation between drill hole grade and the potential
production grade is the focus of the current study. What is
consistently observed is that larger samples provide more
representative results in line with production records, which have
much higher grades than drill hole inferred resources. The
implications are that numerous lower grade drill intersected
resources need to be developed and bulk sampled.
The purpose of the Maxwell Reef trial mining project
commenced in August 2008 is to assess the grade variation
between drill hole intercepts, face samples, stockpile samples
and 1000-tonne stope bulk parcels. This data study will be used
to validate the grade assignment process.

TABLE 1
Comparative review of drill hole intercept, stope face sampling and mill recovered grade.

Grade factoring study (n = 8008)

Drill hole stope intersections Stope face samples Stope production Grade factoring
Reef name No of Grade Grade Per cent No of Ave sample Calculated Average mill Production Factor between
samples mean median assays face grade (g/t) stope grade recovered period drill hole grade
(g/t Au) (g/t Au) (<4 g/t) samples (g/t) grade (g/t) years and recovered
grade
Achilles 62 7.21 2.24 70 2326 31.90 16.30 18.0 - 32.6 1942 - 1955 250 - 450%
AFloor 20 3.26 2.16 78 656 15.80 10.30 16.2 - 18.2 1956 - 1959 500 - 560%
BFloor† 25 4.91 1.39 87 333 23.90 16.2 - 18.2 1956 - 1959 330 - 370%
Burns 51 1.85 1.00 88 766 75.20 10.40 22.1 - 67.7 1934 - 1939 >1200%
Campbells 52 2.38 0.90 92 333 90.10 11.60 22.1 - 67.7 1934 - 1939 >900%
Exeter† 23 6.70 2.31 72 1255 34.90 18.0 - 32.6 1942 - 1955 270 - 490%
Lancaster† 25 3.02 2.77 71 381 26.50 18.0 - 32.6 1942 - 1955 >600%
Stirling 31 5.01 2.16 71 569 55.10 37.10 18.0 - 32.6 1942 - 1955 360 - 650%
VFloor 29 2.19 1.23 88 899 39.00 25.80 16.2 - 18.2 1956 - 1959 >700%
Whitelaw† 96 1.97 0.79 92 35 42.50 9.81 22.1 - 67.7 1934 - 1939 >1100%
XFloor 24 2.13 1.69 87 16.2 - 18.2 1956 - 1959 >700%
Average 438 3.69 1.69 81 7570 39.90 ~24 g/t 1934 - 1959
Average CSG 5601 42.90 18.10 ∼26.5 g/t‡

† Data modelling still in progress.

‡ Sampling of stamp battery tailings dump had minimum average grade = 2.5 g/t Au; therefore production grade ≥ 26.5 g/t Au.

54 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference


GRADE ASSIGNMENT
The following resource estimation is based on a combination of a
drill hole database and stope development records to assign
resource outlines and tonnage calculations. The grade
assignment is based on a proxy system that examined the
geological/structural features of the veins and their relative
relationship with known production and development data.
Ranking of the data assigned grade related to production,
stope/development assay samples and drill hole intersections in
respective order of priority.
The relevance of Table 1 data is significant, as there remain
large extensions to the main historical production stopes. Some
of these reef zones had considerable level and rise development
by GMA, but were never extracted. These reefs are blocked out
on two or three sides and constitute a major component of the
resource included in this report.
Table 2 shows the sensitivity on the gold resource through
variation between the drill hole intercept grade, stope face
sample grade, and the mill production reconciliation report grade
(this study was based on the annual production summary
reports). The assignment of measured and indicated resources
related to a combination of sample density and known historical
development. In the case of measured resources, this required
recent drill hole intersections from the current program to
validate intact resource blocks. In the case of indicated resources,
this was a combination of historical data which in some cases
had development on all three sides of the blocks, but may not
have had a recent drill hole to confirm the block as being intact.
It is important to note that 100 per cent of all development and
stope areas down to the No 10 Level have been accessible. All
these voids were surveyed to be exactly as the historical records
had indicated and on this basis, we would expect that all
indicated resources blocks would be intact. Based on the fact that
all of the results from the current program drilling, underground
surveying and sampling are in full agreement with the GMA
records, it is believed that the historical data has been validated
and that it complies with JORC guidelines.
The key outcome is that drill core intersections are an
excellent guide to the location and orientation of the reef
structures, the presence of key indicators and gold. However,
drill hole grades greatly underestimated the gold resource, and
only later development sampling identified the production
opportunities. Key indicators are described in Goodz et al (1997)
as a combination of a preferred sulfide-carbonate-mica alteration/
mineralisation mineral assemblage in association with structural
features such as planar orientation, foliation, and position to dyke
hostrock morphology (Figure 2).
This study identified resource outlines (polygons) using a
combination of drill hole intersections and development headings
to define the planar orientation of the reef zones. Geological
interpretation was carried out on a series of 1:1000 scale cross-
sections spaced at 15.24 m intervals connecting composited
intersections greater than or equal to 0.5 g/t Au (an inflection
TABLE 2
Grade sensitivity on underground gold resource.
Gold sensitivity for gold resource – comparison of production versus stope versus ddh intercept grade assignment
Name Measured resource Indicated resource
Tonnes Grade Koz Tonnes
(Kt) (g/t) (Kt)
Prodcution grade 21.6 20.90 14.51 259.4
Calculated stope grade 21.6 11.56 8.03 259.4
Drill hole intercept grade 21.6 9.62 6.68 259.4
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
RESOURCE ESTIMATION AND GRADE ASSIGNMENT
was observed at approximately 0.5 g/t Au from statistics carried
out on raw data and therefore was used as the mineralisation
cut-off for geological interpretation). Software Gemcom GEMS
single cut-off value compositing module was used to calculate
composite intervals greater than or equal to 0.5 g/t Au with a
maximum waste inclusion of 2 m. Each composite interval is
made up of a series of consecutive intervals from the raw assays
with values greater than or equal to 0.5 g/t and including up to
2 m waste, and using 0.0 g/t Au for missing samples and
ignoring missing intervals. Assays were extracted within the
interpreted geological solids. Solids were created by extruding
the polygons 7.62 m either side of section. Raw assays were
extracted within these solids for geostatistical analysis.
Figure 6 shows the comparative variation between the solids’
models. Geological solids were created by either tying geological
outlines together between sections (Figure 6a – Achilles reef
zone) or by extruding the closed lines into solids 7.62 m either
side of the cross-section plane (Figure 6b). Mineral resource
calculations used solids created from extruded polygons only.
From the perspective of volume calculations, both methods yield
essentially the same result, but from a sectional analysis, the
extruded solids are more representative of geological
descriptions of the vein sets which follow and intersect along
multiple orientations (Figure 6d) rather than a single planar vein
(Figure 6c). At intersection points, the stope widens to cover
both orientations, but then one orientation has been selected to
continue the narrow production stope.
In defining the resource outlines (polygons), both drill hole
intersections and development headings were used; however, we
have used either development sampling or production records to
assign the grades where available and deemed applicable.
Where production grades have been assigned, the bottom
(lowest) value for recovered grade was used for each period. This
was done because all reporting periods involved blending of
multiple feed sources and therefore the base grade was applied to
all these production areas. Utilising a mill recovered grade was
considered to be a further discounting of the gold resource.
Based on the tailings grades, the mill treatment process between
1934 and 1959 did not recover between 2.5 and 5 g/t Au. This
equates to a discount of 12 - 25 per cent based on an assignment
of recovered production grade instead of using a calculated head
grade.
Detailed metallurgical bulk sample test work in 2007 has
shown that all of the gold is free milling and recoverable
(Ammtec, 2007). Four programs were carried out, where 200 -
500 kg bulk samples were collected of quartz vein and
mineralised dyke hostrock. The samples were then blended and
composited to replicate mining scenarios that would blend ores
from different stopes. This study also looked at the blending
ratios of quartz to hostrock volumes to establish effects on
grinding/crushing properties of the ore types and consumption of
reagents. This work is ongoing in preparation for a feasibility
study on mill processing options.
Inferred resource Subtotal
Grade Koz Tonnes Grade Koz Tonnes Grade Koz
(g/t) (Kt) (g/t) (Kt) (g/t)
15.63 130.33 1733.6 10.43 581.58 2014.6 11.21 726.07
11.42 95.25 1733.6 5.90 329.09 2014.6 6.67 432.02
6.41 53.48 1733.6 2.15 119.90 2014.6 2.78 180.06
55
M D GOODZ, J REA and P JACKSON
FIG 6 - Achilles reef group in plan view (A and B) and composited cross-section view (C and D). The view is 700 m long by 150 m high.
Grade variation between average drill core sampling and
production sampling ranged from 1.85 g/t to 22.7 g/t Au.
Extensive sampling of mineralised (key sulfide-carbonate-mica
mineral assemblages) zones peripheral to gold-bearing quartz
veins consistently provided grades of ~3 g/t Au. A grade of 3 g/t
Au has been applied as a background development and stope
dilution grade, where data is unavailable to calculate stope
dilution. Application of minimum mill production grade, or stope
assay grade where available, had a sensitivity impact of varying
the overall resource grade range between 6 g/t and 11 g/t Au.
GRADE FACTORING
This study reviewed the largest reef deposits of the Morning Star
mine, comparing 438 drill hole intersections through these reefs
with 7570 face samples collected from stope development
headings (Table 1). The results show consistent increases in gold
grades, with increasing sample size and density, with stope
production grades being consistently 250 to 650 per cent higher
than drill hole intersection assays. Grade factoring was even
higher for pre-1940 production, which could reflect the small
size of early diamond drill core (being 2 cm core diameter),
combined with very high core losses (Goodz and D’Astoli,
1997).
The average drill hole intercept grade was between 1.8 and
7.2 g/t Au for the major production reefs in the immediate
stoping areas; however, these reefs had documented production
grades that averaged between 16 and 33 g/t Au. The grade
factoring study has shown that drill hole intercept assay results
greatly underestimate the resource grade for all major production
reefs studied.
This work has substantiated the use of diamond drilling as an
integral tool for identifying vein orientation, key indicator
minerals, and confirming that the veins are gold bearing, but
should not be used for determining an absolute grade value. The
variation between stope face samples and mill records is that the
average recovered grade is 32 per cent higher than the calculated
diluted stope grade. This could be related to grade control
management and/or not all mined material being processed in the
reporting period.
This phenomenon of grade factoring was also measured in a
series of bulk sampling studies were carried out on the Morning
Star surface deposit between 1995 and 1997. In these tests,
trench sampling indicated an average grade of 1.9 to 2.2 g/t Au,
but bulk metallurgical testing yielded grades of 2.6 to 4.1 g/t Au.
56 Ballarat, Vic, 14 - 15 October 2008
TONNAGE IMPACT
The continuity of geological structures and key mineral
indicators provide for a high order of confidence on tonnage
calculations. Solid modelling using extruded polygons has the
risk of stretching true thickness. This has a large impact on
tonnage calculation for narrow vein mining. In the case of the
Morning Star mine, this study was able to carry out an analysis
of the solids generated from extruded polygonal outlines
generated by GEMS with the actual stope surveys. Some
discrepancies occur where interpreted solid deviates from
average reef orientation. Needling patterns perpendicular to each
individual reef zone were used to estimate a true thickness for the
reef zone using the average dip and strike for each reef. When
this was carried out, it substantiated that the computer modelled
outlines had discounted design thickness by an average of eight
to 32 per cent (depending upon reef plane orientation). This
provided for further upside to tonnage figures (see Table 3).
KEY OUTCOMES
In 1997 Goodz et al carried out a preliminary investigation and
generated a 3D-model based upon ~10 per cent of the GMA
archival data. This review was based upon approximately 50 per
cent of GMA archival production records, a drill hole database of
508 holes (~450 diamond), underground refurbishment and
access to all mine workings down to the No 10 Level of Morning
Star Main Shaft workings (305 m vertically below surface [RL
1400 m]). Several underground stopes were mapped and
surveyed and four bulk sampling and metallurgical test work
programs were carried out.
The outcomes were:
• GMA records were of excellent quality and up to date.
• Ground conditions were open and stable, verifying historical
reference on the long-term stability of mine openings in the
dyke hosted mines.
• Bulk sampling grades matched or exceeded expectations
from historical records, with multiple occurrences of visible
gold in existing headings.
• Geological exposures of both quartz vein and mineralised dyke
host rock confirmed gold distribution and geological models.
• The host dyke body remains open in all directions and has
continued to be expanded in size through current drilling
programs.
Narrow Vein Mining Conference

TABLE 3
Tonnage impact validated by comparison with traditional historical stopes.
Comparison of modelled reef thickness versus actual stope thickness measured perpendicular to the reef plan (true thickness)
Ref name Polygon design dhole Stope actual true thickness
calculated true thickness (m)
(m)
Achilles 1.26 1.39
A-floor 1.17 1.4
Burns 1.58 1.85
Campbells 1.37 1.82
Stirling 1.29 1.42
V-floor 1.05 1.2
Whitelaws 1.87 2.03
• All significant historical production reefs (so far reviewed)
have been shown to also be open, from both current drilling
and assessment of underground development headings.
• High-grade reefs in the upper GAP Zone, not developed after
1940, have been validated through face sampling of
development and stope faces.
• New high-grade reefs have been discovered in the upper
GAP Zone, with three reefs (Kenny, Maxwell and Sydney)
having multiple visible gold drill hole intersections, and
adequate geological confidence to be described as newly
named structures.
• Several other gold-bearing new reefs have been intersected,
and modelling and naming of these reefs will be carried out
in due course as drilling proceeds.
• Continued historical data review and modelling, combined
with new drill detail, has lifted the confidence in the usability
and value of historical data and has established some
baseline studies on grade factoring.
• The 200 - 500 kg bulk sampling and production data review
supports that larger sample size studies yield higher recovered
grade, and that there is a consistent upgrade in gold
concentration values from drill hole samples to production
samples. This was also observed in surface sampling when
comparing hand sampling with costean sampling.
• Sampling and data review has shown that drill hole
intersections of major reefs have average grades of 3.69 g/t
Au, which increase to average stope sample grades for the
quartz reefs of 18.1 g/t Au and average production grades of
26.5 g/t (based on a study of the ten major stopes in
production between 1934 and 1959 [n = 8008]).
• These geological models are now being applied to a variety
of regional targets within an 11 km radius of the Morning
Star mine site. For many of these dyke hosted deposits, this
activity makes the first new geologically driven modelling
since these mines operated 50 - 100 years ago.
IMPLICATIONS FOR THE FUTURE
The next step is to complete the Maxwell Reef trial mining project
commenced in August 2008. The purpose of this study is to
directly compare the drill hole intercept grade with that from
progressive face and stockpile sampling, and then to reconcile
with 1000-tonne stope parcel production (recovered grade). The
results will be compared with previous observations and historical
data to validate methods for estimating grade for public reporting.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
RESOURCE ESTIMATION AND GRADE ASSIGNMENT
Tonnage upside Reef plane orientation
(%) (GDA94_north) azimuth/dip
(right hand rule)
10.32 316.02/-49.74
19.66 335.65/-48.88
17.09 285.00/-28.82
32.85 304.76/-30.78
10.08 299.50/-49.62
14.29 324.13/-62.34
8.56 207.13/-12.38
This will be detailed at the conference, as the mining project has
only just commenced and data collection/review will occur over
the next two months (during September/October 2008).
We believe this methodology will provide the basis for grade
assignment going forward and will reduce the risk of
misinterpreting the nugget effect. It is imperative that a full cycle
case study is carried out to understand the grade distribution and
sampling methodology required to validate methods in line with
the JORC guidelines.
Study of the bi-modal grade distribution within the quartz vein
sets and the host dyke rock enrichments is imperative to
successfully applying a dilution model and understanding
optimal block sizes and proposed methods for extraction.
ACKNOWLEDGEMENTS
The authors wish to thank Morning Star Gold NL for open
access to the underground mine workings and unpublished
company data, and their support of this work. Thanks also to
Robin Lonsdale for critical comments.
REFERENCES
Ammtec, 2007. Unpublished company reports Nos 10860 and 10976,
Ammtec Limited Independent Metallurgical Test Laboratory, Perth.
Campbell, J D, 1990. Hidden Gold: The Central Norseman Story, vols 1
and 2 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Goodz, M, 1994. Survey of mineral resource potential for the Shire of
Mansfield and environs, Geological Survey of Victoria UR1994/17,
39 p, Department of Local Government, Melbourne.
Goodz, M D and D’Astoli, D J, 1997. Diamond drill core versus RC
chips: The real sample, chapter 2, in Proceedings Third International
Mining Geology Conference, pp 73-79 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Goodz, M D, Harvey-Kelly, F E L, Cusack, C J and Kentwell, D J, 1997.
Discovering new orebodies – Creation of an exploration database,
Morning Star gold mine, Vic, in Proceedings Third International
Mining Geology Conference, pp 175-180 (The Australasian Institute
of Mining and Metallurgy: Melbourne).
Mason, A A C, 1994. No Two The Same, pp 51-208 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Schwartz, M O, 1986. Interpretation of the Conolly contour diagrams of
fault-related veins, Mineralium Deposita, 21(1):1-11.
VanderBerg, A H M, Cayley, R A, Willman, C E, Morand, V J, Seymon,
A R, Osborne, C R, Taylor, D H, Haydon, S J, McLean, M, Quinn,
C, Jackson, P and Sandiford, A C, 2006. Walhalla-Woods
Point-Tallangallook special map area geological report. Geological
Survey of Victoria report 127, GeoScience Victoria, Department of
Primary Industries, Melbourne.
57
The Ballarat East Goldfield – New Insights on an Old Model
D J Osborne1,2
ABSTRACT
Historical exploration of the Ballarat East goldfield defined two primary
quartz vein lode types. Tracking the intersection of these lodes with
vertical lines of enrichment (or so-called ‘indicators’) was the basic
practice for historical mining. This strategy proved highly successful in
terms of identifying quartz veins associated with mineralisation.
However, it provides little understanding of the timing or processes
involved in the formation of these veins, or the associated emplacement
of gold.
Current underground mapping shows the Ballarat East gold deposit can
be subdivided into four stages of vein formation. The earliest veins (V1
and V2) strike north-south and formed during bedding flexural slip and
limb thrust fault propagation in parasitic chevron folds. V3 veins are
subhorizontal and relate to further fold tightening involving a transition
from east-west compression to north-south extension that initiated
re-shear of bedded slip surfaces and additional strike-slip faults. Sporadic
V4 veins overprint the shallow-dipping sections of V2 and V3 veins and
are associated with weak north-south compression.
The V4 veins mineralogy is distinguished by a high carbonate content
and sphalerite-galena sulfide assemblage. Vein sampling indicates a
significant increase in gold mineralisation where overprinting by the later
deformation/veining has occurred. As a result of these observations, a
new model is emerging that identifies the chronology, structural control
and spatial distribution of respective quartz vein types. This is a major
advance in terms of exploration targeting, resource modelling and mine
planning.
INTRODUCTION
The formation of primary gold deposits within the Victorian gold
province is linked to extensive lithospheric shortening, subcrustal
melting and exhumation during accretion of the Western Lachlan
Fold Belt to the proto-Australian continent. In recent times
geochronological investigations of the relative timing of
deformation and associated gold emplacement have proposed
their occurrence as series of discrete eastward progressing events
(Foster et al, 1998; Ramsay et al, 1998; Bierlein et al, 2001).
The model of Bierlein et al (2001) for example, postulates
gold formation throughout the region is constrained to three gold
forming episodes characterised by differing structural controls or
an association with felsic intrusions (Figure 1). The earliest and
most significant of these occurred in the period between 455 and
435 Ma; during regional folding and thrusting in response to
east-west compression. The second episode of mineralisation at
about 420 to 410 Ma, is linked to intermittent variations in the
regional farfield stress that led to the introduction or
remobilisation of gold into reactivated or newly formed
structures in some deposits. A final episode of mineralisation
related to extensive lower crustal magmatism and middle to late
Devonian granitic plutons occurred around 380 - 370 Ma.
An understanding of the geochronology and paragenesis is
therefore considered vital for successful delineation of resources
(Wilson and Dugdale, 2006) and there are also implication for
efficient extraction and processing. However, this depth of
understanding has yet to be achieved in the majority of Victorian
gold deposits.
1. GAusIMM, Research Geologist, LGL Ballarat Goldfields Operation,
PO Box 1228, Bakery Hill Vic 3354.
Email darren.osborne@lglgold.com
2. The University of Melbourne, Vic 3010.
Email: d.osborne@pgrad.unimelb.edu.au
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
The Ballarat East goldfield is an example, where, despite some
evidence of multiple deformation and felsic intrusion related
mineralisation (eg Baragwanath, 1923; Forde and Bell, 1994;
Bierlein et al, 1999) the observations of the ‘old timers’ are
poorly constrained in terms of their significance with regard to
the emergence of regional chronological models.
The current mining operation of Lihir Gold Limited Ballarat
Goldfields Operation (LGL-BGO) is complimented by a
diamond drilling program that has evolved over the past decade
to generate a high quality three dimensional stratigraphic,
structural and resource model. In light of this information, this
paper aims to compare some of the historical accounts of gold
mineralisation at Ballarat East with the current observations of
gold bearing veins.
GEOLOGICAL SETTING
The Ballarat Goldfield is a vein hosted orogenic gold deposit
situated in the Bendigo Zone of the Lachlan Fold Belt (Figure 1).
The field consists of three separate deposits, being in order of
historical output; Ballarat East, Ballarat West and Nerrina
goldfields (Finlay and Douglas, 1992).
The Ballarat East deposit has a strike-length of approximately
10 km and occurs within the Ballarat Anticlinorium. This is a
north trending domal fold culmination with an amplitude and
wavelength of about 2 km (Figures 2 and 3). The Anticlinorium
is made up of numerous doubly plunging parasitic folds that host
the gold bearing veins. These are close to tight, chevron folds
with axial and form surfaces variably upright/symmetric or
inclined/asymmetric relative to their position in the hinge zone of
the regional scale folds (Taylor et al, 1996). The east verging
First Chance and Sulieman anticlines were the most important
hosts to mineralisation historically.
The host stratigraphy consists of Lower Ordovician marine
turbiditic sediments of the Castlemaine Group (VandenBerg et al,
2000). Only the basal Lancefieldian age sequence of the group
outcrops locally. This is a sparsely fossiliferous interbedded
greywacke, mudstone and minor black shale succession
characterised by variable sandstone/shale ratios (ranging from
10:1 to 1:1) and intervals dominated by mudstone (Taylor et al,
1996).
Facies associations (Boucher, 2004) suggest that the local
stratigraphy comprises lower turbidite fan overbank facies,
consisting of shale topped, thinly bedded, fine to medium
sandstone. Less frequent thicky bedded, medium to coarse sand
channel facies packages also occur, displaying abrupt top and
bottom boundaries of hemipelagic mud/shale. The maximum
thickness of sand or shale beds is up to ten metres with the
thickest examples informally known as the ‘Big Sandstone’ and
the ‘Big Slate’. The Castlemaine Group is estimated to have a
total thickness of ~3 km and is underlain by a basement inferred
to consist of Cambrian age oceanic basalts that abut a
Proterozoic continental fragment known as the Selwyn Block
(VandenBerg et al, 2000; Cayley et al, 2002).
Pervasive shortening (~65 per cent) and structural thickening
(~300 per cent) of the Bendigo Zone turbidites is thought to have
occurred via east-west compression in the period of 455 to
439 Ma during the Benambran Orogeny (Gray and Foster, 2004;
Foster and Gray, 2007). This deformation resulted in the
formation of regional scale folds and subcrustal faults that are the
host and possible pathways for mineralisation (eg Ballarat
Anticlinorium, Williamsons Creek Fault; Figure 2).
59
D J OSBORNE

HF MELBOURNE AUSTRALIA
A AF BENDIGO ZONE
ZONE
CaF MuF WF
STAWELL ZONE Fosterville
Tarnagulla Bendigo Toolleen
LF
MF

CF

Kingston
NSW

PF
Stawell Maldon 37S Victoria
v
Castlemaine
v Percydale
v Mt Piper
v v

v
Lachlan Fold Belt
v
v
Ararat Clunes
Moyston Daylesford Great Rand
Mafeking

MW
v
v v

FZ
GRAMPIANS-
STAVELY ZONE v Ballarat Steels Ck Morning Star
v v
v v
144E A1 Mine
Mt Stavely WCF
v

MF - Moyston Fault, CF - Coongee Fault, CcF - Concongella Fault, LF - Landsborough Fault, PF - Percydale Fault, AF - Avoca Fault, CaF - Campbelltown Fault,
Melbourne
MuF - Muckleford Fault, WCF - Williamson Creek Fault, WF - Whitelaw Fault, HF - Heathcote Fault, MWFZ - Mount lWe lington Fault Zone
Town/ field locality 380 -370 Ma gold
Discon

440 Ma orogenic gold (>30 tonnes Primary Au) 440 Ma orogenic gold (>10 tonnes Primary Au) 420-400 Ma gold
Walhalla
Mortlak

Late Cambrian sediments Structural trend

tinuity

Cainozoic basalt
e

(with & without bedding formlines)

Late Silurian and
Silurian Grampians Group Devonian rhyolite Paleozoic
Cambrian calc-alkaline post-collisional
v v
(with bedding formlines) reverse fault
association (Mount Stavely Volcanic Complex)
Late Devonian pluton
Moornambool Metamorphic Complex Ordovician to Devonian turbidites Fault
(Early Cambrian sediments) (with bedding formlines) Pluton age boundary
SCALE
Ordovician turbidites Early Devonian pluton
Proterozoic-Cambrian volcanic rocks (with bedding formlines) 25 km

B
298
CARBONIF.

STAWELL BENDIGO MELBOURNE open folding

ZONE ZONE ZONE
354
chevron folding
L
DEVONIAN

369
Au
M
major thrusting
384 arrow
Au Au
shows
E sandstone minor thrusting vergence
400 direction
shale
410 ? fault reactivation
quartz Au
(Au)
SIL.

sandstones quartz - rich

Au turbidites
426 Au quartz - rich minor reactivation
turbidites /thrusting
434
siltstone
sandstone
regional
439 shale metamorphism
Au
ORDOVICIAN

Au
contact
450 metamorphism
quartz - rich ?
turbidites
granitic
quartz - rich
plutonism
?
turbidites
quartz - rich
turbidites cherts Au
gold bearing
490 quartz veins
L
498 mafic volcanics mafic volcanics
CAMBRIAN

500 mafic volcanics

M andesites
509 dyke intrusion

E
Avoca FZ Heathcote FZ Mt. Wellington FZ
545
440 Ma 426 Ma 410-390 Ma
550

FIG 1 - Regional geological setting and time space diagram of Victorian gold deposits. (A) Regional geological setting and distribution map
of Victorian gold deposits showing structural zones, stratigraphic and chronological features. (B) Time space diagram showing the sequence
of sedimentary, metamorphic, deformational, plutonic and mineralising events for the western subprovince of the Lachlan Fold Belt in
Victoria (after Miller, Wilson and Dugdale, 2006, fig 1, p 678).

60 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 THE BALLARAT EAST GOLDFIELD – NEW INSIGHTS ON AN OLD MODEL

LOCALITY DIAGRAM

760 000 m E
755 000 m E
750 000 m E
Ballarat Goldfield
LITTLE BENDIGO
VIC
Geological Map OR NERRINA
Bendigo GOLDFIELD (LBG)
BALLARAT MELBOURNE BALLARAT EAST
GOLDFIELD (BEG)

ORIUM
ANTICLINORIUM

ANTICLIN
Monte Christo
BALLARAT WEST
GOLDFIELD (BWG) Dimocks
Yorkshire
Temperence
MN
Nil Desperandandum
Lake Wendouree 5 840 000 m N

Black Hill

TO
CHRIS
IOUM

BALLARAT

MONTE

eek Fault
LINOR

Britannia United
North First Chance
B B’
ANTIC

Last Chance
Llanberris 1

Wlliamson Cr
Central Plateau Extended Sir Henry Loch
North Normanby
A A’
Band & Albion 9 Band & Albion 10

regional scale
Star of the East 2 LGL BALLARAT GOLDFIELDS
South Star

5 835 000 m N

of
RAT

intrusion trace
ON

BALLA
ALBI

WILLIAMS AUSTRALIA

Projected pre-
LEGEND

Quaternary Newer Volcanics

basalt flows

Late Mount Egerton

Devonian Granodiorite

Lower Castlemaine
Ordovician Supergroup

Cross course fault

Regional fault
5 830 000 m N
Regional anticlinorium

old hard rock mine

Current/recent mine Buninyong

Urban development

KINGLAKE RESOURCES

0 1 2

km

FIG 2 - Local geological map Ballarat Goldfield (from Taylor, 1998, fig 1, p 544. Section lines A-A’ and B-B’ are shown in Figure 3).

PREVIOUS INTERPRETATIONS OF Douglas, 1992). Unfortunately, many of the historical reports of

BALLARAT EAST mining are prone to ambiguities due to the informal
nomenclature and a lack of geological understanding. In recent
The Ballarat East goldfield produced ~47 tonnes of primary gold years, accounts of the Ballarat East goldfield have had to rely on
(average 10 grams/tonne) in the period from 1858 to 1918 when this old material (eg Hunt and Ransom, 1983; D’Auvergne,
nining ceased in response to the First World War (Finlay and 1990; Finlay and Douglas, 1992; Forde and Bell, 1994; Taylor et

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 61

D J OSBORNE

West A Ballarat Goldfield East

First
A Cross-Section Sulieman Chance A’
Guiding Star Consols Line Line
Albion Line Line Exchange Line
Line Yarrowee Creek Line(s)

400mRL

0mRL

-500mRL

-1000mRL
LEGEND
0 1km
Shafts Basalt cover
Access drives
Black slate beds
Faults
Mineralised quartz body Thick sand beds

West B Ballarat East Goldfield East

B Cross-Section B’
Natural Surface
NORTH FIRST CHANCE SHAFT

Alluvial Clays Gravel Drift & c.

Surface of Bed Rock

SHAFT

Note- This mine is located north of

a crosscourse which displaces
the Sulieman lode etc. horizontally
CHANCE

Spurs
Auriferous 352’ t
ke
SHAFT

t c
400’ ke Ja
E 425’ J ac her
D ) r t 434’
LO et he Lea
AN Jac
k at
EM r
Le
LI the 502’ 494’
FIRST

520’ SU ea
( L
et
E a ck
rJ
D t)
crosscourse LO ke
AN Ja
c
t he 592½’
EM er
LI ath
623’ L ea
SU Le
(
CHANCE

700’ G” 690’
PU
HE
t “T
ke 790’
J ac
r
he
at
Le
888½’
LAST

0 100m
986’

FIG 3 - Ballarat goldfield cross-sections. (A) Ballarat goldfield section line A-A’ showing first order anticlinoriums and major lines of
mineralisation hosted within parasitic anticlinal folds. Note the relative tightness and easterly vergence of the Ballarat east parasitic folds
(Sulieman and First Chance anticlines). (B) Section of line B-B’ showing historic mine workings and quartz vein lodes (from Brown and
Hogan, 1932, p 30).

62 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 THE BALLARAT EAST GOLDFIELD – NEW INSIGHTS ON AN OLD MODEL
al, 1996; Bierlein et al, 1998; Taylor, 1998; Bierlein et al, 1999;
Bierlein, Cartwright and McKnight, 2001; Wood and Large,
2007). The advancement of our geological understanding from
these contributions has thus been restricted by the ability to
decipher the important features from the outdated mapping and
interpretations of ore controls.
The historical reports suggest Ballarat East gold bearing quartz
veins have a dominant structural control, with most of the larger
orebodies formed by dilation of west-dipping thrust faults known
as ‘leather jackets’ or ‘fissure lodes’ (Gregory and Baragwanath,
1907; Baragwanath, 1923; Brown and Hogan, 1935). The
presence of thick, comminuted fault gouge was a generic features
of these structures and they were also associated with highly
weathered felsic dykes known as the ‘Sediment’ on the Sulieman
anticline and the ‘Pug’ on the First Chance line.
Leather jackets are reported to propagate primarily on the
eastern limbs of the parasitic anticlines as stacked lenses of up to
15 m thickness with widths up to 30 m and up to hundreds of
metre in length (Figure 3). Their continuity up and down dip is
limited by a tendency to dissipate into bedded faults when
reaching the next anticline or syncline (Baragwanath, 1923).
Typical geometry of a leather jacket is reported to have strikes
subparallel to slightly east of bedding and dips ~45° W. They are
associated with fault displacements in the order of 8 - 25 m and
the vertical interval of repetition between faults is approximately
100 m (Baragwanath, 1953).
Shallow east-dipping auriferous veins know as ‘spurs’ were
described as extending away for several tens of metres from the
leather jackets in a ‘pinnate’ arrangement (Baragwanath, 1953).
In addition to the spurs associated with leather jackets, flat veins
also occur elsewhere in sufficient abundance to be regarded as
another important lode type locally (Brown and Hogan, 1935).
These veins were historical termed ‘cross-veins’, ‘flat-makes’, or
simply ‘spurs’ (Baragwanath et al, 1932). The cross-veins range
from centimetres up to a few metres in thickness but may be
several decimetres wide and similar or greater length.
Historical models imply that leather jackets and the
cross-veins carry coarser and richer gold where intersecting
favourable slate horizons termed ‘indicators or ‘indicator beds’
(Baragwanath, 1923; Baragwanath et al, 1932) (Figure 4). The
cross-veins are also suggested to be locally enriched where offset
vertically by bedded reverse faults known as ‘slides’, ‘flucans’ or
‘piths’ and in the longitudinal dimension by oblique faults and
splays known as ‘cross-courses’ or ‘heads’ (Figures 4c and 4d).
An indicator is apparently less than 0.5 cm thick, yet generally
occurs within ‘indicator slates’ that represent favourable beds up
to a few metres thick (Figure 4e). The recognition of indicators
as a useful guide to gold mineralisation occurred in 1871,
whereby Ballarat East became known as the pre-eminent
indicator type deposit of the state (Baragwanath et al, 1932;
Finlay and Douglas, 1992). However, the difficulty of correlating
some indicators along strike caused earlier authors to debate
whether indicators represented specific sedimentary units or
were structural features (Baragwanath, 1923).
Other vein types are less reliably mineralised or may be totally
barren. For example thin, subvertical quartz veins, known as
‘vertical’ veins are reported to be a common occurrence, yet
these veins were regarded as predominantly un-payable (Lidgey,
1893; Gregory and Baragwanath, 1907; Baragwanath, 1923).
These thin vertical veins should not be confused with axial
planar fault hosted vein systems, historically know as ‘vertical
lodes’ (Baragwanath, 1923). Their provenance is uncertain; being
previously described as a form of saddle reef (Lidgey, 1894) or
‘breached anticlinal lode’ (Ransom and Hunt, 1984). However,
these were considered a minor class of structures and saddle
reefs or bedded reefs, such as those at Bendigo, are notably
scarce, though inverted saddle reefs and bedded legs (formed
around synclines) were worked at Ballarat West (Baragwanath,
1923).
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
The crucial aspect to successful mine scale exploration of
Ballarat East mineralisation is thus the ability to define the
respective structural controls and chronology of leather jacket
and flat vein lode types. Determining the relative distribution and
repetition of these veins also has obvious ramifications for
mining, as they appear to be different in terms of scale and
geometry. However, the prevailing uncertainties regarding the
occurrence of cross-veins and indicators has led some authors to
question their significance and suggest that indicators are not
necessarily associated with high gold grades (eg Bierlein,
Cartwright and McKnight, 2001) or are in fact some form of
syn-genetic source rock (Wood and Large, 2007). These
inferences are inconsistent with the extraordinary discoveries of
large gold nuggets associated with ‘ultra-thin’ indicators and
therefore suggest these features are inadequately understood.
TOWARDS A NEW GENETIC MODEL
The Ballarat East goldfield has been the subject of modern
exploration since the 1980s. Initially, Hunt and Ransom (1983)
undertook to review all available literature as a means to
document the geology and exploration potential. The review led
these authors to conclude that the most important structural
feature at Ballarat East is the asymmetrical nature of parasitic
chevron folds, although it was not ascertained at the time how
this feature related to the occurrence of mineralised veins.
A baseline 3D structural and stratigraphic model was also built
from the historical mapping at an early stage. This work during
the 1990s by Ballarat Goldfields NL resulted in the prevailing
genetic model that identified strong rheological and competency
contrasts within the host stratigraphy as a primary structural
control. Mineralised vein systems were proposed to occur mostly
where west-dipping faults transgressed thick sandstones within
the host stratigraphy because shearing of the adjoining shaly
units was required to accommodate deformation around the
thicker, more competent beds. Therefore, it was concluded that
the high dilation potential at the intersection of these bedded
shear zones or so-called ‘vertical shears’ with west-dipping faults
was a fundamental control to formation of a leather jacket, which
led to the targeting of this setting with respect to exploration
drilling.
VEIN CHARACTERISTICS OBSERVED FROM
TRIAL MINING
Resources are defined from 1m composited drill hole assay data
and are classified in the range of Inferred to Indicated resources
according to the JORC code (see Williams and Sykes this
volume). Classification depends on the density of drilling and the
level of confidence in the geological interpretation. For example
a decision to upgrade an Inferred resource (based on 100 m
spaced L2 drilling) to Indicated 1 may be made where there
is some evidence that the geological interpretation is robust (eg
consistency of data, previous mining in area, cross-cut mapping,
etc). In areas where the interpretation is dubious such as remote
or previously un-mined sections, subsequent L3 (50 m spaced)
drilling assays results are generally required prior to considering
any resource upgrade.
Samples comprising two 10 kg portions of the broken rock
from each face, as well as samples from individual veins are
assayed on a routine basis. This is to provide some indication of
the likely grade from each face for the purpose of comparison
with vein characteristic observed during mapping. However, at
present, the decision where to mine and at what scale is based
purely on the drill hole assay data. Ore mining is planned as
scale dependant, mechanical cut and fill or longhole open
stoping, with extent of proposed development reflecting the
modelled resource shape. Resource shapes are generally tabular
with moderate to steep westerly dips indicating west-dipping
fault hosted lodes are most common.
63
D J OSBORNE

A
Transverse Sections of “Indicators”
r
to to
r
r
ca 2 3 to 4 r
1 indi
in
di
ca
ica ca
to
nd di
i in

VEIN
A

QUARTZ
QU
AR
TZ

E
BL
YA
PA
B
“Victoria United” “New Normanby” 0
Llanberris N 4
(Looking North) (Looking South ) “New Speedwell C o ” Scale 5 Feet to an inch
Scale, 10 feet to an Inch Scale, 10 feet to an Inch

B 1 2 3
West East
East West East
West

LA M

IN
LA
I N

D IC
IN
MI
I N

AT
NA

AT
Cross

ED
D I

Vein
TE

OR
DI

Cros ld
No go or
DB

s Vein
C

or “

BE
ED

“F
C A

Flo loo
or”
DE
AT

DD
Cross r”
D

ED
Vein
T O

ld

WI
O R

or “Flo
QZ

QZ
Go Gold or” d
Gol

THI
VE
VE

N
R
IN

IN
C D Longitudinal Section along an Indicator
1 Nu
gg
HEA

ett
yg
South old Nu Surface
D

Nu gg North
gg ett
Shaft cutting several Cross veins

Go ett yg
ld Nu
CRO

yg
HEA

gg o
N

PITC ld
ug

HO ett o ld Nu
HEA

yg gg
ge

F CR
D

ett
SSC

O SS V ol
tt

EINS d
HE

yg
y

vein Nu
D

o
go

z OR gg ld
AD

Quart
O

“FLO ett
ld
INDICATOR

URS

Go ORS yg
ld ” old
E FA

Quar Plane of Indicator

tz ve
ULT

in

Indicator
2 E Relative distances of Nth Woah Hawp indicators to "The Indicator"
Name feet metres direction Indicated thickness (m) comments
Gold The Indicator 0 0
The Streaky Slate 8 2.4 East
Quart
z vein The Telegraph 7 2.1 West
Gold The Blocky Slate 20 6.1 West 6.1
The Western Indicator 90 27.4 West
PITH

The Mundic Grit 111 33.8 West in a 1.8 m slate

The Greasy Head Country 117 35.7 West
Quartz vein Gold The Flucan / 12' Slate 160 48.8 West 3.6
The 18" Sandstone 172 52.4 West 0.45
Gold The 2' Slate 174 53.0 West 0.6
The Big Sandstone 200 61.0 West 10 with several 6" and one 1" slates
The 4' Slates 236 71.9 West 1.2
The 7' Slates 270 82.3 West 2.1

Pith

FIG 4 - Historical cross-sections of auriferous veins relationships to indicators. (A) Sections after Lidgey, 1893. (B) Sections after
Baragwanath et al, 1932, p 31. Note that the quartz veins generally dip east and cross-cut the indicator. (C) Comparison between an
indicator and a pith; after Baragwanath et al, 1932, p 32. (D) Historical long section of an indicator showing the influence of perpendicular
structures to the occurrence of gold after Baragwanath et al, 1932, p 32. (E) Table of known indicator beds in the North Woah Hawp mine
after Baragwanath, 1923, pp 238-239.

64 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 THE BALLARAT EAST GOLDFIELD – NEW INSIGHTS ON AN OLD MODEL
During initial ore mining at Ballarat East (2004 - 2005) an
inclined gradient was maintained for drainage of strike drives. To
maintain contact with west-dipping faults, development tended to
proceed up-dip and therefore pass across bedding with the result
that these headings sometimes encountered a mass of thin,
cleavage-parallel veins. The sample grades encountered in these
faces regularly fell short of the reputed richness of leather jacket
lodes, being more consistent with the historical observation that a
preponderance of thin, vertical veins are usually unmineralised
(Baragwanath, 1923). The initial observations of gold were made
in thicker east-dipping veins where they refracted to a
subhorizontal angle across a black shale inferred to be an
indicator unit (the Mundic slate).
As trial ore mining progressed to lower levels (2006 - 2007),
the spatial and chronological relationships of this vein refraction
characteristic became clearer. In the WH218 level for example,
large masses of east dipping veins propagate as tension gashes
from a thin west-dipping fault (<5 cm) that offsets the veins by
up to a metre. The occurrence of visible gold and high-assaying
samples was again related to the vein rotation through known
indicator units (Four Foot and Seven Foot Slates). Owing to the
characteristic dip change observed in connection with gold, these
mineralised veins are now referred to as ‘rollovers’.
Mining on other levels (PR385, PR485) revealed larger
individual rollover veins, with the PR385 (SOD1) vein
comprising only a flat section that terminated in a hook like
bulge at the rollover point (Figure 5). This vein formed at a
transition to overturned bedding along the faulted contact of
thick sandstone and slate beds. The dip change of the rollover is
crossed by a west-dipping fault, but the fault in this instance is
intermittent and weak. Rollover veins thus appear to originate as
features of indicator beds but may merge into the steeper tension
gashes associated with west-dipping faults.
Subsequent headings (2007 onwards) showing contrasting vein
examples representative of west-dipping fault lodes (leather
jackets) are characterised by fault hosted quartz masses up to
subparallel to a gouge-filled west- dipping fault (Figure 5b).
Shallow east-dipping veins (spurs) also extend away from the
main body of quartz for several tens of metres in a more or less
planar geometry. Upright to east-dipping bedding as well as
dilatancy of the west-dipping fault appears to be characteristic of
orebodies of this type. The vein sulfide mineralogy and wallrock
alteration is typically arsenopyrite and pyrite.
Cross-cutting relationships of rollover veins were established
in the footwall veining of WH218 fault where the flatter
(high-grade) vein sections clearly post-date the steeper portions
(Figure 6b). These veins or vein sections are designated V3 and
V2 respectively, as V2 veins also have a normal shear component
which offsets earlier cleavage/bedding parallel veins (ie V1)
(Figure 6a). V2 and V3 veins are in turn overprinted by thin
(<5 cm) carbonate veins (V4) (Figure 6c). The V4 veins merge
into crustiform and stylolitic carbonate veinlets that are
associated with gold within the V2/V3 hosts. These veinlets
carry a distinctive mineral assemblage containing carbonate +
pyrite+ sphalerite + galena + coarse gold.
V4 veins/stylolites propagate on the dip change and
subhorizontal sections of rollover veins (V3) and also where
rollovers are offset by bedded faults and oblique fault splays/
cross-courses (Figures 4c and 4d). The movement directions
associated with these late vein forming events are identified as
graphitic/pyritic slickensides, which indicated strike-slip
movement on north-south striking faults and dip slip movement
on oblique faults (Figure 7). It has yet to be fully determined how
these vein generations relate to the timing of dilatant west-
dipping fault lodes, although sulfide textures such as rotated/
mantled arsenopyrite porphyroblasts (Bierlein et al, 2000)
suggest the west-dipping lodes are older; possibly coeval with
V2.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
INTERPRETATION OF STRUCTURAL CONTROLS
It is widely known that strain associated with chevron folding in
multilayered sequences is not distributed uniformly, but is
actually accommodated by flexural slip on certain beds (eg
Tanner, 1989). The irregular distribution of these slip surfaces is
also known to be localised at the major facies/rheological
contrasts of the sedimentary succession (Ramsay, 1974; Tanner,
1989; Fowler and Winsor, 1997). At Ballarat East it has
previously been suggested that these slip surfaces (or vertical
shears) are important structures envisaged to have had a long
history.
The key observation from recent work is that movement
indicators on eastern fold limb bedded fault structures show two
distinct stages of bedding parallel shear. The first stage is
evidenced by the shallow to moderate easterly dip of the V3
veins. The dip of these veins and their termination on bedding
contacts suggest they formed as extensional fractures via simple
shear associated with anticlockwise bedding slip during folding.
This sense of shear is characteristic of normal chevron fold
behaviour, whereby overlying units in the stratigraphic package
must be progressively thrust toward the hinge zone so as to
maintain straight fold limbs (Figure 8).
The second stage of bedded fault movement is evidenced by
the lineation and displacement on pith faults that offsets the V3
veins in an inclined west-over-east manner. This is a clockwise
sense of movement that is inconsistent with the expected sense of
transport during flexural slip folding. It occurs where folds are
tight and the bedding is overturned as a discrete package or the
entire limb. This rotation could simply be due to bedding drag
above or beneath a major west-dipping fault, however, this model
does not account for the vertical repetitions of rollovers distal to
large west-dipping faults (Figure 3b).
An alternative explanation for the observed slip reversal and
isolated occurrence of rollovers is that they are a function of
acute fold tightening according to the chevron fold model of
Ramsay (1974) (Figure 9). In this model it is suggested that
extension vein formation is a characteristic of flexural slip that
may be followed by a reversal of shear sense if interlimb angles
are tightened to critical angles approaching the fold lockup
threshold (Ø). This shear reversal point in time is referred to as
the changeover dip (Θ) between region one and region two type
flexural slip. Fold flattening linked to elevated viscosity and
pressure induced dissolution is the eventual strain
accommodation mechanism beyond Θ. Fold flattening also
provides an alternative explanation for the formation of late vein
stylolites (within isolated flat veins) as opposed to high fluid
pressure induced by fault valve behaviour (cf Sibson, Robert and
Poulsen, 1988).
The Ramsay model also proposes limb thrusts (west-dipping
faults in this case) propagate continuously as folds develop,
although their locus is restricted to thicker sand beds within the
succession. This accounts for the apparent variations in longevity
and scale between mineralised west-dipping faults and weaker,
smaller faults, which may have formed during the late shear
reversal event (Figure 9d).
There is an overlap of fold tightening and fault propagation at
Ballarat East, because the strain indicated by fold related rollover
veins and piths suggests further shortening must have occurred in
their formation beyond what was required to form a west-dipping
limb thrust fault. At this stage the kinematic analysis of
respective vein generations associated with rollovers indicates
V1 and V2 formed via east-west compression with vertical
extension (σ3), whereas, the V3 geometry indicates a transition
of σ3 to north- south extension (σ1 was east-west) and V4 weak
north-south compression (σ1 north-south, σ3 vertical).
65
D J OSBORNE

A
west 336’ LEVEl
Sections of east
PUG LODE 336 LEVEL
LAST CHANCE MINE
BALLARAT EAST
0 10m

yke
ic d
fels
433’ LEVEL 433’ LEVEL

e
TRANSVERSE SECTION
175 metres North of Shaft

s
TRANSVERSE SECTION

i
200 metres North of Shaft

r
494’ LEVEL
494’ LEVEL 494’ LEVEL
TRANSVERSE SECTION
115 metres North of Shaft

593’ LEVEL

west east

bedding

quartz lode
spur vein

0 5m leather jacket
Mako fault

C D
west thick sandstone east
indicator slate
pith fault

no scale
no orientation

reverse fault E
west east
rollover vein

bedding
0 2m 0 2m

FIG 5 - Cross-sections of auriferous quartz veins from the Ballarat East goldfield. (A) Historical cross-sections of the Pug Lode leather jacket
in the Last Chance mine (from Baragwanath, 1923, fig 58, p 219). Only access levels are shown, the bulk of the orebody was stoped. Note
the extent of shallow east-dipping spur veins on the 494’ level at the 175 m section. (B) WH386 level access wall photo-mosaic showing
intersection of Mako fault lode. The lode cuts across steeply east-dipping bedding and displays two prominent west-dipping faults (black
lines). The footwall fault is a zone of fault gouge (ie leather jacket), whereas the hanging wall fault is surrounded by quartz. A large,
extensional vein array extends away from the hanging wall similar to the historical section in (A). (C) Rollover vein from PR 385 level (south
ore drive face 017). The vein is associated with a bedded reverse fault and altered, overturned slate package, these features are analogous
to historically termed ‘indicator slates’ and piths. Note that the bedding is younging to the east and progressively overturned from west to
east in the image to dip at and angle of ~80° west. There is also a weak west-dipping fault with minor late offset. (D) Historical drawing of
stopes from the 827 ft 11 in level, Llanberris No 1 Mine (from Gregory, 1907, fig 21, p 30). Note the similarities of vein morphology, sense of
fault offset and west-dipping bedding/cleavage in the footwall between this vein and recent examples (C, E). This vein was described as
occurring in west dipping strata (see text for discussion. (E) Rollover vein in PR385 level south ore drive lift 1. Note this vein occurs in a
different stratigraphic position to (C).

66 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 THE BALLARAT EAST GOLDFIELD – NEW INSIGHTS ON AN OLD MODEL

west east
A
A
Indicator
North South

V 2 veins V 1 vein

0 10 cm 0 10cm
WH218

west east B
B South
North

V 3 vein

V 2 vein V 4 veins
0 10cm
PR385
0 10 cm

south north C
C
South North
V 4 veins

V 2 vein

Au
0 10 cm

FIG 6 - Cross-cutting relationships and geometry of quartz veins. 0 5 cm PR405

(A) Quartz veins WH217 level showing cross-cutting relationships
and variations in geometry betweenV1 and V2 vein generations.
(B) V2 and V3 vein relationship in WH 217 level. Note that the V2
vein has similar east-dip to vein shown in (A) and wispy V4 veins
are also present. (C) V4 relationship to V2 vein in WH189 level.
The east dipping V2 vein and has been reactivated along its
margin during formation of north dipping V4. Gold was locally
visible around all veins shown.
FIG 7 - Movement analysis of fault plane striations. (A) NE striking
dextral-reverse fault WH218 level. (B) NE striking sinistral-normal
fault PR385 level. (C) Minor NW striking reverse cross-course fault
PR 405 level. Arrows show transport direction of hanging wall.
The fault lineations are similarly graphitic/carbonaceous on all
examples although the resolved palaeostresses (grey = tension,
white = compression) are incompatible.

The variability in vein geometries, apparent fault transport and
isolated nature of late strain features is somewhat problematic for
an explanation of regional overprinting deformation events solely
due to change(s) in regional stress tensors. This is because faults
having a similar strike are seen to have opposing strike slip
movement (Figures 7a and 7b) and the V4 veins have been
observed to dip shallowly north or south in different areas of the
mine.
IMPLICATIONS FOR MINING
Current observations confirm that the Ballarat East deposit is
dominated by two distinct lode styles, comprising leather jackets
with associated spurs, and rollover/veins associated with indicator/
pith beds.
These lode types have been suggested to coexist in some
sections along strike (Figure 3b), but may also represent discrete
end members whose distribution could relate to different
stratigraphic/structural settings. For example, Gregory and
Baragwanath (1907, p 18) stated:
The mines in the northern part of the field may be
grouped together as of the leather jacket type.
The typical indicator mines are in the southern
half of the field and are east of the country
traversed by the leather jacket.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 67

D J OSBORNE

BULK STRAIN
σ3

10% shortening
1. Innitiation of folding with bedding-plane slip and incipient formation of saddle reefs

50% shortening
2. Lock-up of chevron folds, propagation of bedding plane faults and formation of thrusts

65% shortening

3. Progressive steepening of all slip surfaces and extension array development

accompanying fold tightening

FIG 8 - The structural evolution of chevron folds, limb thrusts and extensional veins during progressive shortening
(from Sibson and Scott, 1998).

This historic interpretation of segregated lode distribution is
inferred here to be a reflection of the northerly fold plunge and
increased shortening associated with rollover veins. This trend
conforms to the general fold plunge reversal evident in the centre
of the field as indicated from the Vulcan model (Figure 10).
Although seemingly complicated, the structural controls on
respective vein types may thus simply be a function of the fold
asymmetry and plunge, whereby these parameters governed
deformation to have occurred in one way or the other, such that:
• Upright/east-dipping bedding and shallow south plunging
hinges favour leather jacket orebodies. These are fault bound
and have a minimal plunge component if the fault and fold
are co-axial. They most likely propagate up-dip from hinge
closures of the thick sand units (eg limb thrusts). The extent
of dilatational mineralisation in these faults is also influenced
by intersections with vertical shears further limiting their
continuity down dip.
• The combination of overturned, vertically re-sheared bedding
on north plunging folds produces rollover veins. These veins
are possibly in arrays defined by steeply west-dipping
to vertical beds. They are more erratic, nuggetty veins,
governed in three-dimensions by the extent of indicator slates
/pith faults and oblique structures (Figure 3).
The historical miners were very successful at extracting the
various veins by following the indicators. Part of their
prospecting involved noting zones of enrichment for later
development that entailed raises/winzes and drives on subsequent
levels. Trial mining has shown that high-grade veins are
encountered periodically that bear a striking resemblance to
these historically depicted settings and flexible, small scale
mining is the most appropriate method to facilitate the removal
of this lode type.
Qualitative or ‘visual’ grade control as practiced by historical
miners also presents an option to direct the course of
development because of the patchy occurrence of gold.
Moreover, the availability of quantitative bulk sample is likely to
require several tonnes of ore to overcome sample bias from the
coarse nature of the gold, which may be impractical in the
timeframe of ore-drive advance. Experimental mining and
sampling programs aimed at optimising the extraction of these
veins may be developed in light of these new ideas regarding
their formation. For the time being however, the ability to
visually anticipate vein grade is purely conceptual and credence
is still given to the drill assay data even in the rollover type of
vein style.
Leather jacket lodes represent larger resources. This type of
orebody is amenable to mechanised mining although the

68 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 THE BALLARAT EAST GOLDFIELD – NEW INSIGHTS ON AN OLD MODEL

1.0

A
outer arc
Region 1 extension fissures
1
inner arc
C cleavage
shear viscosity
strain
2 moduli

dip è Ô
t1
sigmoidal en-echelon
extension fissures
t2

è = changeover dip, Ô = lockup dip Region 2

B D
extension
fissures

t2
t1
flow toward
t2
hinge zone
t2
t2
t2 west-dipping
fault (weak)
t1 pith fault rollover veins
t2

FIG 9 - Structural aspects of chevron folding in multi-layered sequences with contrasts in thickness and grain size; t1= thick sandstone,
t2 = otherwise thinner homogeneous layering (after Ramsay, 1974, fig 8, p 1745; fig 21, p 1750. (A) Deformation mechanism diagram for
chevron folds during progressive tightening. θ = changeover dip of regions 1 and 2 (see C, D), Φ = fold lockup dip. (B) Incipient limb thrust
formation resultant from accommodation of an anomalously thick layer (C, D) Potential development of small-scale structures in folds under
presence of fault gouge and generally poor ground conditions
conditions imposed in regions 1 and 2.
dictates the use of fibrecrete in the ground support regime. The
requirement therefore is to be able to predict the extent of these
orebodies with reduced geological exposure. The model
presented here suggests that large leather jackets will extend
up-dip from beneath hinge closures of the thick sandstones. It is
also suggested that a later generation of smaller west-dipping
faults occur quite regularly in high strain areas although these are
Big Sandstone not necessarily well mineralised.

CONCLUSION
Understanding how deformation has manifested is crucial to
N revealing the nature and timing of gold deposition; although this
insight has long been inhibited by the lack of direct evidence
Ú

from multiple underground exposures. The outcome of recent

0 500m
Blue Whale Fault
FIG 10 - Vulcan™ 3D model of the Ballarat East goldfield (as
of 2007). The model shows the major Blue Whale fault and Big
Sandstone form surfaces. The white planes are 100 m spaced
interpretative paper sections registered in 3D, which form the
basis of the model. Note the tightness of parasitic folds and
general plunge reversal indicated in the centre of the field. The
model has undergone continual updates in response to new drill
data since 2003.
observation is to provide insight with respect to historical
accounts through new observations of transient vein formation
and deformation that form a structural/chronological framework
for development of a more definitive genetic model. This model
shows that so-called ‘region one’ and ‘region two’ behaviours are
important criteria of the chevron fold-lockup process with
respect to gold mineralisation in late stage flexural slip related
veins, though not necessarily influential in earlier limb-thrust
settings.
The ongoing refinement of the Ballarat East Vulcan™ model
which includes surface modelling of the thicker sand beds, fold
axes and major faults is a valuable tool to visualise the relative
fold tightness, plunge and bedding dip trends. The model

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 69

D J OSBORNE
proposed here suggests that these parameters are useful guides to
the likely style of veining within a particular area. The absolute
age of various mineralised veins as evident from structural
relationships as well as the age and significance of felsic dykes at
Ballarat East are research questions that remain to be resolved.
REFERENCES
Arne, D C, Bierlein, F P, McNaughton, N, Wilson, C J L and Morand, V
J, 1998. Timing of gold mineralisation in western and central
Victoria, Australia: New constraints from SHRIMP II analysis of
zircon grains from felsic intrusive rocks, Ore Geology Reviews,
13(1-5):251-273.
Baragwanath, W, 1923. The Ballarat Goldfield. Memoirs of the
Geological Survey of Victoria, no 14, 257 p (Geological Survey of
Victoria: Melbourne).
Baragwanath, W, 1953. Ballarat Goldfield, in Geology of Australian Ore
Deposits, Fifth Empire Mining and Metallurgical Congress,
Australia and New Zealand 1953 (ed: A B Edwards) (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Baragwanath, W, Mahony, D J, Howit, A M and Kenny, J P L, 1932.
Guide for Prospectors in Victoria, second edition, 91 p (H J Green:
Melbourne).
Bierlein, F P, Arne, D C, Foster, D A and Reynolds, P R, 2001. A
geochronological framework for slate belt-hosted gold mineralisation
in central Victoria, Australia, Miner Deposita, 36:741-767.
Bierlein, F P, Arne, D C, McKnight, S, Lu, J, Reeves, S, Besanko, J,
Marek, J and Cooke, D, 2000. Wall-rock petrology and geochemistry
in alteration halos associated with mesothermal gold mineralization,
Central Victoria, Australia, Economic Geology, 95(2):283-311.
Bierlein, F P, Cartwright, I and McKnight, S, 2001. The role of
carbonaceous ‘indicator’ slates in the genesis of lode gold
mineralization in the Western Lachlan Orogen, Victoria,
Southeastern Australia, Economic Geology, 96(3):431-451.
Bierlein, F P, Foster, D A, McKnight, S and Arne, D C, 1999. Timing of
gold mineralisation in the Ballarat goldfields, central Victoria:
Constraints from 40Ar/39Ar results, Australian Journal of Earth
Sciences, 46(2):301-309.
Bierlein, F P, Fuller, T, Stuwe, K, Arne, D C and Keays, R R, 1998.
Wallrock alteration associated with turbidite-hosted gold deposits.
Examples from the Palaeozoic Lachlan Fold Belt in central Victoria,
Australia, Ore Geology Reviews, 13(1):345-380.
Boucher, R, 2004. Utilising sediments to constrain the structural
reconstruction of the Central Victorian goldfields, AIG News,
76:21-23.
Brown, G and Hogan, E J, 1935. Victoria: Gold and Minerals, 100 p
(Mines Department: Melbourne).
Cayley, R A, Taylor, D H, VandenBerg, A H M and Moore, D H, 2002.
Proterozoic early Palaeozoic rocks of the Tyennan Orogeny in central
Victoria: The Selwyn Block and its tectonic implications, Australian
Journal of Earth Sciences, 49:225-254.
D’Auvergne, P B, 1990. Ballarat East gold deposits, in Geology of the
Mineral Deposits of Australia and Papua New Guinea, pp 1277-1278
(The Australasian Institute of Mining and Metallurgy: Melbourne).
Finlay, I S and Douglas, P M, 1992. Ballarat mines and deep leads, Geol
Surv Vic Rep, 94:38.
Forde, A and Bell, T H, 1994. Late structural control of mesothermal
vein-hosted gold deposits in central Victoria, Australia:
Mineralization mechanisms and exploration potential, Ore Geology
Reviews, 9(1):33-59.
Foster, D A and Gray, D R, 2007. Strain rate in Palaeozoic thrust sheets,
the western Lachlan Orogen, Australia: Strain analysis and fabric
geocronology, The Geological Societ of America, special paper
433:19.
70 Ballarat, Vic, 14 - 15 October 2008
Foster, D A, Gray, D R, Kwak, T A P and Bucher, M, 1998. Chronology
and tectonic framework of turbidite-hosted gold deposits in the
Western Lachlan Fold Belt, Victoria – Results, Ore Geology Reviews,
13(1-5):229-250.
Fowler, T J and Winsor, C N, 1997. Characteristics and occurrence of
bedding-parallel slip surfaces and laminated veins in chevron folds
from the Bendigo-Castlemaine goldfields: Implications for flexural-
slip folding, Journal of Structural Geology, 19(6):799-815.
Gray, D R and Foster, D A, 2004. Tectonic evolution of the Lachian
Orogen, southeast Australia: Historical review, data synthesis and
modern perspectives, Australian Journal of Earth Sciences, 51(6):
773-813.
Gregory, J W and Baragwanath, W, 1907. The Ballarat East goldfield,
memoir no 4, 53 p (Geological Survey of Victoria, East Melbourne).
Hunt, F L and Ransom, D M, 1983. The geology and exploration
potential of the Ballarat East goldfield, special report prepared for
Eastern Copper Mines NL and Kinglake Mining Co Pty Ltd, 48.
Lidgey, E, 1893. Report on the Ballarat East goldfield, special report,
Department of mines, Victoria.
Phillips, G N and Hughes, M J, 1996. The geology and gold deposits of
the Victorian gold province, Ore Geology Reviews, 11(5):255-302.
Ramsay, J G, 1974. Development of Chevron Folds, Bulletin of the
Geological Society of America, 85(11):1741-1753.
Ramsay, W R H, Bierlein, F P, Arne, D C and VandenBerg, A H M, 1998.
Turbidite-hosted gold deposits of Central Victoria, Australia: Their
regional setting, mineralising styles, and some genetic constraints,
Ore Geology Reviews, 13(1-5):131-151.
Ransom, D M and Hunt, F L, 1984. A reinterpretation of the Ballarat East
goldfield, Geological Society of Australia Abstracts, 12:450-451.
Sibson, R H, Robert, F and Poulsen, K H, 1988. High-angle reverse
faults, fluid-pressure cycling, and mesothermal gold-quartz deposits,
Geology, 16(6):551-555.
Sibson, R H and Scott, J, 1998. Stress/fault controls on the containment
and release of overpressured fluids: Examples from gold-quartz vein
systems in Juneau, Alaska; Victoria, Australia and Otago, New
Zealand, Ore Geology Reviews, 13(1-5):293-306.
Tanner, P W G, 1989. The flexural-slip mechanism, Journal of Structural
Geology, 11(6):635-655.
Taylor, D H, 1998. Ballarat Goldfields, Geology of Australian and Papua
New Guinean Mineral Deposits, pp 543-548 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Taylor, D H, Whitehead, M L, Olshina, A and Leonard, J G, 1996.
Ballarat 1:100 000 Map Geological Report, 117 p (Department of
Agriculture, Energy and Minerals: Victoria).
VandenBerg, A H M, Willman, C E, Maher, S, Simons, B A, Cayley, R
A, Taylor, D H, Morand, V J, Moore, D H and Radojkovic, A, 2000.
The Tasman Fold Belt System in Victoria, 462 p (Geological Survey
of Victoria East Melbourne).
Wilson, C J L and Dugdale, L J, 2006. Gold mineralisation in western
Victoria: Its setting, new discoveries and techniques applied to
identify blind ore bodies, Australian Journal of Earth Sciences,
53(5):671-676.
Williams, K and Sykes, M, 2008. Historical challenges, modern solutions
at Ballarat East, in Proceedings Narrow Vein Mining Conference,
pp 229-236 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Wood, B L and Large, R R, 2007. Syngenetic gold in western Victoria:
Occurrence, age and dimensions, Australian Journal of Earth
Sciences, 54(5):711-732.
Narrow Vein Mining Conference
Significance of Geological Control on Assay Data in the Narrow,
High-Grade Gold Quartz Veins at the Sand Queen Mine, Comet
Vale, Western Australia
D Potter1, C Sheriff 2 and P Collins2

ABSTRACT TABLE 1
Historical production from Comet Vale goldfield, 1897 - 1948.
The Sand Queen gold mine is a typical Archaean narrow, high-grade gold
quartz lode deposit within a shear hosted system. Initial mining during Mine Period Ore treated Gold Grade (rec)
1904 - 1948 produced 181 659 ounces of gold from 248 564 tonnes of ore (t) (oz) (Au g/t)
milled at an average production grade of 22.7 g/t Au, which equates to a
minimum of 750 ounces per vertical metre. 1904 - 1920 205 705 167 302 25.3
Sand Queen/
The current mining operation is using the old Sand Queen main shaft 1928 - 1938 42 772 14 583 10.6
Gladsome
to access the Sand George deposit, which is a narrow quartz lode system 1938 - 1948 - 552 -
discovered in the same shear structure along strike to the south of the
Sand Queen mine workings. Current underground resources contain Sand King n/a 36 30.3 26.2
360 200 tonnes at a grade of 11.8 g/t Au for 136 500 ounces. Mining is Sand Prince n/a 95 12.2 4.0
scheduled to produce 20 000 ounces of gold annually.
Lady 1897 - 1906 6728 3261 15.07
Re-opening of the Sand Queen mine and underground development of
Margaret
the Sand George lodes has revived several grade control issues relevant to
a narrow, ‘nuggetty’ gold lode mining environment. These include the Lady Mack 1905 - 1906 363 174 14.89
dependence of assay grade of individual samples upon a number of Happy Jack 1901 - 1920 11 634 7088 18.90
factors, including how and where a sample is collected, how and where it
is analysed, and how grade data is used to estimate an ‘average’ grade. Lake View 1899 - 1932 1855 766 12.84
Ideally, the only way to achieve an acceptable estimate of ‘true grade’ is group
to mine and mill a large parcel of ore (eg >100 tonnes) from each lode or Total 269 000 193 700 22.3
ore type.
In a mining operation producing relatively small tonnages from a
number of separate ‘nuggetty’ gold lodes, availability of resources and
operational constraints may preclude intensive sampling for predictive
CURRENT MINING OPERATIONS
grade control. At the Sand Queen mine, one of the key issues has been to Underground mining operations are focused on the Sand George
establish a reliable system for estimation of gold grade with which to orebody, which is centred some 250 m south of the old Sand
reconcile milling data. Extensive use has been made of detailed
geological data to understand the grade distribution within individual Queen shaft (13 550N local grid). The Sand Queen shaft had
lodes and between lodes, so as not to rely solely on assay results. previously been dewatered and rehabilitated down to the four
level (152 m vertical depth) but most of the old workings are
inaccessible. This shaft now provides access to some remnant ore
LOCATION near the shaft and to the Sand George lodes.
The Sand Queen mine is located 2 km south of the abandoned The current mining operation has produced 5850 ounces of
Comet Vale town site, which is about 100 km north of Kalgoorlie gold from 24 370 tonnes of ore milled at a grade of 7.85 g/t Au.
in Western Australia. The Sand Queen shaft and headframe for Initial mine production included considerable amounts of
the current operations is adjacent to the Kalgoorlie to Menzies development ore and the grades are expected to increase as more
highway. stope ore is mined with the last two parcels averaging 10.7 g/t.

HISTORY Resource model

Gold was discovered at Comet Vale in 1897, but the main
historical workings, the Sand Queen and the adjacent Gladsome
mines, were not discovered until 1904 due to extensive aeolian
sand that masks the western extent of the area.
Until 1920, the Sand Queen and Gladsome operations were
mined as two separate entities before being sporadically mined
as a combined operation from 1928 to 1948. Total production for
the period 1904 to 1938 was 248 560 tonnes at an average
production grade of 22.7 g/t for 181 660 oz down to the six level
(243 m vertical) (Table 1). This production from all six levels
of the old mine (from a depth of 243 m vertical) equates to
approximately 750 ounces of gold per vertical metre over 330 m
of strike. In addition 550 ounces of gold was produced by
tributors during 1938 - 1948 and a further 5000 oz were recovered
by re-treatment of the tailings in the late 1980s.
1. Chief Geologist, Reed Resources Ltd, PO Box 177, Kalgoorlie WA
6430. Email: dpotter@reedresources.com
2. Reed Resources Ltd, 97 Outram Street, West Perth WA 6005.
In view of the distribution and number of data points, thickness
of wireframes and target block size it was deemed that ordinary
kriging would be the most appropriate method of grade
estimation. A variogram model was created from the spatial
continuity analysis which had the following characteristics:
• nugget: 2.14;
• sill 1: 7.57, range 1: 24;
• sill 2: 2.90, range 2: 190;
• nested spherical model;
• anisotropy – major/semi major 1;
• major/minor 5; and
• plunge: -65 degrees to south.
To aid in the Mineral Resource classification process a number
of estimation passes were performed through the resource block
model maintaining the anisotropy ratio of 1:5. At the completion
of each pass the resource classification flag was set; passes one

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 71

D POTTER, C SHERIFF and P COLLINS

and two equated to an indicated classification and pass three to

an inferred classification:
• cat 1 (indicated) on a 30 m × 30 m search radius with min
eight samples,
• cat 2 (indicated) on 60 m × 60 m search radius with eight
samples,
• cat 3 (inferred) on 90 m × 90 m search radius with four
samples, and
• cat 4 (unclassified) run on 200 m × 200 m search radius with
four samples.

TABLE 2
Open pit and underground resources (as at December 2006).
Current underground resources (premining)
Category Tonnes Grade Ounces
Indicated 190 500 12.1 74 400
Inferred 169 700 11.4 62 100
Total 360 000 11.8 136 500

Development and stoping is undertaken utilising airleg

drilling, 12B rocker shovels, and 2.5 t electrical loco motions,
with ore extraction via gallery stopes. The current workforce is
approximately 20 underground and surface personnel.
REGIONAL GEOLOGY
The Comet Vale project is underlain by mafic-ultramafic
volcanic rocks of the Ora Banda Domain and granitic rocks of
the Goongarrie Monzogranite to the west and the Comet Vale
Monzogranite to the north. The Ora Banda Domain is one of six
tectono-stratigraphic domains that make up the Kalgoorlie
Terrane (Table 3) and is host to several large gold deposits,
including the Ora Banda and Mt Pleasant gold camps.
The mafic-ultramafic volcanic and metasedimentary rocks,
and mafic igneous sills, within the Ora Banda Domain are
referred to as the Ora Banda Sequence. Comet Vale is on the
eastern side of the Ora Banda Domain, along a 1 - 5 km wide
arm that extends for about 30 km north of Menzies. This arm of
the Ora Banda Sequence, known as the Menzies Greenstone Belt,
is bound to the west by the Goongarrie Monzogranite and to the
east by the regional scale Bardoc-Menzies Tectonic Zone
(Figure 1).
Ora Banda sequence at Comet Vale
Mafic-ultramafic volcanic rocks in the Comet Vale area are a
continuation of the lower part of the Ora Banda Sequence,
though generally with a reduced thickness (Swager, 1994). The
mafic-ultramafic volcanic sequence at Comet Vale is divided into
FIG 1 - Regional geology and gold deposits of the Comet Vale area.
three formations that are correlated with the Missouri Basalt,
Walter Williams Formation and Siberia Komatiite (Table 3).
Only the Missouri Basalt and Walter Williams Formation crop
out in the vicinity of and along strike from the Sand Queen-
Gladsome mine.
The Wongi Basalt at the base of the Ora Banda Sequence
apparently does not crop out in the project area (Swager, 1994).
Younger formations such as the Big Dick Basalt and Bent Tree
Basalt (Table 3) may underlie the eastern part of the project area.
Missouri Basalt
The Missouri Basalt is the lowermost formation in the Comet
Vale area. It consists of massive, fine to medium grained
metabasalt (now mostly amphibolite) although this may be
porphyritic with plagioclase phenocrysts up to 2 mm long. The
thickness of the formation is not known but is of the order of
500 - 1000 m. The Missouri Basalt is conformably overlain to
the east by ultramafic rocks of the Walter Williams Formation
(Swager, 1994).
The Missouri Basalt may be divided into two distinct units, an
upper and lower basalt, based on geochemical data (Holla, 2001).
The boundary between the two units is about 200 m below (and
structurally above) the boundary with the Walter Williams
Formation. The lower basalt is enriched in Ti and Zr and
depleted in Cr relative to the upper basalt (Holla, 2001).

TABLE 3
Regional tectono-stratigraphic setting for the Comet Vale area.
Kalgoorlie Terrane Ora Banda Domain Comet Vale area Mineralisation
Felsic volcanic and sedimentary unit Black Flag group ? Under Lake Goongarrie Au
? Under Lake Goongarrie Au
Upper basalt unit + dolerite sills Grants Patch group
Big Dick Basalt Au
Ora Banda Sequence Siberia Komatiite Au-Cu, W, Ni ?
Komatiite unit Linger and Die group
Walter Williams FM Au, laterite Ni
Missouri Basalt Au
Lower basalt unit Pole group
Wongi Basalt

72 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference


Walter Williams Formation
This formation is dominated by olivine cumulate rocks (mostly
serpentinised dunite and minor peridotite) that crop out as a low
ridge to the east of the Missouri Basalt and east of the Goldfields
Highway. Dunite is dominant in the western part of the formation
whereas peridotite, with some pyroxenite and gabbro, is
dominant in the eastern (upper) part of the formation (Swager,
1994). The Walter Williams Formation is conformably (?)
overlain by a thick sequence of komatiitic flows that make up the
Siberia Komatiite (Swager, 1994).
The boundary between the Missouri Basalt and Walter
Williams Formation appears to be a zone of intense shearing
within talc-carbonate-chlorite altered ultramafic rock at the base
of the Walter Williams Formation.
Siberia Komatiite
The Siberia Komatiite is a thick sequence of komatiitic flows,
which conformably (?) overlies the Walter Williams Formation
(Swager, 1994). Although the rocks are totally serpentinised,
A-zone (spinifex olivine) and B-zone (cumulate olivine) are
recognisable and indicate an overall east-facing sequence
(Swager, 1994). Fine-grained, layered interflow sedimentary
rocks occur throughout the unit.
Granitic rocks
There are two granitoids within the Comet Vale area. The
Goongarrie Monzogranite forms the western margin of the
Menzies greenstone belt and underlies the western part of the
project area. The Comet Vale Monzogranite is a late tectonic
granitoid that has forcefully intruded the greenstone belt and cuts
into the eastern margin of the Goongarrie Monzogranite. This
massive porphyritic granitoid has a circular shape and truncates
the mafic-ultramafic volcanic sequence in the northern part of the
Comet Vale area (Figure 1).
The basaltic sequence has been intruded by several narrow,
plagioclase ± quartz porphyry dykes (or sills) that generally are
subparallel to the stratigraphy. Most of the porphyry dykes have
been emplaced at the top of the Missouri Basalt, within about
60 m of its boundary with the overlying Walter Williams
Formation. Some narrow porphyry dykes have also been
emplaced within the Walter Williams Formation.
Holla (2001) identified three suites of porphyry dykes at Sand
Queen, based on petrological data and immobile element
geochemistry. There are two feldspar-quartz porphyry dykes with
distinctive Al/Zr ratios, a quartz feldspar porphyry and a sheared
porphyry that has lower Al/Zr ratios when compared with the
other porphyries.
Structural geology
Comet Vale is on the eastern limb of a regional-scale, north-
south trending anticline: the Goongarrie-Mt Pleasant Anticline.
The Goongarrie Monzogranite is in the core of this south
plunging anticline, and is enclosed by rocks of the Ora Banda
Domain.
The mafic-ultramafic rocks in the Comet Vale area therefore
are part of the same sequence that hosts the Grants Patch, Ora
Banda and Mt Pleasant gold camps on the southern and western
limbs of this anticline.
Three significant deformation events are recognised within the
Comet Vale region that broadly correlates with the known
deformation history of the Eastern Goldfields Province (Table 4):
• regional upright folds and shears that formed during D2
and were accompanied by granitoid intrusion (eg Goongarrie
Monzogranite);
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
SIGNIFICANCE OF GEOLOGICAL CONTROL ON ASSAY DATA
• NNW-striking domain-bounding shear zones developed
during D3; the comet vale monzogranite is interpreted to
have intruded syn-D3; and
• NE- and NW-striking faulting occurred during D4.
Emplacement of the late tectonic Comet Vale Monzogranite
has caused bending of the Missouri Basalt and Walter Williams
Formation from a NNW-SSE trend to a more NW-SE trend close
to its southern margin. The Siberia Komatiite appears to have
been deflected in the opposite direction, with a resulting NE-SW
trend.
WNW-ESE trending structures host other smaller gold
deposits in the Comet Vale area, including the Lake View, Happy
Jack and Lady Margaret shears (Witt, 1990).
In contrast, the larger gold deposits at Comet Vale (ie Sand
Queen-Gladsome, Sand George mines) are in a NNW-SSE
trending and steeply west dipping shear zone that is subparallel
to the boundary between the Missouri Basalt and Walter
Williams Formation. This structure apparently is unrelated to
emplacement of the Comet Vale Monzogranite (Wesley, 2003).
Recent geological interpretation of underground mapping at
Sand Queen and drilling at the Sand George prospect has also
identified relatively flat lying thrust (?) faults that have a shallow
west to north-west dips.
Regional gold mineralisation
Gold mineralisation occurred late in the deformation history
exploiting structures and structural intersections developed
during D2, D3 and D4 events. Three structural styles
of Au mineralisation are recognised:
1. NNW-striking Comet Vale shear zone, with boudinaged
quartz veins, resulting in low-grade, gently N-plunging ore
shoots.
2. ESE-striking accommodation structures associated with
intrusion of the Comet Vale Monzogranite.
3. Intersection of NW- and NE-striking faults, which offset
NNW-striking structures by tens of metres in a dextral
dip-slip sense. High-grade S-plunging ore shoots occur at
these fault intersections.
MINE GEOLOGY
The geology of the mine has four major features (Figures 2 and 3):
1. Steep/moderately dipping NNW/N striking quartz lodes
that host the bulk of the economic gold mineralisation.
Currently 19 different lode positions have been modelled,
ten of which contain economic resources.
2. NE brittle – ductile faults.
3. NW ductile shear zones.
4. Flat quartz ± carbonate ± sulfidic vein sets, shears and
fractures.
Current production is focused on three main ore positions
(Figure 3):
1. the main lode and the adjacent Emain lode south of the old
mine where truck sample grades range from 5 g/t to
100 g/t +,
2. the main lodes at Sand George where truck sample grades
range from 1 - 20 g/t, and
3. the upper porphyry lode at Sand George where truck
sample grades average 15 - 20 g/t.
Additional quartz lodes are found in both the HW and the FW
to the above positions.
73

74
Regional deformation event
D4
Regional shortening E-W (Swager,
1997).
Dextral movement on NNW shears and
NE faults (Witt, 1993).
D3
Regional shortening E (NE)-W (SW)
(Swager, 1997) with forceful
emplacement of intrusives (Witt, 1993).
Strike-slip movement on regional shear
zones (Witt, 1993).
D2
Ballarat, Vic, 14 - 15 October 2008
ENE-WSW regional shortening (Swager, thrusting (Swager, 1997). D2 was
1997) and granite pluton intrusions
(Myers, 1997). Shortening resulted in
regional scale thrusting, upright folding emplacement and associated regional
(Witt, 1993) and late-stage transcurrent metamorphism (Swager, 1997).
fault movement.
D1
Compression and syn-tectonic
emplacement of granitoids
(eg Goongarrie Monzogranite) led to
thrust-stacking of greenstones and
granites with a south to north movement
direction (Swager, 1997). This was
followed by regional E-W extension
(post D1, pre D2) evidenced by roll-over
anticlines, synclinal basins with clastic
infill (Swager, 1997).
Narrow Vein Mining Conference
Regional deformation history and its affect on gold mineralisation at Comet Vale.
Eastern Goldfields
Oblique dextral/reverse faults (Witt,
1993).
Strike and reverse slip faults and en
echelon folds (Swager, 1997).
Regional NNW-trending upright
foliations and folds, and domain scale
contemporaneous with a major thermal
event represented by voluminous granite
Recumbent folds and major
thrust-stacking (Witt, 1993; Swager,
1997).
D POTTER, C SHERIFF and P COLLINS
TABLE 4
Ora Banda domain Comet Vale Affect on Au mineralisation
Movement caused reactivation on the NE- and NW-striking oblique thrust/slip Emplacement of Au-bearing
Bardoc Tectonic Zone and Moriaty Shear faults (eg Princess Grace thrust, south-plunging shoots along D4 fault
Zone. NE-fault set, NW fault set). intersections.
Mineralisation and vein emplacement in
west-dipping thrusts at Princess Grace
and Sand Prince West.
Syn D3 intrusion of the Comet Vale E-W mineralised structures north of Vein emplacement and mineralisation in
Monzogranite with accompanying Comet Vale generated by forceful E-W accommodation structures; eg Lady
narrow to broad aureoles of pervasive granitoid emplacement; compression of Margaret, Happy Jack, Lake View
deformation (Witt, 1992). adjacent rocks and development of deposits.
accommodation structures.
Strike-slip movement on Menzies and
Zuleika Shears (Witt, 1993).
Formation of the Goongarrie-Mt Pleasant Activation of Comet Vale shear zone in Syn-tectonic vein emplacement and
Anticline, which is cored by a large close proximity to basalt-ultramafic mineralisation at Gladsome, Sand Queen
granitoid dome (Goongarrie contact. and Sand George.
Monzogranite).
No obvious indications of D1 in the Ora
Banda Domain, although subhorizontal
movements along contacts between
greenstones and early granitoids may
have occurred (Witt, 1993).

FIG 2 - Main lode structures in the Comet Vale shear zone at the
Sand George deposit.
Gold mineralisation
Economic gold mineralisation at the Sand Queen mine is solely
within quartz boudains that vary from 0.1 m to 4.5 m wide
(average 1.2 m) and of highly variable lengths ranging from 5 m
to 50 m. En-echelon stacking occurs both laterally and vertically.
The gold mineralisation shows a shallow northerly plunge,
overprinted (?) by a steeper southerly plunge.
A number of different generations of quartz and associated
mineralisation are apparent. Evidence exists for early low-grade
buck quartz veins that have been overprinted and recrystallised
by later highly sulfidic higher grade quartz veins (Figure 4).
There also appears to be post gold mineralisation quartz veining
and felsic intrusions that ‘stoped’ the gold bearing veins.
Alteration haloes are extremely tight around individual veins
with 5 - 30 cm sheared contacts quickly grading into weakly
foliated wallrock. Five types of alteration have been recognised
underground:
1. proximal potassic consisting of biotite/sericite, silica and
pyrite;
2. proximal bleaching in basalts consisting of carbonate
(siderite?), silica, sericite and sulfides with quartz
carbonate stringers;
3. sericite, carbonate, chlorite bleaching in porphyries
proximal to the vein;
4. silica flooding and bleaching in porphyries resulting in
feldspar destruction; and
5. weak chlorite and carbonate alteration haloes (1 - 2 m).
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
SIGNIFICANCE OF GEOLOGICAL CONTROL ON ASSAY DATA
FIG 3 - Type cross-section (looking north) through Sand George.
Footwall contacts tend to be extremely sharper than the
hanging wall contacts, which tend to have 30 - 50 cm intensely
sheared margins with 1 to 10 cm wide quartz stingers. Further,
later veining appears to have intruded along the hanging wall
contact rather than the footwall. Gold grades in the shears and
the alteration haloes rarely exceed 0.5 g/t.
Key factors controlling grade distribution within the overall
lode system and within individual veins include (Figures 5 - 7):
• Lode positions occur between NW and NE structures. When
driving south mining hits NW structure then ore position
which is then bounded by NE structure to south. These
structures both control and offset gold mineralisation, ie they
have, in places, been reactivated.
• High-grade (HG) to very high grade (VHG) distribution
structurally localised within overall lode/shear system.
• Better grades occur when lodes meet (both vertically and
horizontally) and/or flatten (away from cross-cutting
structures).
• Lower grade (LG) occurs when lodes bifurcate, are dragged
into structures and/or change in strike.
• Better grade appear to preferentially occur along wall rock
contacts at the margins of quartz veins. It is postulated that
up to 80 per cent of the gold occurs within 30 cm of the
contacts of most veins.
75
D POTTER, C SHERIFF and P COLLINS

Vuggy massive
marcasite/pyrite

Folded quartz
vein thrusted
against earlier
veining

Boudains of
•early•quartz vein 28 g/t 2.07 g/t
with in lode

FIG 4 - Ore zone showing different generations of veining, Sand George main lode.

FIG 6 - Lode positions and grade distribution for the Sand

FIG 5 - Fault controls on lode positions and grade distribution,
George area.
2 Level, Sand Queen South.

• Lower grades are found in milky bucky quartz. However,

• Better grades (16 g/t +) are where the vein shows extensive
lamination and recystallisation (of previous quartz) with fine-
grain sulfides.
• Moderate to high grade (5 g/t to 16 g/t) is found where veins
show a green chloritic tinge due to wallrock absorption,
moderate recystallisation and lamination.
some of the best visible gold seen in the mine can be found
in these veins.
• Thus, grade is dependent on the amount of each generation
of vein type present.
• FW lodes do not have economic grades above intersection
with main lode (see Figure 3).

76 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 SIGNIFICANCE OF GEOLOGICAL CONTROL ON ASSAY DATA

Gold is mostly present as free gold and interstitial to quartz in

veins. It is spatially (and genetically?) associated with pyrite/
marcasite, pyrrhotite, sphalerite, galena and chalcopyrite.
However, there is not a direct positive correlation between the
amount of base metals and the gold grade. The gold shows a
remarkable bimodal distribution (Figure 8). On average
70 per cent of the gold is greater than 105 µm. This is reflected in
the excellent recoveries (up to 60 per cent) achieved by gravity
methods.

Sampling and assay techniques

Four main sampling techniques are employed:
1. Channel sampling – across faces and backs of the lodes
± wall rock. This is done using an air grinder with a four
inch diamond blade (where possible) to collect a 2 - 3 kg
sample across the whole of the vein. Channel samples
provide a cleaner more representative sample as it includes
minor dilution from wallrock when compared to chip
sampling.
2. Spot samples of individual features.
3. Truck sampling − 0.5 kg grab samples are taken from each
truck to make up an 8 - 10 kg composite sample.
4. A sample from each stockpile is dollied and panned as a
visual check.
All samples are processed the same way with the entire sample
crushed to nominal 95 per cent passing 75 µm and then split to
400 g before being assayed for gold using a four-hour Leachwell
technique. Fire assays on the residues have shown that the
Leachwell recovers 98 - 99 per cent of the gold.

Mining techniques
Current level development utilises the historical levels, which are
approximately 45 m apart. Where possible, drives along the level
are driven on ore, with each cut a nominal 1.8 m × 1.8 m × 2.0 m.
Sublevels are nominally 9 m apart from back to floor with 3 m
sill pillars; the first sublevel is established 3 m above the level
drive. Rises and mill points are positioned based on geology at
no more than 20 m apart. Thus, stope panels of no more than
20 m × 9 m can be defined with four sublevels between
extraction levels. The actual extraction sequence is based on
geotechnical and access constraints.
Due to the extremely sharp contacts, little overbreak is
experienced with the ore ‘peeling’ off the walls back to the
unaltered, unfoliated wallrock. Dilution levels are thus generally
less than 20 per cent with mining widths as narrow as 600 mm
achieved on 300 mm wide quartz lodes.

Grade control
This mining technique allows for systematic mapping and
sampling of development and stopes that enables individual areas
of the stope to be identified by geological and grade
characteristics. This can then be compared back to geologically
similar areas that have been previously mined to determine grade
potential. Areas of lower grades also tend to be structurally
complexity and unstable and can be predicted from development;
further defined from in stope mapping and left as pillars.
Thus, an area from a stope can be identified as potentially high
grade (>5 /gt) or low grade (3 - 5 g/t) based on geology and the
anticipated dilution rates. Pillars can then be left in areas of
lower grade and/or geotechnical poor area allowing for the
preferential extraction of the higher grade. Mill points are then
FIG 7 - Bulk sampling along the 3 level south ore drive, monitored visually to determine potential excess dilution and
Sand George. related back to the truck sample results.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 77

D POTTER, C SHERIFF and P COLLINS
30.0
25.0
% Au by distribution
20.0
15.0
10.0
5.0
4.9 27.8
0.0
+ 4.75 mm
- 4.75 + 2.36
- 2.36 + 1.18
mm
FIG 8 - Typical gold sizing analysis and bi-modal distribution for the Sand George lodes.
Individual truck samples are dumped separately on the ROM
pad until they are classified. Classification is based over a period
of time looking at all of the available data. QAQC practices are in
place for all stages of sampling with duplicate samples taken every
20 samples with blanks and standards submitted with each batch.
This is essentially a qualitative rather than quantitative
approach to grade control. The logic behind the approach is to
extract as much higher grade (ie >5 /gt) as possible, leaving the
low grade and waste behind. Current approaches for head grade
prediction for milling are based on looking at the geostatistics
from each area of the mine and predicting a grade range rather
than an exact number. Work is currently being done based on
harmonic sequences (eg log normal, binomial and Fibonacci) as
well as more conventional methods.
Reconciliations
Kingsrose, as operators, are currently undertaking a
reconciliation exercise to look at the various stages (ie model
versus defined (sampling) versus truck samples versus mill belts
samples versus mill feed versus actual) to define optimal
sampling and grade prediction methods.
As anticipated, anecdotal evidence suggests that truck samples
are over stating the predicted grade in parts of the mine due to a
sampling bias. This is partially due to differences in
fragmentation from blasting with the smaller sized ore fragments
being preferentially sampled over the larger sized waste
fragments. This becomes more of an issue in areas where waste
dilution is high, particularly during level development.
Currently, milling is done by exclusive toll treatment where
the plant is first fed high-grade ore, then low grade, with a final
flush of mineralised waste.
During toll treatment mill feed belt samples are taken off the
belt feeding the fine ore bin every five minutes to produce a one
hour composite sample of 10 - 12 kg. Samples are processed
exactly the same as the grade control samples (see section on
grade control).
Results for each individual ore category show remarkably low
variance, suggesting the sampling practice is working.
CONCLUSION
Gold grade distribution within this structurally complex deposit
is a function of vein type, timing, location within individual
veins and both local and mine scale structural controls. As such,
78
22.3 9.8 5.3 6.2 12.8 10.9
0.075 mm
0.300 mm
0.150 mm
- 0.600 +
- 0.300 +
- 0.075 mm
- 1.18 + 0.600
- 0.150 -
mm
mm
Size Fraction
individual assay results cannot be used to determine ore
categories alone. The determination of ore classification is
undertaken using a combination of face/spot sampling, truck
sampling, geological mapping and visual inspection of
drawpoints/stockpiles. This is purely a qualitative rather than
quantitative approach to grade where the primary aim is to
determine if a given stope panel is economic or not.
Although this appears to be successful, predicting the overall
grade for mill parcels is proving more problematic with the truck
samples tending to over state the grade. Hence, a detailed
reconciliation exercise is currently being undertaken to look at
the various stages, ie model versus defined (sampling) versus
truck versus mill belts versus mill feed versus produced.
A strong understanding of the geological controls on grade
variation will provide improved resource models, mining
methods, grade control and reconciliation. The work that has
been undertaken in the past has raised confidence in the
assessment of grade and will lower the risk in any future
planning and decision making.
REFERENCES
Holla, L A, 2001. Geological setting of gold mineralisation at the Sand
Queen-Gladsome mine, Comet Vale, Western Australia, Honours
dissertation (unpublished), Curtin University, 176 p.
Myers, J S, 1997. Preface: Archaean geology of the Eastern Goldfields of
Western Australia – Regional overview, Precambrian Research,
83:1-10.
SRK Consulting, 2008. Comet Vale Structural Targeting (unpublished),
Reed Resources Ltd, Western Australia.
Swager, C P, 1994. Geology of the Menzies 1:100 000 sheet, 1:100 000
series explanatory notes, Geological Survey of Western Australia.
Swager, C P, 1997. Tectono-stratigraphy of late Archaean greenstone
terranes in the southern Eastern Goldfields, Western Australia,
Precambrian Research, 83:11-42.
Wesley, G, 2003. Genesis of lode gold mineralisation at the Sand George
deposit, Comet Vale, Western Australia, Honours dissertation
(unpublished), Curtin University, 172 p.
Witt, W K, 1990. Geology of the BARDOC 1:100 000 sheet, Record
1990/14, Geological Survey of Western Australia
Witt, W K, 1993. Gold Deposits of the Menzies and Broad Arrow Areas,
Record 1992/13, pp 67-76, Geological Survey of Western Australia.
Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference
Resource Modelling and Mine Design at Obuasi Mine, Ghana
H Eybers1, J Visser2 and C K Boafo3
ABSTRACT
Obuasi Mine in Ghana, with more than 100 years of mining history,
developed from a surface operation into a deep >2000 m underground
gold mine. The challenge is to extract value from narrow veins left after
extensive mining in the wider ore zones of the mine. Since the merger of
AngloGold and Ashanti Goldfields in 2004 the mine strives to achieve
optimum extraction through an integrated Mineral Resource management
process from geological modelling, resource evaluation to mine design
and scheduling.
The shear hosted gold deposit displays complex secondary shears of
variable thickness and continuity. There is an ongoing process of changes
taking place in the way that geological modelling and resource estimation
is done. Recent developments include the use of computerised systems
that support the whole resource management chain from resource
modelling to reserve optimisation. The change in thought and technology
now allows for targeting narrow veins previously ignored in optimisation
and mine design since moving into mechanised mining. The increased
knowledge and confidence in the resource models now allows for
flexibility in mine design, selection of mining method and scheduling and
extraction of the reserves.
The focus area for change is the way the orebody is evaluated and
extraction designed. Since 2005 major changes were made to the
process. This paper considers the historical approach of resource
management. It looks at the individual process changes and qualifying the
way ahead.
OBUASI MINE
Obuasi Mine is located in south-western Ghana some 200 km
inland from the Gulf of Guinea (Figure 1). The mine has been in
operation since 1897 and currently produces around 296 000 gold
ounces per annum from underground operations. The average head
grade is 5.6 g/t. The declared underground mineral resource as at
December 2007 is 108 Mt at 8.9 g/t and 30.7 Moz. At the same
time the ore reserve is 38.3 Mt at 6.4 g/t and 7.8 Moz. Published
Mineral Resource and Ore Reserves are in accordance with the
minimum standards described in JORC (2004).
GEOLOGY
Regional geology
Obuasi Mine is situated along a north-easterly striking 300 km
wrench fault system known as the Ashanti volcanic belt
(Figure 1). This belt is the most prominent of the five Birimian
Supergroup gold belts found in Ghana and comprises
sedimentary and mafic volcanic rocks. The Birimian was
deformed, metamorphosed and intruded by syn and post-tectonic
granitoids. Folding trends are dominantly north-east. Major
faulting has taken place along the same trends.
The contrast in competency at the contact between the
metavolcanic rocks to the east and the more argillaceous rocks to
the west formed a plane of weakness. During crustal movement,
1. MAusIMM, Resource Manager, MRM Department, Obuasi Mine,
PO Box 10, Obuasi, Ghana.
Email: heybers@anglogoldashanti.com.gh
2. Head of Business Planning, Technical Services Africa Region,
AngloGold Ashanti Limited, PO Box 62117, Marshalltown 2107,
South Africa. Email: jvisser@anglogoldashanti.com
3. Chief Mining Engineer, Business Planning – MRM Department,
AngloGold Ashanti Limited, Obuasi Mine, PO Box 10, Obuasi,
Ghana. Email: cboafo@anglogoldashanti.com
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
this plane became a zone of shearing and thrusting. Gold
mineralisation occurs within these graphite-chlorite-sericite fault
zones. The shears are associated with silica, carbonate and
sulfide hydrothermal alteration and occur in tightly folded Lower
Birimian schists, phylittes and greywackes.
Local mine geology
Gold mineralisation within the Obuasi area has a proven
continuous lateral extent of about 8 km along strike and
continues to an explored depth of over 2000 m below surface.
The main shear dips steeply to the north-west at 70° and is
deformed resulting in an anastomising structural pattern.
Characteristic of this type of shear hosted deposit is the
development of strong pinch and swell structures. There are
variations in thickness, continuity, dip and strike of mineralised
shears. The result is a complex geological structure with
unpredictable geometry and grade distribution.
High-grade ore zones within the shears are surrounded by
barren or very low-grade shear or host rocks. There are two main
mineralisation types (Baidoe-Ansah, 1998). The main historical
source of high grade ore is from quartz veins associated with
coarse, visible free gold. The second type is disseminated
sulphides (arsonepyrite predominant) in metavolcanics and
metasediments. The ore zones are defined by the presence of
mineralised quartz or presence of sulfides.
The narrow secondary shears have become the focus of mining
and will form the core of any future deep level extraction.
RESOURCE MODELLING
Historical processes
For many years mineral resource models at Obuasi were based
on polygonal estimation methodology. The polygons were
constructed based on cross-cut mapping and sampling and fan
drilling between levels and cross-cuts. Cross-cuts are between
15 and 20 m apart.
A resource envelope was created from upper and lower levels
of the resource block and lateral extension was limited to the
extent of the sampling information. The polygonal created did
not cover the possible extend of the mineralisation. Average
thickness of the ore zone was implied, which is in contrast to the
known variability. Not all ore zones were considered in the
geological model for polygonal estimates. Some thin secondary
ore zones were excluded from the resource models.
Weighted average grades from cross-cut sampling and drill
holes were used in polygonal estimation. Low-grade intercepts
(below 3.43 g/t – which was based on 2 dwt/short ton) that are
geologically equivalent to the ore zone were not included for
grade interpolation.
The result was resource models with the recognised tendency
to be overestimated in terms of grade and volume and not
representing the geometry of the ore zone. This had a
downstream effect on tons and grade from mine design to
mining. It was felt then that on a mine wide scale the estimates
would be robust based on mass of data. This however did not
support the monthly grade control and reconciliation.
‘Operating experience’ allowed for suitable mining factors to
be applied for estimates of tonnage and grade. It was difficult to
accurately quantify mined material from the stopes and to
approximate the mined grade.
79
H EYBERS, J VISSER and C K BOAFO

OBUASI

FIG 1 - Birimian Supergroup hosted gold belts in Ghana.

The shortcoming of the resource definition impacted on the
ability to optimise the mining method, dilution, recovery, etc.
The Mineral Resource classification for polygonal estimates
was based on sampling spacing. For Measured Resource a
minimum sampling spacing of 30 m horizontally and vertically
was accepted. To be classified as an Indicated Resource, the
sampling spacing must have been between 30 and 60 m. A
sampling spacing of between 60 m and 120 m was considered as
Inferred.
Process turnaround
The importance of geology in the estimation of tonnage and
grade is highlighted in numerous publications over time (eg
Stone and Dunn, 1996; Dominy et al, 1999).
A comprehensive list of key aspects to consider when
computing a resource model was addressed by Dominy et al
(1999). They highlighted the importance of incorporating vein
geometry, effect of other geological features, grade distribution
and geological data. This all comes together in interpretation and
modelling of the geology. Together with the consideration of
resource estimation methods, mining and processing parameters
it forms the basis for a lower risk assessment of narrow veins.
At Obuasi Mine in 1995 a study was conducted to address the
understanding of geological controls on mineralisation. This was
followed up by a training and technology transfer program to
ensure that geologists effectively integrate the geological model
and methodologies in mining and exploration (Etheridge and
Henley, 1995).

80 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 RESOURCE MODELLING AND MINE DESIGN AT OBUASI MINE, GHANA
As recent as 2002, following a series of external audits and the
decline in gold production, there were moves to change the way
the resource was evaluated. The process change gained impetus
in 2004 when AngloGold and Ashanti Goldfields merged to
become AngloGold Ashanti.
Resource modelling changed from polygonal to block models.
The block models are now constraint based on the geological
model of the shears. The model is developed from extensive
definition and exploration drilling and development mapping and
sampling. Estimates are now based on geostatistics.
Drill holes
The drill holes are of the type core diamond drilling. Exploration
and infill drill hole spacing is typically 40 m along strike, 30 m
to 40 m vertically, and sampled at 1.5 m across strike. The
drilling process is being changed to allow for more definition
drilling from cross-cuts to define the geometry of the various ore
zones. The fan drilling pattern is aimed at closer intersections.
Drill hole spacings of 10 m along strike and dip are now the
target and the changed drilling program is being implemented.
Underground development sampling data
Development sampling consists of reef drive sampling and
cross-cut sampling. Reef drive sampling is not used in estimation
in situations where the full width of the ore zone is not exposed
in the reef drive. Cross-cuts are at intervals of approximately
15 - 20 m. Up to the end of 2007 cross-cut sampling produced
six lines of channel samples per cross-cut. Three horizontal
channel lines across the lode are sampled at 0.5 m interval on the
north and south walls respectively. The process has changed and
only one channel line is sampled on either side of the cross-cut.
The change is based on a variability study. The study showed that
the grade distribution for only one channel sampling is the same
as for all the cross-cut channel sampling. The variability does not
increase with a decrease in sampling. The added benefit of the
reduction in sampling is that extra effort can be made to take a
quality sample.
The sampling is now also extended throughout the entire
length of the cross-cut into the hanging and footwall where no
mineralisation is seen. This is done to identify areas of possible
mineralisation and to assist with geological interpretation. The
various lithological units within the cross-cut are identified and
as far as possible sampling is restricted to the various units
(ie quartz is not mixed with mineralised schist or vice-versa).
All development and drill samples are submitted to the local
mine assay laboratory fire assay analysis.
Database and systems applications
Historically all borehole logs were recorded on logsheets and
geological plans. Since 2002 an in-house borehole database was
used to capture borehole data. Only the sampled intersections
with associated depths, widths, location and assay results were
used for estimation and to identify the mineralised portions of
the shears. Geological data (rock types, mineralisation,
stratigraphic units, etc) were not captured and as such a resource
model lacked the proper geological controls.
In 2005 the database was expanded to include geological
characteristics for use in determining the nature of the orebody. A
project was initiated to capture all historical drill holes (as many as
6000) to be included in the database. This is an ongoing process.
Development sampling data is also included in the Geological
Database Management System. Together with the drill holes a
comprehensive data set is developing to assist in geological and
resource modelling. The more data that is added to the database
the more complex the orebody becomes and is crystallising as
multiple thin shears.
The commercially available range of modules in Datamine is
used in the development of a resource model.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
Geological model – wireframing
Due to the limited geological information in the database there are
currently no geological wireframes. The existing wireframing is
restricted to identifying the mineralised envelope pertaining to a
resource grade cut-off. The wireframing is typically constructed
using the development and drill hole sampling information.
Wireframes are colour coded to represent the individual ore zones
that are modelled. In some areas of the mine there are up to eight
individual zones, which is a huge step forward in understanding
and modelling the geology of Obuasi Mine.
There are isolated high-grade areas outside of the main ore
envelopes. These grades are associated with small ore zones
splaying from the major ore zones and are not easily identified in
the drill holes. These grades are normally located some distance
from main shears.
An indicator kriging run is done to find these high-grade areas
and the indicator kriging identifies areas of grade above 2.2 g/t.
The wireframe is adjusted to include these samples in
collaboration with the geological interpretation of the shear
boundaries. Any other mineralised indicators are considered as
secondary targets and constraint to their own unique zone to
prevent the overestimation of the surrounding areas.
Estimation
Geostatistics is used to estimate into the block models and
ordinary kriging was selected as the best method for estimation.
The colour coding of the wireframe is used to define each
individual ore zone as a kriging zone (KZONE). The drill hole
and development samples are zoned per KZONE and estimation
parameters are determined for each zone. As the model is
constrained based on a resource cut-off grade, the wireframes
distinguish between the high grade ore and surrounding barren
waste. These hard boundaries are used to constrain the estimation
inside the ore zone. A top cut is applied to the sampling data
when modelling the semi variogram. The top cut is determined
from the grade distribution and in general is not more than
0.5 per cent of the data.
The areas that were highlighted by indicator kriging are also
estimated to improve on the geological understanding and to
produce potential mining blocks.
Due to the scale and complexity of the orebody, the method of
estimation and estimation parameters are still areas for
improvement and work is ongoing to optimise the geostatistical
process and controls.
Classification of the Mineral Resource is a function of
increasing confidence in the estimate and is affected by parameters
such as continuity of the mineralisation and the density and quality
of the sampling. The Mineral Resource classification analysis is
done using the ‘15 per cent error with 90 per cent confidence’ rule.
This technique is based on estimating the average grade above
a specified cut-off with less than 15 per cent relative error at
90 per cent confidence. The criterion for a Measured Mineral
Resource is the ability to estimate one month’s production to this
level, while the criteria for an Indicated Mineral Resource would
be one year’s production.
An investigation during 2007 revealed that only a limited
number of blocks were deemed to be classified correctly. The
problem is associated to the fact that the orebody is not sampled
on a regular grid and therefore there are limited areas where the
theoretical spacing is achieved within each individual kriging
zone.
The recommended solution which has been used at Obuasi in
the past is to make use of a nearest neighbourhood estimate for
each block in the resource model. The borehole spacing method
uses a proximity search method to classify the resources into
indicated and inferred.
81
H EYBERS, J VISSER and C K BOAFO
Blocks which lay within a search of 35 m along strike and
35 m down dip of any samples within that same krig zone are felt
to represent a spacing of 60 m × 60 m. Any blocks that fell
within this range are classified as Indicated. Blocks within 20 m
range are classified as Measured. All other blocks that are not
classified as Measured or Indicated, but received a gold grade,
were classed as Inferred.
A minimum of four samples within the search ellipse are
required for the classification to be applied.
Additional resource model attributes
The resource model is also depleted where mining took place. In
recent years a cavity monitoring system (CMS) is used and the
wireframe obtained from this is used to classify the resource as
mined or unmined. Survey data for historically mined areas are
used in an attempt to develop wireframes to deplete the model.
Resource to reserve
With a global resource model in place for all the ore zones the
next step is to determine the economic resource available and to
convert it into a mineable reserve. The process of converting the
resource into a reserve follows a series of steps eliminating
sterile resource created by economic criteria, safety factors,
geotechnical constraints, mine design and extraction constraints.
The first step is to determine a resource cut-off. Pre 2004 the
resource cut-off of 3.43 g/t was used across the board with no
evidence of economic criterion applied. With the new
comprehensive resource model, a grade tonnage profile can be
calculated from which a cut-off grade can be determined. Based
on gold price, costs and profit margins, a required mining grade
can be obtained. The required mining grade is ‘dialled’ into the
grade-tonnage profile as an average grade above cut-off and the
resource cut-off and available volume can be established.
The Mineable Reserves Optimiser (MRO) is a Datamine based
application that analyses a resource model. Using a floating stope
method it delineates an optimal mineral resource based on user
supplied criteria. It creates and evaluates three-dimensional
envelopes of material taking into account factors such as the cut-off
grade, minimum size, shape and orientation of the mining units.
The resource cut-off obtained and the minimum mining unit
(MMU) are used as criteria for the optimisation. Currently the
minimum mining unit (MMU) used for mine design on Obuasi
mine is 20 m × 5 m × 15 m. The units represent the:
Global R esource Block Model
Mineralised wireframes from drilling
and cross-cut sampling
FIG 2 - Development of a planning resource model at Obuasi.
82 Ballarat, Vic, 14 - 15 October 2008
• 20 m strike intervals of the cross-cuts,
• 5 m minimum width required for current mining equipment
for the selected mining method, and
• 15 m level intervals.
The result of the resource optimisation is a resource model
containing areas flagged as either an economic resource or
inventory.
The second step in converting the resource to reserve is to
eliminate areas that will be abandoned. These areas can be pillars
based on rock engineering recommendations, areas not
accessible due to age of infrastructure and the galamsey activity
(illegal small-scale mining). A resource model further classified
in terms of available economic resource is now available.
Step three is the last step in creating a resource model that can
be used for mine design and further optimisation for the reserve
component. During this step all shaft pillars and crown pillars are
identified in the model. These are areas of the resource that can
not be mined for geotechnical reasons but at a later stage can be
extracted. The integrated model obtained from this process is the
planning resource model (Figure 2). The model comprises only
the economic resource that can be designed and scheduled which
also considers other mining issues.
Previously CAD planning capabilities were not in place to
generate a reserve component. The Mine 2-4D package is now
employed to do mine design and to generate reserves from the
resource model. The whole Mineral Resource Management
process is now integrated from geological modelling, resource
evaluation, rock engineering to mine planning. There is still
room for improvement in all areas and work is ongoing. The
system allows for auditable resource and reserve figures backed
by reproducible processes.
SENSITIVITY ANALYSIS
With the extensive mining over the years the primary thick, high-
grade ore zones are diminishing and there is a need for change in
methodologies to mining the thinner secondary ore zones.
The value-add of an integrated resource model is the ability to
interrogate the resource by doing scenario planning. For the first
time on Obuasi Mine, an attempt can be made to quantify the
narrow ore zones and to do sensitivity and selectivity analysis.
Available Planning Resource
Model
Narrow Vein Mining Conference
 RESOURCE MODELLING AND MINE DESIGN AT OBUASI MINE, GHANA

Methodology 15 500 000

MMU Selectivity - Block 1
5.0

An initial study was conducted to determine the grade sensitivity 15 000 000
through optimisation with different stope dimensions. As 4.8
14 500 000
indicated previously, the MMU used for mine design on Obuasi
Mine is 20 m × 5 m × 15 m, which represents the strike, width 14 000 000 4.6
and depth dimensions. The focus of the study is to only change

Tonnes
13 500 000

g/t
the width of the MMU in the narrow ore zones. For the study two 4.4
13 000 000
areas of the mine were considered where thin ore zones appear.
Mining Block 8 is an area with six individual ore zones and 12 500 000 4.2
Block 1 is a less complicated area with three ore zones. The ore 12 000 000
zones have variable thicknesses (1 - 20 m) and display the typical 4.0
pinch and swell geometry and short range change in strike and 11 500 000

dip (Figure 3). 11 000 000 3.8

1 2 3 4 5
MMU width (m) Tonnes G/t
N

MMU Selectivity - Block 8 Lower

23 500 000 6.4

23 000 000
6.2
22 500 000
6.0
22 000 000

Tonnes 21 500 000 5.8

g/t
21 000 000 5.6
Ore lodes - Block 8 0 40
20 500 000
m 5.4
20 000 000
5.2
N 19 500 000

19 000 000 5.0

1 2 3 4 5
MMU width (m) Tonnes G/t

FIG 4 - Impact of change in mining width.

0 40
Ore lodes - Block 1
m
FIG 3 - Obuasi ore lode geometry.
Using the floating stope method, different stope widths were
optimised to determine the change in selectivity. Stope widths
of 5 m, 4 m, 3 m, 2 m and 1 m were optimised with a float
increment (subcell) of 1 m.
The output is a block model with cell sizes specified by the
float increment and it defines the volume and grade of the best
stope configuration. The wider stope width needs a thicker ore
zone to achieve the head grade and will mine more waste in thin
zones. The result is increased tons at lower grade. The narrower
stope width enables more thin ore to be mined while achieving
the head grade. This produces less waste which mean less tons
but at a higher grade.
A deficiency in the software is the ability for the floating stope
and block cells to follow the geometry of the reef. Currently it
only optimises in a fixed direction and dip and is not optimum
for the type of orebody being mined. As this is an initial study,
the results are only an indication of the change in grade that can
be achieved. Further investigation is underway to optimise the
stope width for any change in dip and strike.
Analysis
In Block 1 the average resource grade (excluding dilution)
achieved for a stope width of 5 m is 3 g/t for all the ore zones at a
head grade of 2.2 g/t. The grade improves five per cent to 4.1 g/t
when the stope width is reduced by 1 m to 4 m. If it was possible
to mine a 1 m wide stope the grade would improve by 26 per
cent. To mine an idealised 3 m wide stope would see a nine per
cent reduction in tons and an improvement in grade of 11 per
cent (Figure 4).
The analysis was performed on all three ore zones together. If
the study is expanded to analyse the effect on the individual ore
zones at different cut-offs, a further improvement in grade can be
expected at variable stope widths.
In Block 8 Lower there is no difference in a stope width of 4 or
5 m but narrower widths do have an improvement in grade and
lower tons. The reason that there is no change can be attributed
to the overall thicker ore zones present and the undulation of the
ore zones. A five per cent improvement in grade and six per cent
drop in tons is possible at a 3 m wide stope width (Figure 4).
The reason why the grade improvement in Block 1 is more
than double than in Block 8 can be attributed to the difference in
geology. The undulations in the ore zones and the variable
thickness over short distances in Block 8 would suggest that
a blanket 3 m wide stope will not be optimal. The importance
of geological control is highlighted in this analysis. This
understanding forms the basis of the continuous improvement of
the resource model in narrow vein areas on Obuasi Mine.
OBUASI UNDERGROUND MINING METHODS
History
Introduction
Ancient mining in Obuasi included techniques such as trenching,
scooping up and panning gold-bearing materials found in soils
and areas of heavily weathered gold-bearing rock. The
techniques of shallow mining started in pre-colonial days.
With the introduction of deep mining, mining operations were
historically concentrated in relatively narrow high-grade veins
using cut-and-fill as the predominant method. Until the mid-
1980s, approximately 70 per cent of underground mining was
undertaken using the cut-and-fill method and 30 per cent using
sublevel caving and other methods.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 83

H EYBERS, J VISSER and C K BOAFO

The cut-off grade below which mining was not economically
feasible was 11 g/t during that period, largely because of
constraints in milling capacity. In 1990, Obuasi mine commenced
a study to evaluate its ore reserve at cut-off grades similar to mines
in other parts of the world.
The Obuasi mine determined that, at a cut-off grade of 3.4 g/t,
the underground ore reserve in the exposed workings and
explored areas would increase from seven million tonnes to
31 million tonnes, with average mining widths increasing from
some 6 m to over 10 m in the sulfide areas.
This determination enabled the Obuasi mine to consider the
introduction of mechanised mining methods, which enable larger
scale mining and increased productivity.
The range of mining methods employed up to year 2002
includes mechanised open stoping (60 per cent of total);
mechanised cut-and-fill (ten per cent of total); sublevel caving
retreat and reclamation (12 per cent of total); pillar extraction
(<1 per cent of total); ‘conventional’ captive cut-and-fill (<1 per
cent of total); vamping or the cleaning of lateral ore transfer
ways spillage (<1 per cent of total); and stope preparation (16 per
cent of total) (Figure 5).
Current mining methods
The range of mining methods currently employed in Obuasi
includes mechanised open stoping (75 per cent) and sublevel
caving (25 per cent). The mechanised open stoping methods are
transverse, longitudinal and narrow vein longitudinal retreat.
Table 1 shows the mining method and orebody width.
Sublevel caving
A methodology carried out only in Block 5 and 6 and Block 2. It
is based on the principle that the ore is fragmented by blasting,
whilst the overlying host rock fractures and caves. It is a ‘top
down’ method, with the ore being extracted level by level
downwards through the orebody. Mining is longitudinal, with the
number of reef drives dictated by the width of the orebody. The
distance between reef drives is 9 m skin-to-skin.
Cross-cuts are developed at 30 m intervals. Depending on the
width of the orebody, more than one reef drive may be developed
parallel to the strike of the orebody, with a minimum of about
100%
TABLE 1
Mining method and orebody width.
Mining Open stope Open stope Narrow vein Sub level
method transverse longitudinal longitudinal caving
retreat
Average +10 m 5 - 10 m 2-6m +6 m
width
6 m pillar between them. A slot raise is mined at the block-end of
the stope in a cross-cut close to the hanging wall. Slot extension
holes are drilled in the cross-cut across the width of the orebody.
The sublevel vertical spacing is 15 m and cross-cuts are
developed at 30 m intervals. A slot raise is mined at the
block-end of the stope in a cross-cut close to the hanging wall.
Slot extension holes are drilled in the cross-cut across the width
of the orebody.
Sublevel open stoping – transverse
The mining method used predominantly in Block 1 and Block
8L, but also in other areas, where the width of the orebody is in
excess of 10 m.
The design is based upon the extraction of a single lift, with a
primary and secondary stope configuration. Access is via
cross-cuts from the permanent footwall drive, usually placed in
the footwall. Following the backfilling of a stope void, the filling
horizon becomes the next extraction level.
The stope dimensions are typically 15 m on strike for both the
primary and secondary, and are extracted vertically to the full
height of the sublevel. From the cross-cuts, a slot drive 15 m long
is developed parallel to and close to the hanging wall. A slot
raise is mined up to the drilling level at the stope block-end. The
crown and hanging wall are supported with cement grouted cable
bolts from the drives.
Slot extension holes are drilled from the slot drive on the
hanging wall and fan holes drilled down to the mucking level
from the cross-cut at a burden of about 2 m. During blasting the
slot extension on the hanging wall is blasted to create the initial
free face. The rest of the holes are blasted into the slot. Drilling
is downwards using electro-hydraulic drill rig with a 76 mm
diameter hole size.

90%

80%

70%
% Production Tonnes

60%

50%

40%

30%

20%

10%

0%
1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2,008

2,009

2,010

LOM
2,007

Years

Open Stoping Sublevel Prep Vamping Square sets Stope Prep. Pillar Ext. Reclamation
Conv. C & F Mech. C & F Sub-Level

FIG 5 - Underground mining methods.

84 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference


Key points
• The cross-cut is 6 m wide within the reef for providing
improved cable bolting coverage and blasthole drilling.
• No support tees are required. Only one tee is mined at the
end the cross-cut for drilling and blasting the slot. ‘Tee’ is a
reef drive along strike in a transverse open stope.
• A reinforced pillar is created below the level. It is supported
with backfill and provides stability to the access cross-cut for
mining the up dip stope.
• The secondary stope pillar is presupported from the Primary
cross-cut.
• The cable bolt layout for support of the crown will be done
for individual stopes depending on ground conditions.
• The toe of the last blasthole is 4 m below the level.
• The part of ore in the pillar, below the level, is lost.
Sublevel open stoping – longitudinal
This method is predominantly used in the South of the mine, in
Block 8 Upper and Sansu; it is selected for areas where the
transverse stoping method is not required, ie in areas where the
orebody width is less than 10 m.
From a drive, usually placed in the footwall, cross-cuts at 30 m
intervals are developed to access the ore and define both the F/W
and H/W contacts. The stopes are panelled out 30 m long on
strike. A reef drive is developed and slot raises placed at the two
extremities of the panel.
Slot extension holes are drilled parallel to the slot raise.
Production holes (76 mm) are drilled in the form of a fan.
The two extremity slot extension holes are blasted to create
free faces, large enough for the fan holes to be blasted into.
Blasting of the fan holes progresses inwards closing in at the
central cross-cut.
Key points
• The drive is 4 m wide and is located near the middle of the
reef for providing improved cable bolting coverage and
blasthole drilling. The drive may be moved for ground
stability considerations.
• No support tees are required.
100%
90%
80%
70%
% OF MINING METHOD
60%
50%
40%
30%
20%
10%
0%
2008
Sub-Level Caving
Narrow Vein Mining Conference
RESOURCE MODELLING AND MINE DESIGN AT OBUASI MINE, GHANA
• Reinforced pillars are created below the level. They are
further supported with backfill and provide stability to the
access cross-cut for mining the up dip stope.
• The cable bolt layout for support of the crown will be done
for individual stopes depending on ground conditions. The
cable bolts in the pillar will be extended to at least 2 m into
the walls.
• The toe of the last blasthole is 4 m below the level.
• The part of ore in the pillar, below the level, is lost.
Future mining methods
Open stoping is planned to be the predominant mining method
on the mine, while sublevel caving method will be reducing from
the current 25 per cent to less than ten per cent towards the end
of life (Figure 6). The introduction of open stoping longitudinal
retreat mining will be introduced during 2008.
Longitudinal and transverse stoping
The stopes are mined either as longitudinal or transverse based
on the width of the ore. Ore widths greater than 10 m are treated
as transverse and widths less than 10 m are treated as
longitudinal, due to geotechnical considerations.
The development regime and stope strike length for
longitudinal stopes has been reviewed and in certain areas the
15 - 20 m distance between cross-cuts for about 300 m strike
orebody has been replaced by a single access cross-cut for the
same strike length resulting in longer stopes on strike. This
change resulted in a significant saving in waste development
metres (Figure 7) and mining will now be conducted on a trial
basis.
Open stoping longitudinal retreat
The improved geological modelling techniques allow
identification of medium- to high-grade narrow veins for the
implementation of an appropriate narrow mining methodology.
The current resource breakdown by width suggests that
35 per cent of the resource at an average gold grade of 7.7 g/t lies
between 1.5 to 6 m (Figure 8). Currently, orebody with widths
less than 6 m is classified as narrow vein and therefore a revised
open stoping longitudinal retreat method will be implemented
(Figure 9).
2009 2010 LOM
YEARS
OS - Longitudinal OS-Long Retreat(Narrow Vein) OS - Transverse
FIG 6 - Mining method trend.
Ballarat, Vic, 14 - 15 October 2008 85
H EYBERS, J VISSER and C K BOAFO

Previous

Proposed

FIG 7 - Development saving using revised longitudinal method.

Resource per ore lode width category

The biggest challenge in using narrow vein mining as
16% 10 explained by Steward (2005) will be:
14% 9.5 • with narrow-vein mines there is usually a lot of extra
9 material (dilution) handled;
12%
8.5
10% 8 • lack of precision mining, ie taking too much waste out in an
Oz

effort to get the vein of gold;

g/t

8% 7.5
6% 7
• high dilution reduces the profitability of an operation as it
6.5
4% increases costs and results in poor use of resources; and
6
2% 5.5 • the narrower an orebody, the more exposed it is to the risks
0% 5 associated with achieving precision in mining and inaccurate
2.25

3.75

5.25

6.75

8 .25

17.5

22. 5

27.5

32.5
0.75

10.5

13.5

Width (m) - Class Midpoint
FIG 8 - Resource percentages by lode thickness.
The current Obuasi reserves include traditional narrow veins
above 50 level and the deeps project below 50 level (unproven).
Mining the narrow veins with longitudinal retreat caters for both
crown and back stability due to known traditional ground and
stress effects. Geotechnical criteria and ground control
parameters are taking into consideration in the mine design.
The current issues which have brought about this type of
mining method in some mining blocks are:
• high waste development (waste/ore ratio >50 per cent),
• low productivity (<7000 t/month) per stope or mining block,
• high unplanned dilution (up to 30 per cent) due to hanging
wall failure,
• high cost $/t,
• narrow veins treated as thick reef (developed same way up to
10 m width), and
• timely information on reef width and variation.
The ore to waste ratio comparing the two methods indicates
a potential saving of 25 per cent at the 6 m cut-off point
(Figure 10) as a result of less waste metres that will be required
to extract the shear zone.
dilution prediction.
. Tonnes Grade
Key points
• The development rate is critical for high production,
development ends need to be at least six months ahead of the
current working stopes to ensure quality resource drilling and
then to establish a proper geological model. The geological
model is critical to proper design and control of the mining
plan.
• Stability of the reef drive is crucial for the success of the
method. Ground support precautions are required.
The fill management
• Waste rock: all waste must be dropped in stopes after initial
set-up development. A few cut-and-fill lifts can consume part
of the initial development waste. An internal pass may be
dedicated to this, if necessary. The ideal place will be to use
the waste for topping up the stopes already filled with
hydrafill. This will provide better roadway for equipment.
• Hydrafill: hydrafill will be poured from an extremity slot
raise and also through the ramp access via a vertical or
inclined hole. This will be reviewed on a stope by stope basis
and in certain cases the stope span will be reduced. Drainage
holes must therefore be provided.
• Bulkheads: hydrafill bulkheads have to be built before
mining past the rock pass accesses or as per geotechnical
recommendations.

86 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 RESOURCE MODELLING AND MINE DESIGN AT OBUASI MINE, GHANA

CROSS CUT TO REEF

AT 10M VERTICAL SPACING

SHAFT ACCESS

38L

BSVS
SHAFT

RAMP
SHAFT ACCESS
VENT RAISE
CROSS CUT TO REEF AT
VENT RAISE 10M VERTICAL SPACING

FIG 9 - Longitudinal retreat mining for narrow vein.

60%
Waste/Ore Ratio

50%
Longitudinal

40%
W aste/O re r atio

30% Thin reef

R e duc tion

Transverse
20%

10%

0%
0 5 10 15 20 25 30 35 40 45
<-------Longitudinal|| Transverse------------------------------------------------------------------------------------->
Reef width m

FIG 10 - Longitudinal retreat waste tonnes to ore tonnes generation.

• Connection to ventilation raises may not be flat cross-cut but

a raise at say 45°. This will avoid the need to build a fill wall
and resulting leakages from it.
• Waste in startup: quickly mine some cut-and-fill lifts and
backfill the initial waste in them. Otherwise existing passes
will have to be used to dispose of initial waste generated in
the area. First sublevel can then be higher than normal. This
will also increase the storage capacity of the passes.
Storage in rock passes
• If using tracked cars for gathering from all internal passes,
the combined storage should be large enough to fill the train.
• If using trackless for gathering from internal passes, the
required minimum capacity for the main pass should be at least
1.5 times the train capacities. It can be achieved by increasing:
• height,

Vent and rock pass management • diameter,

• number of passes, and
• Shallower vent passes for reducing speed and dust, and
• dropping ore to a lower tracked level.
• keep the rock pass at safe distance of 20 m from reef.
Thick portions of reef: it is desirable to mine reef wider than Ventilation
10 m with transverse stoping. A footwall subdrive will be required
Through ventilation is desirable but ramp access and reef drives
for that length. It will be desirable to locate the access ramp at the
development will need to be done under forced ventilation. It is
boundary of thin and thick reef. This will make footwall drive essential that their total length is within the 180 m maximum
easily accessible. Transverse stopes can be mined in double lifts, if legal distance. Upper portions of rock passes can be used for
required. providing through ventilation, where possible.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 87

H EYBERS, J VISSER and C K BOAFO

TABLE 2
Narrow vein equipment selection.

Development Activity Dimension (m) Equipment type

Drilling Mucking/cleaning/haul
Ramp 4×4 Jackleg ST 3.5(2.8 t bucket)
Footwall drive 4×4 Jackleg ST 3.5(2.8 t bucket)
Lateral Cross-cut to reef 3×3 Jackleg ST 2D(1.8 t bucket)
Reef drive 3×3 Jackleg ST 2D(1.8 t bucket)
Access to orepass, vent raise 3 × 3 (sublevels) Jackleg ST 2D(1.8 t bucket)
Access to orepass, vent raise 4 × 4 (main levels) Jackleg ST 3.5(2.8 t bucket)
Orepass 1.8 × 1.8 Alimak/stoper ST 2D(1.8 t bucket)
Vertical Waste pass 1.8 × 1.8 Alimak/stoper ST 2D(1.8 t bucket)
Vent raise 2.4 × 2.4 Raisebore
Longhole drilling Activity
Support holes Wagon/skid mounted pneumatic, simba for longholes
Production holes Wagon/skid mounted pneumatic, simba for longholes

Definition drilling CONCLUSION

Advance definition drilling is essential for locating reef drives
within the good quality reef. This can be achieved by drilling
definition holes from the available cross-cuts a level below.
Existing cross-cuts to the ventilation raise, rock passes and ramps
may be used for this purpose.
Vent ducting
Small size loaders must be used for reef drive development
leaving sufficient space for ventilation ducting. The loaders will
only operate in one direction; therefore the ventilation ducting
must be placed on the side opposite to the driver canopy.
Charge density
The charge density in the boundary blastholes must be reduced
by using:
• low density explosive,
• spaced cartridge explosive,
• small diameter holes, and
• small diameter cartridge explosive.
Blasting sequence
The boundary holes must be timed in such a way that the effective
burden is reduced for these holes. This can be achieved by timing
them last. This measure will help reduce wall damage and
dilution.
Equipment selection
Today, a selection of standard ‘slim size’ machines is available,
allowing mechanised mining in 2 m wide drifts. The range of
slim-size machines includes the face drill rig for narrow drifting,
matched with same size longhole rig. The small drifter jumbo and
longhole rig, complemented with the 2 m3 bucket LHD loader is
all that is needed for mechanised mining of the 2.0 m wide vein.
The equipment selection for open stoping longitudinal retreat
that will be considered is shown in Table 2. More work is
required to ensure that the appropriate equipment is selected to
ensure high productivity with the lowest amount of dilution.
A vast narrow vein reserve can be mined using modern
techniques, and the right mining equipment will undoubtedly add
value to the Obuasi asset; the major challenge will be the change
in mindset, the establishment of advance grade control, up to
date geological model and the implementation of a sound mining
plan that will allow the mining of these narrow veins safely and
profitably.
Continuous improvement in geological understanding,
research into and refinement of modelling and estimation
processes will minimise the business risk.
ACKNOWLEDGEMENT
The authors would like to express their gratitude to AngloGold
Ashanti management in particular Obuasi mine for their
assistance and permission to publish this paper.
REFERENCES
Atlas Copco, 2008. Application underground mining (accessed: 1 June
2008).
Baidoe-Ansah, W, 1998. An application of the Conolly (Contour) diagram
to the GCS quartz vein; Obuasi Mine, MSc thesis (unpublished),
University of London.
Dominy, S C, Annels, A E, Wheeler, P, Simon, G and Barr, S P, 1999.
Geology in the resource and reserve estimation of narrow vein
deposits, Exploration and Mining Geology, 6(4):317-333.
Etheridge Henly Williams, 1995. Structural controls on gold
mineralisation in Ashanti Gold Mine, Etheridge and Henley
Geoscience Consultants, in-house report (unpublished), Ashanti Gold
Mine Corporation Ltd.
JORC, 2004. Australasian code for reporting of Mineral Resources and
Ore Reserves (The JORC Code), The Joint Ore Reserves Committee
of The Australasian Institute of Mining and Metallurgy, Australian
Institute of Geoscientists and Minerals Council of Australia.
Stewart, P C, 2005. Minimising dilution in narrow vein mines, PhD
thesis, Julius Kruttschnitt Mineral Research Centre, The University
of Queensland.
Stone, J G and Dunn, P G, 1996. Ore Reserve Estimates in the Real
World, second edition, special publication 3, p 160 (Society of
Economic Geologists: Littleton).

88 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

Gold Particle Characteristics in Narrow Vein Deposits –
Implications for Evaluation and Metallurgy
S C Dominy1,2, Y Xie3 and I M Platten4
ABSTRACT
Narrow vein gold mineralisation generally contains both fine (<100 µm)
and coarse (>100 µm) gold particles. The in situ size and shape,
deportment, distribution and abundance of the particles controls deposit
sampling characteristics, grade distribution and metallurgical properties.
Particles can range from individual disseminated grains, clusters of
particles and masses above 1 cm in size. At each end of the coarse-gold
to fine-gold spectrum, the samplability of a deposit ranges from
relatively simple for fine-grained disseminated gold particles, through to
extremely difficult for very coarse particles. Deposits generally show a
background grade of mineralisation dominated by disseminated fine
gold particles. Drilling will resolve the background grade well, though
is likely to considerably understate the grade related to the coarser gold
particles. Gold mineralogy has a significant impact on metallurgy, with
key factors affecting recovery being particle size, deportment, liberation
parameters and silver content of gold minerals. For coarse gold particles
the focus should be on liberation and the early removal of the particles
from the mill circuit by gravity concentrators and gold traps. Finer
particles will require more grinding to liberate prior to flotation or
cyanide extraction. The determination of gold particle sizing is thus
required to minimise evaluation and geometallurgical uncertainty and
lower project risk.
INTRODUCTION
Narrow vein gold deposits can be defined as having a true width
of less than about four metres. They represent a globally
significant source of gold, which is often exploited by
junior/small-medium sized companies. Their evaluation poses
particular problems because of inherent geometrical, geological
and grade complexities. The veins display strong grade
anisotropy with multiple grade populations generally observed.
Economic grades are associated with discrete ore shoots, which
show a high-nugget effect and are either structurally and/or
lithologically controlled.
The nugget effect has two principal components: the
geological or in situ nugget component and the sampling nugget
component. The geological nugget component is related to the
natural distribution of the gold within the mineralisation and
short-scale variability (Platten and Dominy, 2003; Dominy,
Platten and Raine, 2003). The sampling nugget component is
related to errors induced by inadequate sample size, preparation
methods and analytical procedures (Dominy, 2004; 2007).
Small-scale geological variability, and gold particle size and
distribution affect both the geological and sampling nugget effect
1. FAusIMM(CP), Executive Consultant and General Manager (UK),
Snowden Mining Industry Consultants Limited, Abbey House,
Wellington Way, Brooklands Business Park, Weybridge Surrey KT13
0TT, England. Email: sdominy@snowdengroup.com
2. Adjunct Principal Research Fellow, School of Science and
Engineering, University of Ballarat, University Drive, Mount Helen
Vic 3353.
3. Professor of Applied Mineralogy, Department of Civil and
Environmental Engineering, University of Science and Technology
Beijing, Xueyuan Road, Haidian District Beijing 100083, China.
4. Principal Geologist, Snowden Mining Industry Consultants Limited,
Abbey House, Wellington Way, Brooklands Business Park,
Weybridge Surrey KT13 0TT, England.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
(Dominy, Platten and Raine, 2003; Platten and Dominy, 2003;
Dominy and Platten, 2008). Further discussions on the nugget
effect, its cause, effect and problems can be found in Platten and
Dominy (2003), Dominy et al (2001; 2003) and Dominy (2004).
The particle size characteristics of gold are relevant to both
resource estimation and extraction. This information is required
at an early evaluation stage when understanding of the deposit is
changing most rapidly. Exchange of information and integration
of testing procedures ensures most effective use of funds
(Dominy, Xie and Platten, 2008).
Sampling is challenging in gold deposits, particularly in
narrow vein deposits that contain coarse gold particles
(Sketchley, 1998; Dominy et al, 2000; Roberts, Dominy and
Nugus, 2003; Dominy and Petersen, 2005; Johansen and
Dominy, 2005; Cintra et al, 2007; Gonzales and Cossio, 2007;
Dominy, Platten and Xie, 2008). The use of diamond drilling and
fire assays (ie small sample supports) in this environment often
results in an understatement of grade. In addition, the undercall
is augmented by wide-spaced drilling that gives rise to an
information effect. The information effect refers to the fact that,
even during production, the real mining block grades are not
known due to a lack of information (eg samples). In some cases,
other parameters such as quartz texture, mineralogy, structural
features and geochemistry can be used to provide an indication
of gold grade potential – the so-called gold proxies (Dominy and
Johansen, 2004).
Larger diameter drill holes improve the likelihood of obtaining
a representative sample mass in theory, but rarely in practice
unless scaled-up to tens of tonne bulk samples (Dominy, Platten
and Xie, 2008). A move from NQ to HQ diameter core may have
little effect, unless the entire sample is taken for assay. This will
be problematic from the perspective of verification sampling at a
later date. The proportion of coarse gold has a direct impact on
the effectiveness of sampling, where more coarse gold will
require bigger samples and more specialised protocols.
Short-term variation is characteristic of narrow vein settings
and coupled with relatively small tonnages, is likely to result in
significant changes to the character of mill feed and reliability of
estimation procedures. Monitoring of gold grain size parameters
during exploitation is useful to both disciplines.
The size, shape/deportment, abundance and distribution of
gold particles have a fundamental control on deposit evaluation
from the perspective of both drilling and sampling. In addition,
these characteristics will also exert a strong control on the
metallurgical extractability of gold. This contribution discusses
the issues of gold particle sizing and distribution based on the
authors experience within narrow vein systems.
MATERIAL CHARACTERISATION
The need for characterisation
A key precursor to deposit evaluation is the characterisation of
the mineralisation to determine in situ gold particle sizing and
deportment (Dominy, 2007; Dominy, Platten and Xie, 2008;
Dominy, Xie and Platten, 2008). It is important to ensure that
sampling protocols fit the mineralisation type and are not just
stated to be so-called ‘industry best practice’.
91
S C DOMINY, Y XIE and I M PLATTEN

Holistic testing
The measurement of gold particle sizes and their distribution can
be gained through observation of drill core, rock samples, etc.
Extensive mineralogical studies as part of wider paragenetic and
geometallurgical investigations are required (Chryssoulis and
McMullen, 2005; Angove, 2005).
A holistic evaluation should include (Dominy, 2007; Dominy,
Xie and Platten, 2008): inspection of drill core, exposure, etc;
mineralogical study; metallurgical testing; progressive
crush/grind, screening and assaying; duplicate/repeat sample
determination, fire assaying to extinction and comparative assay
tests; and heterogeneity testing (eg Gy 50 Piece test). FIG 1 - Range of gold particle sizes from the Ballarat East Project,
Each technique above provides information about aspects of Australia (photo: Ballarat Goldfields Limited).
the gold particle size and distribution that is relevant to the
design of a routine sampling program for the deposit. No
individual technique provides sufficient information on its own
and some data may only be interpreted using observations
from other sources. Integration of observations from several
techniques will give the most robust assessment for the design
of the sampling program. Holistic programs are part of wider
geometallurgical studies, and will support resource evaluation
through increased geological knowledge and sampling
optimisation (Dominy, Platten and Xie, 2008; Dominy, Xie and
Platten, 2008).

GOLD PARTICLE CHARACTERISTICS

Size of gold particles

General sizing
Many styles of gold mineralisation are dominated by the
presence of coarse (>100 µm) and often very coarse (>2000 µm)
gold particles. In some cases 40 per cent or more of the gold is
greater than 100 µm in size. Other systems are dominated by
very-fine (<10 µm) disseminated gold particles (eg Carlin style
mineralisation). The size of particles affects the sampling and
metallurgical characteristics of the deposit, and can range from
individual disseminated grains from 100 µm to 2000 µm in size
through to ‘nuggets’ greater than 2000 µm in size (Figures 1
and 2). Masses of gold, thin continuous sheets, dendritic growths
and/or wires can be also be seen (Dunn, 1929). One of the largest
single masses of gold was discovered at the Hill End goldfield
(Australia) weighing 286 kg, other large masses were recorded
from the Australian Bendigo and Ballarat goldfields (Dunn, 1929).
Coarse gold particles >300 µm are generally visible in hand
specimen, core, etc. The proportions of different particle size
populations vary considerably between deposits and internally
within them, showing dominantly fine gold, dominantly coarse
gold or mixed coarse/fine gold populations (Table 1). For
example, in the Orion gold mine (Canada) two styles of
FIG 2 - Coarse gold particles from the Bendigo Project, Australia.
Particles from the St Anthony’s Reef after treatment in the New
Moon plant. The relatively unmodified particles on the left were
collected in the boiler box (prior to grinding), whereas the hammer
mill modified the rounder grains on the right (photo: Bendigo
Mining Limited).
mineralisation exist in close association: quartz veins and
breccias dominated by clusters of visible gold and pyritised
basalt dominated by fine gold.
Gold particle size reflects a variety of geological factors:
original mode of deposition of host and gold, original void space
between host crystals, exsolution or recrystallisation of gold and
host, later deformation of host minerals and redistribution of
gold. Discrete size populations may exist in complex structures
with multiple veins and/or lengthy history of deformation and
recrystallisation (eg fine gold within intact parts of sulfide grains
and coarse gold within fractures in sulfide grains).
Studies have shown that gold can occur as large dispersed
masses throughout the host mineral(s). Visually, gold particles
may appear to be single (though potentially clustered), but in fact
are connected through the host mineral. This is typical of some

TABLE 1
General characteristics of gold particle size end members.

Gold particle Size range Size range Size range Visible gold GRG Sampling K Grade range Example
scenario 100 µm 1000 µm recorded (%) constant (g/t Au) projects
(g/cm1.5)
Fine-gold <5% 0% Up to 100 µm. No <10% Up to 1000, Up to 10 g/t Au, Carlin,
dominated Mostly <50 µm generally <500 commonly USA
2 - 5 g/t Au
Mixed 20 - 70% <5 - 30% Up to 5000 µm. Possibly, rare >20 - 70% >1000 to 15 000 Up to 30 g/t Au, Vivian,
fine-gold Mostly <1000 µm to common commonly Australia
coarse-gold 5 - 15 g/t Au
Coarse gold >70% >30% Up to 10 000 µm Yes, often >70% 2000 to 250 000 Up to 120 g/t Au, Bendigo,
dominated or larger masses abundant commonly Australia
5 - 15 g/t Au

92 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 GOLD PARTICLE CHARACTERISTICS IN NARROW VEIN DEPOSITS – IMPLICATIONS FOR EVALUATION AND METALLURGY
very high-grade deposits (eg Central Norseman and Gympie in
Australia, Sixteen-to-One in the US and Campbell-Red Lake in
Canada). In three-dimensions, individual grains may reach a few
millimetres in size, though the larger connected mass may be
tens of centimetres in size. This has clear implications for
sampling, as the gold particle size between the in situ mass, and
field sample and pulverised material can be very different.
In one case investigated by the authors, the in situ connected
mass was close to 10 cm3 in volume. In drill core it was restricted
to the core diameter (48 mm), whereas after pulverising (to
P80-75 µm) particles were between 120 µm and 1350 µm in size.
The latter emphasises the problems of pulverising in the presence
of coarse gold particles (Royle, 1989; Wooldridge, 1998;
Dominy et al, 2000).
Predominantly fine-gold deposits may contain sufficient
coarse gold to affect sampling. Springer (1983) reported on a
microscopic study of gold particles from 50 Canadian deposits.
It found that some 75 per cent of the gold particles observed
were between 0.1 µm and 100 µm in size; the most common size
being between 40 µm and 50 µm. Some 25 per cent of the total
population fell into the coarse gold type with sizes ranging
between 100 µm to 10 000 µm.
Coarse gold particle sizing
The expectation that a coarse gold problem can be always
recognised through the presence of visible gold (generally
particles >250 µm in size) is misguided. In many cases visible
gold particles can be elusive, particularly when only drill core is
available. Such particles can be either of the ‘fine’ coarse type
(eg 100 µm to 350 µm) which are intrinsically difficult to see or
as very coarse (eg >1000 µm), but rare particles that will have a
marked impact on sampling. These very coarse particles are
surprisingly not always recognised during logging and mapping.
Given that coarse gold-bearing systems are the most
challenging to sample effectively, there are two principal type
end-members.
The most difficult coarse gold systems to sample and evaluate
are those that contain highly dispersed coarse gold, and where
there is often large amounts of very coarse gold. Gold can be
present as either isolated particles and/or as clusters of particles.
Key examples are Bendigo and other Central Victorian deposits
and the Norseman systems. As little as 20 per cent of diamond
drill intersections in the deposits can contain economically
indicative gold grades (>1 g/t Au).
FIG 3 - Gold particle size curves for selected coarse gold deposits.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
The less challenging coarse gold systems are those such as
sections of the Red Lake mine (Canada) where the gold
distribution, whilst still often being dominated by coarse/visible
gold material, is more ‘evenly’ distributed through ore
shoots/zones. A notable feature is that a high number of drill
holes will intersect economic grades (>4 g/t Au). At Red Lake
some 85 per cent of all drill holes intersecting the High Grade
Zone contained visible gold, and 40 per cent of all assay grades
are generally above 30 g/t Au.
Gold particle sizing can change within a mineralised zone. For
example, in the Norseman mines there is a general 50 per cent
upgrade between the resource and produced grade (Olsen, 2004).
At the Norseman Bullen mine, the size of the gold ‘nuggets’ and
their relative rarity generally increase with depth. Hence the
resource grade became lower with depth as the ‘rare nugget
events’ were difficult to intersect with drill holes. However,
production grades were locally up to 800 per cent with respect to
the original resource grade.
The size of gold particles exerts a strong control on the
characteristics of a sample, and how effective sampling and
assaying will ultimately be. The proportion of coarse gold has a
direct impact on the effectiveness of sampling, where more
coarse gold will require bigger samples and more specialised
protocols. In general terms, where more than ten per cent of
contained gold is above 100 µm in size, particular care should be
taken during sample collection, preparation and assaying.
Gold particle size curves
Figure 3 shows gold particle size curves for eight deposits that
contain notable quantities of coarse gold. The Bendigo and
Ballarat deposits show the greatest proportion of coarse gold
above 100 µm (both >90 per cent), though Bendigo shows a
higher proportion of >1000 µm particles.
Table 1 shows the general characteristics of the gold particle
size end-member models and the fine/coarse gold mixture.
Gold particle shape
Gold grain shape affects the release and sampling of grains.
In situ particles can be simple, equant and rounded shapes, rods,
plates or complex forms with deep re-entrants in their outline
(Platten and Dominy, 2003: Figure 4). Gold grains filling narrow
fractures may form plates or thin sheets (<1 mm thick but
extending 10 mm to 100 mm in fracture plane).
93
S C DOMINY, Y XIE and I M PLATTEN

Gold particle distribution

Gross particle location

Gold particles can be located anywhere within a mineralised
narrow vein structure. However, in some deposits there are clear
local controls on gold within the vein itself. For example, at
Bendigo gold is often associated with pressure solution sutures
(eg stylolites) and/or within quartz-shale contacts (Dominy,
Platten and Raine, 2003). The latter is also observed in the Wattle
Gully deposit (Australia). Potter, Sheriff and Collins (2008) note
that at the Sand Queen mine (Western Australia), in 80 per cent
of cases gold particles occur along or close to the wall rock
contacts. In the sulfide-locked fine-gold environment, Platten and
Dominy (2007) reported the accumulation of gold-rich sulfides
in the footwall of the B Lode, Hadleigh Castle mine
(Queensland, Australia). Highly localised controls on gold
particle distribution are shown to exist.

Clustering effects
In the majority of cases, gold particle distribution within vein
structures can be considered in terms of clustering at different
scales (Dominy and Platten, 2007). Clustering can be recognised
on three levels: the micro-, meso- and macro-scale (Table 2).

TABLE 2
FIG 4 - Idealised sketches of typical gold particle shapes: Summary of types of clusters and relative scales.
(a) rounded anhedral; (b) rounded with some facets; (c) peg; Scale type Example Impact on sampling Scale
(d) plate; (e) rod; (f) equant, anhedral with concave facets; of clusters protocol design
(g) composite form with type f combined with rods and sheets to
Micro Clustering of grains Definite impact Less than
give deep re-entrants in outline; (h) irregular anhedral with
on the subhand 10 cm
concave facets; (i) thin sheet between anhedral columnar quartz specimen scale
grains; (j) plate spanning entire width of fracture, plate sides
defined by fracture walls; (k) bladed films formed on late lineated Meso Clustering of groups Potential impact 10s cm to
of ‘micro’ clusters few m
surfaces; and (l) very small equant grains define a thin sheet-like
volume between quartz grains. Individual particle sizes are Macro Large-scale clustering For drill grid design >10s to
generally between 100 µm and 2000 µm (from Platten and of clusters and/or and placement 100s m
meso-clusters on the
Dominy, 2003).
ore shoot scale

Macro-scale clusters effectively form ore shoots that are

Gold particle deportment structurally and/or geochemically controlled. Overall ore shoot
Deportment addresses the association of gold particles with other grades are higher than the surrounding vein, which may be
mineral phases. For example, gold particles may be enclosed barren. Ore shoots represent a primary macro-cluster, whereas
within or located around the edges of arsenopyrite or pyrite any contained smaller subshoots will represent secondary macro-
grains. From a sampling perspective deportment is not critical, clusters. In some cases small high-grade zones may comprise
unless gold is associated with telluride minerals or electrum and groups of clusters over a few metres – meso-clusters (Figure 5).
recovery through a bottle roll assay (eg accelerate cyanide
extraction using LeachWELL) is applied.
Where the sulfide containing fine gold is a relatively rare
component in a dominantly quartz vein, but contains significant
amounts of the gold (or millimetre-scale sulfide-rich vein in
non-auriferous host rock), subsamples must contain a
representative proportion of the sulfide grains. This can be
considered in two ways:
1. effectively there is a potential nugget effect in the sulfide
distribution in the sample, which carries with it an effect on
the gold grade in any subsample; and
2. sulfide grains represent clusters of fine gold particles and
produce heterogeneity at the scale of the primary sulfide
particles in the in situ rock and at the scale of fragments of
crushed sulfide in processed sample.
This suggests that the size of finest gold grains relative to the
primary grain size of host sulfides and size of fragments at
successive stages of crushing may be significant in some cases. FIG 5 - A hand specimen of ore from the Garrard Reef at Bendigo,
This overlaps with data needed to ensure effective metallurgical Australia showing a cluster of coarse gold particles on the
recovery of ultrafine gold. centimetres scale (photo: Bendigo Mining Limited).

94 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 GOLD PARTICLE CHARACTERISTICS IN NARROW VEIN DEPOSITS – IMPLICATIONS FOR EVALUATION AND METALLURGY
Studies by Koch and Link (1971) recognised clustering in ore
shoots at the Homestake gold mine (South Dakota, USA) and
were likely to be describing meso- to micro-scale clustering.
Homestake contains stratabound to stratiform gold mineralisation
hosted in iron-rich sediments within Archean metavolcanics. The
gold is locally hosted in disseminations and vein or stringer
systems. The shoots were approximately 75 m along their strike
and extended over 50 m down dip. Koch and Link provided no
physical description of the clusters; though concluded that to
enable the clusters to be properly resolved, more diamond
drilling was required. By implication, the clusters are likely to be
at the macro-scale.
The smallest-scale of clustering is on the 15 cm or less scale.
Individual gold particles cluster in groups, potentially leading to
locally extreme gold grades (Figures 5, 6 and 7). Clusters can be
described as being either 3D or 2D. The 3D clusters are those
that have a 3D form, often close to being spherical in nature
(Figure 6), whereas 2D clusters are linear in form with two long
dimensions (Figure 7). Micro-scale clustering has the greatest
effect on sampling. The occurrence of clusters (or larger
connected masses of gold) has clear implications, as the effective
gold particle size between the in situ mass, and field sample and
pulverised material can be very different.
FIG 6 - Gold cluster (8 mm maximum dimension) contains gold
particles from 30 µm to 1600 µm diameter hosted in quartz with
minor sphalerite. Careful observation shows that some of the
particles form small 3D networks of gold. Represents a small
high-grade cluster related to the background mineralisation from
the Chidlaw Lode, Gwynfynydd mine, Wales, United Kingdom
(width of view 2 cm).
FIG 7 - Core photograph from Bendigo, Australia. The intersection
reveals a linear cluster of very coarse gold particles associated
with stylolitic sutures in quartz (photo: Bendigo Mining Limited).
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
Observation indicates that an enhanced coarse gold problem
can be present in some deposits due to the presence of clusters at
the micro-scale. These give the effect of individual coarse gold
particles (Figures 6 and 7: Dominy and Platten, 2007; Pitard,
2007). Clusters may also occur in fine gold deposits where they
produce an apparent- or pseudo-coarse gold effect. This is
signalled by an unexpected high nugget effect from variography
and poor correlation between field sample duplicates. Clusters
can be accounted for through the definition of an agglomerated
particle size when optimising sampling protocols.
Key conclusions – clustering effects
Current research leads to a number of key conclusions in relation
to gold particle clustering.
Firstly, clustering is a remarkably common feature in many
narrow vein and other types of gold deposit. It is a feature that
often goes unnoticed, or at least not seen for its significance.
Secondly, the presence of extreme gold assays can often be
attributed to clusters of gold particles (Figures 5, 6 and 7). There
is much debate as to how these extreme grades should be treated
during estimation, ranging from top-cutting to indicator
methodologies (Bird, 1991; Healey, 1993; Dominy et al, 2003).
However, the key issue is that these assays usually represent a
real high-grade population (assuming good sampling and
assaying practices are employed) that may contain a substantial
proportion of the total gold inventory. It is therefore important to
understand the underlying short-scale geological and grade
continuity controls of these grades before manipulating them
during estimation (Dominy, Platten and Raine, 2003; Dominy
and Platten, 2008). In some cases, top-cutting may not be
appropriate and it would be better to model the zone around the
samples separately.
Finally, the sample characterisation stage should include an
assessment of clustering, in addition to the standard descriptions
of the gold particle size distribution to permit effective protocol
optimisation (Dominy, Xie and Platten, 2008).
Background gold grade and relation to gold particle
size
In some gold deposits, there is evidence to show that the fine and
coarse gold particles may be part of separate paragenetic stages.
In general, the fine gold particles are likely to be relatively
disseminated through the orebody (or ore shoot) and responsible
for a ‘background’ grade of between 0.5 g/t Au and 4 g/t Au. The
coarse particles are more likely to be dispersed in nature and
probably clustered. In addition, the coarser particles are more
likely to be related to higher grade zones or shoots. General
observations by the authors of numerous mesothermal/lode-gold
vein deposits suggest that this background fine/gold versus
coarse gold relationship is common. It should be noted that the
background grade is not the same as the mineralisation indicator
grade, though in some cases they may have close values.
Pitard (2007) proposed a mathematical approach (the
Ingamells method) to the estimation of background gold content.
The authors apply a combination of statistical analysis of assays
(from heterogeneity studies and drill core, etc), supported by
mineralogical and metallurgical determination.
By defining the theoretical particle distribution and sizing for
one tonne of ore, it is possible to model the probabilities of
achieving certain numbers of particles in a given sample mass
using a Poisson distribution (Dominy, 1997).
For example, given that one tonne contains approximately
100 coarse gold particles (∼2500 µm particles) representing a
grade of 15 g/t Au, the mean number of gold particles in 1 m
of NQ core (~5 kg) sample is 0.5. This gives a 61 per cent
probability that it contains no gold particles, a 30 per cent
probability that the core will contain a single particle and a
95
S C DOMINY, Y XIE and I M PLATTEN
seven per cent probability that it will contain two gold particles.
It is easy to see how a tonne of ore with an expected grade of
15 g/t Au can be severely understated. If clustering occurs, the
effect will be more pronounced. Using the above case where
100 gold particles cluster into just two-cluster groups, then the
probability that the whole core sample will contain zero particles
is 99 per cent, one particle one per cent and three particles
zero per cent.
If one tonne of ore contains 800 000 60-µm fine ‘background’
gold particles, they will contribute about 1.7 g/t Au to the total
grade. In this case the probability of missing gold particles in the
same sample is effectively zero. The core will have a Poisson
mean of 4000 contained gold particles. Thus for a total grade
16.7 g/t Au, drilling will generally resolve the background
fine-gold grade well, but poorly resolve the grade of the coarse
particles (assuming 100 2500-µm particles) which contribute
89 per cent of the total gold inventory.
Identification of the background grade and associated gold
particle sizing characteristics is important, since it may represent
the only part of the mineralisation that can be drilled and
sampled effectively. For example, if a deposit with a mean grade
of 10 g/t Au has a background grade of 2 g/t Au related to fine
gold and the rest of the inventory (8 g/t Au) is related to coarse
gold, then 70 per cent of the contained gold can potentially be
severely under-called by drilling. In addition, the break-even
cut-off grade of the deposit is important. In another example, the
background grade is 6 g/t Au, the cut-off grade 4 g/t Au and the
mean grade 12 g/t Au. In this case drilling is likely to evaluate
the background grade which is above the cut-off grade
effectively. The coarse gold population will yield up to an
additional 100 per cent of gold above the background. In this
example, the situation is manageable, whereas in the former
example the cut-off grade (assuming 4 g/t Au) was within the
high-grade fraction thus making evaluation more challenging.
Gold particle abundance
Together with particle size, the abundance of gold particles has a
direct control on local gold grade. A greater abundance of gold
particles will lead to a higher grade. For example, a single 0.5 cm
diameter gold particle in one tonne of rock will yield a grade of
1.2 g/t Au, whereas 100 particles either dispersed through the
rock or clustered will yield a grade of 120 g/t Au.
SAMPLING APPLICATIONS
Introduction
Sampling is a critical component to all stages of a mining project
as it forms the basis for mineral resource and ore reserve
estimates. It includes the sampling of in situ material and broken
rock. In all cases, the aim is to gain a representative sample to
accurately determine the grade of the material in question. Field
sample collection is followed by sample reduction in both mass
and fragment size to provide an assay charge for analysis.
Errors can be introduced during sample reduction and
homogenisation, especially in the presence of coarse gold
(Royle, 1989; Dominy et al, 2000; Dominy, 2004). Sample
reduction relates to the method by which samples are reduced in
mass for further preparation and analysis; with inadequate jaw
crushing prior to sample splitting being a common source of
error. Errors can be introduced at each stage in the sample
reduction and splitting process, not only because of the selection
of an inadequate sample volume, but also because of
contamination and a poorly homogenised sample pile, which
may result in non-representative subsamples. Further errors may
also be introduced by poor analytical procedures and data
transcription.
96 Ballarat, Vic, 14 - 15 October 2008
Application of Gy Sampling Theory
Introduction
A commonly applied tool to optimise sampling protocols during
evaluation is the Gy Sampling Equation for the Fundamental
Sampling Error (‘FSE’: Gy, 1982; Pitard, 1993; François-
Bongarçon, 1993; François-Bongarçon and Gy, 2001).
The Gy Equation is the only sampling theory commonly
applied in the mining context, and addresses key sampling
questions. These include:
• what weight of sample should be taken from a larger mass of
ore, so that the FSE will not exceed a specified variance;
• what is the possible FSE when a sample of a given weight is
obtained from a larger lot; and
• before a sample of given weight is drawn from a larger lot,
what is the degree of crushing or grinding required to lower
error to an acceptable level?
The sampling constant (K, as defined in the Gy Equation: Gy,
1982; Pitard, 1993) is dependent on the microscopic properties of
the minerals. Specifically, K is proportional to gold particle
shape and size, and inversely proportional to gold grade. As the
liberation (particle) size increases, the K value also increases. A
large K value is related to samples with larger gold particle sizes
and a lower grade. It has the dimensions of g/cmα for gold, where
α usually takes the value of 1.5 (François-Bongarçon and Gy,
2001). The value of K is an important and generalised measure
of ore characteristics, where values below 150 g/cm1.5 pose little
or no sampling challenge and above 5000 g/cm1.5 may lead to
extreme issues (Dominy and Petersen, 2005).
In situ sample size
The Gy Equation has traditionally been used to optimise
sampling protocols for broken rock (eg material fragmented by
blasting, crushing or grinding), but there is no reason why it
cannot be used to indicate the in situ sample mass. The
variability modelled is the ‘in situ nugget variance’ rather than
the FSE. In this special case, the liberation term of the formula
does not exist, and the gold particles take the value for the
particle fragment size of the lot mass. Thus for a given accepted
variability (eg ±15 per cent), the sample size can be determined.
For example, a mineralised vein of grade 5 g/t Au and gold
fragment size of 1000 µm requires an in situ sample mass of
about 600 kg to be taken. The same mineralisation dominated by
fine 50 µm gold particles requires a 5 kg sample.
Like any mathematical approach, this methodology must be
considered to be an estimate of the likely required sample mass
and should be tempered with realism and geological knowledge.
Gold particle size – the liberation diameter
Sampling liberation diameter – definition
A problem of applying the Gy Equation is the definition and
determination of the liberation diameter (François-Bongarçon
and Gy, 2001). In the strictest sense the liberation diameter (dL)
is defined as the particle size of a grind that liberates 95 per cent
of the mineral of interest (Gy, 1982). However, in the presence of
coarse gold this diameter, strictly applied, can be sub-100 µm
and lead to inadequate sample protocols.
Gold particle populations commonly include a wide range of
sizes, commonly several orders of magnitude and spanning the
coarse-fine boundary. Once fine gold represents a substantial
quantity of the total gold present, it is likely to define the dL.
Clearly this information is important for ore processing, but is of
Narrow Vein Mining Conference
 GOLD PARTICLE CHARACTERISTICS IN NARROW VEIN DEPOSITS – IMPLICATIONS FOR EVALUATION AND METALLURGY

limited use for calculating K as the particles most responsible for Bendigo example above, the background mineralisation requires
the nugget effect are ignored. For sampling procedures in gold an an in situ sample of about 20 kg, whereas the coarse gold
alternative definition of dL is required. (>500 µm) mineralisation requires an in situ sample of close to
For example, in a recent case the dL value was found to be 15 t. The run-of-mine ore indicates an in situ sample size of 10 t.
85 µm and yielded a K value of 300 g/cm1.5. However, the ore
contained abundant visible gold up to a maximum of 1000 µm in
Gold particle shape – the shape factor
size, yet a low K value indicated a relatively simple sampling The shape factor (‘f’) as defined by Gy (1982) is a dimensionless
protocol. In reality, the protocol used was inadequate for the factor that gives a measure of how different the shapes of the
coarse gold fraction. Based on a mineralogical/metallurgical gold particles are from an ideal cube. It is dependent upon the
study, K was recalculated for a dL value representing the mineralogical characteristics of the ore and to the degree of
95 per cent passing (dLmax) value for the gold particles. The dLmax grinding – hence gold liberation. As the f value reduces, so does
value of 750 µm yielded a K of 8000 g/cm1.5. Thus for gold the sampling constant K (assuming all other parameters remain
deposits, the dL value should be defined as dLmax to represent the the same).
coarse fraction. If gold particle clustering is observed, then the The value for a perfect cube is 1.0 and for needles 0.1. Most
combined clustered-particle liberation diameter (dLclus) needs to minerals have a shape factor of between 0.2 (eg flaky particles)
be defined (Dominy and Platten, 2007). and 0.5 (eg approximately spherical particles). The value 0.5 is
generally taken as the default value for in situ particles (eg
non-liberated gold particles in rock fragments, etc), whereas
Practical issues of the liberation diameter liberated/pulverised gold particles are likely to have a value
For the purpose of sampling protocols, it should be noted that the approaching 0.2.
dLmax value and sample fragment size reduce in parallel. Figure 2 shows coarse gold particles from Bendigo, liberated
Crushing rarely has a notable effect, except on ores with very from their host as a result of processing. Those liberated by
coarse gold particles (potentially >0.5 cm in size). However, crushing retain a shape indicative of their in situ form, whereas
pulverising to P90 -75 µm can have a marked effect on the gold those liberated by hammer milling have been transformed to a
particle sizing. For fine particle sized ores, pulverising will spherical shape.
liberate gold and lead to some gold particle size reduction. For
coarse gold dominated ores, pulverising will result in gold Gold particle abundance – gold grade
particle size reduction, but it is possible that >100 µm particles Sampling calculations should be undertaken at an appropriate
will still dominate the pulp. This results in high pulp- gold grade likely to be either the run-of-mine grade or breakeven
heterogeneity and provides the well-known challenges of cut-off grade. Optimisation around the cut-off is often considered
preparing and assaying coarse gold ores (Royle, 1989; Berkman, to be most appropriate to discriminate between ore and waste.
1998; Dominy et al, 2000; Dominy and Petersen, 2005). Table 3 summarises gold particle characteristics and key
descriptive terminology.
In situ issues of the liberation diameter
There is observation and modelling-based evidence to suggest TABLE 3
that the quantity of coarse gold and dLmax value increases with Key characteristics for the description of gold particles.
increasing gold grade (Dominy, 1997, 2004; François-Bongarçon
and Gy, 2001; authors’ unpublished data). For example, at Gold particle Descriptor Comments
characteristic
Bendigo a good correlation was noted between increasing bulk
sample grade and more visible gold particles/particle clusters Size Coarse versus fine Determined by observation and
seen in faces (Quigley, 2006). The real situation may be more particles; size in various methods. Is defined as
complex and involve a change in the ratio between coarse and microns dLmax.
fine gold particles and/or the role of clustering (Dominy and Shape Spherical to Determined by observation and
Platten, 2007; Dominy, Xie and Platten, 2008). With increasing irregular or various methods. Is defined as f
grade dLmax increases as the background gold grade (dominance massive (shape factor).
of fine gold particles) is breached. In many cases, as the quantity Distribution Disseminated Can be determined by
of coarse gold increases so clustering becomes more prevalent versus clustered observation and various methods.
thus dLmax becomes dLclus. If clustered, defined as dLclus.
Should also consider gross
As noted previously, a common feature of many narrow gold location of particles such as
veins is the presence of low-grade background mineralisation. A related to stylolites, vein margins,
heterogeneity study at Bendigo (Johansen and Dominy, 2005) etc.
yielded a background grade of 1.3 g/t Au with a dL value of
Abundance Rare (very Can be relatively determined by
~50 µm, which underpined a mean grade of 10 g/t Au and a dLmax low-grade) to observation. Gold grade can be
or dLclus approaching 10 000 µm. This is backed up by both abundant defined from sampling, etc.
historical and modern drilling data that suggest that low-grade (high-grade)
reefs and low-grade sections of high-grade reefs yield a grade of
approximately 1.5 g/t Au (in the 0.5 g/t Au to 3 g/t Au range). A
theoretical model for Bendigo ore presented in Johansen and Concluding comments
Dominy (2005) suggests that there are about 70 >500 µm The Gy Equation thus accounts for gold particle size (dLmax or
particles that account for 70 per cent of the gold inventory in one dLclus), shape (f) and abundance (grade).
tonne of ore. These gold particles are the most difficult to sample The FSE is sensitive to the largest particles present in any
effectively. Poisson statistics show that there is a 70 per cent sample (represented by either dLmax or dLclus), whereas the finer
probability of not hitting any >500 µm particles, with a generally lower-grade background population particles often
probability of hitting one (25 per cent), two (four per cent) or result in a relatively small FSE. Attention must be paid to the
three (0.5 per cent) particles. basic field sampling step, as large errors can occur at this time (eg
In some cases it appears that the dL value approximates to the delimitation error, extraction error, etc). Errors are cumulative
background fine gold-grade. By default the lower grade fine-gold and will continue through the sampling process (Gy, 1982;
population will require smaller samples to be taken. In the Pitard, 1993; Dominy, 2004; 2007).

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 97

S C DOMINY, Y XIE and I M PLATTEN
The above observations confirm the challenge of applying the
Gy Equation to narrow vein gold systems, and reinforce the need
for full material characterisation.
Theoretical approaches rarely emphasise the need for
verification in the context of actual mineralogy and tests for
consistency of results in practice. When Gy Sampling Theory is
applied in the context of gold sampling, it should be tempered
with a good understanding of local geological and gold particle
size variability to maintain reality.
METALLURGICAL ISSUES
Overview
Gold ores are commonly classified as either free-milling or
refractory. Typically, free-milling ores are defined as those where
over 80 per cent of gold can be recovered by conventional
cyanide leaching. Refractory ores are those that give a low
recovery from conventional cyanide leaching.
Gold mineralogy has a significant impact on gold metallurgy.
Key mineralogical factors affecting gold recovery are particle
size, deportment, liberation parameters and silver content of
gold minerals. In addition, the proportion of very fine or
submicroscopic (<1 µm) gold is important. Submicroscopic gold
refers to gold contained in the structure of other minerals in a
minor or trace quantity.
Gold particle size issues during metallurgical processes relate
to liberated gold after grinding. In general terms, gold ores are
milled to P80 between 100 µm and 300 µm. As such the true
in situ particle size does not control metallurgical recovery per
se, though it has a marked influence given that coarser gold
particles will require more effective grinding to reduce particle
size. For coarse-gold particle dominated mineralisation, the focus
should be on liberation and the early removal of coarse particles
from the mill circuit by gravity concentration or ‘gold traps’
(McLean et al, 2007). Finer gold particles will require more
grinding to liberate prior to flotation or cyanide extraction.
Metallurgical tests, like assays, are undertaken on samples of
mineralised rock. These can be from in situ material, cores and
chippings. They can be single samples or composites of several
individual samples. Samples will be split into subsamples to
apply different tests. Samples and subsamples need to be
representative of the range of output from the mine and mill feed.
The factors and protocols required to produce good assays, are
also applicable to samples for metallurgical testing.
Gold particle size and cyanide-based extraction
Liberated gold particles are the key to effective extraction by
cyanide. The material to be leached is ground to a size that
optimises gold recovery and comminution costs, typically
between P80 -150 µm to P80 -45 µm (Marsden and House, 2006).
In a few cases, whole ore is being ground to P80 -20 µm to 25 µm
for optimal processing, either by oxidative pretreatment and/or
leaching. Agitation leaching is rarely applied to material at sizes
coarser than approximately 150 µm because it becomes more
difficult to keep coarse solids in suspension, and abrasion rates
increase.
In some refractory (fine gold) systems, ultra-fine milling has
been used to produce a P80 -7 µm to 10 µm product prior to
agitation leaching and/or oxidation pretreatment.
Some narrow vein systems have attributes that do not lend
themselves to high recovery through the direct cyanidation route
(Laplante and Gray, 2005). Coarse gold and gold associated with
complex sulfide minerals tend to complicate the cyanidation
process. The presence of coarse gold particles increases the
residence time required to achieve high recoveries by
cyanidation. Complex metallurgy can result in coatings on gold
particle surfaces that render them impervious to cyanide. Fine
98 Ballarat, Vic, 14 - 15 October 2008
gold locked into the sulfide crystal lattice (eg submicroscopic
gold) or attached to a sulfide particle will generally report to the
tails stream. The above issues will be exacerbated where the
grind product is coarse.
Gold particle size and flotation-based extraction
Like other methods, particle size is an important parameter in
flotation and size limits exist at which minerals will not float.
The high particle-density of gold and its malleable and ductile
properties that favour platy particles compound this effect.
During grinding gold particles can be impregnated with
non-floatable particles (eg quartz) that will inhibit flotation
(Dunne, 2005). In addition, some gold particles can be rendered
passive by steel grinding media.
Dunne (2005) notes that the maximum size limits for gold
particles to float are between 5 µm and 200 µm. Fine gold
particles down to 3 µm have been floated in laboratory
experiments (Allison and Dunne, 1985), whereas coarse particles
to 300 µm and 700 µm have been floated. Flotation of 590 µm
particles using industrial scale unit float cells has been achieved.
Gold particle size and gravity recovery
Gravity extraction of gold provides a convenient and
environmentally friendly process option in many cases. Deposits
dominated by coarse gold particles can yield gravity recoveries
of above 70 per cent. At Bendigo (Australia) for example,
recoveries of above 75 per cent were achieved from very coarse
gold (>60 per cent above 1000 µm in size) dominated ores
(McLean et al, 2007).
The key controlling parameters for gravity recovery are:
particle size, degree of liberation, mineral density differential,
particle shape, composition and hydrophobicity. The parameter
gravity recoverable gold (‘GRG’) is determined in the laboratory
to measure the potential for gravity recovery in an ore sample
(Laplante, 1996; Subasinghe, 2007).
Recent innovations in extreme gravity recovery have led to the
development of devices that routinely recover liberated fine-gold
(<100 µm) without the use of chemical methods (Van Kleek,
2000; Laplante and Gray, 2005). Centrifugal-type concentration
systems have been at the forefront of this development,
potentially recovering 80 per cent or more based on a
predominant gold distribution of 50 µm and above. Very fine
gold (<5 µm) is poorly recovered by both gravity, unless locked
in recovered sulfides. A modern approach to gravity recovery is
to concentrate gold as both liberated gold and gold-bearing
sulfides such as arsenopyrite. The gold in the mixed concentrate
can be extracted using an intensive leach reactor (Gray and
Hughes, 2007).
The efficiency of collection of gold by gravity concentrators is
controlled by the settling ratios of the mineral components of the
system in question and it has been shown that the shape of the
particles has a marked effect on settling velocities. Liberated
particles of gold become progressively flattened with grinding,
yielding flakes which can be difficult to recover by gravity
processes.
The key issues resulting from flaky particles are: floating due
to hydrophobic nature; particle orientation that leads to
hydrodynamic effects (‘particle surfing’) which can prevent
particle settling (Knipe and Chryssoulis, 2005); and particle
geometry that controls settling velocity. Laplante et al (1995)
undertook experimental tests and showed that spherical tungsten
particles of 25 µm could be recovered effectively, whereas larger
flaky gold particles were lost.
This effect will be enhanced considerably when the in situ
gold particles are very coarse. The thickness of a gold particle
has been shown to be the single most important factor in
determining behaviour in a gravity separation operation.
Narrow Vein Mining Conference
 GOLD PARTICLE CHARACTERISTICS IN NARROW VEIN DEPOSITS – IMPLICATIONS FOR EVALUATION AND METALLURGY

GOLD PARTICLE MODELS 2005, 2006; Dominy and Petersen, 2005; Dominy et al, 2006).
The operation is based on an Indicated Mineral Resource of
A major part of characterisation is to determine the in situ gold 535 000 t (18 g/t Au) and an Inferred Mineral Resource of
particle sizing relationships within an orebody. Studies have 1.5 Mt (17 g/t Au) for a total of 1.1 Moz gold (Crew, 2007).
shown that in many gold deposits, different grade-particle The principal orebody is the Main Vein structure, which varies
scenarios can exist in different or similar domains. For example, in width from 0.05 m to 1.6 m (mean 0.7 m) and contains
there may be substantial differences between oxides versus significant gold grades (up to 5000 g/t Au over 0.8 m). Gold
primary ores. Thus for each mine or vein system a theoretical particle size ranges from a few µm up to 8 mm, with coarse
gold particle model (‘GPM’) can be defined. A series of GPM visible gold being common in the very high-grade (>30 g/t Au)
case studies follow. sections of the vein.
Crushing and screening tests were undertaken on Nalunaq ore
Gwynfynydd gold mine, Wales, United Kingdom to investigate gold particle sizing, the results of which are shown
on Table 5 (Dominy and Petersen, 2005). The mean percentage
This example is taken from the former high-grade Gwynfynydd
of coarse gold (proportion of >105 µm) for the four samples was
gold mine in North Wales. Mineralisation is primarily associated
56 per cent. As expected, the proportion of coarse gold
with the footwall of a 3 m to 5 m wide mother vein. The quartz
(>105 µm) broadly increases with grade. Notably the very coarse
vein contains sphalerite, galena, pyrite and chalcopyrite, with
fraction (>825 µm) percentage does not change significantly with
minor carbonates. Gold is generally coarse-grained (up to 3 mm),
grade.
though highly variable, and locally associated with light brown
coloured sphalerite. Within the veins over 65 per cent of the gold
reported to the >300 µm fraction and >80 per cent to the TABLE 5
>150 µm fraction. Less than ten per cent generally reported to the Results of tests to determine the gold particle size distribution in
<50 µm fraction. samples from the Nalunaq gold mine. Figures should be
Gold was extracted by a small gravity plant, yielding a considered minimum values as some gold particle size reduction
will have occurred during processing.
94 per cent to 98 per cent recovery (Dominy and Phelps, 2002).
During operation, a detailed study was undertaken to investigate Sample number #LR-5205 #NAL-RCO #LR-5525 #NG23A
the controls on gold mineralisation and the nature of the gold
Sample head grade 111 70 45 18
particle sizing and distribution (Dominy, Phelps and Camm,
(g/t Au)
1996; Platten and Dominy, 2003; Dominy and Platten, 2007).
Work included underground geological mapping, mineralogy/ Fraction size (µm) %Au %Au %Au %Au
microscopy, processing of numerous 50 kg to 100 kg and tens of >850 3 1 2 2
tonnes sample lots, and observation of extracted gold particles. >425 - 850 15 6 6 7
A number of different gold particle-grade scenarios were
>250 - 425 22 17 6 21
identified ranging from very high-grade coarse gold to low-grade
fine gold. Most importantly, particles were seen to cluster in all >105 - 250 17 24 14 13
ore types (Platten and Dominy, 2003; Dominy and Platten, <105 43 54 72 57
2007). Table 4 shows the characteristics of the five different % coarse gold 57% 46% 28% 43%
GPMs at Gwynfynydd. The low-grade clustered scenario (>105 µm)
provided the highest K value (130 000 g/cm1.5), indicating that it
has a high sampling difficultly.
From a metallurgical perspective, gold recovery was found to Gold particle clustering is observed at Nalunaq and often
be lower in the fine gold dominated areas (background grade; occurs with particles >250 µm in size. Isolated coarse gold grains
are widespread. Unlike many nuggety deposits, Nalunaq is
1 g/t Au to 4 g/t Au), reducing to between 80 per cent and
fortunate to be high grade, relatively consistent throughout
85 per cent. Lower efficiency was related to the relatively higher
sections of the Target Block.
level of fine <50 µm gold particles in these areas and the inability
of the plant to recover them. Local small clusters were observed In some restricted (0.01 m3 to 0.03 m3) extreme high-grade
within the background gold mineralisation (Figure 6). zones, the volume of rock may contain up to 20 per cent or more
gold. This would yield a very large effective particle diameter of
upwards of 5 cm (Table 6). However, these gold-rich zones are
Nalunaq gold mine, Southern Greenland relatively rare and highly spatially restricted.
The Nalunaq mine yields approximately 80 000 oz gold per The highest sampling difficulty (K value) ore was related to
annum from an underground narrow vein operation (Snowden, the clustered run-of-mine material with a greater relative particle

TABLE 4
Gold particle scenarios at the Gwynfynydd gold mine.
Scenario Comment on gold particle sizing Grade dLmax or dLclus K value Indicative
(g/t Au) value (µm) (g/cm1.5) in situ sample
mass
Very high-grade coarse gold Large gold particles, often locally forming a more 625 10 000 3000 165 kg
continuous 3D network of gold particles (20 oz/t)
ROM coarse gold Individual isolated particles, or clusters of 20 - 50 16 3700 26 000 1.4 t
particles
ROM coarse clustered gold Clusters of 50 - 150 250 µm to 1000 µm particles 16 3000 19 000 1.1 t
Low-grade ‘background’ fine Individual isolated particles, or more dispersed 3 500 7000 380 kg
gold clusters
Low-grade ‘background’ Clusters of 150 - 1000 250 µm to 700 µm particles 3 3500 130 000 7.1 t
clustered fine gold

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 99

S C DOMINY, Y XIE and I M PLATTEN

TABLE 6
Gold particle scenarios at the Nalunaq gold mine.

Scenario Comment on gold particle sizing Grade dLmax or dLclus K value Indicative
(g/t Au) value (µm) (g/cm1.5) in situ sample
mass
Extreme high-grade coarse Very strong clusters of large gold particles, almost 130 000 50 000 160 10 kg
gold forming a more continuous 3D network of gold
particles (10 - 20% gold)
High-grade coarse gold ore Individual isolated particles or clusters of 20 - 50 60 300 5000 280 kg
particles up to 850 µm in size
ROM coarse gold Individual isolated particles 18 850 2500 140 kg
ROM coarse clustered gold Clusters of 20 - 50 particles to 850 µm particles 18 3000 16 800 1t
Low-grade ‘background’ fine Individual isolated particles, or more dispersed 8 250 1000 55 kg
gold clusters

size (Table 6). A background level of more dispersed gold
particles was always present, though clustering proved to be
important.
The mean gold recovery in Nalunaq ore has been 97 per cent,
with 66 per cent attributed to gravity, 25 per cent to cyanide
leaching and six per cent to flotation. In some areas of the mine,
there has been a reduction in the quantity of coarse gold with less
gravity recovery (57 per cent) and more flotation recovery
(14 per cent).
Crystal Hill mine, Tarnagulla, Victoria, Australia
The Tarnagulla goldfield is located in the Central Victorian
goldfields. During 1995 to 2000, Reef Mining NL mined and
milled approximately 58 000 tonnes at a head grade of 30 g/t Au
from the Nick O’Time ore shoot (McCarthy and Faulkner, 2008).
The ore shoot is a vertical pipe like body extending from 70 m
below the surface. High grades are hosted in laminated veins that
range up to 1.5 m in width around a central core of low grade
quartz which ranges up to 7 m in width.
Gold generally occurs in association with pyrite, arsenopyrite,
galena, sphalerite and chalcopyrite, with minor tetrahedrite.
Isolated gold specks occur in the laminated quartz, with
relatively abundant visible gold. Gold, pyrite and arsenopyrite
also occur in stylolitic sutures developed around the massive
quartz core of the shoot.
Gold particles are relatively disseminated and up to 7000 µm
in size. Local clustering occurs of visible gold particles, up to
5 mm, with rare masses of gold on the centimetre scale observed.
Previous studies provided a broad particle size distribution
within run-of-mine ore (Table 7).
TABLE 7
Gold distribution by size fraction for 30 g/t Au run-of-mine ore at
Tarnagulla.
Gold particle size Mass percentage Fraction grade
(µm) (% Au) (g/t Au)
>425 46 14
300 - 425 11 3
100 - 300 22 7 for ROM ore (Table 9). This yields a K value of 660 g/cm1.5
53 - 100 11 3 based on an ore grade of 10 g/t Au, which indicates a small
<53 10 3
Gold extraction during mine operation was dominated by
gravity methods and yielded a recovery of close to 80 per cent.
Cyanide leaching methods accounted for almost 20 per cent
recovery, giving a total life-of-mine gold recovery of 99 per cent.
Local areas of very coarse gold-dominated mineralisation
yielded a gravity recovery of 97 per cent, with 90 per cent of
gold particles >100 µm in size, and 20 per cent >850 µm in size.
The run-of-mine clustered ore showed the highest K value at
Tarnagulla (Table 8). Mining consistently yielded grades of
>15 g/t Au, with a mean of 30 g/t Au. Similar to Nalunaq, whilst
considered a ‘nuggety’ deposit, the coarse gold was relatively
disseminated, often as clusters, through the laminated veins.
Cononish gold project, Scotland, United Kingdom
The Cononish mine comprises a narrow quartz-vein system, the
Cononish Vein, which has been trenched, diamond drilled and
developed underground along one horizon (Snowden, 2008).
The vein is a steep <6 m wide quartz vein emplaced in
metasediments. It completely postdates metamorphism and the
associated foliations. The Cononish Vein and other associated
quartz veins trend subparallel to the nearby, NE-SW trending,
Tyndrum Fault.
Cononish has a ‘diluted’ Measured Mineral Resource of
53 000 t (17.9 g/t Au), an Indicated Mineral Resource of
63 000 t (10.1 g/t Au) and an Inferred Mineral Resource of
285 000 t (11.2 g/t Au) at a 3.5 g/t Au cut-off for a total of
155 000 oz gold (Scotgold, 2008; Snowden, 2008).
Pyrite is the main sulfide phase within the Cononish Vein. It
occurs disseminated as very fine grains in grey quartz, but in
white quartz forms macroscopic aggregates. In polished section
it is seen as euhedral grains, but has also been commonly
subjected to shattering or brecciation. This fracturing may be
related to the development of pyritic shears and the deformation
associated with formation of the breccias with grey quartz
matrix. There is more than one episode of pyrite deposition. This
deposit, unlike the other examples, has abundant silver, with
silver grades four times gold grades.
A hand specimen and polished thin-section mineralogical
study recorded the presence of over 250 gold particles ranging in
size from <10 µm to 550 µm (Dominy, Xie and Platten, 2008;
Snowden, 2008). Gold particles ranging in size from <10 µm to
550 µm were seen, plus rarer >1000 µm particles. Some ten per
cent of particles were >100 µm in size. Based on the observed
particle sizes, an estimated a dLmax value of 250 µm was defined
sampling difficulty.
Metallurgical testing on Cononish material showed that gold
could be recovered by various methods including gravity,
flotation and cyanide leaching. The highest recoveries were
obtained from the cyanide leach tests (93 per cent to 96 per cent),
with generally good recoveries also from the flotation tests
(89 per cent to 94 per cent). GRG values ranged from 34 per cent
to 66 per cent, reflecting the variable presence of coarser gold.

100 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 GOLD PARTICLE CHARACTERISTICS IN NARROW VEIN DEPOSITS – IMPLICATIONS FOR EVALUATION AND METALLURGY

TABLE 8
Gold particle scenarios at the Tarnagulla Crystal Hill gold mine.

Scenario Comment on gold particle sizing Grade dLmax or dLclus K value Indicative
(g/t Au) value (µm) (g/cm1.5) in situ sample
mass
High-grade coarse gold ore Individual isolated particles or clusters of 20 - 40 120 3500 4 000 180 kg
particles to 1000 µm in size
ROM coarse gold Individual isolated particles 30 1000 2000 110 kg
ROM coarse clustered gold Clusters of up to 50 particles to 1000 µm particles 30 4000 19 000 850 kg
Low-grade ‘background’ fine Individual isolated particles, or more dispersed 8 1000 9000 400 kg
gold clusters

TABLE 9
Gold particle scenarios at the Cononish gold project.

Scenario Comment on gold particle sizing Grade dLmax or dLclus K value Indicative
(g/t Au) value (µm) (g/cm1.5) in situ sample
mass
Single particle, high-grade Larger gold particles 1500 µm 60 1500 1600 100 kg
Single particle, ROM grade Individual isolated particles 10 250 660 40 kg
Single particle, low-grade Individual isolated particles or more dispersed 3 75 360 20 kg
clusters
Clustered particles, ROM Clusters of 50 - 100 100 - 250 µm particles 10 500 1900 115 kg
grade
Clustered particles, Clusters of 50 - 100 50 - 100 µm particles 3 300 2900 180 kg
‘background’ low-grade

The highest sampling difficulty with a K value of 2900 g/cm1.5
was at 3 g/t Au based on low-grade clustered ore (Table 9). In
reality, it is likely that the occurrences of the coarsest gold
particles will be as discrete clusters (possibly over centimetres or
tens of centimetres) and cause higher local grades. Gold particles
are intimately associated with galena and pyrite and related to a
late stage shear.
Summary
Table 10 shows a summary of the gold particle characteristics of
the previous case studies, together with selected examples. The
GPM provides a useful basis to describe the nature of gold
particles in a deposit. Additional information can be presented in
the form of Table 11.
CONCLUSIONS
The size/shape, deportment, distribution and abundance of gold
particles affect both the sampling/evaluation and metallurgical
characteristics of narrow vein deposits. Gold particles can range
from individual disseminated grains, through to greater than
2000 µm in size. At each end of the coarse-gold to fine-gold
spectrum, the samplability of a deposit ranges from relatively
simple for fine-grained disseminated particles, to extremely
difficult for coarse nuggety mineralisation. The proportion of
coarse gold has a direct impact on the effectiveness of sampling,
where more coarse gold will require bigger samples and more
specialised protocols.
In some gold deposits, there is evidence to show that the fine-
and coarse-gold particles are part of separate paragenetic stages.
In general, the fine gold particles are likely to be relatively
disseminated through the orebody and responsible for a
‘background’ grade of potentially between 0.5 g/t Au and 3 g/t
Au. The coarse particles may be more clustered and related to
higher grade zones. General observations by the authors on
numerous mesothermal/lode-gold vein-type deposits suggest that
this background fine-gold versus coarse gold is relatively
common. Coarse gold may represent in excess of ten per cent of
the total gold inventory.
Gold particle clustering is of major significance and often
leads to an enhanced coarse gold problem. This can be due to the
presence of gold particle clusters that give the effect of individual
coarse-gold particles and lead to extreme gold grades and high
in situ variability. These clusters, which may be distributed on
the centimetre-scale or greater, give rise to high-grade variability
in field samples. Clusters are readily recognised in coarse gold
deposits; however, they also occur in fine gold deposits where
they produce an apparent coarse gold style.
The Gy Equation is the only sampling theory commonly
applied in the mining context to control the FSE. It has been
traditionally used to optimise sampling protocols for broken
rock, but it can be used to indicate the in situ sample mass. A
problem of applying the Gy Equation is the definition and
determination of the liberation diameter. In the case of gold, this
diameter represents the 95 per cent passing (dLmax) value for the
gold particles. If gold particle clustering is observed, then the
agglomerated clustered-particle liberation diameter (dLclus) needs
to be defined.
The FSE is sensitive to the largest ‘nuggets’ present in any
sample, whereas the finer generally lower-grade background
population particles often result in a relatively small FSE.
Gold particle distribution and abundance within the
mineralisation have a great effect on gold recovery. This will be
related to both the abundance and size of the in situ gold
particles. Different domains within a mineralised body are likely
to have different gold particle size properties and will need to be
characterised early on in a project to optimise recovery. This
characterisation ties in with resource evaluation needs
(eg sampling) and should be part of wider geometallurgical
studies. The authors suggest holistic studies leading to the
definition of gold particle models for each mine or vein system
to aid evaluation and metallurgical needs (Dominy, Platten and
Xie, 2008; Dominy, Xie and Platten, 2008).

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 101

 S C DOMINY, Y XIE and I M PLATTEN
102
TABLE 10
Summary of gold particle characteristics for selected projects and those noted in the text.

Project ROM grade dLmax or dLclus Percentage Percentage gold K value Indicative Comments References
[worst sampling value (µm) gold 100 µm 1000 µm in size (g/cm1.5) in situ
case grade] in size [maximum sample
(g/t Au) particle size mass†
observed]
Bendigo, 10 [3] 10 000 90% 60% 190 000 10 t Extreme coarse gold deposit yielding very high K value. Johansen and Dominy
Australia [10 000 µm] Clustering observed of both finer and very coarse particles. (2005)
Visible gold common. High GRG of +80%.
Gwynfynydd, 15 [3] 3500 80% 35% 130 000 1.5 t Coarse gold deposit characterised by very restricted high Dominy et al (2001)
Wales [3000 µm] grade pockets grading 10s oz/t Au. Clustering very common.
Lower grade material shows small clusters yielding a high K
value. High GRG of +90%.
Sand Queen, 12 [nd]‡ 4750 83% 60% 64 000 3t High proportion of coarse and visible gold. Bimodal Potter, Sheriff and Collins
Australia [6000 µm] population of gold particles sizes, likely to indicate (2008)
high-grade versus lower grade background population. High
GRG of 60%.
Ballarat, Vic, 14 - 15 October 2008

Nalunaq, 18 [8] 3000 50% <5 16 800 1.2 t Relatively ‘disseminated’ coarse/visible gold particles. Local Dominy and Petersen
Greenland [7000 µm] clustering and very high grade clusters. GRG of >70%. (2005)
Tarnagulla, 30 [8] 3500 80% 30 10 600 0.7 Relatively ‘disseminated’ coarse/visible gold particles. Local Dominy et al (2001)
Australia [7000 µm] clustering and very rare extreme massive gold. High GRG of
80%.
Yellowjacket, 15 [4] 1000 50% 10 7100 0.2 t Disseminated, sometimes visible gold particles, related to Dominy, Platten and Xie
Canada [3000 µm] specific associations with some clustering. Small areas of (2008)
‘sheet gold’ related to fault slickensides. GRG of 70 - 80%.
Cononish, 12 [3] 300 10% <<1 2900 50 kg Minimal visible gold (>300 µm). Considered to be a strong Dominy, Xie and Platten
Scotland [1200 µm] pseudo coarse gold effect due to clustering. Gold closely (2008)
associated with galena. GRG of about 30 - 40% (including in
sulfide concentrate).
Red October, 8 [4] 500 30% <<2 2600 140 kg Relatively fine-grained deposit, with much gold associated Dominy, Xie and Platten
Australia [2000 µm] with pyrite. Occasional visible gold. GRG of about 30%. (2008)
Bindal, Norway 5 [5] 250 40% <<1 2000 80 kg Relatively fine-grained deposit, with much gold hosted in Dominy, Platten and Xie
[1500 µm] arsenopyrite. GRG of 80% (including in sulfide concentrate). (2008)

† Estimate based on ROM grade and dLmax value calculated via the Gy Equation modified for the ‘in situ nugget effect’.
‡ nd = not determined
Narrow Vein Mining Conference
 GOLD PARTICLE CHARACTERISTICS IN NARROW VEIN DEPOSITS – IMPLICATIONS FOR EVALUATION AND METALLURGY

TABLE 11
Key information to define the particle size characteristic of a gold deposit. Example from the former Wattle Gully mine, Australia.

Characteristic
Gold particle size dLmax: 6000 (µm)
>100 µm: 80%; >1000 µm: 35%; >10 000 µm: <5%
Distribution As isolated particles, clusters and masses through ore shoots and local very high-grade zones (>2 oz/t Au)
Clustering observed, dLclus: 10 000 (mm)
Deportment Free gold in quartz and locally associated with margins and fractures in arsenopyrite. Also observed at vein margins,
quartz shale contacts and along stylolitic sutures.
Abundance (grade) ROM grade of 11 g/t Au, locally much higher.
K value (g/cm1.5) 82 000 [dLclus: 39 000]
Gravity recoverable gold (GRG) >85%
Indicative in situ sample size dLmax: 4.5 t [dLclus: 9.5 t]
Background grade Range between 0.5 g/t Au to 2 g/t Au dominated by <100 mm particles.
Visible gold Commonly observed up to 0.5 cm in size, as individual particles and clusters towards 1 cm in size.
Comment Very high-nugget coarse gold dominated system. Large primary sample sizes are indicated in order to be
representative. Large numbers of drill or face samples required to define grade plus use of supporting indicators
(‘proxies’). Core samples generally understate grade and may yield a mean grade of <3 g/t Au in a 10 g/t Au to 12 g/t
Au zone.

ACKNOWLEDGEMENTS Dominy, S C, 2004. Fundamental sampling error and its relationship to

Gold particle studies by the authors have benefited from the
support of numerous companies, including: Bendigo Mining
Limited, Castlemaine Goldfields Limited, Goldfields Limited,
Charters Towers Gold Mines Limited, Clogau St David’s Gold
Mines Limited, Crew Gold Corporation, Nalunaq Gold Mine
A/S, Reef Mining NL, Scotgold Resources Limited and Welsh
Gold PLC. The Dolgellau (Gwynfynydd mine) and Tarnagulla
goldfields are currently being explored by Victorian Gold
Limited; the Nalunaq mine is operated by Crew Gold
Corporation and the Cononish project operated by Scotgold
Resources Limited. The use of mineralogical facilities at the
University of Science and Technology Beijing (USTB) is
acknowledged. In addition, thanks are due to Andy Ross
(Snowden Group) and The AusIMM conference reviewers, Nigel
Grigg and Steve Olsen, for their comments on the manuscript.
The opinions expressed in this paper are those of the authors, and
not necessarily those of Snowden, USTB or named operating
companies.
REFERENCES
Angove, J, 2005. Metallurgical testwork: Gold processing options,
physical ore properties and cyanide management, Advances in Gold
Ore Processing: Developments in Mineral Processing, No 15,
pp 97-107 (Elsevier BV: Amsterdam).
Berkman, D, 1998. Coarse gold problems, ALS Minerals Division
Newsletter, 7(1):1-4.
Bird, H H, 1991. Dealing with coarse-gold and cutting factors, or dealing
with the nugget effect in practice, in Proceedings Symposium on
Sampling and Ore Reserves, pp 35-54 (Prospectors and Developers
Association of Canada: Toronto).
Chryssoulis, S L and McMullen, J, 2005. Mineralogical investigation of
gold ores, Advances in Gold Ore Processing: Developments in
Mineral Processing, No 15, pp 21-71 (Elsevier BV: Amsterdam).
Cintra, E C, Scabora, J A, Viegas, E P, Barata, R and Maia, G F, 2007.
Coarse gold sampling at Sao Francisco mine, Brazil, in Proceedings
Third World Conference on Sampling and Blending 2007,
pp 187-197 (Fundacao Luiz Englert: Porto Alegre).
Crew, 2007. Updated mineral resource and mineral reserve for Nalunaq
gold mine, press release dated 20 April 2007, Crew Gold Corporation,
3 p.
Dominy, S C, 1997. Sampling and estimation of gold deposits,
unpublished research report, 230 p (University of Greenwich:
Chatham and Welsh Gold PLC: Dolgellau).
the nugget effect in gold deposits, in Proceedings EGRU Mining and
Resource Geology Symposium, EGRU contribution No 62, pp 30-45
(James Cook University: Townsville).
Dominy, S C, 2007. Sampling – A critical component to gold mining
project evaluation, in Proceedings Project Evaluation Conference
2007, pp 89-96 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Dominy, S C, Annels, A E, Platten, I M and Raine, M D, 2003. A review
of problems and challenges in the resource estimation of high-nugget
effect lode-gold deposits, in Proceedings Fifth International Mining
Geology Conference 2003, pp 279-298 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Dominy S C and Johansen G F, 2004. Reducing grade uncertainty in
high-nugget effect gold veins: Application of geological and
geochemical proxies, in Proceedings PACRIM Congress 2004,
pp 291-302 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Dominy, S C, Johansen, G F, Annels, A E and Cuffley, B W, 2000.
General considerations of sampling and assaying in a coarse gold
environment, Transactions of the Institutions of Mining and
Metallurgy, Applied Earth Science, 109:B145-B167.
Dominy, S C, Johansen, G F, Cuffley, B W, Platten, I M and Annels, A E,
2001. Estimation and reporting of mineral resources for coarse
gold-bearing veins, Explor Min Geol, 9:13-42.
Dominy S C and Petersen J S, 2005. Sampling coarse gold-bearing
mineralisation – Developing effective protocols and a case study
from Southern Greenland, in Proceedings Second World Conference
on Sampling and Blending 2005, pp 151-165 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Dominy, S C and Phelps, R F G, 2002. Gold extraction at the
Gwynfynydd mine, North Wales, United Kingdom: A case study of
environmentally sustainable mineral processing with a national park,
in Proceedings Green Processing Conference 2002, pp 225-232 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Dominy, S C, Phelps, R F G and Camm, G S, 1996. Geological controls
on gold grade distribution in Chidlaw link zone, Gwynfynydd mine,
Dolgellau, North Wales, Transactions of the Institutions of Mining
and Metallurgy, Applied Earth Science, 105:B151-B158.
Dominy, S C and Platten, I M, 2007. Gold particle clustering – A new
consideration in sampling applications, Transactions of the
Institutions of Mining and Metallurgy, Applied Earth Science,
116:B130-B142.
Dominy, S C and Platten, I M, 2008. Continuity risk in narrow reef
gold deposits – Implications for evaluation and exploitation, in
Proceedings Narrow Vein Mining Conference 2008, pp 31-50 (The
Australasian Institute of Mining and Metallurgy: Melbourne).

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 103

S C DOMINY, Y XIE and I M PLATTEN
Dominy, S C, Platten, I M and Raine, M D, 2003. Grade and geological
continuity in high-nugget effect gold-quartz reefs: Implications for
resource estimation and reporting, Transactions of the Institutions of
Mining and Metallurgy, Applied Earth Science, 112:B239-B259.
Dominy, S C, Platten, I M and Xie, Y, 2008. Bulk sampling of complex
gold deposits: Material characterisation and programme design and
management, in Proceedings Sampling Conference 2008, pp 41-57
(The Australasian Institute of Mining and Metallurgy: Melbourne).
Dominy, S C, Sides, E J, Platten, I M and Dahl, O, 2006. Estimation and
exploitation in an underground narrow vein gold operation – Nalunaq
mine, Greenland, in Proceedings Sixth International Mining Geology
Conference 2006, pp 29-44 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Dominy, S C, Xie, Y and Platten, I M, 2008. Characterisation of in situ
gold particle size and distribution for sampling protocol optimisation,
in Proceedings Ninth International Congress for Applied Mineralogy
(ICAM) 2008, pp 175-186 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Dunn, E J, 1929. Geology of Gold (South Africa, Australia and New
Zealand), 303 p (Charles Griffin and Company Limited: London).
Dunne, R, 2005. Flotation of gold-bearing ores, in Advances in Gold Ore
Processing: Developments in Mineral Processing, No 15, pp 309-344
(Elsevier BV: Amsterdam).
François-Bongarçon, D M, 1993. The practice of the sampling theory of
broken ores, Canadian Institute of Mining, Metallurgy and Petroleum
Bulletin, 86(970):75-81.
François-Bongarçon, D M and Gy, P M, 2001. The most common error in
applying ‘Gy’s Formula’ in the theory of mineral sampling and the
history of the liberation factor, in Mineral Resource and Ore Reserve
Estimation – The AusIMM Guide to Good Practice, pp 67-72 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Gonzales, P and Cossio, S, 2007. A review of sampling protocol for a
gold ore based on liberation study, in Proceedings Third World
Conference on Sampling and Blending 2007, pp 163-174 (Fundacao
Luiz Englert: Porto Alegre).
Gray, S and Hughes, T, 2007. Gravity, flotation and intensive leaching
process for gold recovery improves environmental and cost
outcomes, in Proceedings World Gold Conference 2007, pp 229-238
(The Australasian Institute of Mining and Metallurgy: Melbourne).
Gy, P M, 1982. Sampling of Particulate Materials – Theory and Practice,
p 431 (Elsevier: Amsterdam).
Healey, C M, 1993. Performance of reserve estimation techniques in the
presence of extremely high-grade samples, Jasper gold mine,
Saskatchewan, Explor Min Geol, 2:41-47.
Johansen, G F and Dominy, S C, 2005. Development of sampling
protocols at the New Bendigo gold project, Australia, in Proceedings
Second World Conference on Sampling and Blending 2005,
pp 175-183 (The Australian Institute of Mining and Metallurgy:
Melbourne).
Koch, G S and Link, R F, 1971. Sampling gold ore by diamond drilling in
the Homestake mine, Lead, South Dakota, Report RI-7508, United
States Department of the Interior, Bureau of Mines report of
investigation, p 36.
Knipe, S W and Chryssoulis, S L, 2005. Flaky gold: Problems with
recovery and mineralogical quantification, J Materials, 56(7):58-62.
Laplante, A R, 1996. A laboratory procedure to determine the amount of
gravity recoverable gold – A case study at Hemlo gold mines, in
Proceedings 28th Annual Canadian Mineral Processors Conference
1996, pp 69-82 (Canadian Institute of Mining and Metallurgy:
Montreal).
Laplante, A R and Gray, A H, 2005. Advances in gravity gold technology,
in Advances in Gold Ore Processing: Developments in Mineral
Processing, No 15, pp 280-306 (Elsevier BV: Amsterdam).
Laplante, A R, Huang, L, Noaparast, M and Nickolepotolous, N A, 1995.
A philosophers stone: Turning tungsten into gold – The use of
synthetic ores to study gold gravity separation, in Proceedings 27th
Annual Canadian Mineral Processors Conference 1995, pp 379-394
(Canadian Institute of Mining and Metallurgy: Montreal).
Marsden, J O and House, C I, 2006. The Chemistry of Gold Extraction,
p 651 (Society for Mining, Metallurgy and Exploration Inc: Littleton).
104 Ballarat, Vic, 14 - 15 October 2008
McCarthy, P L and Faulkner, L, 2008. Tarnagulla gold mine, in
Proceedings Narrow Vein Mining Conference 2008, pp 237-248 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
McLean, E J, Hayward, T E, English, D H and Smith, S J, 2007. Process
design aspects at Bendigo Mining’s gold operations, in Proceedings
Ninth Mill Operators’ Conference 2007, pp 175-184 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Olsen, S, 2004. Personal communication.
Pitard, F F, 1993. Pierre Gy’s Sampling Theory and Sampling Practice,
p 488 (CRC Press: Boca Raton).
Pitard, F F, 2007. The in-situ nugget effect: A major component of the
random term of a variogram, in Proceedings Third World Conference
on Sampling and Blending 2007, pp 91-110 (Fundacao Luiz Englert:
Porto Alegre).
Platten, I M and Dominy, S C, 2003. The occurrence of high-grade gold
pockets in quartz reefs at the Gwynfynydd mine, Wales, United
Kingdom: A geological explanation of the nugget effect, Explor Min
Geol, 10:249-272.
Platten, I M and Dominy, S C, 2007. Asymmetrical filling of inclined
crustified hydrothermal gold veins: An example from Hadleigh
Castle mine, North Queensland, in Proceedings Ninth Biennial SGA
Meeting 2007, pp 735-738 (Irish Association for Economic Geology:
Dublin).
Potter, D, Sheriff, C and Collins, P, 2008. Significance of geological
control on assay data in the narrow, high-grade gold quartz veins
at the Sand Queen mine, Comet Vale, Western Australia, in
Proceedings Narrow Vein Mining Conference 2008, pp 71-78 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Quigley, P, 2006. Bendigo mine grade control, in Proceedings Sixth
International Mining Geology Conference 2006, pp 233-240 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Roberts, L S, Dominy, S C and Nugus, M J, 2003. Problems of sampling
lode-gold deposits: Case studies from Australia and North America,
in Proceedings Fifth International Mining Geology Conference 2003,
pp 387-400 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Royle, A G, 1989. Splitting gold assay pulps containing coarse gold
particles, J Leeds Univ Mining Assoc 1989, pp 63-68.
Scotgold, 2008. First JORC Code compliant mineral resource for
Cononish gold-silver deposit, Scotland, press release dated
13 May 2008, Scotgold Resources Limited, p 6.
Sketchley, D A, 1998. Gold deposits: Establishing sampling protocols
and monitoring quality control, Explor Min Geol, 7:129-138.
Snowden, 2005. Independent mineral resource estimate: Nalunaq gold
mine, Greenland, unpublished report by Snowden Mining Industry
Consultants Limited for Nalunaq Gold Mine A/S (author: S C
Dominy, Project No 5122), 76 p.
Snowden, 2006. Independent mineral resource estimate: Nalunaq gold
mine, Greenland, unpublished report by Snowden Mining Industry
Consultants Limited for Nalunaq Gold Mine A/S (author: S C
Dominy, Project No L102), 123 p.
Snowden, 2008. Geological review and mineral resource estimate:
Cononish gold project, Scotland, United Kingdom, unpublished
report by Snowden Mining Industry Consultants Limited for
Scotgold Resources Limited (authors: S C Dominy and I M Platten,
Project No L136), 147 p.
Springer, J, 1983. Invisible gold, in The Geology of Gold in Ontario,
pp 240-250 (Ontario Geological Survey: Toronto).
Subasinghe, G K N, 2007. Evaluating an alternative to the gravity
recoverable gold (GRG) test, in Proceedings Ninth Mill Operators’
Conference 2007, pp 235-239 (The Australasian Institute of Mining
and Metallurgy: Melbourne).
Van Kleek, D M, 2000. Knelson Concentrators extreme gravity, in
Proceedings Seventh Mill Operators’ Conference, pp 277-279 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Wooldridge, R, 1998. Sample preparation and assaying of coarse gold
ores, in Proceedings AusIMM Annual Conference 1998, pp 143-154
(The Australasian Institute of Mining and Metallurgy: Melbourne).
Narrow Vein Mining Conference
The Python – An Underground Processing Plant for Narrow
Vein Mining
T R Hughes1 and N J Grigg2
ABSTRACT
Gekko Systems has developed an underground pre-concentration plant
(Python) to reduce mine operating costs.
Many ores liberate favourably using either high pressure grinding rolls
(HPGR) or vertical shaft impactors (VSI). Using these comminution
circuits for mineral liberation; rather than for the final recovery process
and by utilising a high mass pull, high recovery process route
(gravity/flotation), a narrow, low head, compact processing plant was
developed that can sensibly be operated underground.
The Python underground processing plant can provide a number of
advantages for narrow vein mines including lower haulage costs, improved
mine call factor and lower operating costs. The focus on fine crushing
reduces the power requirement of the plant to approximately 8 kWh/tonne.
The Python underground processing plant has been designed to be
towed down the decline or lowered down a shaft, and is capable of being
installed in sloping, non-linear drives for placement close to the working
face.
The overall benefit to mining companies of this concept has been
estimated to be a saving in mining costs and a significantly smaller
environmental footprint.
BACKGROUND TO PROJECT
During 2004/2005 Gekko Systems was awarded an AusIndustry
research grant to both investigate and commercialise the idea of
underground processing of primarily gold bearing ores.
1. FAusIMM, Research and Development Manager, Gekko Systems Pty
Ltd, 321 Learmonth Road, Ballarat Vic 3350.
Email: timh@gekkos.com
2. MAusIMM, Sales Manager, Gekko Systems Pty Ltd, 321 Learmonth
Road, Ballarat Vic 3350. Email: nigelg@gekkos.com
FIG 1 - Conceptual underground processing plant layout.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
The key concept of the project is to fine crush and
pre-concentrate the ore underground and as close to the working
face as possible with a combination of gravity and flotation
methods. The valuable component is then the only material
removed from underground to the surface while the non-valuable
tailings remain underground for backfilling (see Figure 1).
Gekko Systems has built a prototype processing plant (Python)
to demonstrate the concept and successfully commissioned the
unit on surface in September 2007.
PYTHON PROCESSING PLANT
The principle of the Python processing plant involves size
reduction, screening, gravity and flotation pre-concentration
underground and as close to the working face as possible. In order
to achieve this, a low tonnage Gekko Modular unit has been
designed to be installed in drives at each of the working stopes,
such that as the ore is blasted it is removed by a LHD or similar
unit and delivered directly into the Python’s grizzly feed hopper.
The limited handling minimises the possibility of fines losses into
footwall cracks and the operating costs of internal haulage.
The ore is then crushed and ground to minimum economic
liberation size for gravity and flotation pre-concentration. The
liberation size is determined by testing the ore and by
interpreting typical gravity yield recovery curves for different
type ores. A typical yield recovery curve for a Witwatersrand ore
is presented in Figure 2.
These curves represent the un-optimised single pass gravity
recovery. A rougher/scavenger type application has been recently
commercialised, which increases the recovery achievable in
practice. The addition of a flash flotation module further increases
the recoveries of the fine (-150 µm) fractions.
105
T R HUGHES and N J GRIGG
FIG 2 - Laboratory gravity yield/recovery curves for a typical
Witwatersrand ore.
Gekko are proposing that only the pre-concentration step be
performed underground. The conventional understanding of
gravity concentration is the utilisation of extremely high grade,
very low mass pull equipment (eg centrifugal concentrators);
however the Python exploits the use of a mass pull of ten to
35 per cent using inline pressure jigs resulting in very high
recoveries, typically in the +90 per cent region. This results in
concentrate grades of three to ten times the mined grade, and
removes secondary issues such as security concerns which
normally accompany ‘gravity concentration’. For the Python the
risks are no greater than those associated with any conventional
mining process.
FIG 3 - Python – underground processing flow sheet.
106 Ballarat, Vic, 14 - 15 October 2008
The schematic flow sheet for the Python Processing Plant is
presented as Figure 3.
The Python utilises a jaw crusher and vertical shaft impactor
(VSI) in a closed circuit to achieve a -5 mm product suitable for
the rougher gravity concentrator, an inline pressure jig (IPJ1500).
The tails from the IPJ are screened to remove any oversize back
to the VSI with the undersize product reporting to flash flotation.
The rougher IPJ concentrate is cleaned using an IPJ1000 with the
cleaner concentrate and flotation concentrates transported to the
surface for final treatment. The cleaner tail is recirculated back to
the rougher IPJ and the flash flotation tails are available for
backfill.
The configuration of the plant can be changed to suit individual
ore types and mining methods.
PYTHON SPECIFICATIONS (PYTHON200)
The current Python processing plant (Figure 4) has an estimated
capacity of between 10 - 20 t/h run of mine ore feed.
• Target grind size (P80 = 500 to 800 µm).
• For soft ores – vertical shaft impactor used. For hard ores or
less than 500 µm grind, high pressure grinding rolls can be
used.
• Installed power is 8 kWh/t, excluding pumping of concentrates
and tails to final destinations.
• Labour requirement is estimated as one to two dedicated
operators, one to drive loader (LHD) and operate plant front
end, and a concentrator circuit operator. (This is unlikely to
change as the size of plant increases, unless loader driver
becomes a full time job.)
• Plant dimensions: 2.4 m wide × 5 m high × 67 m long. Plant
can be split in two, same width and height but in two sections
35 m and 32 m long and installed on two mining levels with
piping and power cables run between them.
• Plant is to be designed to be towed down the decline or
lowered down a vertical shaft.
Narrow Vein Mining Conference
 THE PYTHON – AN UNDERGROUND PROCESSING PLANT FOR NARROW VEIN MINING

FIG 4 - Python processing plant.

ADVANTAGES OF THE PYTHON
There are a number of advantages in pre-concentrating the ore
underground, which were the driving forces for the development
of the Python. These are summarised as follows:
• improvement in mine call factor (MCF) due to less handling
points for the ore en-route to the plant;
• reduction in tramming and hoisting costs due to the movement
of lower tonnage;
• no necessity for backfill to be produced on surface and sent
back underground;
• reduced surface plant costs as lower tonnes will be treated at
a higher grade (30 per cent of mass);
• considerable reduced power consumption over conventional
processing (estimate of underground installed power 8 kWh/t
versus 14 - 16 kWh/t using conventional milling power
consumption);
• minimal underground infrastructure required due to plant
capable of being installed in sloping, non-linear drives for
placement close to the working face;
• no detoxification required on the backfill product; as it has
not been exposed to toxic chemicals; and
• general environmental advantages, lower power, less surface
disturbance, noise, etc.
Some site specific issues that Gekko have identified that will
need to be addressed before full implementation underground
include:
• the concept of lower recovery at the face being offset by an
improved MCF and lower operating costs;
• heat generated by the underground processing facility needs
to be managed;
• mine layout will need to be designed around the principles of
underground processing, for pumping of concentrates, swell
factor, backfill, etc; and
• the use of multiple units will need to be considered, ie are
1 × 100 t/h units centralised better than 2 × 50 t/h units
localised.
Gekko Systems strongly believe that the Python underground
processing plant will be of benefit to mining companies with
substantial savings in mining costs, both capital and operating
and a significantly smaller environmental footprint.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 107

Mechanised Ore Removal from the Floor of Narrow, Flat Stopes
S L Tombs1, A Buckingham2 and R M Hancock3
ABSTRACT
The underground Sunrise Dam gold mine commenced in October 2003 to
extract ore from a series of orebodies with various dips, widths, and ore
grades over a strike length of 1.5 km. To successfully operate flat stopes
it was identified that a means to extract ore left behind after blasting in
stopes with dips ranging from 15 to 40 degrees was needed.
Various options were available, which included a bulk water cannon,
hand held high-pressure water jet, remote controlled small dozer, several
mini excavators and finally a specialised remote controlled excavator
from the demolition industry.
Trials were commenced for the bulk water cannon option in 2006 that
proved that the concept worked, however, the system was too
cumbersome to set up for use in cycle with production. Hand washing
down using mine (low pressure) water with a two-inch bullhose proved of
limited value. The small, one metre high remote controlled dozer as used
in South Africa was reviewed, but the steepness of most of the stopes
precluded this machine. The hand held high-pressure water jet option is
still being investigated; with safety factors still being addressed.
Several mini excavators were trialled to clean out the flat stopes with
generally poor results, the machines being either too large to fit in the
drives and pull ore down or too small with too short reach to effectively
and efficiently pull ore out of the stope. In early 2007 a Brokk 330 remote
controlled excavator was trialled on site with the low profile of the
machine and unique arm configuration proving an effective tool for
pulling ore from the footwall of short, flat stopes.
A Brokk 330 remote controlled excavator was purchased in late 2007
and is currently in operation.
INTRODUCTION
AngloGold Ashanti (AGA) own and operate over 20 gold mines
throughout the world with the Sunrise Dam Gold Mine being the
only operating mine in Australia. Currently the 33 per cent
owned Boddington Gold Mine is under construction and AGA’s
75 per cent owned Tropicana Project in the feasibility stage. The
Sunrise Dam Gold Mine is situated some 1000 km east of Perth
and is operated as a fly-in, fly-out operation with Downer the
open pit contractor and Barminco the underground contractor.
Underground mining commenced at Sunrise Dam in October
2003 and stoping in the SSZ ‘flat’ stopes in May 2005, with
37 km of development and 1.6 Mt of ore mined at an average
grade of 6.8 g/t completed to July 2008.
Gold mineralization at SDGM is structurally controlled and
vein hosted. The understanding and interpretation of the
structural characteristics has improved substantially with the
increased exposure created by the open pit excavation and the
construction of the underground mine. This structurally complex
series of orebodies exhibits a variety of ductile and brittle
deformation fabrics that influence the nature, geometry and
distribution of the mineralisation. Mining is undertaken on a
variety of different orebodies, ranging from narrow and flat
high-grade to large and steep medium-grade.
1. MAusIMM, Underground Manager, Sunrise Dam Gold Mine,
AngloGold Ashanti Australia.
Email: STombs@AngloGoldAshanti.com.au
2. Contract Mining Engineer, GPO Box F377, Perth WA 6001.
3. Senior Planning Engineer, Sunrise Dam Gold Mine, AngloGold
Ashanti Australia. Email: RHancock@AngloGold Ashanti.com.au
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
STOPING
The shallow dipping lodes have been extracted using a jumbo or
long hole stoping methods, dips varying from 15 to 50 degrees.
Shallow dipping stoping has accounted for 50 per cent of the
mines stoping tonnes, however, this ratio will drop to about
30 per cent when bulk stoping commences.
The jumbo stopes are formed in 3 separate blast rounds (cuts)
as seen in Figure 1, which advance the excavation a nominal
2.5 m up or down dip. One 2.5 m cut is made up dip from the
lower drive, and after rock bolting is then extended to support the
up dip cut, a second up dip cut is bored out but not fired. A 2.5 m
cut is then made down dip from the upper ore drive, and then the
second up dip cut is fired to complete the stope. This sequence
ensures that the drilling of the second up dip cut does not bore
into unidentified misfires created by the down dip cut.
FIG 1 - Jumbo stoping section – Mako lode.
The long hole stoping areas were opened up by developing
parallel ore drives along strike or along an apparent dip angle. The
dip stopes were then drilled using a long hole drill with holes
aligned parallel to the stope axis along the ore dip, normally
dumped forward by 20 degrees as shown in Figure 2. The holes
were drilled through to the upper ore drive, so that alignment and
location could be verified before stope charging. Hole lengths of
up to 12 m, depending on the separation of the ore drives and the
dip of the orebody. Final extraction ratio could reach as high as
85 per cent. A minimum mining width (true width) of 1 m is
required for the stope design. An additional 0.8 m of overbreak is
assumed (0.3 m footwall; 0.5 m hanging wall).
The common problem with both methods was removing ore
from the footwall of the mined stopes as the floor of the stope
was too shallow to allow the ore to rill and access to the mined
stope was not possible.
POSSIBLE METHODS FOR ORE REMOVAL FROM
FLAT STOPES
Three systems were reviewed for ore removal from the flat stope
footwalls. AngloGold Ashanti initially trialled a water cannon
system, reviewed a small remote controlled dozer and finally
chose a mini excavator. A comparison of the methods reviewed
can be seen in Table 1.
Water cannon trial
Approval was granted for a water cannon trial in early 2006,
AGA hired two large open pit diesel pumps – one to feed the
water cannon from two ten-tonne tanks and the other to pump
111
S L TOMBS, A BUCKINGHAM and R M HANCOCK

TABLE 1
Comparison of the different systems available for removing ore.

System Water cannon Reef Shark dozer Brokk 330 D excavator Mini excavator
Set up Two shifts per stope panel One hour One hour One hour
Remote operation Yes Yes Yes No
Productivity 15 t/hour 15 t/hour (?) 30 t/hour 5 t/hour
Ease of transport Tie up loader and IT for 30 minutes 30 minutes 30 minutes
up to two shifts
Versatility Rigid use Only one use Adapt to rock breaker, scale Minimal other uses
bad areas, remove bridges
Costs Estimated $400 000 Estimated $450 000 $320 000 Estimated $150 000
Ease of use Requires training – Requires training and skill in Requires training and skill in Requires training and skill in
not difficult operation of remote equipment operation of remote equipment operation of excavators
Acceptance Very negative Medium – Anglo would High – Anglo would operate Medium – Anglo would
operate operate
Ability to use in cycle Not suited Ideal for use in cycle Ideal for use in cycle Ideal for use in cycle
with stoping

FIG 3 - Section through stope showing 30 per cent ore left in rill.

FIG 2 - Long hole stoping section.
return water from the sump back to the tanks. The system was
operated as a closed circuit. The remote controlled 100 mm
aperture water cannon operated at 50 litres per second at a
pressure of 15 bar.
The trial proved that it was possible to clean our stopes
hydraulically at a rate of 15 tonnes per hour; however, the nature
of the trial set-up meant that the water cannon only operated
for 25 per cent of the time and was being moved for 75 per cent
of the project time. The trial cost $200 000 and resulted
in 3000 tonnes of additional ore being recovered – a large
component ($90 000) of this cost was incurred through hourly
hire of the contractor’s personnel and equipment.
The outcome of this trial was that the method was too slow,
held up stoping operations and needed to be a mobile system to
be used ‘in cycle’.
Remote controlled dozer
Two suppliers of remote controlled dozers were contacted,
however, AGA did not proceed with that system due to the
variable dips found in the stopes that would make traction
problematic and the use of a winch likely to assist in pulling the
machine back up the stope.
Mini excavator
Two manually operated mini excavators were trialled
underground, both had issues with the boom configuration
limiting the height up the stope that the mini excavator could
reach, this coupled with the requirement for remote operation led
to an alternate small excavator being considered, the Brokk 330.
The Brokk was trialled in a series of stopes in Mako that had
already been ‘cleaned’ with a small mini excavator and pulled
out an additional 300 tonnes of high-grade ore over the two-day
trial at a rate of 30 tonnes per hour.
The purchase of the Brokk 330 was justified on the assumption
that AGA could recover ore from the flat stopes at a rate of
30 tonnes per hour, recover ten per cent of ore blasted with an
average grade of 7 g/t and use AGA staff as well as those of
Barminco Mining Contractors to operate the machine. The mine
had 1.1 million tonnes of ore to be mined from ‘flat’ stopes at the
time the decision to purchase the Brokk 330 was made.
OPERATION OF THE BROKK 330
REMOTE EXCAVATOR
Table 2 details the performance of the Brokk excavator over a
four-month period this year. The machine is not in constant use
and has not been able to recover ten per cent of stope ore from
the flat stopes as anticipated and neither has it extracted ore at a
steady rate of 30 tonnes per hour (recovery rates varied from
15 to 50 tonnes per hour). Ore from flat stoping comprised about
20 per cent of ore production from the mine during the period
April through to July 2008.

112 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 MECHANISED ORE REMOVAL FROM THE FLOOR OF NARROW, FLAT STOPES

TABLE 2 production rate of 30 tonnes per hour (only 28 t/h achieved) –

mainly due to a shortage of trained operators and double handling
Operating statistics over a four-month period.
of the ore in some of the larger stopes.
April May June July Totals There have been challenges in keeping the Brokk 330
2008 2008 2008 2008 operating due to a shortage of trained operators, this will be
Operating hours 40 5 8 74 127 rectified in 2009 when four trained technicians are transferred to
Days operated 9 2 2 17 30
the underground from the open pit – all are experienced backhoe
operators. The machine is currently in Perth getting a series of
Stopes cleaned 5 1 2 5 13 modifications completed that will minimise the damage caused
Brokk ore extracted 1180 160 330 1930 3600 by rocks being pulled back towards the machine, this combined
Total ‘flat’ stope ore 10 019 9418 13 581 11 650 44 668 with the dealer sending a maintenance fitter to site monthly will
improve the availability of the Brokk 330.
Brokk as a per cent of 12% 2% 2% 17% 8%
stope An example of the benefits of having the Brokk 330 on site
can be seen in Figure 4. This shows a section through the Sunrise
Shear South (SSS) 2006 B stope where the green line represented
the ore rill after bogging and the purple line the CMS outline
There have been many issues adversely affecting the use of the
after clean up with the Brokk 330. The 555 ore tonnes took
Brokk 330 excavator, many are detailed below:
33 hours to remove with the Brokk 330 at a rate of 17 tonnes per
• issue with retaining trained operators from both the client hour and cost of $5300 or $9.50 per tonne – stope grade was
and contractor; 19.2 g/t and an additional 295 ounces produced.
• focus of mining technicians on other stoping activities;
• focus on higher grade stopes, lower grade stopes may not get
cleaned out;
• normally only require to operate the Brokk 330 for short
periods – stop/start;
• risk assessment of some stopes indicated that it was
inappropriate to place the Brokk in the stope due to poor
hanging wall conditions;
• machine damaged (minor) when pulling ore down towards
itself; and
• Brokk WA dealer changed hands.

CONCLUSION
The purchase of the Brokk 330 excavator has assisted in additional
ore extraction from the flat stopes and with operating costs in the
region of $5 per ore tonne and 7500 tonnes of ore extracted this
year the machine has paid for itself many times over. The machine
has fallen slightly short of expectations of ten per cent additional FIG 4 - SSS 2006 B stope cross-section before and after the
ore (only eight per cent achieved) from flat stopes and a Brokk.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 113

Recovery of Broken Ore Stocks Left on the Mine Floor and Other
Applications of the Underground Mobile Supersucker
K Biegaj1
ABSTRACT
Use of conventional LHD units and scrapers on an uneven floor
underground leaves behind a significant portion of metal (or precious
stones) contained in the already broken ore stocks. Depending on the
mining method, this material is often never recovered (ie it is wasted)
when utilising conventional mining methods and equipment. The portion
of metal or precious stones contained in the conventionally
‘unrecoverable’ floor material is included in the initial Ore Reserves
estimate when a mining project is being evaluated and capital funds are
sought for its development. The application of a Mobile Supersucker, a
powerful vacuum cleaner previously not available in Australia, has
permitted development of a floor material recovery system. This enables
cost effective reclamation of the ‘conventionally’ unrecoverable material,
allowing for significant additional revenue and enhancement of the return
on invested capital. This paper describes the construction of the Mobile
Supersucker and outlines the challenges faced during the development of
the floor material recovery system and presents physical and financial
results achieved on gold and nickel mines in Australia over the last few
years. Other applications of the Mobile Supersucker include an
alternative hoisting system for underground deposits, underground
exploration winzing, mining of narrow and ultra-narrow veins and surface
sump clean-down are also outlined briefly.
INTRODUCTION
The Mobile Supersucker was initially designed and constructed
in Kalgoorlie, Western Australia in 2000, to carry out winzing
from underground operations mainly for geological exploration
and bulk sampling in difficult/nuggety orebodies.
A typical South Australian opal mining diesel powered
single-stage fan supersucker was placed on 5 t excavator tracks,
equipped with driving capabilities and fitted with all statutory
safety features to make it suitable for underground use in a
modern trackless mine.
The idea of using a Mobile Supersucker for underground
winzing originated in 1999 at the Central Norseman Gold
Corporation in Norseman, Western Australia, a mining region
with narrow veins containing a high percentage of coarse gold.
Surface winzing with similar supersuckers has already been
successfully carried out for bulk sampling purposes on iron ore
mines in Western Australia (Biegaj, 2000, 2002; Biegaj and
Dominy, 2003), although these were not mobile. However, the
concept of underground winzing with a new tool (ie the Mobile
Supersucker) has not yet been accepted by the mining industry in
Australia. This is perhaps due to the process being perceived as
outdated technology in a world of mechanisation, as it still
involves manual labour. At the same time the demand for skilled
airleg miners is on a steep rise in Australia.
As a result of the lack of response from the industry for
winzing applications, significant modifications were made to the
original machine to adapt it for a new application in the recovery
of broken ore stocks left underground after conventional mining
activities. The modifications included the installation of a
swinging boom and the addition of a 200 mm diameter flexible
rubber hose with a suction head. This allows vacuuming
operations to be performed by an operator whilst situated in front
1. MAusIMM, Director, Ausvac Mining Pty Ltd, 18 Cambey Way,
Brentwood WA 6153. Email: ausvacmining@bigpond.com
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
of the machine. The recovered floor material is delivered via a
flexible hose to the bin/hopper situated at the rear of the machine
and then discharged into a bucket of a bobcat/skid steer loader
and then trammed to the nearest underground stockpile.
The first trial of the Mobile Supersucker in a new application
was conducted in May 2003 at the Placer Dome Mt Pleasant
underground gold mine near Kalgoorlie. Subsequently, there have
been several more projects conducted on various narrow vein type
high grade nickel and gold mines in Australia.
The following are the minimum excavation sizes and
ventilation requirements to safely accommodate the underground
Mobile Supersucker:
• width – 3.0 m,
• height – 3.2 m (2.9 m in travelling mode), and
• airflow – 8.0 - 10 m3/s (@ 0.05 m3/s per kW, depending on
size of the auxiliary loader/bobcat used).
CONSTRUCTION OF MOBILE SUPERSUCKER
Mobility is an essential feature of the Supersucker in an
underground floor material recovery application. A self-propelled
steel tracked carrier drives the machine. Although it is not
envisaged to utilise this ability on a daily basis, the supersucker
can propel itself over a distance of a few kilometres on a gradient
as steep as one in five. For longer distances, the supersucker has
been equipped with heavy-duty detachable wheels and a tow bar.
The towing is normally carried out by an IT vehicle or a bigger
size LHD available on a mine.
The basic parameters of the supersucker are as follows:
• width – 2.2 m,
• height – 2.95 m (2.65 m in travelling mode),
• length – up to 9.5 m (with a clean-down boom/jib fully
extended),
• weight – 8.0 t,
• vacuum – single stage 42 inch diameter centrifugal fan, and
• suction hose – flexible 200 mm diameter.
The machine has been fully equipped for underground use as
an independent mobile diesel powered unit with a roll-over/
falling objects protection canopy for the operator (ROPS/FOPS),
an AFFF automatic fire suppression system and exhaust-
conditioning device (catalytic converter).
The construction is based on many years of operational
experience with mobile equipment in Australian underground
metalliferous mines. It is simple, robust, compact and designed
for heavy-duty applications in a harsh underground environment.
The supersucker is best suited for a decline accessed mine.
However, where required, it can be partially dismantled and
lowered down a shaft.
Vacuum-lifting distance
A distance of 50 m of a mixed vertical and semi-horizontal
vacuum lift was achieved on an airleg stope clean-down job
carried out on a nickel mine in Western Australia.
115
K BIEGAJ
INITIAL UNDERGROUND TRIALS IN
VACUUMING APPLICATION
First trial
The Mt Pleasant Gold Mine was a typical Australian decline-
accessed mine from a portal situated near the pit’s bottom. The
Mobile Supersucker was, on that occasion, self-propelled to the
level where the material recovery was to be performed. The
supersucker worked in conjunction with a Case bobcat/skid-steer
loader used to deliver the material discharged from the
supersucker’s bin to the nearest stockpile underground. The
recovered floor material was then loaded and trucked to the
surface by the company.
Whilst a number of technical challenges were experienced
with the bobcat, there were no major technical problems with
the prototype Mobile Supersucker apart from a minor fault with
the hydraulic valve on the jib/boom controls that was
promptly rectified.
The trial continued for three weeks and was not successful.
There was abundance of floor material in the ore drive where the
trial was conducted but the tonnage recovered was small (less than
50 tonnes with grades varying from 0.3 - 30 g Au/t, based on 12
samples of floor material taken from the floor prior to vacuuming).
Part of the ore drive was situated below an already extracted
high grade room-and-pillar airleg stope. The ore drive size was
well in excess of a minimum size to accommodate the
supersucker (ie it was a 5 cu yard LHD size ore drive). The floor
of the ore drive was highly compacted and there were serious
challenges with vacuuming it. The hand picking/conditioning of
the floor with a pelican pick prior to vacuuming was not
productive. In addition, the vacuuming job itself and the
delivering of ore to the nearest underground stockpile with a
bobcat was found to be a two-person task in terms of
productivity and operator’s physical effort.
Second trial
After significant modifications to the machine were effected
following the first trial, in August – September 2003, a second
trial was performed at the Otter-Juan decline nickel mine in
Kambalda owned by the Goldfields Mine Management. The total
duration of the trial was just over a month, including
transportation of supersucker to and from the underground. The
initial scope of work was to clean-down a typical low-dipping
room and pillar airleg stope located close to the bottom of the
mine (~1000 m below surface). The material left in the stope to
be vacuumed was very well fragmented high grade massive
nickel ore. The Company also provided an offsider to the
operator. The machine was towed down the mine by an IT
vehicle. Following the first trial referred to above, it was
recognised that self-propelling the machine on tracks was too
slow and would adversely impede on the mine’s ‘normal’
activities. As a result, the machine was equipped with two
heavy-duty detachable wheels and a carefully designed towing
bar. This proved to be an invaluable solution, especially for a
deep mine like Otter-Juan.
The machine was positioned above the stope in a cross-cut and
remained stationary throughout the whole stope clean-down. On
that occasion, ore discharged from the supersucker’s bin was
delivered to the nearest stockpile above with a reliable Cat skid
steer loader. Due to physical effort involved, a smaller six inch
diameter vacuuming hose was used, with a system of wire ropes
installed across the stope, which was up to 30 m wide and 40 m
long down-dip. Specially designed rollers were used to swing the
vacuuming hose across the stope. The suction hose was dropped
down to the stope via a 10 m long semi-vertical raise access.
116 Ballarat, Vic, 14 - 15 October 2008
Due to the very high specific gravity (SG) of the vacuumed
material (4.0 t/m3), some challenges were experienced with
vacuum-lifting to the level above, where the supersucker was
positioned. These were overcome by lifting up the engine’s
revolutions. Even with the assistance of a capable offsider, it was
not possible to vacuum more than 8 - 9 t of stope floor material
during the ten-hour shift.
The supersucker was then driven down below the stope to
perform vacuuming in an ore drive and later on, vacuuming the
same stope in a downwards suction mode. Once again, even with
two operators it was not possible to vacuum more than
eight to nine tonnes during a ten hour shift. Whilst working in
the stope, moving around the vacuuming hose presented a
physical challenge despite the special attachments used.
Similarly as on the first trial when vacuuming in an ore
drive, the compaction of the floor material prevented
better productivities.
Overall, 161 nickel ore tonnes were recovered in a month long
trial at an overall grade of 8.6 per cent Ni as sampled by the
company geologist. There were 19 true vacuuming shifts during
the trial. The rest of the available time was used for days off,
mobilisation/demobilisation, shifting equipment between the
levels and minor breakdowns of the equipment.
For the first time, the trial was a success both technically
and financially.
Third trial
Shortly after the second trial, a third trial was conducted at the
Bendigo Gold Mine in Victoria during September – October
2003. The scope of work was to vacuum the floor of
development ore drives at the Swan Decline Mine in Bendigo. As
on the previous job, the company provided an offsider to the
operator. The total duration of the trial was 12 days, inclusive of
transportation of the equipment to and from underground and set
up. A small 151D Toro LHD mobilised from Western Australia
was used to collect the ore from the supersucker’s bin. In an
attempt to loosen the floor prior to vacuuming, an industrial
high-pressure cleaner was hired. It proved to be a failure and a
safety hazard. In total less than 70 t of floor material were
recovered and delivered to the nearest underground stockpile.
Whilst there was up to 0.4 m of broken ore on the floor in the ore
drives, due to its compaction it was not possible to recover any
more tonnes per shift than on the previous trials. The trial was
not a success.
Breakthrough
Following a discussion with an engineering company in late
October 2003, a solution to the biggest challenge was offered, ie
that of loosening the floor prior to vacuuming. A practical trial
was promptly conducted at one of the local gold mines and
proved to be a breakthrough. A small 1.5 t excavator with a
ripper attachment normally used on earthworks/maintenance jobs
on railway lines was now successfully conditioning/loosening the
floor in an ore drive.
Subsequently, a seven day trial during December 2003 to
January 2004 at the Cosmos Nickel Mine was undertaken. The
floor recovery vacuuming was conducted in an ore drive
excavated in a massive nickel ore. For the first time
productivities up to 20 tonnes per 12 hour shift were achieved
with one operator (Figures 1 and 2).
A total of 61 tonnes of 7.35 per cent Ni ore was recovered with
25 hours or ~3 shifts of pure vacuuming.
The Cosmos trial proved to be a breakthrough and a success.
Since on that occasion the company did not provide an offsider
for the operator, it was clear that a second operator was needed
and preferably a second shift (with two operators), to realise
Narrow Vein Mining Conference
 RECOVERY OF BROKEN ORE STOCKS LEFT ON THE MINE FLOOR

FIG 1 - Conditioning of floor in ore drive – Cosmos Nickel Mine, FIG 3 - Supersucker in ore drive, Mincor’s Redross Nickel Mine
Western Australia, January 2004. near Kambalda, Western Australia, February 2007.

FIG 2 - Supersucker towed to the surface from the Ilias Decline at FIG 4 - Floor in ore drive before and after Supersucker vacuuming
Cosmos Nickel Mine of Jubilee Mines, Western Australia, – Mincor’s Redross Nickel Mine near Kambalda,
January 2004. Western Australia, February 2007.

full the benefits from the floor recovery system, especially with
the nickel grades mined at the Cosmos Mine.

Subsequent vacuuming projects

In October 2005, a profit sharing contract was signed with
Lightning Nickel in Kambalda, Western Australia, to perform a
floor material recovery at the Gibb South mine section located at
the Long Shaft-Victor Underground Complex. A total of
193 tonnes of 4.89 per cent nickel ore were recovered during a
nearly three week trial with a single operator.
Mechanical problems were experienced with the hired
bobcat/backhoe type loader, which negatively impacted on the
overall tonnes recovered and duration of the project. Finally, a
reliable Cat skid steer loader was hired to complete the task.
The Gibb South trial was a success. It is also worth mentioning
that the ore was received, sampled and treated by a third party
(BHP Billiton), and that there were no disagreements over the
tonnage or grade recovered.
FIG 5 - Discharge from the bin into the Cat skid steer loader –
In January – February 2007, another profit share contract was
Mincor’s Redross Nickel Mine near Kambalda, Western Australia,
signed with Mincor Resources at the Redross Nickel mine
in Wigdiemooltha, Kambalda. A total of 145.3 tonnes of February 2007.
2.84 per cent nickel ore were recovered during a three week trial.
The floor vacuuming work was performed in an ore drive with and significantly lower grade recovered compared with the
second lift taken (Figures 3, 4 and 5). Despite lower ore tonnage previous trial, both the Company and Ausvac made a profit.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 117

K BIEGAJ
Table 1 shows the benefits from mine floor vacuuming – a
comparison between the recovered mine floor material containing
nickel and gold is shown with the nickel and gold prices as of the
30 May 2008. The mining, carting and milling costs used in the
table are typically experienced at the mines operating in Western
Australia. Ore drives of a typical/common size and length were
used in this example. To calculate/simulate tonnages of valuable
ore material left on the floor after conventional bogging
(mucking), various thicknesses were considered. As an example,
there was very little material left on the floor at the Redross
mine, (no more than 0.12 m). This adversely impacted on the
productivities achieved; especially with one operator only, ie there
was a need to re-position the supersucker more often than would
have been the case if a thicker layer of floor material had been
present. From the author’s experience on previous projects, there
were instances of mine floor material being up to 0.4 m thick and
containing high grade material (Table 1).
OTHER APPLICATIONS OF THE MOBILE
SUPERSUCKER
Other applications of the Mobile Supersucker include:
• Alternative access to underground deposits. The method
provides an environmentally friendly, alternative mine access
excavation/mucking method (other than conventional ones
used in a vertical shaft or a decline) aimed at reduction of
excavated waste, reduction of noise and diesel gas emissions,
especially in highly populated and environmentally sensitive
areas (Biegaj, 2000).
• Underground supersucker winzing for exploration and bulk
sampling in complex gold veins (Biegaj, 2000; Biegaj and
Dominy, 2003).
• Surface or underground sump clean-down (Figures 6 and 7).
Following a rather unsuccessful underground trial at the Mt
Pleasant Gold Mine in May 2003, a clean-down of a surface
sump filled with mud was performed at the same mine when
the supersucker was brought back to the surface. The sump
was 2.2 m × 2.2 m × 2.5 m deep. That trial was regarded as a
success, taking less than eight hours in total to thoroughly
clean the sump and the surrounding area. It proved to be an
economically viable method of cleaning surface or
underground sumps/wet areas, especially those containing
high values of gold, nickel or other precious metals or stones.
FIG 6 - Surface sump clean down – before. Placer Dome
Mt Pleasant Gold Mine near Kalgoorlie, Western Australia,
May 2003.
• Mining of the bottom part of ore zones/shoots that do not
economically warrant further decline development or a pit
extension (Maher, 2000 - 2003).
118 Ballarat, Vic, 14 - 15 October 2008
FIG 7 - Surface sump clean down – after. Placer Dome
Mt Pleasant Gold Mine near Kalgoorlie, Western Australia,
May 2003.
CONCLUSIONS
Whilst initial tests of the Mobile Supersucker for mine floor
recovery did not appear to be promising, after a number of trials
and a 500 per cent improvement to the system’s productivity it is
now able to deliver.
In a longer-term contract arrangement, it is envisaged that this
would involve using two operators on the machine and working
two shifts per day. With 40 t of vacuumed floor material per shift
and working two shifts per day with productive 20 days per
month (allowing for maintenance and break downs), 1600 tonnes
of floor material could be recovered. Refer to Table 1 for the
revenue that would be available, depending on the floor material
grade – the revenue that is normally left behind and lost after the
completion of conventional mining operations. It is self-
explanatory and those figures should no longer be ignored by the
directors of mining companies and mine management.
It is strongly recommended to, wherever possible, incorporate
the recovery of mine floor material into the ‘normal’ mining
cycle and start to reclaim capitally developed and already broken
ore to provide more value to the shareholders.
As a comparison, industrial waste on the surface has been
properly dealt with for a number of years and it is time for
underground mining to follow other industries.
ACKNOWLEDGEMENTS
Special thanks are extended to a number of individuals. Marc
Blythe and Steve Tombs of then Placer Dome’s Mt Pleasant
Gold Mine for allowing and sponsoring the first trial of the
supersucker in a vacuuming application at their mine in
May 2003. Andy Ziersch for providing the idea and first
practical trial of the mine floor conditioning prior to
vacuuming. Paul Maher (Maher Mining Contractors Pty Ltd)
for his creative and practical contribution into the design and
construction of the first underground Mobile Supersucker. Rob
Miller (Miller Fabrications Pty Ltd) is thanked for his inventive
and technical input during the construction of the machine,
subsequent modifications after the first underground trial and
generously allowing use of his world-class workshop in
Kalgoorlie for these activities. Leszek Sternal (Positron Pty
Ltd) is thanked for numerous discussions, practical solutions to
electrical challenges and the idea of making the supersucker a
truly mobile machine.
Narrow Vein Mining Conference
 RECOVERY OF BROKEN ORE STOCKS LEFT ON THE MINE FLOOR

TABLE 1
Nickel versus gold mine floor material recovery financial comparison.
Nickel – nickel mine A calculation
Nickel price (US$/t) 22 850
Exchange rate 0.94
Ni credit to mine A 65%
Mill recovery 85%
Mine A costs ($/t)
Loading and trucking to surface 8.0
Carting to the mill 6.0
Milling 28.0
Diesel fuel @ $1.0/L with 140 L/shift @ 10 t/shift 14.0
Other 6.0
Total ($/t): 62.0
Revenue calculations – X level
SG 3.0
Length 140.0
Width 4.0
Thickness of floor material (m) 0.1 0.2 0.3 0.4
Ore tonnes 168 336 504 672
Y level
SG 3.0
Length 200.0
Width 4.0
Thickness of floor material (m) 0.1 0.2 0.3 0.4
Ore tonnes 240 480 720 960
Total ore tonnes 408 816 1224 1632
Total Ni T @ credit of – grade (%) 65% 65% 65% 65%
1.5 4.0 7.956 11.9 15.9
2.0 5.3 10.608 15.9 21.2
2.5 6.6 13.26 19.9 26.5
3.0 8.0 15.912 23.9 31.8
3.5 9.3 18.564 27.8 37.1
4.0 10.6 21.216 31.8 42.4
4.5 11.9 23.868 35.8 47.7
5.0 13.3 26.52 39.8 53.0
Value of Ni in A$ – grade (%)
1.5 96 699 193 399 290 098 386 797
2.0 128 932 257 865 386 797 515 729
2.5 161 165 322 331 483 496 644 662
3.0 193 399 386 797 580 196 773 594
3.5 225 632 451 263 676 895 902 526
4.0 257 865 515 729 773 594 1 031 459
4.5 290 098 580 196 870 293 1 160 391
5.0 322 331 644 662 966 993 1 289 323
Value of Ni less mine A costs – grade (%)
1.5 71 403 168 103 264 802 361 501
2.0 103 636 232 569 361 501 490 433
2.5 135 869 297 035 458 200 619 366
3.0 168 103 361 501 554 900 748 298
3.5 200 336 425 967 651 599 877 230
4.0 232 569 490 433 748 298 1 006 163
4.5 264 802 554 900 844 997 1 135 095
5.0 297 035 619 366 941 697 1 264 027

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 119

K BIEGAJ

TABLE 1 cont …
Gold – gold mine B calculation
Gold price (US$/oz and per g) 902 27.25
Exchange rate 0.94
Gold credit to mine B 100%
Mill recovery 92%
g/ounce 33.1
Mine B costs ($/t)
Loading and trucking to surface 8.0
Carting to the mill 3.0
Milling 25.0
Diesel fuel @ $1.0/L with 140 L/shift @ 10 t/shift 14.0
Other 6.0
Total ($/t): 56.0
Revenue calculations – X level
SG 2.5
Length 140.0
Width 4.0
Thickness of floor material (m) 0.1 0.2 0.3 0.4
Ore tonnes 140 280 420 560
Y level
SG 2.5
Length 200.0
Width 4.0
Thickness of floor material (m) 0.1 0.2 0.3 0.4
Ore tonnes 200 400 600 800
Total ore tonnes 340 680 1020 1360
Total Au ounces @ credit of – grade g/t 100% 100% 100% 100%
5.0 51.4 102.7 154.1 205.4
10.0 102.7 205.4 308.2 410.9
15.0 154.1 308.2 462.2 616.3
20.0 205.4 410.9 616.3 821.8
25.0 256.8 513.6 770.4 1027.2
30.0 308.2 616.3 924.5 1232.6
35.0 359.5 719.0 1078.5 1438.1
40.0 410.9 821.8 1232.6 1643.5
Value of gold in A$ – grade g/t
5.0 45 341 90 681 136 022 181 362
10.0 90 681 181 362 272 044 362 725
15.0 136 022 272 044 408 066 544 087
20.0 181 362 362 725 544 087 725 450
25.0 226 703 453 406 680 109 906 812
30.0 272 044 544 087 816 131 1 088 175
35.0 317 384 634 769 952 153 1 269 537
40.0 362 725 725 450 1 088 175 1 450 900
Value of gold less mine B costs – grade g/t
5.0 26 301 52 601 78 902 105 202
10.0 71 641 143 282 214 924 286 565
15.0 116 982 233 964 350 946 467 927
20.0 162 322 324 645 486 967 649 290
25.0 207 663 415 326 622 989 830 652
30.0 253 004 506 007 759 011 1 012 015
35.0 298 344 596 689 895 033 1 193 377
40.0 343 685 687 370 1 031 055 1 374 740

120 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 RECOVERY OF BROKEN ORE STOCKS LEFT ON THE MINE FLOOR

REFERENCES
Biegaj, K, 2000. Alternative access, mining and hoisting for underground
deposits, in Proceedings MassMin 2000 Conference, pp 911-914
(The Australasian Institute of Mining and Metallurgy: Melbourne).
Biegaj, K, 2002. Mobile supersucker winzing – Attaining proven reserves
prior to capital commitment, in Proceedings 2002 Underground
Operators’ Conference, pp 217-224 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Biegaj, K and Dominy, S, 2003. Lowering orebody risk in complex gold
veins – Application of underground mobile supersucker winzing, in
Proceedings Fifth International Mining Geology Conference,
pp 273-278 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Maher, P, 2000-2003. Personal communication.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 121

Thermal Fragmentation – An Innovative Process for Mining
Narrow High-Grade Precious Metal Veins
D Brisebois1

ABSTRACT fragments 0 - 13 mm in size. The leftover rock between

fragmented holes is broken loose using soft explosives and a
The mining of high-grade narrow vein deposits is a predominant field of
activity in the precious metal sector. These types of deposits are located
narrow mining corridor with widths of 30 cm to 1 m is thus
throughout the globe and have a significant presence in mining extracted. Since the waste walls are left intact, the dilution factor
operations. The principle factor that has undermined the profitability and and the inefficiencies associated with traditional mining methods
effectiveness of mining such ore zones is the substantial dilution that are greatly reduced.
occurs when blasting with explosives during extraction.
In order to minimise ore dilution, the thermal fragmentation mining LEVEL
L EVEL 45
method enables the operator to extract a narrow mineralised corridor,
30 cm to 1 m wide (according to the true width of the ore vein). By
inserting a strong burner powered by diesel fuel and air into a pilot hole
previously drilled directly into the vein, a thermal reaction occurs spalling

gra de

grade
Low grade

de
the rock and enlarging the hole to 30 - 100 cm in diameter. This patented

Low gra
method produces highly concentrated ore, resulting in 400 - 500 per cent

Low

Low
less dilution when compared to conventional mining methods. The
mining method was conceived with the aim of mining ounces, not tonnes.
This method of extraction, reduces the environmental impact of mining
operations since much smaller quantities of rock are displaced,
stockpiled, and treated using chemical agents. The fully mechanised L EVEL 70
LEVEL
equipment operated by a two-person team, maximises the effectiveness of
skilled personnel, increases productivity, and reduces risks of injuries.
Important economic benefits are generated for large- and small-scale
mining operations alike. For large-scale operations, the mining method
increases mining reserves, lowers the cost per ounce, and increases the
total profitability of existing mine operations. The method also
significantly lowers costs and risks of developing and mining small-scale
deposits, requiring less investment capital, shortening the payback period,
thus reducing the lending risk. Combined with a small modular
processing unit able to process between 100 - 300 tons per day, the
method substantially lowers the financial break-even point for small-scale
mining operations in a narrow vein environment.
The company has been working with the technology at its Rocmec 1
mining site; results achieved and future developments are discussed.

INTRODUCTION
The mining of high-grade narrow vein deposits is a predominant
field of activity in the precious metal sector. These types of
deposits are located throughout the globe and have a significant
presence in mining operations. The principle factor that has
undermined the profitability and effectiveness of mining such ore
zones is the substantial dilution that occurs when blasting with
explosives during extraction. There is a clear need to develop
mining techniques that minimise ore dilution and development
costs in order to make the mining of high-grade narrow vein ore
deposits more profitable. The thermal fragmentation mining
method has been conceived to mine a narrow mineralised
corridor in a productive and cost efficient manner in order to
solve this particular challenge. The following describes this
mining method in depth and outlines its successes in improving
the extraction process of such orebodies.
THERMAL FRAGMENTATION MINING METHOD
A strong burner powered by diesel fuel, is inserted into a six inch
pilot hole drilled into the vein using a conventional longhole
drill. The burner spalls the rock, quickly increasing the diameter
of the hole to 30 - 100 cm (Figure 5) and producing rock
1. President and CEO, Rocmec Mining, 1 Holiday Avenue, Suite 501,
Tour Est, Pointe-Claire QC H9R 5N3, Canada.
Email: dbrisebois@rocmecmines.com
FIG 1 - The method.
The burner
The burner powered by diesel fuel and compressed air, creates a
thermal cushion of hot air in the pilot hole which produces a
thermal stress when coming in contact with the rock (Figure 2). A
spalling effect occurs (Calaman and Rolseth, 1968), and the rock
is scaled off the hole walls and broken loose by the compressed
air. This technology has been used for over 40 years in large-scale
open pit operations for the enlargement of blastholes.
The fragmented rock
The process of fragmenting the rock is optimal in hard, dense
rock. The spalling process produces rock fragments 0 - 13 mm in
size. Figure 3 illustrates the difference in size of the fragmented
ore in comparison to ore that has been extracted using
conventional blasting methods. The finely fragmented ore
requires no crushing before entering the milling circuit and can
be more efficiently transported since it consumes less space than
ore in larger pieces.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 123

D BRISEBOIS
FIG 2 - The burner.
Conventional Thermal
Ore Fragmentation
FIG 3 - The fragmented rock.
Drift development and stope layout
Drift development is performed directly into the ore at intervals
of 15 to 20 metres in accordance with the geology of the
orebody. Using a re-suing method, the ore is blasted and
recovered in the first cut then the waste is blasted and hauled
away in the second cut.
Following the creation of two sublevel drifts, a pilot hole is
drilled between the two levels and enlarged by way of thermal
fragmentation. The mining layout is applicable for low dip veins,
as the unit also operates horizontally (drilling flat).
The unit is designed to operate in a compact underground
environment in a drift or sublevel as small as 1.5 m wide by
2.5 m high (Figure 4). The unit designed and manufactured by
the Kubota and its approximate dimensions are 125 cm wide,
275 cm long and 165 cm high.
Ore extraction – selective and continuous
The thermal fragmentation mining method allows for selective
ore extraction; high-grade sections can be prioritised and
extracted first. The method extracts a narrow corridor (30 cm to
1 m wide) leaving waste walls on each side of the mineralised
zone intact and minimising the damage caused to the drift
structure. The area mined can be easily rehabilitated using cables
and panels to cover the narrow opening in the hanging wall of the
undercut drift. Mine residue is then used to fill the extracted zone
and a cement floor is laid, if needed, on the footwall of the
overcut drift to permit future access. Using this method, the
stability of the rock is maintained and access to the orebody is
never compromised.
124 Ballarat, Vic, 14 - 15 October 2008
FIG 4 - Stope layout.
FIG 5 - Fragmented hole.
Furthermore, thermal fragmentation is a continuous mining
method; it uses no explosives and is operated in a continuous
chain, with one person first drilling a pilot hole, followed by a
second who enlarges it by way of thermal fragmentation.
Mining ounces not tonnes
The method produces highly concentrated ore, resulting in
400 - 500 per cent less dilution when compared to conventional
mining methods. Table 1 compares the quantity of rock extracted
when mining a 50 cm wide vein using the thermal fragmentation
mining method as opposed to a shrinkage mining method.
Table 1 shows that approximately four times less rock needs to
be mined for the equivalent mineralised content. This method of
extraction allows mine operators to solely extract mineralised
zones thus significantly reducing dilution factors and as a result,
optimising mine operations.
REDUCING ENVIRONMENTAL IMPACT
There is a growing need to develop sustainable mining methods
that minimise the environmental footprint left behind by mining
operations. In all corners of the world, local populations have
expressed their concerns about mine operations being established
in their region, as awareness about the negative impact of
industrial activities has spread.
Narrow Vein Mining Conference
 THERMAL FRAGMENTATION – AN INNOVATIVE PROCESS FOR MINING NARROW HIGH-GRADE PRECIOUS METAL VEINS

TABLE 1 TABLE 2
Tonnage calculation; comparing thermal fragmentation and Productivity (one thermal fragmentation unit); comparing thermal
shrinkage. fragmentation and shrinkage methods.
Tonnage calculation of a Thermal Shrinkage Tonnage calculation of a Thermal Shrinkage
40 m by 20 m ore block fragmentation method 40 m by 20 m ore block fragmentation method
Width in situ (m) 0.5 0.5 Width in situ (m) 0.5 0.5
Mining width (m) – final result 0.5 1.8 Mining width (m) – final result 0.5 1.8
Planned dilution 0% 260% Planned dilution 0% 260%
Height (m) 20 20 Height (m) 20 20
Length (m) 40 40 Length (m) 40 40
Density 2.8 2.8 Density 2.8 2.8
Total volume (t) 1120 4032 Total volume (t) 1120 4032
Number of workers 2 2
While developing the thermal fragmentation mining method, Productivity per 12 hour shift (t) 30 30
important efforts were made to address and reduce the Tonnes extracted per 24 hours 60 120
environmental effects mine operations have on the surrounding
Days needed to extract ore block 18.7 33.6
areas. Using the method, mine development is performed directly
into ore resulting in less waste rock being extracted and
displaced into large piles at the surface. By solely extracting the Table 2 demonstrates that for the equivalent amount of mineral
mineralised zone, only the necessary excavations are made. As content, it takes approximately half the time to mine the ore zone
shown in Table 1, four times less rock needs to be mined for the using the thermal fragmentation mining method than when using
equivalent mineral content. a shrinkage mining method. Furthermore, since less rock needs
As a result of less rock being mined, fewer tonnes need to be to be mucked, and hauled from the stope, fewer personnel are
processed at the mill to extract the precious metals. The quantity needed for handling the ore. In total, we estimate that to extract
of chemical agents needed in the process is greatly reduced. an equivalent amount of precious metal ounces, the method
Furthermore, the quantity of energy needed to process the ore is reduces the personnel needed for mine operations by 30 - 50 per
also greatly diminished since less rock is sent to the mill. The cent when compared to conventional mine operations.
reduced quantity of energy for hauling and processing the ore
results in fewer greenhouse gases being emitted. The mining Mechanisation and employee safety
residue that remains once the precious metal contents are removed
Each unit is completely mechanised, reducing the risk of injuries
is four times less abundant, using the example above, meaning
and strain caused by manual manipulation of heavy equipment.
much smaller tailing areas need to be constructed, maintained, and
The operator stands at a safe distance from the mining stope,
rehabilitated once mining operations have ceased. The space
virtually eliminating the risk of flying debris and falling loose
needed to host the mine site is greatly reduced, the alterations to
rock from the waste walls. Furthermore, unlike shrinkage mining
the landscape are significantly diminished, and the result is a
methods, smaller excavations are made (0.5 m compared to 2 m)
cleaner and more responsible approach to mine operations.
so the occurrence of falling loose rock is greatly diminished.
The company is moving forth with an environmentally Each unit is equipped with sensors capable of identify the
responsible approach at its Rocmec 1 mine site. A small presence of seven harmful gases (such as carbon monoxide,
processing mill (75 t/d) will be installed in an underground sulfides, etc), important safety measures when mining in an
section of the mine, using a gravity and flotation circuit, no underground environment. When designing the equipment,
chemical agents will be used to treat the ore. The mining residue feedback from members throughout the organisation was
created will remain underground and will be returned to the gathered, and the final result is a tool that is well adapted to the
stope, thus eliminating the need for tailings at the surface. Such needs of those who operate the machinery and leads to a safe and
innovations are made possible by drastically reducing dilution comfortable work environment.
through the use of the thermal fragmentation mining method.
ECONOMIC BENEFITS FOR EXISTING MINE
MAXIMISING EFFECTIVENESS OF SKILLED OPERATORS
PERSONNEL
By rendering a greater number of narrow mineralised zones
Given the current shortage of skilled personnel in the mining economical to extract, the mining method has the potential to
community, it is essential to find ways to increase productivity convert a substantial portion of the mineral resources of an
per worker while improving working conditions in order to operating company into mineral reserves. A large number of
attract and retain skilled miners. It is with this thought in mind mines currently in operation today contain narrow, precious
that the thermal fragmentation mining method was conceived metal veins throughout the orebody, but unless these veins are of
and developed and with which we will continue to refine the significant width (usually 1 m or greater) or very high grade they
equipment. are often overlooked. As the mine operator develops the zones to
be extracted, high-grade, narrow orebodies are often uncovered,
Productivity but not extracted since it is uneconomical to mine such orebodies
using conventional mining methods (shrinkage, long hole, room
The work group required to operate one thermal fragmentation and pillar, etc). Table 3 demonstrates the cost savings per ounce
unit consists of a two person team (one thermal fragmentation of using the thermal fragmentation mining method in comparison
operator, one drilling operator). Table 2 shows the time needed to to the long hole method. The study was done by Canadian
extract an ore block using the thermal fragmentation mining Institute of Mining using 2001 exchange rate figures (Poirier
method in comparison to using a shrinkage mining method. et al, 2003).

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 125

D BRISEBOIS

Table 3 shows, it is approximately 45 per cent less costly to
mine a narrow vein orebody using the thermal fragmentation
mining method that using a conventional mining method. Overall
profitability of mine operations is increased since more precious
metals can be economically mined for the same level of
development expenditures.
SMALL-SCALE MINE OPERATIONS
(LESS THAN 300 T/D)
Small-scale producers have a tightly limited milling capacity to
process mineral ore and require high-grade mineral content to
cover expenses. Since there is little extra capacity to process
low-grade ore, reducing dilution when extracting narrow vein
orebodies is critical for the profitability of the project. The
thermal fragmentation mining method offers the possibility of
dramatically increasing the productivity of such small-scale
mining operations. As shown in Table 1, the ratio of tons
extracted for an equivalent in mineral content is approximately
4:1, meaning that for the same mineral content, nearly four
tonnes need to be extracted to produce the equivalent of one
tonne using thermal fragmentation. Hence, a 200 t/d operation
using thermal fragmentation has the capacity to produce as much
precious metals content as an 800 t/d operation using
conventional mining methods.
Project financing
This increase in productivity has important implications for
project financing. The low capital requirements needed to start a
project that utilises the thermal fragmentation mining method,
when compared to traditional mining operations, means less
money is at risk for an equivalent potential benefit (smaller mill,
less development costs). This also results in a shorter payback
period and a lower lending risk. The substantial start-up costs for
mining operations as well as the scarcity of capital available in
the marketplace have always been important barriers for the
mining industry. By introducing a technology that aims to lower
the financial barrier and increase the profitability of mining
start-ups, the goal is to bring a greater number of high-grade
precious metal orebodies into production in a financially,
sustainable manner and with a lower investment risk.

TABLE 3
Estimate cost comparison between underground thermal fragmentation and longhole.

Tonnage calculated on the basis of a 60 m by 60 m reserve block Thermal drilling 3024 t Long-hole drilling 3024 t
Grade in situ (g/t) 35.00 35.00
Width in situ (cm) 30 30
Minimum width (cm) 30 140
Planned dilution 0% 367%
Geological reserves 3024 14 112
Reserve grade (g/t) 35.00 7.50
Mining
Wall dilution 5% 35%
Stope recovery 79% 90%
Ore development 544 2540
Planned mining reserve 1961 14 606
Grade (g/t) 33.25 4.88
Mill recovery 96% 96%
Produced ounces 2013 2198

Thermal drilling Long-hole drilling

Unit cost $/M Total cost $/M Unit cost $/M Total cost $/M
Development
Drifts 1000.00 180 000.00 1000.00 180 000.00
Subdrifts 1000.00 120 000.00 1000.00 120 000.00
Raises 1000.00 60 000.00 1000.00 120 000.00
Drawpoint 1000.00 1000.00 60 000.00
Mining cost ($/t) 113.50 222 600.00 19.00 277 516.00
Mucking 8.00 15 690.00 4.00 58 424.00
Transportation 12.00 23 535.00 6.00 87 636.00
Milling 16.00 31 380.00 20.00 292 122.00
Environmment 2.00 3922.00 2.00 29 212.00
Backfilling 5.00 73 030.00
Total 657 127.00 1 297 940.00
$ per tonne 335.06 88.86
$ per ounce 326.49 590.59
US$ per ounce 0.65 212.22 383.88

126 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 THERMAL FRAGMENTATION – AN INNOVATIVE PROCESS FOR MINING NARROW HIGH-GRADE PRECIOUS METAL VEINS
THE TECHNOLOGY AT WORK – ROCMEC 1 MINE
In 2005, Rocmec Mining purchased the Russian Kid
underground mine located near Rouyn-Noranda in Abitibi,
Quebec (renamed Rocmec 1), to showcase its unique patented
thermal fragmentation mining method and use the technology to
verify the grades contained within the mineralised structures in
the most cost efficient manner. Since late 2006, a 44 000 tonne
bulk sampling program is underway at the Rocmec 1 mining site,
and several thousand tons of ore have been extracted by way of
Thermal Fragmentation in high-grade narrow vein zones.
Development work or drifting was completed on two
superimposed levels (110 and 130 m) and the mineralised
structure between these levels was extracted by way of thermal
fragmentation. The company has analysed the performance data
and is quite pleased with the results. A mining corridor was
extracted with precision, minimising dilution of the mineralised
zone. As a result, important cost savings as well as impressive
productivity gains were realised.
In addition, a number of mine site visits have taken place and
industry participants have demonstrated an important interest
towards the new mining method. Discussions with mining
producers are ongoing and negotiations to introduce this new
mining method within their mining sites are underway. These
discussions were undertaken with the purpose of offering
services to mining companies currently in production and extract
ore using the thermal fragmentation mining technology in their
existing operations.
Through exploration efforts the company has newly uncovered
promising ore structures and future developments at the Rocmec
1 mine include further underground drilling to define the gold
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
bearing potential of the property. The company is installing a
small-scale processing mill to process the ore on site and bring
mine operations to full-scale production using the thermal
fragmentation mining method.
CONCLUSION
Many variations and adjustments have been made to
conventional methods of mining narrow precious metal veins, but
the serious shortfalls brought upon by dilution remain. The
thermal fragmentation mining method is a new and innovative
way of mining narrow vein orebodies and a foremost solution to
solving the problem of ore dilution. It uses a unique tool, a
powerful burner, to mine with precision, a narrow mineralised
corridor in an effective and productive manner. The technology is
positioned to meet the growing challenges of skilled labour
shortages, tougher environmental guidelines, and the depletion of
traditional large-scale ore deposits mined using conventional
methods. As the technology continues to develop and spread
through the mining community, the objective remains to optimise
the productivity and profitability of mining narrow high-grade
precious metal orebodies and to make a substantial, lasting
contribution to this sector of activity.
REFERENCES
Calaman, J J and Rolseth, H C, 1968. Surface Mining, first edition,
Chapter 6.4, pp 325-337 (Society for Mining, Metallurgy and
Exploration Inc: Littleton).
Poirier, S, Fecteau, J M, Laflamme, M and Brisebois, D, 2003. Thermal
rock fragmentation – Applications in narrow-vein extraction, CIM
Bulletin, 96(1071):66-71.
127
Monorail Technology – A Rapid and Cost-Effective Method of
Decline Development
E Chanda1 and B Besa2
ABSTRACT
Australia is a world leader in the design and operation of mines accessed
by the decline method. Literature indicates that a number of metalliferous
underground mines using the decline system are also increasing steadily.
This increase has generated a great deal of interest in underground
haulage systems in future due to the huge costs involved in the
development of declines. Studies have also shown that the current method
of accessing orebodies by the conventional decline method has proved to
be expensive, especially for narrow vein deposits. This is because decline
development costs are influenced by mine planning and design
parameters as well as available haulage systems that the mining engineer
can choose from. The current access method suggests that more waste
material is being extracted because of the size of decline openings
adopted (5.5 mW and 5.5 mH). Also the conventional 1:7 decline
gradient and turning radius of 20 m makes decline lengths, as well as the
duration of development, excessively longer. This paper discusses results
of a mine design case study with monorail technology as a method of
accessing and developing the decline. The proposed system is
economical, has high advance rates and offers a significant reduction in
capital expenditure. The study compares costs, development metres and
the duration of decline development using the monorail system with the
conventional (truck haulage) access method. Results obtained indicate
that the monorail technology has the potential to reduce the decline
length by over 62.6 per cent and decline costs by 63 per cent for the case
study. The monorail system also reduces the total capital development
costs by 22 per cent for the same case study. Due to rapid development
resulting from increased advance rates and the shorter decline length
coupled with smaller decline openings, the time to access the deposit
reduced by 71.8 per cent. This case study demonstrates that the
economics of narrow vein deposits can be improved significantly by
using the monorail system of decline development.
INTRODUCTION
The monorail technology
A monorail works in a manner similar to the classic mine rail
transport systems (Figure 1). It shares many of the advantages of
floor mounted rail, but overcomes the bulk of that system’s
limitations. The monorail train carries up to six containers with
total payload up to 30 tonnes including the weight of the container.
Since the dead weight of one container is one tonne it means that
when all monorail containers are loaded, the system has hoisting
capacity of 24 tonnes. The monorail train has capabilities to
negotiate gradients of more than 30° and vertical and horizontal
curves of 10 m and 4 m radius respectively (Scharf, 2005; Chanda
and Roberts, 2005). When exceptionally steep grades have to be
negotiated, a seamless transition to a rack-and-pinion arrangement
is effected. An inherent safety feature is a braking system which
also acts on the ‘I’ beam web and is spring applied in the event of
power being lost. Normal service braking is carried out by putting
the electric motor into regenerative mode, sending power back into
the system. The configuration of the monorail drill-load-haul
system to be used in decline development is shown in Figure 2.
1. MAusIMM, Head – Mining Engineering Department, Western
Australian School of Mines, Curtin University of Technology,
Locked Bag 22, Kalgoorlie WA 6433.
Email: e.chanda@curtin.edu.au
2. Research Fellow, Mining Engineering Department, Western Australian
School of Mines, Curtin University of Technology, Locked Bag 22,
Kalgoorlie WA 6433. Email: b.besa@curtin.edu.au
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
FIG 1 - Monorail in a mine tunnel (courtesy Scharf DBT).
The system consists of the pneumatic loader that works in
conjunction with the monorail drilling system fitted with two
independent drilling units.
The loading system consists of a suction pipe that is connected
to the storage hopper. The loading mechanism is based on the
principle of vacuum lift system or supersucker (Biegaj, 2002;
Dorricott and Jones, 1984; Jones, 1989). The high pressure fan
connected to the storage hopper creates sufficient negative
pressure inside the hopper that enables transport of blasted
material from the development face into the hopper to take place.
Once the hopper is filled up, the suction pipe is disconnected
from the hopper and the monorail train pulls the loaded hopper to
the position of an empty monorail container where automatic
discharge of rock fragments takes place (Figure 3). According to
studies (Chanda et al, 2008), the efficiency of the monorail
suction system depends on the fragmentation of the material
being sucked. Therefore, this study has determined that the
optimal diameter of rock fragments that the monorail pneumatic
loading system can effectively suck is 110 mm. However, there
will always be rock fragments greater than 110 mm in the muck
pile. Larger fragments will be secondary broken using a
mechanical hammer. This process has the potential to increase
the mining cycle.
Since the maximum payload per container is four tonnes
(Scharf, 2007; Stefan, 2007), the storage hopper is also designed
with a capacity of four tonnes to allow material from the hopper
to be loaded in each monorail container in one pass. Figure 4
shows the configuration of the monorail drilling system fitted
with two independent drilling units. The system has its also own
power supply attached to it with two horizontal and two vertical
hydraulic stabilisers to act as supports during drilling operations.
The operation of the monorail drilling system is such that once
the development face is blasted, drilling of the top part of the
face commences whilst the loading system simultaneously cleans
the development face. Thus, the operation of the monorail
drilling system depends on the efficiency and operations of the
suction loading system. This configuration reduces the total
cycle time of drilling, blasting and cleaning the development face
thus potentially increasing the daily advance of the decline.
JUNDEE CASE STUDY
The Jundee operations are situated approximately 800 km
north-east of Perth in Western Australia. The operations are
129
E CHANDA and B BESA

M uck pile

FIG 2 - Configuration of monorail drill-load-haul system for decline development.

FIG 3 - Movement of hopper during loading mechanism.

owned by Newmont Mining Corporation, which is one of the
world’s largest producers of gold. The Jundee operation began
operations in 1995 and is composed of two underground mines
as well as several satellite open pits about 30 km south of the
operation. It produced 313 000 ounces of gold in 2006 and
reported 1.48 million ounces of gold reserves at year-end.
The Nim3 deposit (Figure 5) is the third largest underground
resource of Jundee – Nimary gold field, situated beneath Nim3
open cut, immediate west of Barton Deep deposit.
A number of structures; Nim3_Lyons, Nexus, Midas/Money
Line, Hughes Extended and Colloform structure collectively
forms the Nim3 deposit (South Deeps) and was the primary
source for Nim3 open cut/underground operations. Figure 6
shows the orebodies of the South Deeps mineralisation.
In 2004, investigations were carried out to determine the
potential of South Deeps narrow deposits by designing capital
developments to the deposits using the conventional 1:7 decline
gradient. A conceptual mine design was completed for accessing
the four optimised areas in the South Deeps (Figure 7). The
decline portal to access these areas was designed in the southern
centre of the Nim3 pit. Capital development consists of the main
decline, fresh air drives, return air drives and level cross-cuts.
From the portal the decline continued to the south curving right
past the optimised section of the Colloform mineralisation then

130 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 MONORAIL TECHNOLOGY – A RAPID AND COST-EFFECTIVE METHOD OF DECLINE DEVELOPMENT

View A

FIG 4 - Configuration of monorail drilling system.

FIG 5 - Nim3 deposit of Jundee operations.

FIG 6 - Nim3 pit and South Deep deposits at Jundee mine (Newmont, 2004).

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 131

E CHANDA and B BESA

The Nexus structure

The Nexus structure strikes north and dips moderately to the west
at about 40°. The structure generally lies close to the basal contact
with a thick overlying dacitic porphyry body (the Nexus Dacitic
Colloform Porphyry), which appears to be largely concordant with the local
Midas / Moneyline stratigraphy. The Nexus orebody has approximate strike length of
1.1 km. About 400 m long of the southern portion of the structure
is drilled with wide spaced drilling and geological confidence is
low. The mineralisation at Nexus structure is patchy within the
corridor between 96 550 mN and 96 800 mN. High-grade Nexus
mineralisation that occurs beneath the Nim3 pit could be modelled
only for a short strike length of 125 m. The structure beneath
Nexus
Nim3 trends north and dips steeply at 70° to the west.
Gateway
Mining method and access to Nexus structure
Sublevel open stoping mining method is used for the extraction
of the Nexus deposit. This is because the same mining method is
applied in similar areas of the mine and has proved to be
successful. The orebody characteristics also favour this mining
method. Horizontal levels (cross-cuts) from the decline to the
FIG 7 - Conceptual mine design for accessing South Deep deposit are developed at 10 m intervals using the monorail system.
deposits (Newmont, 2004).
Access
continuing towards the Midas and Moneyline structures before The access to Nexus deposits is via a 4.0 mH × 4.0 mW size of
turning left and declining parallel to the Nexus structures. Access to monorail decline that is 212 m long. The decline starts from the
Gateway was through the hanging wall of the Nexus mineralisation. box-cut entry portal (located on the southern centre of the
Following the optimisation of the deposits, resources were existing Nim3 pit). The access joins Nexus main decline at
found to be far from becoming potentially economic. Table 1 elevation 2390mL (Figure 8).
indicates detailed development costs for Nexus deposit using The main decline to Nexus deposit is developed with gradient
conventional truck and haulage method (Newmont, 2004). 20° from 2440mL and spirals down to 2140mL. According to the
manufacturers of monorail train, the minimum decline dimension
for one monorail train application is 3.0 m × 3.0 m. However,
TABLE 1 decline dimensions of size 4.0 m × 4.0 m are used in this design
Cost analysis of capital development to Nexus deposits so as to accommodate other mine services such as ventilation
(conventional method). tubing, air and water pipes, cable, etc. Since a monorail train can
negotiate curves of 4 m minimum horizontal track radius
No Description Development cost (Scharf, 2005; Chanda and Roberts, 2005; Chanda et al, 2007), a
(% of total) curve radius of 6 m is used in this application. This is to avoid
Cost of decline access unnecessary stresses on the beam and on the rollers, thus causing
1 Incline from portal to Nexus Decline 6.2 excessive wear on the track and damage to the roller bearings.
Also varying lengths of straight ramps are used to provide best
2 Decline from 2440mL to 2140mL 23.8 access to the orebody. Table 2 summarises the design parameters
Subtotal 29.9 used for the monorail decline.
Cost of cross-cut access to Nexus deposit
1 Cross-cuts to Nexus deposits 56.4 TABLE 2
Design parameters for Nexus deposits.
Subtotal 56.4
Cost of ventilation network No Development type Units Size
1 Fresh air intake cross-cuts 4.3 1 Decline gradient Degrees 20
2 Fresh air intake raise 2 2 Curve radius m 6
Subtotal 6.3 3 Nexus decline development m 4.0 × 4.0
1 Exhaust cross-cuts 4.8 4 Development density factor t/m3 2.85
2 Exhaust raise 2.5
Subtotal 7.3 Waste handling
Grand Total 100 The monorail drill-load-haul system is designed specifically for
development of declines. During operations of this system, waste
material from decline development ends will be removed by
means of a monorail pneumatic loading system. The pneumatic
PROPOSED MINE DESIGN FOR MONORAIL loader will suck rock fragments from the decline development
APPLICATION face into the monorail hopper and once the hopper is full
material is automatically discharged into the monorail
In an effort to improve the economic viability of South Deeps containers. However, removal of waste material from access and
deposits, monorail technology is used to design the decline to ventilation headings will be done using conventional truck
these deposits. Nexus deposit is used as a case study area for haulage system. This means that waste material from the access
designing decline access for monorail application. Mine design and ventilation heading will be removed and loaded into
for the Nexus deposit was completed in Datamine. monorail containers using diesel load-haul-dumps (LHDs).

132 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 MONORAIL TECHNOLOGY – A RAPID AND COST-EFFECTIVE METHOD OF DECLINE DEVELOPMENT
Nexus
NexusDeposit
Deposit
FIG 8 - Decline design from the portal to Nexus deposits.
It is also anticipated that after blasting the development face,
rock fragments will have oversized material that will not be
sucked by the pneumatic sucker. The oversized materials will
create problems during pneumatic loading operations and as such
they will need to be reduced to manageable sizes by secondary
breaking at the face. However, to minimise oversized rock
fragments and improve rock fragmentation, this study has
determined that blasted rock fragments should have a maximum
diameter of 110 mm.
Since the decline will be developed with a size of 4 mH × 4 mW,
the tonnage per metre is 46 t/m (with density of 2.85 t/m3).
According to simulation results of the monorail system, the
system will have an advance rate of 12 m per day based on two
shifts per day and 12 hours per shift, meaning that a total of
547 tonnes of material will be handled by the pneumatic loading
system per day. Practically, this tonnage is quite substantial to be
sucked by the pneumatic loading system per day. Since this study
is theoretical, practical test of the monorail pneumatic loading
system is required so as to establish its true productivity (tonnes
per day). Nevertheless, the key to achieving this advance rate is
the simultaneous drilling, loading and transport operations of the
monorail system.
Total cycle time for the monorail drill-load-haul system.
Description Value (hours)
Time to clean the decline face 3.2
Time to drilling decline 3.5
Charging, rail extension, blasting and ventilate 1.5
Total drill-blast-load-haul cycle time 5.0
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
Portal
Portal
Incline
Inclinefrom
from
portal to Nexus
portal to
declineNexus
decline
NexusDecline
Nexus Decline
Ore handling
Removal of the ore from stopes may be accomplished by means
of LHDs or it is being suggested that the monorail system be
extended to the cross-cuts to lode. If the LHDs are used to
transport ore from the stopes to the monorail system via the
cross-cuts, special arrangements have to be made to allow the
LHD load into the monorail containers in the decline. However,
if the monorail system is extended into the cross-cuts, LHDs will
remove the ore from the stopes to the stockpile located at the end
of the cross-cut, and from here ore will be loaded into the
monorail containers using the pneumatic system. Alternatively,
LHD may be used to load ore from the stockpile into the
monorail containers but this will just increase the loading time
and needs to be avoided.
Monorail cycle time
Table 3 shows the total cycle time for the monorail drill-load-
haul system. According to the table, the total drill-blast-load-haul
cycle time for the system is five hours (ie 3.5 + 1.5). Results of
the simulation also indicate that cleaning and drilling of
TABLE 3
Comment
Include loading of material in containers, travel to surface, dumping on surface and
back underground.
Including drilling of support holes and holes for rail installation
Charging and rail extension to be done simultaneously
For two blasts (ie ten hours) this leaves two hours for delay and other activities in a
12 hour shift.
133
E CHANDA and B BESA

the development face would take 3.2 hours and 3.5 hours
TABLE 4
respectively. Thus, for a 12 hour shift, the number of blasts will
be 12 ÷ 5 = 2.4 (≈2) per 12 hour shift. Access development parameters for Nexus structures.

No Description Rock Length

Design of cross-cuts and ventilation access tonnage (t) (m)
Incline from portal to Nexus decline 9400 212
Horizontal development headings exiting the Nexus decline
provide access to stopes and draw points of the Nexus deposits. 1 Decline from 2440mL to 2140mL 41 600 895
Horizontal cross-cuts from the decline towards the Nexus Total 51 000 1107
deposits are also designed with dimensions 4.0 m × 4.0 m and at Cross-cuts (a total of 31) 227 600 4991
10 m interval from 2440mL to 2140mL. Results of the design are 2
Total 227 600 4991
shown in Figure 9.
Fresh air intake access 9800 640
Fresh and exhaust air ventilation drives 3 Exhaust access 27 000 1674
Total 36 800 2300
Figure 10 shows the fresh air intake access to Nexus deposits.
Fresh air intake access (ie intake cross-cuts and raises) are
designed with dimensions 2.5 m × 2.5 m as compared with
3.0 m × 3.0 m (for intake raises) and 4.5 m × 4.5 m (for intake Conventional development versus monorail
cross-cuts) used in conventional design. Smaller dimensions for system
fresh air intake access are used because the monorail system uses
electricity and it is anticipated that there will be less diesel fumes Total development metres for the monorail system are compared
with development metres obtained using conventional decline
in the decline during operations. Fresh air intake access is
access to the Nexus deposit. Figure 11 shows that the total
designed on the footwall side of the deposit to take advantage of decline development metres (ie from the portal to 2140mL)
the competent ground. According to the designs, fresh air enters would reduce from 2963 m using conventional development to
Nexus decline from 2390mL and is pumped down to 2140mL by 1107 m using monorail technology. This represents 62.6 per cent
means of booster fans. Several fresh air intake cross-cuts supply reduction in total development metres. Also because of the
fresh air to intermediate levels. Exhaust airways from Nexus reduction in the size of decline development, ie from the
deposits are also designed on the footwall side with dimensions conventional 5.5 mH × 5.5 mW to 4 mH × 4 mW, the total
of 2.5 m × 2.5 m (Figure 10). Table 4 shows the total tonnage and tonnage of material to be moved per metre reduces from 91 t/m
development metres of the design. to 46 t/m. This means that the total tonnage of waste material to

Portal
Portal
Nexus
NexusDeposit
Deposit

Incline from
Incline from
portal to
portal toNexus
Nexus
decline
decline

X-Cuts
X-Cuts to
to Nexus
Nexus
Deposit
Deposit

Nexus
Nexus Decline
Decline

FIG 9 - Decline design from the portal to Nexus deposits.

134 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 MONORAIL TECHNOLOGY – A RAPID AND COST-EFFECTIVE METHOD OF DECLINE DEVELOPMENT

To
To surface
surface

Fresh
FreshAir
air Portal
Portal

Nexus Deposit
Deposit

Incline from
Incline from
portal to Nexus
portal to Nexus
decline
decline
Exhaust
Exhaust
Ventilation drives
ventilation drives
X-Cuts to
X-Cuts to Nexus
Nexus
Deposit
Deposit

Fresh
FreshAir
air
intake
Intake

Nexus
Nexus Decline
Decline

FIG 10 - Design of fresh air intake and exhaust to Nexus structures.

6000

5000
D evelopm entm etres

4000

3000

2000

1000

-
Decline Dev Cross-cut Dev Fresh air intake access Exhaust air access

Conventional (m) 2963.00 5638.00 828.00 970.00

Monorail (m) 1107.24 4991.10 628.92 1673.93

FIG 11 - Conventional versus monorail development metres.

be moved would reduce from 269 000 (2963 × 91) tonnes in
conventional development to approximately 51 000 (1107 × 46)
tonnes using monorail technology giving a reduction of 81 per cent.
Figure 11 also indicates that the total horizontal cross-cut
development metres would reduce by 647 m, ie from 5638 m to
4990 m (a reduction of 11.5 per cent). Although this represents a
modest reduction in the total cross-cut development metres,
the total tonnage reduces by 55.6 per cent (from 513 000 to
227 600 tonnes) as shown in Figure 12. The reduction in total
tonnage is attributed to reduced size of the cross-cuts.
Fresh air intake development metres also reduce from 828 m
to 640 m giving a reduction of 22.7 per cent. The reduction in the
size of the fresh air intake developments, ie from 3 m × 3 m to
2.5 m × 2.5 , resulted in the reduction in total tonnage of material
to be moved from 21 500 tonnes to 9300 tonnes (a reduction of
56.8 per cent).

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 135

E CHANDA and B BESA

600 000

500 000

400 000
Tonnage
300 000

200 000

100 000

-
Fresh air intake
Decline Dev Cross-cut Dev Exhaust air access
access
Conventional (Tonnes) 269 633.00 513 058.00 21 528.00 91 447.94
Monorail (Tonnes) 51 010.02 227 594.62 9797.88 27 086.27

FIG 12 - Conventional versus monorail tonnage to be removed.

Further, Figure 11 shows that the total exhaust airway
development metres increase from 970 m to 1674 m representing
an increase of 72.6 per cent. The increase in exhaust development
metres results from the fact that LHDs will be used for ore
handling in stoping areas hence more fumes from these areas
needs to be exhausted. Figure 12 shows a corresponding reduction
in tonnes to be removed for total exhaust airway development.
Cost analysis of capital developments to Nexus
deposit
Generally, the costs of capital development to access the deposit
as well as mine services is considered to be preproduction capital
costs and are typically the largest component of mining capital
costs. Capital developments also requires the longest time period
of any mine activity in preparing the mine for production. In this
section an analysis of the preproduction costs associated with
decline development to the Nexus deposits using monorail
technology as well as conventional method is undertaken as
indicated in Table 5.
Development capital costs for monorail application were
calculated on first principle, ie development length multiplied by
development cost per metre. This means that after determining
the development metres for the decline, cross-cuts and ventilation
access, the development costs were calculated by multiplying the
development metres by the development cost per metre (see
Table 4). The costs used for determining monorail costs are the
same as those used in conventional development according
Newmont 2004 Report. According to Table 5, it would cost
approximately A$2.7 m to develop a 212 m long incline from
entry portal to the main Nexus decline as well as the
development of 895 m long main Nexus decline from 2440mL to
2140mL. The total cost of developing horizontal cross-cuts from
the decline to Nexus deposit would be approximately A$12 m
while a total of A$4.4 m would be spent to develop ventilation
access drives to the area.

TABLE 5
Cost analysis of capital development to Nexus deposits.

No Description Length (m) Average development cost Development cost†

(A$/m) (A$’000’000)
Conventional Monorail Conventional Monorail Conventional Monorail
Cost of decline access
1 Incline from portal to Nexus decline 622 212 2400 2400 1.5 0.5
2 Decline from 2440mL to 2140mL 2341 895 2480 2480 5.8 2.2
Subtotal 2963 1107 7.3 2.7
Cost of cross-cut access to Nexus deposit
1 Cross-cuts to Nexus deposits 5638 4991 2450 2450 13.8 12
Subtotal 5638 4991 2450 2450 13.8 12
Cost of ventilation network
1 Fresh air intake cross-cuts 608 133 1650 1650 1.0 0.2
2 Fresh air intake raise 230 507 2170 2170 0.5 1.1
Subtotal 838 640 1.5 1.3
1 Exhaust cross-cuts 680 1180 1760 1760 1.2 2.1
2 Exhaust raise 290 470 2070 2070 0.6 1.0
Subtotal 970 1650 1.8 3.1
Grand total 24.4 19.1

† Development costs are based on information from Jundee Mine Planning Group Report, 2004.

136 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 MONORAIL TECHNOLOGY – A RAPID AND COST-EFFECTIVE METHOD OF DECLINE DEVELOPMENT

Conventional versus monorail technology costs TABLE 6

Figure 13 indicates that the total cost of decline development Capital costs for purchase of monorail train (prices by
reduces from A$7.3 m in conventional development to A$2.7 m Scharf as at 14 March 2008).
using monorail technology representing a 63 per cent reduction. No Unit A$ Comments
The reduction is attributed to the reduced number of
development metres due to steeper gradient as well as the 1 Train 1 200 000 Price by Scharf
reduced size of the decline. Development costs for horizontal 2 Containers 16 000 Price by Scharf
cross-cuts also reduce by A$1.8 m with application of monorail 3 Monorail tools 25 000 Price by Scharf
technology. Compared with conventional method, this represents
a reduction of 13 per cent in total cross-cut development cost. 4 Shunting trolley 49 000 Price by Scharf
Figure 13 also shows that the total cost of ventilation access 5 Dispenser (bolt installation) 26 000 Price by Hilti
development would increase by 33 per cent. This is due to Total 1 316 000
increase in exhaust development metres by 72.6 per cent based
on the redesigned mine layout. However, the increase in
development costs is minimal compared with the savings that
will be made from the total development costs. TABLE 7
Capital costs for monorail installation per metre (prices by
Scharf as at 14 March 2008).
Cost analysis for purchase and installation of
monorail system to Nexus deposit No Unit A$/metre Comments
1 Rail 125.00 Price by Scharf
Monorail capital costs and operating costs 2 Electrical components 250.00 Price by Scharf
Capital costs for monorail operations consist of installation of a 3 Bolts (2 bolts/3 m section) 72.00 Price by Hilti
monorail and purchase of a monorail train. Table 6 shows capital 4 Rail suspension components 75 Price by Scharf
costs, for the purchase of a monorail with two drive cabins, four 5 Labour 51.04 Estimated
drive units and six lifting beams each with a payload of
30 tonnes (Sharf, 2007). Table 7 shows the capital costs per 6 Jumbo drill (bolt installation) 4.45 Estimated
metre for monorail installation. The installed power on the train Total 577.49
is 232 kW. The total cost of monorail installation in the Nexus
decline can be calculated as follows:
Thus, the total cost of decline development to Nexus deposit
Total cost (monorail installation) = [Cost of monorail train] + together with purchase and installation of the monorail train is
[Installation cost per metre × decline length] computed as follows:
From Table 6, the total cost of purchase of one monorail train
(with all accessories) is A$1 316 000 while the total installation = Cost of capital developments + total monorail cost
cost per metre was estimated at A$577 (Table 7). Also from the = 19 100 000 + 1 955 000
results of the mine design, the total decline length from the Nim3
entry portal to Nexus deposit is estimated as 1107 m. Therefore, = A$21 055 000
the total cost for monorail installation is estimated as indicated
below: Therefore, approximately A$21 M is required to develop and
install the monorail system to access Nexus deposit. When
Total cost (monorail installation) = [1 316 000] + [577 × 1107] compared with the total cost of using the conventional haulage
system as evaluated by Newmont, 2004, there is a saving of
= A$1 955 000 A$5.3 M for capital development to access the same deposit.

16 000 000

14 000 000

12 000 000
C osts (A $)

10 000 000

8 000 000

6 000 000

4 000 000

2 000 000

-
Fresh air Exhaust air
Decline Dev Cross-cut Dev
intake access access dev
Costs costs
dev costs Costs
Conventional (A$) 7 300 000 13 800 000 1 500 000 1 800 000
Monorail (A$) 2 700 000 12 000 000 1 300 000 3 100 000

FIG 13 - Conventional versus monorail development costs.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 137

E CHANDA and B BESA

This represents a 22 per cent reduction in total capital cost for
access development to Nexus deposit (Figure 14). Operating
costs for monorail operations include maintenance of parts,
power supply as well as labour costs. Table 8 shows operating
costs for the monorail train according to Scharf (Stefan, 2007).
TABLE 8
Operating costs for monorail train (estimates by Scharf).
No Description Operating costs
(A$/m)
1 Maintenance parts
2 Power
3 Maintenance labour
Total
Duration of decline development to Nexus
deposits
It is envisaged that the designed monorail drill-load-haul system
will have an advance rate of 12 m per day based on simulation
studies of the monorail pneumatic loading system compared to
9 m per day in conventional development. The increase in the
advance rate is due to the reduced size of the decline and
increased number of blasts that are possible per shift resulting
from reduced time of cleaning and drilling the development
face. The total length of the decline for the Nexus deposit is
1107 m. Therefore, the time taken to mine the decline is 1107 ÷
12 or 93 days. Theoretically, it would take 93 days to access the
Nexus orebody by means of a decline using monorail technology.
In contrast, it would take 2963 ÷ 9 or 330 days to access the
Nexus orebody using conventional decline development.
12.00 INSTALLATION OF THE MONORAIL IN THE
34.40 DECLINE
2.59 It has been determined that the monorail technology can reduce
48.99 the cost of capital development to exploit the Nexus deposit. This
section looks at installation of the monorail system from Nim3
box entry portal to the bottom of the decline a distance of
1107 m. Sections showing monorail installation have also been
indicated in this section.
The monorail system runs on an ‘I’ beam rail suspended from
the roof by chains attached to suspension bolts. Installation of a
monorail is a combination of three major activities:

30 000 000

25 000 000

20 000 000
C osts (A $)

15 000 000

10 000 000

5 000 000

-
Conventional Monorail
Total Cost of capital
deevelopment (A$) 24 400 000.00 19 100 000.00

FIG 14 - Comparison of total capital development to cost to Nexus deposits.

350

300

250

200
D ays

150

100

50

0
Conventional Monorail
Duration of decline 330 92.3
development (days)

FIG 15 - Duration of decline development to Nexus deposits.

138 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 MONORAIL TECHNOLOGY – A RAPID AND COST-EFFECTIVE METHOD OF DECLINE DEVELOPMENT

1. drilling support holes, phases of the monorail installation process (Oguz and Stefanko,
1971). A badly installed monorail track will develop unnecessary
2. roof bolting, and tensions in the system, which can lead to trouble and large power
3. rail placement. consumption.

Drilling support holes Rock bolting and support

The Hilti OneStep rock bolts are used as the suspension bolts
In decline development (just like in horizontal development),
(Figure 18). With the monorail system, the rock bolt installation
holes for rock bolt installation are drilled to a depth of 2 m
and rail extension in the decline will be done simultaneously as
and at 3 m interval perpendicular to the decline roof surface
the face is being charged. Since supporting the face is critical
(Figure 16). Improper drilling of holes for rock bolt installation
during decline development, this will be done before face drilling
reduces the lifetime of the rock bolts (Figure 17). According to
commences. The monorail system will drill holes that will be
Oguz and Stefanko, 1971, drilling time per support hole (manual
supported before face drilling commences. The bolts used for
drilling) for rock bolt installation (including all kinds of work
monorail installation has an ultimate strength of over 320 kN,
and delay) is approximately 16 minutes. Net drilling time
a diameter of 38.5 mm, length of 2 m and requires a 41 mm
excluding delays takes approximately six minutes per hole. It
diameter hole. Each bolt requires the purchase of a dispenser and
should also be noted that drilling of support holes is likely to
an intensifier for the installation. Resin containing rapid-curing
reduce when drilling with hydraulic booms. Preparatory
activities for drilling such as lining and marking the hole, setting
up the stoper and collaring the hole are the most important

Roof bolts perpendicular

to roof surface

Carefully
Decline roof measured chains
surface

Monorail

FIG 16 - Drilling of rock bolts for monorail installation (Scharf, 2007). FIG 18 - Hilti OneStep anchor bolt (Hilti Corporation, 2004).

FIG 17 - Drilling of holes for rock bolt installation (Scharf, 2007).

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 139

E CHANDA and B BESA

adhesive is contained within the bolt (Scharf, 2005). Bolt CONCLUSION

installation requires one operator and can be installed in
40 seconds after the hole is drilled. The rock bolt for monorail
application must take static and dynamic loads. Normally,
dynamic load with angular transmission will cause deflection of
rock bolts. This will reduce the lifetime of the rock bolts by a
factor of ten or more (Scharf, 2005). Therefore, to avoid this, a
bracket is with preloaded roof bolt is used.
Rail placement and alignment
Installation of the rock bolt is followed by attachment of a special
eyebolt on the threaded end of the bolt. A shackle provides easy
connection of the chain to the eyebolt (Figure 19). From the
shackle down, the distance should be carefully measured to obtain
the length of the chain for a horizontal track installation.
Measurement of the chain is done by connecting a new rail
section to the one which is already installed permanently. The
front end of the new rail section is lifted until it is horizontal and
in line with the others. While one man holds the rail in this
position, another man takes the measurement from the shackle to
the hook on the top flange of the rail. Measured lengths of the
chain are then cut using oxygen burner. During monorail
installation at curves, the rail sections are blocked rigidly
providing more stable track structure. Figure 20 shows a typical
configuration of the monorail train in a drive.
This study indicates that development of decline access to Nexus
deposits using monorail technology is feasible. Compared with
conventional decline development, results have shown that the
monorail system has the potential of reducing the decline length
to Nexus deposits by over 62.6 per cent and decline costs by
63 per cent. Further, the study indicates that with the monorail
system, there is a potential of reducing total capital development
costs by 22 per cent. Also, due to rapid development by the
monorail drill-load-haul system, the shorter decline length
coupled with smaller decline openings, the duration of decline
development reduces by 71.8 per cent. This case study
demonstrates that the economics of narrow vein deposits can be
improved significantly by using the monorail system.
It is also a known fact that the operations of the monorail
drill-blast-load-haul system will have an effect on safety as well
as on mine regulations in Western Australia. Therefore, it is
recommended that further investigations be made in order to
assess the regulatory requirements for the operations of the
monorail system in metalliferous underground mining in Western
Australia. The authors further recommend field verification of
the results of this study by installation of a prototype monorail
system at a mine site whereby experiments maybe conducted on
the operation of the system.

A B

FIG 19 - Rock bolt installation and rail installation (courtesy Scharf, 2007).

A B
Vent bag 906
541

1841
1300

3000
1159

650 1100 500 1100 450

1200 1600 1000

3800

FIG 20 - Configuration of monorail train in a drive: (A) one train, (B) two trains (courtesy Scharf).

140 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 MONORAIL TECHNOLOGY – A RAPID AND COST-EFFECTIVE METHOD OF DECLINE DEVELOPMENT
ACKNOWLEDGEMENTS
The authors would like to thank Minerals and Energy Research
Institute of Western Australia (MERIWA) and Newmont for
financial support of the monorail research project at Western
Australian School of Mines (WASM) and Scharf Mining
Solutions for providing technical data and cost information on
the Electric Monorail Haulage System (EMTS).
REFERENCES
Biegaj, K, 2002. Mobile supersucker winzing – Attaining proven prior to
capital commitment.
Chanda, E K, Besa, B and Mahinda, K, 2008. Evaluation of monorail
haulage system in underground metalliferous mining, final report
M382, Minerals and Energy Research Institute of Western Australia
(MERIWA), 150 p.
Chanda, E K, Darcey, W, Mahinda, K and Besa, B, 2007. Application of
electro-monorail mine haulage systems in underground decline
development, in Proceedings 16th International Symposium on Mine
Planning and Equipment Selection (MPES 2007), Bangkok, 11 - 13
December.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
Chanda, E K and Roberts, M, 2005. Evaluation of a monorail haulage
system in metalliferous underground mining, in Proceedings
Hoist and Haul Conference, pp 39 - 44 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Dorricott, M G and Jones, O, 1984. Vacuum lift systems for the transport
of broken rock in shafts, paper presented to Regional Conference on
Gold Mining, Metallurgy and Geology, The Australasian Institute of
Mining and Metallurgy Perth and Kalgoorlie Branches.
Hilti Corporation, 2004. FL-9494 Schaan, Principality of Liechtenstein
[online]. Available from: <http://www.hilti.com>.
Jones, O I, 1989. Pneumatic conveying in mines, paper presented to Third
International Conference on Bulk Materials, Storage, Handling and
Transportation, Newcastle, 27 - 29 June.
Newmont, 2004. Jundee Mine Planning Group, South Deeps Mining study
report.
Oguz, S and Stefanko, R, 1971. Evaluation of a monorail mine haulage
system, special research report, The PennState University, Department
of Mining Engineering, 106 p.
Scharf, 2005. EMTS monorail electric powered transport system brochure
(Sharf).
Scharf, 2007. EMTS monorail, electric powered transport system brochure
(Sharf).
Stefan, M, 2007. Personal communication, Scharf, Germany.
141
Augusta Gold-Antimony Mine
M McCarthy1, A Fowler2 and S Marshall3

ABSTRACT Following a lull in production, the Costerfield mine started

The Augusta underground mine at Costerfield in Central Victoria has
been in operation since mid 2006; however, AGD Operations Pty Ltd have
explored and undertaken open pit operations in the area for over 20 years.
The high-grade, narrow gold-antimony orebody is being mined using
both mechanised and hand-held methods, to produce up to 70 000 tonnes
per annum at ten grams per tonne gold and six per cent stibnite. AGD
Operations Pty Ltd processes the ore on site using gravity and flotation to
produce two products; a concentrate which is sold to Chinese smelters
and a gravity concentrate which is refined off-site. The circuit achieves up
to 30 per cent recovery of gold from gravity methods. There are many
challenges in the underground mine, including an orebody with an
average width of 300 mm, very poor rock mass conditions in the host
rock and variable dips in the ore along strike, between 45° and 90°. The
upper levels of the mine have been developed ready for stoping; however,
their depths range from 20 m to 70 m below surface in both oxide and
transitional material. A major fault structure consisting of mud has been
encountered underground and is associated with the orebody causing very
poor ground conditions, water inflow and requiring additional
reinforcement. The northern limit of the mine is defined by the proximity
of historical workings in the South Costerfield mine, which is being
continuously dewatered.
producing again between 1934 and 1939. Stibnite concentrate
(1623 tonnes) was produced during this period under the
management of Gold Exploration and Finance Company of
Australia (WMC) (Hazeldene, 2007).
A series of different operators explored and mined the area
from 1939 to 1981 including the South Costerfield Antimony and
Gold Company, Victorian Antimony Mines Ltd, Mid East
Minerals, Metals Investment Holdings and Forsayth Mineral
Exploration and Costerfield Mining Pty Ltd. During this time
production was comparatively limited (Hazeldene, 2007).
In 1981 the Victorian Mines Department explored the area and
discovered mineralisation under shallow alluvial cover that was
later defined as part of the Augusta Deposit. Following
discovery, and up until 2005, Federation Resources and
Australian Gold Development (AGD) delineated an ore reserve
with Probable status of 204 000 tonnes at 10.9 grams per tonne
gold and 5.7 per cent stibnite. Open cut mining down to 10 m
below surface began in early 2006 and underground development
began in mid 2006 (Hazeldene, 2007).

INTRODUCTION GEOLOGY AND MINERALISATION

Various gold-antimony deposits have been mined in the The Costerfield Project is located in the Melbourne Zone of the
Costerfield gold-antimony field, in Central Victoria, since its Lachlan Fold Belt and is divided structurally from the Bendigo
discovery in the 1860s. Production has, however, been sporadic, Zone to the west by the crustal-scale discontinuity, the Mount
being highly dependent on metal prices and extractive William Fault (Vandenberg et al, 2000). Mineralisation in the
technology developments (Hazeldene, 2007). The AGD group west of the Melbourne Zone is characterised by an association
has been actively exploring the area since the 1980s. During the between gold and antimony (Vandeberg et al, 2000). The
mid 1990s AGD mined the upper levels of the Brunswick
Augusta Deposit comprises six mineralised veins that
Deposit for oxide gold by open cut, then in 2001, began
predominantly dip to the west, south-west at approximately
delineating the Augusta Deposit under cover at the southern end
60° and strike north, north-west. They are composed of massive
of the field (Hazeldene, 2007). Mining began in 2006.
stibnite with quartz, which occurs either intermingled with the
stibnite, as a stibnite/quartz breccia, or along the margins of the
LOCATION AND HISTORY stibnite as laminated quartz. Gold is commonly microscopic
The Costerfield Project is centred on the Costerfield township in (<20 microns) and is generally included within stibnite grains.
Central Victoria, located about ten kilometres north-east of Free gold has been observed in the weathered zone and also
Heathcote and 50 km east of the City of Bendigo. Heathcote is in some recent drill holes, below 100 m depth (Figure 1).
serviced from either Bendigo or Melbourne via the McIvor/
Northern Highway.
Small antimony deposits were first discovered and worked in
1853 in the Costerfield area. It was not until 1860 that
prospectors Coster, Field and Youle discovered the main
Costerfield Deposit comprising antimony sulfide (stibnite) with
visible gold. The Costerfield mine and its associated deposits,
Minerva and Bombay produced 14 761 ounces of gold and
21 804 tonnes of 40 per cent stibnite concentrate between 1860
and 1883 (Bannear, 1993).
Between 1904 and 1925, mines in the Costerfield area
produced 114 131 tonnes of ore for 63 445 ounces of gold and
25 768 tonnes of 49.8 per cent stibnite concentrate. Development
of the Costerfield mine reached 309 m depth.

1. AAusIMM, Mining Manager, AGD Operations Pty Ltd, PO Box 667,

Heathcote Vic 3523. Email: mmccarthy@agdops.com.au
2. Head Geologist, AGD Operations Pty Ltd, PO Box 667, Heathcote
Vic 3523 Email: afowler@agdops.com.au
3. Senior Mining Engineer, AGD Operations Pty Ltd, PO Box 667,
Heathcote Vic 3523. Email: smarshall@agdops.com.au
FIG 1 - ‘E’ lode, MH070 128.8 m downhole. Sample: P553101.
Assay results: 113 g/t Au, 30.6 per cent Sb. Photo: 7465, digital
image. Objective: 20x. Light: reflected. Nicols: uncrossed.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 145

M McCARTHY, A FOWLER and S MARSHALL
The mineralisation is interpreted as occurring in two episodes:
first was a gold-in-arsenopyrite phase, which was then followed
by a stibnite-rich fluid that re-mobilised the gold and included it
within the stibnite (McArthur, 2005).
Geometrically, the veins are arranged in an en-echelon pattern,
and are generally subparallel. The largest of the veins, ‘E’ lode is
continuously mineralised along a strike length of approximately
400 m, and is on average 298 mm with a grade of 52.4 grams per
tonne gold and 29.2 per cent stibnite (Figures 2 and 3).
FIG 2 - ‘E’ lode, two level, north drive: 4567 mN in Costerfield mine
grid, looking north.
FIG 3 - Schematic cross-section through the Augusta Deposit on
4400 mN in Costerfield mine grid, looking north.
146 Ballarat, Vic, 14 - 15 October 2008
Two veins are currently being mined: ‘E’ lode and ‘W’ lode.
Both contain ore shoots that plunge to the north at approximately
35° (Figures 4 and 5). Maximum lode width and grade observed in
an ore shoot was 1750 mm of massive stibnite grading 28.6 grams
per tonne and 33.6 per cent stibnite. The ore shoots are spatially
correlated with where the lodes intersect major cross-cutting,
un-mineralised structures that trend northeast-southwest. These
major cross-cutting structures are highly sheared, containing up to
a metre of pug material and generally disrupting the hanging wall
and/or footwall across several metres width. The cross-cutting
structures do not significantly displace the lode structures in a
strike slip sense (dip slip displacement unknown). The maximum
strike-slip displacement observed is approximately 6 m; however,
the mineralisation is consistently observed to continue through and
around this displacement, suggesting that the mineralisation
post-dates the main phase of movement on these north-west
striking faults.
Grade and lode width appears to be distributed in an
en-echelon arrangement. The ore shoot in ‘W’ lode, which is to
the west and in the hanging wall of ‘E’ lode, is offset to the
south, and continues significantly further down plunge than the
ore shoot in ‘E’ lode. Furthermore, on a level, ‘W’ lode starts to
become ore grade at the approximate northing where ‘E’ lode
becomes subeconomic.
The mineralisation is hosted by the ubiquitous and largely
featureless, Silurian Costerfield Siltstone. Regionally, the
siltstone is interpreted to form the core of an Antiformal Dome,
with the gold-antimony deposits occurring along the axis of the
dome (Edwards et al, 1998). In the Augusta mine, the hinge
of the major antiform has not been observed. The bedding
predominantly dips to the west-south-west, in a similar
orientation to the lodes, suggesting the Augusta Deposit was
precipitated on the western limb of the north, north-west striking
antiform. There are exceptions where bedding is observed to be
rotated in the vicinity of faults. Mapping in the ore drives shows
a gentle warping of bedding near the lode structures with fold
hinges plunging down the dip direction of the bedding,
approximately west, south-west. This suggests minor north-south
compression, possibly related to the mineralising event.
RESOURCE ESTIMATION4
An updated Mineral Resource estimate for the Augusta mine has
been prepared. The data set is as of 30 June 2008 and includes all
recent diamond drilling and face sample information and is
reported depleted of development and stoping.
The summary Mineral Resource Table 1 is as follows and
excludes the estimate for ‘N’ lode that was completed in 2007.
There has been no new data added to ‘N’ lode that can yet be
incorporated into a Mineral Resource estimate.
Diamond drill data has been loaded from the site databases
and validated for downhole surveys and assay information. Face
samples have been spatially located to represent 3D channel
samples in their correct location. All mineralised intercepts have
been coded by lode and composited to the full thickness of the
lode in each location. True widths have been calculated for the
diamond drill intercepts. For the face sample information the true
width is as measured at the face by the geological team.
4. The information in this report as it relates to Resource Estimations
was compiled by Mr Dean Fredericksen, MAusIMM, who is a
consultant geologist to AGD operations, with sufficient experience
relevant to the style of mineralisation and type of deposit under
consideration and to the activity which he is undertaking to qualify as
a Competent Person as defined in the 2004 Edition of the ‘Australian
Code for Reporting of Exploration Results, Mineral Resources and
Ore Reserves’. Mr Fredericksen consents to the inclusion of this
information in the form and context in which it appears in this report.
Narrow Vein Mining Conference
 AUGUSTA GOLD-ANTIMONY MINE

FIG 4 - Long section of the ‘E’ lode block model estimated by D Fredericksen in July 2008. Only blocks above 4.6 g/t Au equivalent
are displayed. In Costerfield mine grid.

FIG 5 - Long section of the ‘W’ lode block model estimated by D Fredericksen in July 2008. Only blocks above 4.6 g/t Au equivalent
are displayed. In Costerfield mine grid.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 147

M McCARTHY, A FOWLER and S MARSHALL
TABLE 1
Summary of ‘E’ and ‘W’ lode mineral resources.
Augusta Grade
Category Tonnes Au ppm Sb%
Measured 66 000 18.2 10.5
Indicated 180 000 11.7 5.9
Inferred 85 000 4.8 2.5
Total 331 000 11.4 6.1
The Mineral Resource estimate has been constructed using a
2D estimation methodology. All data has been transformed to a
common 2D plan. Metal accumulations (true width *Au ppm and
true width *Sb per cent) are calculated for each face sample or
diamond drill intercept. Block models for the face sampled or
developed areas of the mine are 2.5 m strike by five metres high.
In the diamond drill areas the dimensions are 20 m strike by 10 m
high. Two separate models were estimated for each data set then
combined together for reporting and mine planning purposes.
The true thickness and metal accumulations have been estimated
by ordinary kriging. Variogram and search ellipses were
determined using the face sample information and scaled
appropriately for use in the diamond drill areas. Gold and
antimony grades are back calculated by dividing the
accumulation by the estimated true thickness.
For each of ‘E’ lode and ‘W’ lode a series of structural
domains were modelled that define areas where the true dip and
strike are similar. For each of these domains (seven ‘E’ lode
domains and four ‘W’ lode domains) a strike and dip correction
factor is calculated that can be utilised to calculate the corrected
lode volume for each block. Volumes, tonnes and grades diluted
to a minimum width of 1.2 m have been calculated for each
block. Density has been estimated based upon an historical
regression calculation with Sb per cent that has been consistently
applied at Augusta mine.
The Mineral Resource has been classified based upon the
quality of the estimates (slope of regression of the estimate), the
drilling density and more importantly demonstration of
continuity of the lodes. Areas that have been accessed by
development top and bottom have been classified as Measured
Resource.
The Mineral Resource has been reported above a cut-off of
4.6 grams per tonne gold equivalent (AuEq) (Aueq = Au + Sb *1.9)
and at a minimum mining width of 1.2 m. The Mineral Resource
is depleted for mining based upon surveys of underground
workings completed 30 June 2008 and estimates of the material
yet to be recovered from stoping areas.
In addition to the resource quoted above, there are some
additional tonnes in the ‘N’ lode and ‘C’ lode; however, currently
there are no plans to mine these lodes.
MINE DEVELOPMENT
The Augusta mine is accessed via a boxcut, which is formed
from the original open pit developed in 2006. A concrete pillar
was installed in the base of the open pit prior to underground
development commencing. AGD Operations Pty Ltd worked
with mining contractors to operate a road header for the initial
50 m of decline and one level to ‘E’ lode.
The initial mining of ‘E’ lode for the 400 m strike length was
an exploration exercise to prove up continuity of the lode along
strike, since initial exploration drill spacing was 40 m. ‘W’ lode
is 200 m in strike length and was not considered to be a
significant contribution to production, but has since proven more
consistent in lode width and contains higher grades than ‘E’ lode.
148 Ballarat, Vic, 14 - 15 October 2008
Contained metal
Au eq Undiluted true Au (oz) Sb (t)
thick (m)
38.2 0.42 39 000 7000
23.0 0.42 67 000 10 700
9.5 0.38 13 000 2100
23.0 0.41 119 000 19 800
The initial mining in the upper levels (one and two level) was
by air leg boring and bolting with some air scraper work,
constructing a 2.4 m wide by 2.8 m high drive, and bogging with
Toro 151 loaders. Water flowing from the orebody created mud
that was a significant hazard, especially in ‘E’ lode on two level,
where after 200 m of advance, the floor level had washed away
up to 500 mm deep and required road base to be continually
brought in. During this time, the four metre wide by four metre
high, one in seven, spiral decline progressed with a single-boom
Atlas 126 jumbo, and cable bolting was required to be both
drilled and installed out of a tele handler basket. Bogging was
performed by an Elphinstone R1700G loader.
In July 2007, a twin boom Atlas Copco M2D was
commissioned for the decline, as well as a single boom Atlas
Copco H104 for the ore development. This enabled cable bolt
drilling to be performed with the jumbo and increased rates in
decline and cross-cut development. At the same time, the decline
was redesigned to include long straights and a 1:6.5 gradient. A
Wagner MT 426 Truck is used to cart ore and waste to surface.
Difficult ground conditions were experienced at the entrance
to four level ‘E’ lode. The entrance to four level ‘E’ lode was at
the intersection of a large fault and the rock mass conditions
were so poor that drill and blast methods were not used. An
excavator was taken underground and cut the profile of both
north and south ‘E’ lode drives for some 10 m until the ground
conditions improved. An excavator with a cutterhead installed
was also trialled, and was able to cut the host rock easily. The
intersection required shotcrete, which was applied using dry
application methods, as well as concrete re-enforcement of the
pillars, to ensure that it remained stable for the life of the level.
The decline is designed to be developed to ten levels (120 m
below surface) and further exploration drilling in 2008 has
extended the life of mine to 18 level (190 m below surface) with
‘W’ lode open at depth.
DILUTION AND GRADE CONTROL
The mine classifies ore as stope, high and low grade and
stockpiles them separately. Grade is controlled by the shift
supervisor in close consultation with the mine geologist and shift
samplers. For the ore drives, samplers visit each face on a daily
basis and advise the shift supervisor of the predicted grade based
on measurements of the lode width, quartz content and the width
of the drive (refer to Figure 6).
For stoping, the grade and width of the lode in the stope panel
is estimated by the mine geologist when it is designed. The
estimation is based on the face samples in the drives above and
below the stope panel. As the stope is mined, the samplers, mine
geologist and production engineer monitor the stope width and
dilute the lode grade accordingly.
ORE DRIVE DEVELOPMENT
In late 2007 to reduce dilution, part-facing was trialled in ore
development. This involved positioning the ore on one side of the
Narrow Vein Mining Conference
 AUGUSTA GOLD-ANTIMONY MINE

FIG 6 - Example of a grade control chart currently being used underground for a shanty-back ore drive profile.

drive, and using the ore contact as the hanging wall. The waste
was bored out and fired, and the waste cut width dimension being
such that the 1.4 m wide Toro buckets could bog the majority of
the waste from the face (refer to Figure 7). The ore was then
stripped in. Experimentation with boring on the hanging wall
contact or in the ore was trialled, as the former resulted in large
lumps of stibnite greater than the process plant grizzly aperture
of 250 mm, and would require secondary breakage with a rock
breaker. The increase in grade delivered to the mill was
significant; however, full face mining has since resumed due to a
number of issues. The number of faces on line to enable efficient
part facing is at least ten, to enable flexibility with the mid shift
and end of shift firings being undertaken. The ore stripping cut
took less than an hour to bore and charge, but the face would
remain idle until firing time. The drives would generally result in
widths greater than the 2.4 m design, to enable the loader bucket
to fit in beside the ore. This required more waste movement. The
positioning of the ore as the hanging wall also posed problems
for the production drill rig, as it required at least 250 mm
additional room to drill on the correct angle for stoping. After the
part facing trial, the ore drive width dimension was reduced to
2.2 m and part facing is now implemented only when the ore
width in the face would result in low grade being delivered to the
mill.
‘W’ lode development is a challenge as its average dip is 50°.
Sublevels are developed so that at least a 1:1 pillar is the result
along the dip of the ore. Shanty back profiles are mined. There is
significant survey control required when developing ‘W’ lode to
ensure the drives do not interact. All ore drives are normally
developed under geological control but in ‘W’ lode a laser is
used to control both the development line and grade. The level
FIG 7 - Part face example from ‘W’ lode, one level at 4361 mN in
drives are advanced at least 20 m in front of sublevels as an
Costerfield mine grid, looking south.
additional control to ensure the drives do not interact.

drill rig being used is a Boart Longyear Stopemate which has a

STOPING
Geotechnical consultants in both the feasibility stage and initial
development stage were sceptical whether short bench stoping
over any length of hanging wall expanse would be possible. The
mine was scheduled to assume that short bench stopes were
possible and a level spacing of 10 m floor to floor on ‘E’ lode
was planned. Trial stoping commenced on ‘E’ lode in January
2008, in an area that ranged from 67° to 65° dip. The production
shorter feed cylinder installed so that in the narrow drives it can
achieve 360° movement. ‘E’ lode on three level south was stoped
in the initial six months of 2008, with the trial parameters
changing as stope performance was assessed. Initially drill
patterns were dice five, with holes in the footwall, hanging wall
and in the ore. Fifty per cent strength ANFO was used for all
firings including the initial slot. For the first stope in three level
‘E’ lode, the initial slot dimensions were 1.2 m square, and this

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 149

M McCARTHY, A FOWLER and S MARSHALL
and subsequent firings were encouraging as they resulted in a
stope width matching design of 1.2 m. However, once the stope
was open for approximately 8 m along strike, parallel beds of
mudstone within the hanging wall would slide off under gravity,
and cause in some cases stope widths of in excess of 3 m.
Minimal overbreak occurred on the footwall. The loader
operators were able to bog these large slabs and stockpile them
separate to the ore in some cases. Extensive cable bolting of the
hanging wall was trialled, at less than favourable orientations
achievable from the drill and extraction drives. W strapping and
cable bolt tensioning was completed and stopes fired with this
extra support installed. However, in most cases the hanging wall
continued to fail under gravity around the cables. Geotechnical
consultants indicated that the transitional and weathered nature
of the rock mass was partly the cause of cable bolts being unable
to hold the hanging wall.
Stoping in ‘ W’ lode where the rock mass is generally blockier
than ‘E’ lode has had success in achieving better than the
predicted stope widths of 1.2 m wide. Current drill and blast
methods are successfully achieving 700 mm average stope width
over 15 m open along strike (refer to Figure 8). Hole spacing is
400 mm with one hole rings drilled 100 mm into the footwall.
Generally holes must be lined with poly as they squash after
24 to 48 hours of being drilled. In each firing the next hole
behind is damaged and is providing a shadowing effect, so
essentially a burden of 800 mm is being fired each firing. Hole
diameter is 57 mm charged with a 50 per cent strength ANFO
product. Ore is bogged and backfilled immediately with waste. A
void must be created on the free face to allow successful firing of
the narrow 700 mm wide stope. Further work with different
backfill materials including cemented fill, is currently ongoing in
an attempt to increase recovery and reduce dilution.
FIG 8 - Example of a stope taken looking down from W4 sublevel
at 4335 mN. Stope was 3.5 m high from backs to floor.
DEWATERING
The South Costerfield mine included workings on the modern
‘E’ lode and is dewatered to 65 m below surface, the mine
workings being approximately 60 m from the northernmost
extent of the Augusta mine workings. An evaporation dam with
40 ML capacity holds any excess mine water, surface storm
water and South Costerfield shaft water and is set up with sprays
to increase evaporation. The mine has recently reached below the
depth of the South Costerfield shaft pump and has experienced
between five to eight litres per second inflow from both the
decline and ore headings, as water is transferred along the lodes
and through fault zones. The host mudstone does not transmit
water. The current mine plan is to push the decline to ten level
and cross-cut to the lode structures in an attempt to dewater the
lodes prior to extensive mining of the lower levels.
150 Ballarat, Vic, 14 - 15 October 2008
VENTILATION
Primary ventilation is provided by a 110 kW axial fan installed
underground at one level. An overpass arrangement was designed
and constructed so that air lock ventilation doors would not be
required, as access to the magazine requires personnel to travel
past the primary fan chamber and return airway infrastructure.
Primary ventilation upcast shafts are approximately 3 m in
diameter and also act as the second means of egress. The primary
fan has been installed with the capability of reversing it in the
event it is required to be downcast. Secondary ventilation is
currently being supplied by 75 kW axial fans and a 110 kW
centrifugal fan located on surface, which was the dust extraction
fan when the road header was in operation. Ventilation testing is
performed by the geology samplers along with remote re-entry
time monitoring and Nox personal Draeger tube testing of loader
operators. Results are communicated to operators personally and
published on notice boards.
GEOTECHNICAL CHALLENGES
Augusta mine previously used contract geotechnical expertise for
monthly inspections and ground support design and has recently
employed a part-time geotechnical engineer. Mesh is installed to
grade-line in all areas, and in poor ground faces are required to
be meshed before boring and charging can occur. Bedding is
generally parallel with the lode and cross-cutting structures cause
hanging wall failure in stopes. These structures are identified
during initial mapping and have a quartz/calcite infill, making
blocks slip under gravity when exposed. The lodes are associated
with a major mud shear/fault zone which dips north-west. The
faults contain puggy shear zones and polished fault surfaces.
Regions where the ‘mud’ fault are intersected are also
hazardous due to the mud crumbling after it has dried out back
from the advancing face, which usually requires rehabilitation. It
is anticipated that mining and stoping ‘W’ lode at depth will be
problematic and a current study is whether or not dry fibrecrete
can be utilised for safe and efficient narrow vein mining.
PROCESS PLANT
Process plant design head grade is six per cent stibnite and
15 g/t gold and a throughput 75 000 tonne per annum (15 t/h).
The grizzly aperture is 250 mm. The grinding circuit consists of
two ball mills and achieves p80 of 106 micron. Gravity circuit
recovery is up to 30 per cent of the total feed gold, and
incorporates spirals on the cyclone underflows from both milling
circuits. The spiral concentrate reports to a Knelson concentrator
which is tabled. The flotation section incorporates two stages of
cleaning to produce a concentrate assaying 52 per cent Stibnite
and 60 grams per tonne gold. Flotation concentrate is filtered and
bulka bagged then shipped to customers in China. Penalties
occur for any concentrates below 52 per cent Stibnite.
Seventy-five per cent of the concentrate produced is sold under
contract and the operation is currently reviewing options for the
remainder to be sold to a different customer.
TECHNICAL SUPPORT
The mine technical staff consist of head geologist, mine
geologist and two samplers who work dayshift only, senior
mining engineer and recently a mine production engineer and
two mine surveyors working an eight and six roster.
PERSONNEL RECRUITMENT AND RETENTION
AGD Operations Pty Ltd has a labour and equipment hire
contract with a mining contractor, and the contract specifies that
personnel are to be sourced from the local area. The towns of
Narrow Vein Mining Conference

Seymour, Heathcote and other smaller communities are targeted
for potential recruitment of heavy mobile equipment operators
who are usually well suited to underground mining. A personnel
bus is run from Bendigo for the shift workers who work seven on
seven off, 12 hour continuous day and night shifts. All
supervision, training and personnel development is provided by
AGD Operations and employee turnover is very low. The
underground mining crews are between eight and ten people who
are required to be multi-skilled. This provides an opportunity to
learn hand-held mining methods, loader and truck operations,
development and production charging, production drilling and
jumbo operations in a relatively short time compared to larger
mines.
CONCLUSION
Augusta mine is successful in mining a difficult orebody in
transitional material.
The ‘W’ lode stoping is proving better than ore reserve
assumptions of a 1.2 m stope width and a significant reduction in
dilution has been achieved. Further work is planned to quantify
dilution being bogged against fill and trialling different filling
methods to reduce dilution. Mechanisation of the underground
mine has a further step to go, to remove the operator from hand-
held bolting if at all possible. Geotechnically the orebody is
challenging and may involve an improved method of ground
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
AUGUSTA GOLD-ANTIMONY MINE
support at depth. The ‘W’ lode orebody is the widest and richest
lode and is currently open along strike and at depth.
ACKNOWLEDGEMENTS
The authors wish to acknowledge AGD Operations Pty Ltd and
its parent Cambrian Mining plc for enabling this paper to be
published.
REFERENCES
Bannear, D, 1993. North Central Goldfields project, historic mining sites
in the Heathcote (Waranga South) Mining Division, Part two: Site
Gazetteer, Department of Conservation and Natural Resources, North
West Area.
Edwards, J, Wohlt, K E, Slater, K R, Olshina, A and Hutchison, D F,
1998. Heathcote, and parts of Woodend and Echuca, 1:100 000 Map
Area Geological Report, Geological Survey report No 108.
Hazeldene, R K, 2007. MIN 4644, EL 3310, and 4848, Costerfield
Project, Annual technical report for the period 1 July 2006 to
30 June 2007, AGD Operations Pty Ltd.
McArthur, G, 2005. AGD Operations Pty Ltd Augusta Project,
Costerfield Sb/Au mineragraphy.
Vandenberg, A H M, Willman, C E, Maher, S, Simons, B A, Cayley, R A,
Taylor, D H, Morand, V J, Moore, D H and Radojkovic, A, 2000.
The Tasman Fold Belt System in Victoria (Geological Survey of
Victoria).
151
Strategies for Minimising and Predicting Dilution in Narrow Vein
Mines – The Narrow Vein Dilution Method
P C Stewart1 and R Trueman2

ABSTRACT
An improved ability to predict dilution for narrow vein deposits enables
more accurate economic comparisons between mechanised longhole
stoping and other narrow vein mining methods, as well as reducing
uncertainty surrounding dilution estimates for deposits at the prefeasibility/
feasibility stage. This paper proposes narrow vein dilution prediction and
minimisation strategies for both greenfield sites and operating mines,
respectively. The narrow vein dilution method (NVD method) is a tool for
more accurately predicting longhole narrow vein dilution. The method is
based on a new concept called benchmark stoping width. Benchmark
stoping width is based on probabilistic analysis of overbreak distributions
from a typical longhole narrow vein mine. The premise for the NVD
method is that dilution associated with longhole stoping can be predicted
using the benchmark stoping width. In addition to dilution prediction, the
NVD method also includes strategies for narrow vein dilution minimisation
generally, including; filling, cablebolting, stress relaxation, extraction
sequencing and blast overbreak.
FIG 1 - Annual operating cost for 25 cm of dilution for typical
narrow vein mine (assumes unit dilution direct operating cost =
INTRODUCTION $25/tonne and 500 000 tonnes ore per annum).
Dilution can be defined as the contamination of ore by non-ore
material during the mining process (Wright, 1983). Dilution is Stability charts are commonly used for stope design and
associated with indirect and direct costs (Pakalnis, 1986; Elbrond, dilution estimation for both large open stoping and narrow vein
1994; Pakalnis, Poulin and Hadjigeorgiou, 1995; Villaescusa, stoping. Figures 2, 3 and 4 show three of the most commonly
1998; Bock, 1996; Bock, Jagger and Robinson, 1998; Revey, used stability charts. These charts plot stability number (N or N’)
1998; Scoble and Moss, 1994). An improved ability to predict against hydraulic radius (HR or S). Hydraulic radius is defined as
dilution enables the economic risks associated with unplanned the ratio of the area to the perimeter for the stope wall under
dilution to be reduced. consideration (Laubscher and Taylor, 1976). The stability
Brewis (1995) defines narrow mining to be the working of number N or N’ is calculated according to Equation 1, where A,
mineral deposits typically no more than two to three metres wide, B and C are empirical factors taking into account induced stress
with a dip exceeding 50 to 55 degrees (an angle at which broken acting parallel to the middle of the stope wall, joint orientation
ore can be expected to flow). Historically, conventional narrow and gravity, respectively. Q’ is the Q classification index value
vein mining has been associated with high operating costs and low (Barton, Lien and Lunde, 1974), with the stress reduction factor
capital costs (Brewis, 1995; Paraszczak, 1992; Robertson, 1990). (SRF) and joint water reduction factor (Jw) set to one.
Over the past 20 years there has been a general trend away from
conventional small-scale narrow vein mining methods towards N or N’ = Q’ × A × B × C (1)
mechanised mining methods. Longhole stoping is the dominant
narrow vein mining method in both Australia and Canada. While
longhole stoping has lower mining costs per tonne and higher
production rates than conventional mining methods, longhole
stoping has been associated with increased dilution. Therefore,
mining method selection is relevant to minimising narrow vein
dilution.
The economic performance of narrow vein mines is particularly
sensitive to depth of overbreak or slough. Using typical narrow
vein dilution costs of $25/tonne; mining direct operating costs
(mucking and trucking $7/tonne) and milling ($18/tonne), and
assuming 500 000 tonnes per annum ore production, Figure 1
demonstrates how the cost of 25 cm of dilution increases as vein
width decreases. Based on this simplified economic model, a mine
with a 25 cm average vein width incurring on average 25 cm more
dilution than expected, would incur more than $6.25 million per
annum in unforeseen operating costs. Furthermore, this model
does not take into account opportunity costs associated with
capital utilisation.

1. MAusIMM, Senior Rock Mechanics Engineer, Golder Associates,

611 Coronation Drive, Toowong Qld 4066.
Email: penny_stewart@golder.com
2. Strata Engineering, PO Box 724, Charlestown NSW 2290. FIG 2 - Modified stability chart (after Potvin, 1988).

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 153

P C STEWART and R TRUEMAN
FIG 3 - The extended Mathews stability graph (after Mawdesley, Trueman and Whiten, 2001).
FIG 4 - Empirical dilution design graph after (Clark and
Pakalnis, 1997).
The Mathews stability graph (Mathews et al, 1981), the
modified stability chart (Potvin, 1988) and the extended
Mathews stability chart (Trueman et al, 2000; Mawdesley,
Trueman and Whiten, 2001) databases contain fewer than nine
narrow vein stopes. For this reason, there was some concern
about the applicability of these charts to narrow vein
underground stopes. This paper discusses how stability chart can
be applied for narrow vein stope design, while proposing a new
method for dilution prediction and minimisation (the NVD
method). The NVD method is based on:
154 Ballarat, Vic, 14 - 15 October 2008
• analysis of hanging wall and footwall overbreak distributions
from a typical longhole narrow vein gold mine (Barkers
mine, Western Australia); and
• addressing the causes of dilution as identified by analysis
of over 500 narrow vein case studies, and an additional
640 relevant non-narrow vein case studies.
APPLICABILITY OF STABILITY CHARTS FOR
NARROW VEIN MINES
A study (Stewart, 2005) of 525 narrow vein case studies from the
narrow vein Barkers mine and 146 relatively narrow (85 per cent
between four to 12 m wide) case studies Trout Lake and Callinan
mines (Wang et al, 2002) found that dilution for stopes plotting
in the stable zone of stability charts is insensitive to N or N’ and
HR. And, that the relationship between narrow vein dilution and
stability chart parameters N or N’ and HR is best described by a
logistic relationship. Figure 5 illustrates the difference between a
logistic regression model and a linear regression model for
a single independent variable X. The S-shaped logistical
distribution reflects that the dependent variable is relatively
insensitive (flat section of S curve) to model parameters for a
range of values before becoming highly sensitive to model
parameters (steep section of S curve) over a small range of model
parameters, and then reverting to being relatively insensitive
again (flat section of S curve).
Stability charts have two independent variables, N or N’ and
HR. Therefore, if the probabilities of the stability chart were
similarly plotted, the graph would be three-dimensional, with
probability of stability plotted as a surface. The insensitivity of
dilution to N or N’ and HR in the stable zone can be related to the
flat section at the top of the S shaped curve shown in Figure 5.
As shown in Figures 6 - 9, Barkers stopes plotting in the stable
zone demonstrate a poor correlation between dilution (in this
Narrow Vein Mining Conference
 STRATEGIES FOR MINIMISING AND PREDICTING DILUTION IN NARROW VEIN MINES – THE NARROW VEIN DILUTION METHOD

3.00

2.50

2.00

Corrected Overbreak (m)

1.50

1.00

0.50

0.00
0 2 4 6 8 10
-0.50

-1.00
HR

FIG 5 - Comparison of linear and logistic regression models.

FIG 8 - Scatter plot showing the poor correlation between HR and
overbreak for 115 Barkers case studies mined before October
2000 (6040 mL to 6120 mL). These case studies all plot in the
3.00 stable zone of stability charts.
2.50

3
2.00
Corrected OB (m)

1.50 2.5

1.00 2
Corrected Overbreak (m)

0.50 1.5

0.00
1
0 1 2 3 4 5 6 7 8
-0.50
0.5
-1.00
0
N
0 5 10 15 20 25 30 35 40
-0.5
FIG 6 - Scatter plot showing the poor correlation between N and
overbreak for 115 Barkers case studies mined before October -1

2000 (6040 mL to 6120 mL). These case studies all plot in the HR
stable zone of stability charts.

3
3.00
2.5
2.50
2
Corrected Overbreak (m)

2.00
Corrected Overbreak (m)

1.5
1.50
1
1.00
0.5
0.50
0
0.00 0 2 4 6 8 10
0 5 10 15 20 25 30 35 40 -0.5
-0.50
-1
-1.00 HR
N

FIG 7 - Scatter plot showing the poor correlation between N and
overbreak for 410 Barkers case studies mined between October
2000 and February 2003 (6120 mL to 6040 mL). These case
studies all plot in the stable zone of stability charts.
FIG 9 - Scatter plots (the second scatter plot only shows case
studies with HR <10) showing the poor correlation between HR
and overbreak for 410 Barkers case studies mined between
October 2000 and February 2003 (6120 mL to 6040 mL). These
case studies all plot in the stable zone of stability charts.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 155

P C STEWART and R TRUEMAN

case corrected overbreak) and stability charts parameters (N or Potvin’s (1988) study of large open stoping indicated that in
N’ and HR). Similarly, Figure 10 plots the difference between most cases, blast induced dilution could not be isolated as a cause
actual and predicted metres of hanging wall ELOS for Trout of instability. However, within the context of large open stopes,
Lake mine and Callinan mine case studies. Equivalent linear 0.5 to 1.0 m blast related overbreak would probably not be
overbreak/slough (ELOS) is one method of quantifying dilution considered unstable. However, 0.5 to 1.0 m dilution would have a
and is estimated as shown in Figure 11. Over 50 per cent of the significant economic effect on narrow vein mines. Furthermore, it
case studies had more than 0.5 m more dilution than predicted. has been argued that because blast damage is rock mass quality
These case studies provide further evidence of poor correlation dependent (low Q’ stopes would have higher levels of blast
of dilution (in this case ELOS) to stability charts parameters. damage than high Q’ stopes), blast damage is indirectly taken into
account by the inclusion of Q’ in existing stability charts
10 0
(Mathews et al, 1981; Potvin, 1988; Mawdesley, Trueman and
E L OS |act-p red|<0.5
Whiten, 2001). This could explain why, even though blasting
parameters are not included in most stability charts, they generally
.

E L OS 0 .5<|a ct-
0.5 m pre d|<1 .0 provide a predictive accuracy of approximately 80 per cent. That
Modified Stability Number N'

E L OS 1 .0<|a ct- is, 80 per cent of stope case studies plot in the correct stability
1.0 m pre d|<2 .0
chart zone (Mawdesley, 2002). Statistical analysis of 115 Barkers
10
E L OS |act-p red|>2.0 mine case studies found that blast pattern had a statistically
2.0 m
significant effect on overbreak (Stewart, 2005). An ‘in-line’ three
blasthole pattern performed significantly better than both the
E LO S |ac t.-pred.| < 0.5m 70 ‘staggered’ and ‘dice five’ patterns. These findings further justify
c as es
the need for explicit consideration of blast overbreak in the case of
E LO S |ac t.-pred.|> 0.5m
1 72 c as es narrow vein mining.
2 4 6 8 10 12
H R (m )
Feasibility studies require realistic estimates of dilution.
Benchmarking is a useful method for estimating mining
parameters where for reasons of complexity no analytical or
FIG 10 - Trout Lake and Callinan narrow stope case studies numerical solutions exist.
showing difference between actual and predicted dilution
(after Wang et al, 2002).
BENCHMARKING NARROW VEIN DILUTION
Narrow vein dilution has been benchmarked using probabilistic
analysis of 115 case studies from the now closed Barkers mine,
at Kundana Gold Operations (Stewart, Trueman and Lyman,
2008). These benchmarks are only applicable to geotechnically
stable stopes. If a stope wall plots in the unstable zone on a
stability chart then these benchmarks are not applicable. The aim
of the benchmarks is to predict dilution associated with using
longhole stoping for narrow vein deposits and should not be used
without reference to the Barkers mine geological conditions
described in this paper. For example, the Barkers case studies are
unaffected by faulting or pinching and swelling of the vein.
Dilution due to these factors is not taken into account by the
benchmarks, and would best be estimated by wireframing a
realistic ore volume similar to what might be done for reserve
estimation.
Each case study represents either a stope hanging wall or stope
footwall. Ideally, benchmarking includes data from a large
FIG 11 - Equivalent linear overbreak/slough (after Clark and number of sites and conditions. While several hundred additional
Pakalnis, 1997). case studies were sourced from at least three other mines, these
case studies could not be included in the database either because
The implications of this study for narrow vein dilution of biases in the database or because of incomplete or unreliable
prediction and minimisation can be summarised as follows: records.
Kundana Gold Operations are located 25 km west-northwest
• While stability charts can be effective for predicting whether of Coolgardie, approximately 595 km east of Perth in Western
a stope will be geotechnically stable or unstable, they are Australia as shown in Figure 12. The 115 case studies analysed
unreliable for predicting the amount of dilution in narrow come from panels located between the 6120 m sill drive and the
stopes. 6040 m sill drive. Site geologists recorded stope stability data
• The stability chart relationship between N and HR including stope geometry and estimation of overbreak from the
empirically demonstrates that geotechnical parameters N or vein to the final stope walls in stope record sheets. Overbreak
N’ are stope size dependent. In contrast, dilution associated from vein to final stope wall was measured using a hand-held
with drill and blast parameters such as drill pattern, drill hole laser distance measurement device.
deviation, explosive type and confinement do not appear to Barkers mine is considered an excellent benchmark mine
be stope size dependent. because dilution at the mine is relatively unaffected by rock mass
conditions (Figure 6 and Figure 7), and within the limits of the
BLAST OVERBREAK database dilution can be considered independent of stope span
(Figure 8 and Figure 9). While stress damage had a significant
Dilution related to blasting parameters such as blast design, drill effect on deeper stopes at Barkers mine (Stewart, Slade and
hole deviation and quality control have been termed ‘blast Trueman, 2005), the shallower Barkers database used for
overbreak’. Under this definition blast overbreak is quite literally determining benchmark stoping widths is highly unlikely to be
breakage of the rock over the limits of the design. affected by stress damage.

156 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 STRATEGIES FOR MINIMISING AND PREDICTING DILUTION IN NARROW VEIN MINES – THE NARROW VEIN DILUTION METHOD

Zuleika Shear

Meekatharra W iluna Jundee

Big Bell
Bronzewing
MENZIES
Mt Magnet

Leonora

Paddington BROAD ARROW

KUNDANA
Kalgoorlie
Southern
Cross St Ives Kundana
Perth Gold Mine KALGOORLIE
Norseman

Yilgarn Craton COOLGARDIE

N
Granite
100 Greenstone 0 25km

FIG 12 - Location of Barker mine, Kundana Gold Operations (after Slade, 2004).

The average vein width in the first Barkers database was 0.3 m
(0.2 m to 0.4 m). Rock mass classification based on scanline
mapping of sill drives indicates that the rock mass for the first
Barkers database ranges from fair to good (Brunton and
Trueman, 2001). It is reasonable to expect that mines with poor
rock mass conditions would experience higher levels of blast
overbreak than that incurred at Barkers. Further case studies
would be required to establish the effect of a poor rock mass on
blast overbreak.
As illustrated in Figure 13 the Barkers mining method is a
combination of the bottom-up Modified Avoca method using
development waste as fill (tight filling) and longhole open
stoping with small rib pillars.
The Barkers drill and blast patterns are typical of narrow vein
longhole stoping practices. The benchmark narrow vein stoping
widths discussed in this paper are based on the blast patterns shown
in Figure 14. All holes were drilled at 64 mm diameter using an
Atlas Copco H157 longhole rig, and were less than 15 m long.
All holes were blow-loaded up-holes. Hanging wall holes were
generally loaded with a low density ANFO (Sanfold 50).
However, when blasting against fill, the first two holes were
loaded with 100 per cent ANFO. Footwall and in-line drill holes
were loaded with ANFO. All holes were initiated with long delay
non-electric detonators. Holes were double primed with one
booster 2.5 m from the toe and one booster halfway between the
collar and the first booster.
Benchmark stoping widths
Stewart, Trueman and Lyman (2008) developed a probability
based benchmarking method to estimate benchmark stoping
widths for staggered, dice five and in-line longhole drill patterns.
The stability graph accuracy of 80 per cent is considered
reasonable and implies that stopes are generally designed
assuming a 20 per cent probability of overbreak or failure
significantly exceeding design. On this basis, it was assumed that
benchmark stability stoping width should include 80 per cent of
case studies. The benchmark stability stoping width (BSW) is
then defined as:

BSW0.3 = vein width + ( x fw + x hw )80 (2)

where:
( x fw + x hw )80 is the total overbreak function and was determined
by mathematically combining the fitted hanging
wall and footwall distributions (Stewart, Trueman
and Lyman, 2008)
Eighty per cent of stopes are expected to have total overbreak
FIG 13 - Schematic of Barkers mining method (long section). less than ( × fw + × hw )80 .

design design design design design design

hangingwall footwall hangingwall footwall hangingwall footwall

Vein
Vein
0.2m Vein Vein

0.95
0.
0.5m 0.
0.5m 0.9 m
0.4

Staggered
Staggered/Zig-zag Dice5 In-line 3

FIG 14 - Blast pattern and stope design outline without consideration of practical stoping width considerations.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 157

P C STEWART and R TRUEMAN

The ‘benchmark stability stoping width’ for each pattern
defines realistic planned dilution limits. These limits provide the
basis from which true unplanned dilution can be assessed.
Table 1 contains the ‘benchmark stability stoping width’ for each
blasting pattern.
In addition, the probabilistic overbreak model was used to
derive benchmark average stope widths for each of the patterns.
Benchmark average stoping width, in conjunction with vein or
ore width, can be used to estimate total dilution (planned and
unplanned). These were calculated using by adding the average
total overbreak to the average vein width of 0.3 m. Table 1 of the
NVD method lists benchmark average stoping widths and
provides details on how benchmark average dilution can be used
to estimated dilution.
Stewart, Trueman and Lyman (2008) assumed that benchmark
stoping widths are not site-specific and can be considered
operating condition specific. In this case, operating condition
specific refers to standard narrow vein longhole mining methods
and equipment such as those used at the Barkers mine. In other
words, benchmark stoping widths are primarily a function of the
longhole stoping method, and not the geotechnical parameters
that are responsible for unplanned dilution. Based on this
assumption, the benchmark stoping widths determined for the
Barkers mine are applicable to similar narrow vein longhole
stoping mines. Because vein widths ranged from 0.2 m to 0.4 m
the accuracy of both benchmark stope widths can be considered
to be ±0.1 m. Benchmark average stoping width and benchmark
stability stoping width form an important component of the NVD
method.
Unfortunately, despite concerted efforts over a number of
years to source data from additional narrow vein mines it has not
been possible to validate these findings at another mine.

TABLE 1
Narrow vein dilution (NVD) method: prefeasibility/feasibility stage.

A1. Geotechnical stability

A1.1 Stability graph assessment1.
A1.2 Stress relaxation potential assessed for tabular stope geometries; especially if principal stress is perpendicular to strike. In cases of full and
tangential stress relaxation set stress factor A = 0.7.
A1.3 If geometry is irregular or complex (eg pillars or multiple lift retreat) then hydraulic radius can be estimated from radius factor2.
A2. Fill requirements
A2.1 Tight filling3. Treat fill as solid rock.
A2.2 Continuous filling with lag between stoping and fill4.
HR = HRafter blasting + HR after backfilling.
A3. Cablebolting
A3.1 Assess if stope is supportable using modified stability chart1.
A3.2 Stable HR can be increased by pattern cablebolting. Modified stability chart can be used to assess possible increase HR achievable with
pattern cablebolting5.
A3.3 Consider effect of backfilling on production cycle. Cablebolts have potential to improve productivity by decreasing delays associated with
regular filling requirements6.
A3.4 Cablebolts can be used for areas with localised instability potential.
A4. Dilution prediction
A4.1 Total dilution for stopes plotting in the stable zone
Benchmark total dilution for stable stopes can be estimated as follows:
Total dilution = (Benchmark average stoping width10 – vein width)/vein width.
The benchmark average stoping widths for common narrow vein blast patterns10:
• in-line = 1.3 m,
• staggered = 1.5 m, and
• dice-5 = 1.7 m.
For vein widths >1.2 m, an allowance of approximately 0.6 m of dilution is required to ensure the probability of ore loss is less than five per
cent (from cumulative overbreak probability distributions). This estimate is based on the average dilution allowance required for a 0.7 m vein
using an in-line pattern, ie 1.3 m - 0.7 m requires a 0.6 m dilution allowance to ensure ore loss probability is less than five per cent. It is
reasonable to expect that a similar allowance would be required for wider veins.
A4.2 Dilution assessment for stopes plotting in the unstable or failure zone of a stability chart
Isoprobability contours can be used to estimate the percentage of stopes expected to exceed the benchmark stability stoping width 7 8.
eg The probability of unplanned dilution for a stope plotting on the stable-failure boundary is approximately, 20 per cent12.
Benchmark stability stoping widths11.
Vein width < 0.7 µ12 using an in-line pattern 1.6 m
Vein width >0.7 m12 using either staggered or dice-5 pattern 2.1 m
A4.3 Undercutting of stope walls
For stope walls with N’ or N <5 empirical evidence suggests that undercutting will lead to stope wall failure (Potvin, 1988).
For stope walls with N’ or N >5 the HR of undercut stope walls can be estimated by assuming that the stope height is infinite13.
A4.4 Extraction sequence
Evaluate potential for pre-stoping stress history to cause dilution. Possible methods include: pillar stability chart7,
σn > 0.8 × UCS8 and σ1 - σ3 > 0.39
If stress damage potential is high then increase dilution by between 0.1 to 0.3 m14.
Layered orebodies should be assessed for stress relaxation potential.

Note: see Table 3, limitations and assumptions, which contains notes as per superscript numbers.

158 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 STRATEGIES FOR MINIMISING AND PREDICTING DILUTION IN NARROW VEIN MINES – THE NARROW VEIN DILUTION METHOD

THE NARROW VEIN DILUTION METHOD
The narrow vein dilution method (NVD method) is a tool for
more accurately predicting longhole narrow vein dilution. In
addition to dilution prediction, the NVD method also includes
strategies for narrow vein dilution minimisation generally,
including; filling, cablebolting, stress relaxation, extraction
sequencing and blast overbreak. The premise for the NVD
method is that dilution associated with the longhole stoping
mining method can predicted using the benchmark stoping
widths discussed in this paper. The NVD method remains to be
validated at other sites. However, NVD method represents a
potential improvement over the relatively arbitrary stope widths
commonly used to predict dilution.
In much of the literature to date dealing with the subject of
dilution and stope stability, the terms overbreak and stope
stability are represented as being intrinsically related. However,
the case studies presented in this paper suggest an alternate
interpretation, and that the parameters causing narrow vein
dilution can be separated into two groups:
1. geotechnical instability (stope size dependent), and
2. blast overbreak (independent of stope size).
This finding is very important because it offers a simple way
to assess whether an operating mine’s dilution is likely to be
related to geotechnical stability or blast overbreak. Blast
overbreak can be distinguished from geotechnical causes of
dilution by analysing whether dilution occurs independently of
stope size.
The NVD method addresses geotechnical and blasting
overbreak related stability separately. In addition, the method
addresses issues associated with different types and levels of data
available at two different stages in a mine’s life:
1. prefeasibility/feasibility stage, and
2. operating mine.
The NVD method is comprised of three tables. Table 1 contains
the prefeasibility/feasibility NVD method. Table 2 contains the
operating mine dilution minimisation method. Table 3 contains the
assumptions and limitation of the NVD method.
The NVD method also incorporates improvements to the
stability graph method that are particularly relevant to narrow
vein mining, including; use of radius factor for complex
geometry (Milne, Pakalnis and Felferer, 1996) and updated
cablebolt recommendations (Diederichs and Kaiser, 1999).

TABLE 2
Narrow vein dilution (NVD) method: minimising dilution at an operating mine.

B1 Geotechnical stability
B1.1 Determine stable HR for geotechnical domains as per A.1 to A.3
Stable HR predictions should be based on revaluation of Q’ based on underground mapping and observations of rock mass behaviour.
Observe any time dependent rock mass properties.
B1.2 Site-specific chart (if necessary)
If necessary, develop site-specific stability chart using at least 150 case studies, of which at least ten per cent are unstable. A site-specific
chart would be considered necessary if the geotechnical stability is consistently poorly predicted by stability graphs. For example the Mt Isa
bench stability chart takes into account bedding spacing and blasting parameters (Villaescusa, 1996).
B1.3 Kinematic analysis
If wedge and block failure appears to be dominating stability, a more detailed kinematic stability assessment is warranted. Kinematic stability
can be assessed using scan-line mapping results in conjunction with some type of 3D joint network model, eg Stereoblock (Grenon and
Hadjigeorgiou, 2003).
B2 Minimising dilution
There is significant evidence that dilution can be minimised by ensuring that all personnel involved in stoping and ore development are aware
of its importance to a mines economic performance. For example, mine productivity can be measured in terms of weight of metal mined.
B2.1 Stope database
Maintain stope dilution database15 and plot dilution versus stope span. This will provide a good indication of whether dilution is caused by
geotechnical causes or drill and blast causes. Empirical evidence suggests that reducing stope span will only reduce dilution if the causes of
dilution are geotechnical.
If dilution is independent of stope span, then dilution is unlikely to be caused by geotechnical factors. Stress damage is the only geotechnical
cause of dilution that would not be span dependent.
If dilution is span dependent, then dilution is likely to be related to geotechnical factors.
B2.3 Minimising geotechnical causes of dilution
More detailed structural analysis using scanline mapping and stereonets.
Assess relaxation potential.
Cost benefit analysis for cablebolting and/or stope span reduction including fill cycle times.
Ore drive development under good geological control including appropriate drive profile.
Stress damage related dilution can be minimised by evaluation of extraction sequence against damage criterion, increasing number of rings
fired per blast and where practical avoiding shrinking central pillar extraction sequence.
B2.4 Minimising drill and blasting related dilution
Select appropriate blast pattern. An in-line pattern for veins <0.7 m. For vein widths greater than 0.7 m there is a five per cent risk of ore loss
so in this case a staggered or dice-5 pattern should be selected. However, in some ground conditions an in-line pattern may encounter issues
associated with bridging.
Survey drill holes and analyse results in a systematic manner16.
Drill and blast trials should be randomised trials17.
Blast damage minimisation18.
Smoothwall blasting19.

Note: see Table 3, limitations and assumptions, which contains notes as per superscript numbers.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 159

P C STEWART and R TRUEMAN

TABLE 3
Notes: Narrow vein dilution (NVD) limitations and assumptions.

1. Use either the extended Mathews stability graph (Mawdesley et al, 2000) approach or the modified stability chart (Potvin, 1988). The
extended Mathews has most data coverage and N is simpler to evaluate. The modified stability chart is in widest use and has a supportable
zone delineated.
2. Radius factor provides a more accurate assessment of the distance to abutments and is therefore better suited to irregular shape geometry
(Milne, Pakalnis and Felferer, 1996). Radius factor (RF) is half the harmonic radius (Rh) that is calculated using the equation below, where rθ
is the distance to the abutments measured from the surface centre:
Rh 0.5
RF = = n
2 1 1
∑
n θ = 1 rθ

3. Tight filling means that there is no volume between the fill and the next rings/slot to be fired.
4. Continuous filling (as opposed to tight fill) means there is a gap between the fill and the next rings to be fired. Milne, Pakalnis and Felferer
(1996) propose that the effective final span converges to the sum of the opening after blasting plus the opening span after backfilling.
However, due to insufficient data, Milne, Pakalnis and Felferer (1996) were unable to confirm this hypothesis. However, the results of
extensive hanging wall monitoring (extensometers) at the Winston Lake mine did suggest that the common practice of treating a moving
backfill abutment like a rock abutment is overly optimistic (Milne, Pakalnis and Felferer, 1996).
5. Pattern cablebolting recommendations: Diederichs and Hutchison (1996) covers various design methods. Diederichs and Kaiser, (1999)
includes charts for cases of stress relaxation.
6. However, cablebolt effectiveness can be reduced by low rock mass stiffness and stress relaxation (Diederichs and Kaiser, 1999).
7. Brow stress damage would correspond to category three for pillar yield (fracturing in pillar walls) and the unstable region of the Confinement
Formula Stability Graph (Lunde, 1994). The main advantage of this approach over the deviatoric stress approach is the very large database of
rock mass conditions incorporated into the database means better prediction of stress damage when there is no opportunity to calibrate to
underground observations.
8. Further validation required.
9. May have effect on dilution below this criterion damage limit, eg Barkers case studies.
10. It has been assumed that benchmark average stoping width is not site-specific, and can be considered operating condition specific. In this case
operating condition specific refers to standard narrow vein longhole mining methods and equipment. In other words, benchmark stability
stoping width is primarily a function of mining method and equipment, and not the geotechnical parameters that are responsible for
unplanned dilution. Based on this assumption, the benchmark stability stoping widths determined for Barkers mine are applicable to similar
narrow vein longhole stoping mines. It should be noted that rock mass classification based on scanline mapping of sill drives indicates that
the rock mass for the Barkers 1 database ranges from fair to good. It is reasonable to expect that mines with poor rock mass conditions would
experience higher levels of blast overbreak than that incurred at Barkers.
For stopes plotting on the stable-failure boundary the probability of a stope width exceeding the benchmark stability stoping width is
approximately 20 per cent12. This estimate assumes parallel holes and standard longhole drill and blast practices.
Standard practice:
• low impact explosives used on the hanging wall,
• blast pattern suited to the vein width,
• 51 mm to 64 mm diameter holes <15 m long,
• hanging wall hole stand-off distance appropriate to the rock mass,
• some drill hole surveys,
• irregular reporting of dilution and or stope reconciliation,
• some blast trials, and
• ore drive development under geological control.
Implementation of best practices and improved quality control has the potential to achieve narrower stoping widths. In these cases the
probability of a stope width exceeding the stable stoping width would be significantly less than 20 per cent.
Each of the practices listed has been proven to be effective in dilution minimisation at least one mine:
• Regular and systematic drill hole survey and analysis.
• Stope survey and stope reconciliation.
• Reporting dilution as a key performance indicator (KPI).
• Drill and blast continuous improvement projects based on properly design randomised drill and blast trials. Design modifications based on
statistically significant results obtained from stope reconciliations.
• Blast damage modelling based on analysis of PPV.
• Tight geological and mining control on ore-drive development to minimise undercutting.
• Cablebolting of localised instability.
• Smooth wall blasting, eg presplit.
• Revaluation/calibration of feasibility study stable stope dimensions.
Sub-standard practice:
• no drill hole surveys,
• inappropriate blast pattern selected for vein width,
• only reporting tonnes as KPI,
• no evaluation or reporting of dilution, and
• no geotechnical mapping to reassess feasibility study design.

160 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 STRATEGIES FOR MINIMISING AND PREDICTING DILUTION IN NARROW VEIN MINES – THE NARROW VEIN DILUTION METHOD

TABLE 3 cont …
11. Like benchmark average stoping width, it has been assumed that benchmark stability stoping width is not site-specific and is a function of
mining method and equipment. An in-line pattern can be used for vein widths less than 0.7 m. If an in-line pattern is used for a vein width of
0.7 m then the probability of ore loss is approximately five per cent. However, depending on geological and rock mass conditions an in-line
pattern can result in bridging.
12. Inside the stable zone the probability of unplanned dilution decreases from 20 per cent the further a stope plots inside the stable zone and
conversely, inside the failure zone the probability of unplanned dilution increase from 20 per cent the further below the stable-failure
boundary. The extended Mathews isoprobability contours (Mawdesley et al, 2000) can be used to estimate the probability of unplanned
dilution at a particular position on the extended Mathews stability graph. While the probability of failure is affected by distance from the
stability graph stable-failure boundary, the amount of unplanned dilution in narrow vein stopes does not appear to be related to position on the
stability graph.
13. This recommendation is based on the assumption that an undercut stope wall cannot obtain support from arching to the abutments and
therefore, assuming the stope height is infinite is a good approximation of the effect of removing lower abutment support. This
recommendation is based on engineering assessment of the destabilising effect of undercutting and has not been validated with case studies.
However, in the absence of alternatives this approach seems to be reasonable, and is likely to be more realistic than assuming that
undercutting has no effect at all. Comment: in one of your chapters undercut footwalls were related to non-undercut hanging walls.
14. Stress damage related unplanned dilution ranged from 0.1 m to 0.3 m and is based on data from one mine (Barkers mine) and therefore
requires further validation. However, the upper limit of 0.3 m could be a function of mineability of ore drives. Barkers mine ore drives had
high levels of support including shotcrete and cablebolting. Based on these levels of support it is reasonable to infer that mineability was
moderately difficult. If stress related conditions in the ore drives were more difficult than Barkers, then stress damage related dilution could
exceed 0.3 m. However, more case studies are required from other mines to validate this proposition.
15. Ideally, a narrow vein stope dilution database should record the following information;
• stope dimensions,
• undercutting,
• failure mechanism,
• drill and blasting patterns including explosive types and initiation sequence, and
• support of stope walls.
16. Drill hole inaccuracy is a significant cause of dilution (Aplin, 1997; Revey, 1998).
17. Randomised trials mean that changes to drill and blast parameters are undertaken randomly without consideration for any other parameters.
The data collated from randomised drill and blast trials is less likely to be affected by systematic bias. This is the basis of experimental
design.
18. It seems that because the relationship between blast damage and overbreak is rock mass dependent there is no generic model for predicting
overbreak using blast damage modelling. However, individual mines could compare overbreak stability to damage modelling and use this
information in blast design. The value of blast damage modelling is in the ability to compare the blast damage potential of alternative blast
designs. Peak particle velocity (PPV) is a good predictor of blast damage (Singh, 1998; Villaescusa et al, 2003). Blast damage contributes to
dilution. Blast damage modelling involves monitoring blast vibration to determine site-specific constants k andα. Once k and α have been
determined PPV can be estimated as follows:
PPV = K (a)a
where:
a is defined as the Holmberg-Persson term
k and α are the rock mass and explosive specific attenuation constants (Holmberg and Persson, 1980)
Once the PPV limit for blast damage has been determined it is possible to design blast that do not exceed the PPV limit. However, scatter in
delay timings and inconsistencies in rock mass attenuation properties can significantly affect the reliability of actual PPV versus design PPV.
19. Pageau et al (1992) found that by implementing smooth wall blasting, including a presplit, they were able to reduce dilution at the Richmont
mines, Francouer mine Quebec from ten to 15 per cent to five per cent.

Multiple lift retreat stoping and the leaving of rib pillars often
results in complex geometry. The cablebolting recommendations
of Diederichs and Kaiser (1999) are particularly relevant to narrow
vein stoping because they include separate recommendations for
cases of stress relaxation.
Predicting dilution for narrow vein longhole
stoping
For stopes plotting in the ‘stable’ zone of a stability chart, an
estimate of average stoping width for each blast pattern can be
used as the basis for predicting narrow vein dilution. Details are
provided in Table 1. If stopes are designed in the ‘unstable’ zone
of a stability chart, then stoping widths will on average exceed
the average stoping width for a particular blast pattern. Dilution
for stopes plotting in the stable zone can be estimated from vein
width and benchmark average stoping width as follows:
Expected stope width − vein width)
Total dilution =
Vein width
Deposits with more complex geometry and structural
influences are likely to incur more unplanned dilution than a
simple tabular vein with gradual changes in strike and dip. The
Barkers average stoping width is a benchmark from which
planned dilution in more complex geological formations can be
estimated.
Minimising dilution
The NVD method has the potential to improve the process of
selecting the optimum narrow vein mining method, while also
listing ways to minimise dilution in the case where a higher
dilution method is more economic overall. The dilution
prediction method for mines at the feasibility study stage
(Table 1) can be used to assess the appropriateness of the mining
method and determine the likely filling and/or cablebolting
requirements if narrow vein longhole stoping is the preferred
mining method. The dilution prediction methods are limited to
mechanised narrow vein longhole stoping. However, elements of
the NVD method will be applicable to narrow vein mining not
(3)
using longhole stoping.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 161

P C STEWART and R TRUEMAN
In the case of an operating mine the NVD method proposes a
back analysis of existing stope data that enables a mine to first
identify whether dilution is most likely caused by geotechnical
factors or alternatively is blast overbreak related. Table 2 contains
details of this approach. Once it has been established whether the
primary cause of dilution is geotechnical or blasting overbreak
related, a number of strategies are suggested to minimise dilution.
Assumptions and limitations
It is important to note that the applicability of benchmark stoping
widths is qualified by a comprehensive set of assumptions and
limitations. These assumptions and limitations are listed in
Table 3. For example; deposits with shallow dip, more complex
geometry and structural influences are likely to incur more
dilution than steeply dipping simple tabular veins with gradual
changes in strike and dip.
Benchmark stoping widths are applicable for narrow vein
longhole stoping with standard drill and blast practices. Best
practice or substandard practice would result in lower and higher
levels of dilution, respectively. The NVD method clearly defines
what is meant by ‘standard practice’, ‘best practice’ and
‘substandard practice’.
Benchmark stoping widths are applicable for vein widths up to
1.2 m. The benchmark stability stoping widths determined for
various vein widths are only applicable to narrow vein longhole
open stoping. It has been assumed that benchmark stoping widths
are operating condition-specific and not site-specific. For this
reason, the benchmark stoping widths are generally applicable to
sites with similar operating conditions and geological formation.
The vein widths used to determine the benchmark stoping
widths ranged from approximately 0.2 m to 0.4 m. Because vein
width was not recorded an average of 0.3 m has been applied for
all three blast patterns and this means that benchmark stoping
widths have are associated with a ±0.1 m error. It is important to
note that staggered and in-line patterns were generally used for
the narrower veins widths while dice-five was generally used for
the wider veins. These differences would result in minor biases
with an estimated ±0.1 m error.
The validity of the benchmark stoping widths at mines with
similar operating conditions to Barkers remains to be validated.
That said, given that benchmark stoping widths are largely a
function of operating conditions such as equipment and blast
pattern, it is proposed that benchmark stoping widths can be
considered operating condition specific and not site-specific.
Irregular geology will impact on expected dilution and
adjustments to expected dilution should be made accordingly.
CONCLUSIONS
Barkers, Callinan and Trout Lake mine case studies suggest that
overbreak is not continuously related to N and HR, and that the
relationship between HR and overbreak, and N and stability is
best described by a logistical relationship. Based on this result it
was concluded that the caused overbreak is unlikely to be related
to geotechnical parameters if a stope plots well inside the stable
zone. And therefore, in this scenario reducing stope size would
not reduce dilution. This interpretation implies that the causes of
narrow vein dilution can be separated into two unrelated causes:
1. geotechnical instability (stope size dependent), and
2. blast overbreak (unrelated to stope size).
Benchmark stoping widths can be used to define realistic
planned dilution limits. These limits provide the basis from
which true unplanned dilution can be assessed. Benchmark
average stoping width, in conjunction with vein or ore width, can
be used to estimate total dilution. Benchmark stoping widths are
primarily a function of the longhole stoping method (operating-
162 Ballarat, Vic, 14 - 15 October 2008
condition specific). On this basis, the benchmark stoping widths
determined for Barkers mine are applicable to narrow vein
longhole stoping with similar operating conditions and remain to
be validated at another mine. It is recognised that complex
geology (eg cross-cutting structures) will require adjustment to
both the benchmark stability stoping width and the benchmark
average stoping width.
The NVD method is a tool for predicting narrow vein dilution
based on the benchmark average stoping widths. In addition to
dilution prediction, the NVD method also includes
recommendations and strategies for narrow vein dilution
minimisation generally, including; filling, cablebolting, stress
relaxation and managing stress damage and blast overbreak. It is
envisaged that improved dilution prediction will lead to more
accurate comparisons of the expected cost of dilution in longhole
stopes compared to other mining methods while recognising that
optimal mining method selection will, in some instances, accept
a higher level of dilution if it results in a higher overall NPV for
an operation. While the NVD method remains to be validated at
other sites, the method represents a potential improvement over
the relatively arbitrary stope widths commonly used to predict
dilution.
ACKNOWLEDGEMENTS
The work presented in this paper would not have been possible
without the financial support of the AMIRA BART II sponsors
initially and finally the ongoing support of the Julius Kruttschnitt
Mineral Research Centre.
The close collaboration of site personnel from the Barkers
mine at Kundana Mining Operations made it possible to collate
a highly valuable narrow vein database. Numerous other
operations provided data for this project, including Mount Isa
Lead mine, Kanowna Belle and several of the Kambalda nickel
mines. However, for various reasons relating to dependencies
between variables, this data could not be used for this project.
Despite the data not being used, the authors would like to express
their appreciation for the support and collaboration offered by
personnel at these operations.
REFERENCES
Bock, I, 1996. Selective blast mining in gold mines, Journal of the
Southern African Institute of Mining and Metallurgy, pp 183-254.
Bock, I, Jagger, L and Robinson, R E, 1998. An economic model for gold
and platinum mining using selective blast mining, Journal of the
Southern African Institute of Mining and Metallurgy, 98(2):75-84.
Brunton, I and Trueman, R, 2001. Rock mass characterisation at Barkers
underground operation, KKMRC/AMIRA/BART II project report,
pp 1-14 (Julius Kruttschnitt Mineral Research Centre: Brisbane).
Clark, L and Pakalnis, R, 1997. An empirical design approach for
estimating unplanned dilution from stope hangingwalls and
footwalls, in Proceedings 99th Canadian Institute of Mining and
Metallurgy (CIM) Annual General Meeting, Calgary (Canadian
Institute of Mining and Metallurgy: Montreal).
Diederichs, M S and Kaiser, P K, 1999. Tensile strength and abutment
relaxation as failure control mechanisms in underground excavations,
Int J Rock Mech Min Sci, 36(1):69-96.
Elbrond, J, 1994. Economic effects of ore losses and rock dilution,
Canadian Institute of Mining and Metallurgy (CIM) Bulletin,
87(978):131-134.
Grenon, M and Hadjigeorgiou, J, 2003. Open stope stability using 3D
joint networks, Rock Mech Rock Engng, 36(3):183-208.
Henning, J G and Mitri, H S, 1999. Examination of hanging-wall stability
in weak rock mass, CIM Bulletin, 92(1032):40-44.
Holmberg, R and Perrson, P A, 1980. Design of tunnel perimeter blasthole
patterns to prevent rock damage, Transactions of the Institution of
Mining and Metallurgy, 89:A37-A40.
Hutchinson, D J and Deiderichs, M S, 1996. Cablebolting in
Underground Mines, 406 p (BiTech Publishers Ltd: Richmond).
Narrow Vein Mining Conference
 STRATEGIES FOR MINIMISING AND PREDICTING DILUTION IN NARROW VEIN MINES – THE NARROW VEIN DILUTION METHOD
Laubscher, D H and Taylor, H W, 1976. The importance of geomechanics
classification of jointed rock masses in mining operations, in
Proceedings Symposium on Exploration for Rock Engineering,
Johannesburg (Springer Wien: New York).
Lunde, J, 1994. Hard rock pillar strength estimation an applied empirical
approach, MSc thesis, Univeristy of British Columbia.
Mathews, K E, Hoek, E, Wyllie, D C and Stewart, S B V, 1981. Prediction
of stable excavation spans for mining at depths below 1000 metres in
hard rock. CANMET Report. DSS Serial No OSQ80-00081, DSS File
No 17SQ (CANMET: Vancouver).
Mawdesley, C, 2002. Predicting rock mass cavability in block caving
mines, PhD thesis, Julius Kruttschnitt Mineral Research Centre, The
University of Queensland, Brisbane, Australia.
Mawdesley, C, Trueman, R and Whiten, W, 2001. Extending the Mathews
stability graph for open-stope design, Transactions of the Institutions
of Mining and Metallurgy, Mining Technology, 110:A27-A39.
Milne, D, Pakalnis, R C and Felferer, M, 1996. Surface geometry
assessment for open stope design, in Proceedings Second North
American Rock Mechanics Symposium (Balkema: Rotterdam).
Pageau, C, LeJuge, G E and McKenzie, C K, 1992. Mining a narrow ore
body – innovative mining and blasting methods, in Proceedings
AusIMM Annual Conference, pp 53-56 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Pakalnis, R C, 1986. Empirical stope design at Ruttan mine, PhD thesis
(unpublished), Department of Mining and Minerals Processing,
University of British Columbia, Vancouver.
Pakalnis, R C, Poulin, R and Hadjigeorgiou, J, 1995. Quantifying the cost
of dilution in underground mines, Mining Engineering, 47(12):
1136-1141.
Potvin, Y, 1988. Empirical stope design in Canada, PhD thesis
(unpublished), Department of Mining and Minerals Processing,
University of British Columbia, Vancouver.
Revey, G F, 1998. Effects and control of overbreak in underground
mining, Mining Engineering, 50(8):63-67.
Scoble, M J and Moss, A, 1994. Dilution in underground mining:
Implications for production management, Mineral Resource
Evaluation II: Methods and Case Histories, Geological Society
Special Publication, 79:95-108 (Geological Society Publishing House:
London).
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
Singh, S P, 1998. Effects of rock mass characteristics on blasthole
deviation, CIM Bulletin, 91(1016):90-95.
Slade, J, 2004. Seismic characteristics of faults at the Kundana Gold
mine, Eastern Goldfields, Western Australia, Department of Mining
Engineering and Mining Surveying, Curtin University of Technology,
Western Australian School of Mines, Kalgoorlie, p 60.
Stewart, P, 2005. Minimising dilution in narrow-vein mines, PhD thesis
(unpublished), The University of Queensland, Brisbane.
Stewart, P, Slade, J and Trueman, R, 2005. The effect of stress damage on
dilution in narrow vein mines, Ninth Underground Operators’
Conference, pp 211-224 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Stewart, P and Trueman, R, 2004. Quantifying the effect of stress
relaxation on excavation stability, Transactions of the Institutions of
Mining and Metallurgy, Mining Technology, 113:A107.
Stewart, P, Trueman, R and Lyman, G, in press. Development of
benchmark stoping widths for longhole narrow-vein stoping,
Transactions of the Institutions of Mining and Metallurgy, Mining
Technology.
Trueman, R, Mikula, P, Mawdesley, C and Harries, N, 2000. Experience
in Australia with the application of the Mathew’s method for open
stope design, CIM Bulletin, 93(1036):162-167.
Villaescusa, E, 1996. Excavation design for bench stoping at Mt Isa mine,
Queensland, Australia, Transactions of the Institutions of Mining and
Metallurgy, Mining Technology, A1-A10.
Villaescusa, E, 1998. Geotechnical design for dilution control in
underground mining, in Proceedings Seventh International Symposium
on Mine Planning and Equipment, pp 141-149 (Balkema, Calgary).
Villaescusa, E, Onederra, I and Scott, C, 2003. Blast induced damage and
dynamic behaviour of hangingwalls in bench stoping, FRAGBLAST
International Journal for Blasting and Fragmentation, 8(1):23-40.
Wang, J, Milne, D, Yao, M and Allen, G, 2002. Factors influencing open
stope dilution at Hudson Bay mining and Smelting, in Proceedings
Fifth North American Rock Mechanics Symposium, Toronto.
Wright, E A, 1983. Dilution and mining recovery – Review of the
fundamentals, Erzmetall, 31(1).
163
The Use of Geotechnical Instrumentation to Optimise an
Engineered Mine Design at Beaconsfield Gold Mine, Tasmania
A R Penney1, P B Hills2 and R J Walton3
ABSTRACT
Geotechnical instrumentation has been implemented as a key step for
monitoring and optimising an engineered mine design for managing
seismicity at the Beaconsfield Gold Mine. Measurements of
displacement, stress change, ground control element load and seismic
response of the rock mass allow for back analysis and calibration of
critical parameters for feedback into the engineering design loop.
Instrument clusters are installed in strategic locations prior to the
commencement of stoping in order to determine the nature of the
response of the rock mass to stoping. A selection of instruments including
instrumented cable bolts and rod extensometers, resistance wire
extensometers, hollow inclusion stress cells, vibrating wire stress meters
and resistance wire extensometers are used to gather data through a data
logging system. The rate at which data is recorded is adjusted to provide
more information during periods of anticipated rapid change such as
around stope firings or during periods of high seismic activity, and less
information during periods of anticipated quiescence.
The twofold purpose of the instrumentation is to calibrate the
numerical modelling output which forms the basis of the engineered mine
design, and to provide an alert mechanism where the behaviour of the
rock mass is not what was anticipated, or where the impact of that
behaviour may compromise the integrity of the engineered ground control
system.
The paper describes the location of instrument clusters with respect to
stoping, the type of instruments employed and the method of data
acquisition. It provides examples of results obtained and illustrates the
way in which that information is used to optimise the mine design and
manage seismicity.
BACKGROUND
The Beaconsfield Gold Mine in Northern Tasmania began to
exhibit mining induced seismicity during sill driving of the
orebody at a depth of 760 m below surface in 2002. The
increasing occurrence and degree of seismicity over the
following four years saw a number of strategies introduced and
work plans implemented for the purpose of seismic management.
Increasingly, the focus of these strategies and plans was directed
at reducing the risk to personnel, equipment and the mining
operation itself, which was posed by the increasingly seismically
active environment.
Geotechnical instrumentation was adopted as a management
tool from the onset of seismicity, although principally, it was
directed at measuring and monitoring the seismicity itself and
measuring the in situ stress field. Stress change monitoring was
introduced when in situ stress measurements were completed in
2003 and 2006, but principally this was done to monitor far-field
changes over time rather than focusing on local changes and the
immediate impact on the day-to-day mining operation.
1. MAusIMM, Geotechnical Geologist, Allstate Explorations NL,
PO Box 58, Beaconsfield Tas 7270.
Email: AdrianP@allstateexp.com.au
2. FAusIMM, Technical Services Manager, Allstate Explorations NL,
PO Box 58, Beaconsfield Tas 7270.
Email: PeterH@allstateexp.com.au
3. Principal Consultant, Coffey Mining Pty Ltd, 1/21 Howleys Road,
Notting Hill Vic 3168. Email: Rob_Walton@coffey.com
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
A complete review of all aspects of the mining operation at
Beaconsfield followed a well publicised seismically induced
rockfall accident on Anzac Day 2006. At the behest of
Workplace Standards Tasmania, a comprehensive peer reviewed
mine design and safety management process, a ‘Case to Manage
Underground Safety’ (or Case for Safety) was implemented. In
essence, from a purely mining perspective, the Case for Safety
involved geotechnical design from first principles and adoption
of a mining method suitable for the safe operation of the mine in
the prevailing geotechnical environment. In particular, the Case
for Safety led to the development of a remote stoping method in
the west zone of the Beaconsfield Gold Mine to exclude
personnel from areas of greatest seismic risk. Enhanced use of
geotechnical instrumentation was introduced to the operation to
verify assumptions made and monitor progress.
GEOLOGY AND SETTING
The Beaconsfield Gold Mine is centred on the Tasmania Reef, a
steeply-dipping tabular quartz-carbonate vein hosted by a
sequence of siliciclastic sediments. The sediments, which dip
towards the east, comprise the Salisbury Hill Formation of
conglomerates and sandstones overlain by the Eaglehawk Gully
Formation of sandstones, siltstones and limestones with
occasional pebble bands. All siliciclastic rocks have been
metamorphosed to quartzites although they retain much of their
sedimentary character. The Tasmania Reef is essentially planar
over a strike length of 350 - 400 m, with an average width of
2.5 m and occupies a shear which cuts across stratigraphy
orthogonally. The west zone of the Tasmania Reef is hosted by
the Salisbury Hill Formation, and it is the brittle nature of the
conglomerate horizons within that formation in particular, which
are prone to mining induced seismicity. The geology of the mine
is discussed by Hills et al (2001).
GEOTECHNICAL DESIGN
The basis of first principles geotechnical design adopted for the
Beaconsfield Gold mine is generally that outlined in the
Canadian Rockburst Handbook (Kaiser, McCreath and Tannant,
1996). The Case for Safety studies (Pfitzner, 2006; King,
Thomas and Scott, 2007; Scott and Reeves, 2007; Sidea, Scott
and Reeves, 2007) progressively developed the design process
through the increasingly geotechnically complex environments
of decline development, ore driving, stoping in the generally
aseismic east zone and finally, stoping in the west zone. The
process was summarised by Reeves (2008) and peer reviewed in
application by Kaiser (2008). Scott, Penney and Fuller (2008)
examine the application of this process to stoping in the west
zone.
REMOTE STOPING METHOD
A key outcome of the Case for Safety process was the realisation
that seismicity associated with stoping in the west zone of the
Beaconsfield Gold Mine would continue and that the magnitude
of those seismic events was likely to equal that which occurred
on Anzac Day 2006. Further, it was realised that no support
system could be designed to adequately ensure the safety of
personnel working in the western ore drives, should they be
165
A R PENNEY, P B HILLS and R J WALTON

present when such seismic events occurred. Consequently, the

Western Case for Safety (Scott and Reeves, 2007) was premised
on the development of a remote stoping method which totally
excluded the presence of personnel from ore drives in the west
zone where stoping has commenced. The development and
application of the remote stoping method devised, is discussed
separately by Hills et al (2008). With this change, there was a
requirement that geotechnical assumptions and other decisions
behind this move are adequately assessed. These assumptions
were to be continually checked and modified as data is collected,
analysed and further observations are made.

DESCRIPTION OF PROJECT AND

INSTRUMENTATION REQUIREMENTS
The need to evaluate the installed ground control system and the
rock mass around the excavations with a degree of confidence
and reliability as the mining front advances continues to be a
challenge facing all mining operations. Areas of ground control
failure and the failure to understand the rock mass behaviour can
result in development and production delays, equipment damage,
reserve loss, or injuries. Instrumentation including microseismic
systems, stress/strain meters, instrumented support elements and
extensometers are critical components in understanding this
behaviour, and measuring the effectiveness of installed ground
control systems. This in turn allows for optimisation of the
engineering design process through back analysis and ongoing
calibration of engineered designs and has the potential to lead
directly to improved operating costs as well as safer operating
conditions (Bawden et al, 2007). This role for instrumentation is
incumbent in the Case for Safety process adopted at the
Beaconsfield Gold Mine.
Microseismic monitoring and the surrounding issue of seismic
management was discussed in some detail by Hills and Penney
(2008), and are not reiterated. This paper describes the use of ®
FIG 1 - SMART bungee cable bolt configuration.
clusters of instruments to monitor the behaviour of the rock mass
and the installed ground support in response to stoping. It
focuses on the site selection and analysis of the western • 1 m and 2 m resistance wire extensometers (RWE) (Windsor
instrumentation installed in the 940 Stoping Block footwall and Worotnicki, 1986), selected for their high precision and
drives as outlined in the Western Case for Safety. accuracy for the measurement of pillar deformation between
the footwall drive and the stope; and
Description of instruments and loggers used • uniaxial vibrating wire stress meters (Dutta, 1985), selected
All instruments have been selected due to their relative robust to provide data on stress increases perpendicular across the
nature, ease of use and general acceptance across the mining drive backs may also be used but are not essential.
industry. The instruments utilised in these clusters are: Data acquisition from the instrumentation in these clusters is
• CSIRO Hollow Inclusion Stress Measurement Cell achieved using dataTaker® DT85 and DT515 data loggers. Data
(Worotnicki and Walton, 1976), selected for its capability to is transferred from the loggers to either a USB memory stick, or
measure stress change in three dimensions; direct to a laptop computer via USB connection.

• SMART® cable bolts (Hyett et al, 1997), specifically
designed to match the bungee (yielding) cable bolts designed
for use at the mine following the Anzac Day 2006 rockfall
(Scott, Penney and Fuller, 2008) with node points at 0.4 m,
0.8 m, 2.0 m, 4.5 m, 6.7 m and 7.1 m (Figure 1)4;
• multi-point borehole extensometer (MPBX) (Windsor and
Worotnicki, 1986), selected to compliment the SMART®
cable with node points at 1 m intervals over a 6 m length;
®
4. The SMART cable design modified for use at Beaconsfield
incorporates a 1.5 m length of Garford bulbed strand at the toe
anchor position. Three Garford nodes occur in this section of the
cable, and the measurement head takes the place of the barrel and
wedge swage block on the standard Beaconsfield bungee cable. A 4.5
m long section of plain strand encased in pvc tubing provides the
debonded capacity to allow for deformation (particularly due to
dynamic load), and a 1.3 m section of plain strand at the collar to
provides for surface anchorage. Self aligning barrel and wedge
assemblies and domed plates are employed as surface fixtures.
Layout of instrumentation and criteria for site
selection
A typical instrument cluster for the footwall drives is shown in
Figure 2. The majority of the instrument installations, and
therefore observations, are made from the drive backs. This is the
critical area identified from geotechnical mapping, numerical
modelling and empirical methods for static and dynamic failure.
Clusters are located in strategic locations in footwall drives
along the strike of the orebody. In a standard cluster, an HI Cell
is installed at a depth above the backs of around 3 m to measure
stress change and provide a quantitative assessment of observed
stress-induced damage. SMART® cables and MPBXs are installed
into the centre of the drive backs within a metre of the HI Cell
and at a separation from each other of not more than 1.5 m. The
cluster allows the impact of measured stress change to be
correlated against measured deformation of the rock mass and
the corresponding response of the support element. RWEs are
installed into the wall between the FW Drive and reef at an

166 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 THE USE OF GEOTECHNICAL INSTRUMENTATION TO OPTIMISE AN ENGINEERED MINE DESIGN AT BEACONSFIELD GOLD MINE
incline of approximately 30°. At this orientation, the RWE is
perpendicular to the reef footwall, and provides insight into the
behaviour of the minimum 6 m pillar between the reef and drive.
A HQ diamond drill observation hole is drilled to a depth of 7 m
in the backs to allow visual observation by borehole camera of
any developing fractures or deformation, and is located within
1.0 m of the MPBX. The observation hole also allows for
visual observations of the depth of fracture caused by either
stress redistribution or pre-existing discontinuities (Figure 2).
Occasionally, a cluster is augmented by the inclusion of an HI
Cell installed approximately 9 m into the wall of the footwall
drive away from the orebody.
Generally one cluster is installed in each geotechnical domain.
This has the benefit of allowing for determination of all changes
during stope block extraction. Instrument locations are also
selected following the criteria outlined by ISRM suggested
methods (ISRM Commission on Standardization of Laboratory
and Field Tests, 1978):
• instruments installed in an area where expected hazard zones
exist (hazards from stress, geotechnical domain, seismicity, etc);
• instruments installed in areas to provide the best possible
coverage of all stages of extraction of the stoping block;
• instruments installed in positions that can be re-accessed to
make any repairs to damaged cables (when it occurs), and
that allow visual inspections to be made to validate visual/
measured responses and calibrate all other damage mapping;
• loggers located in a secure area and not placed in zones
where access exclusions are expected to occur; and
• instruments installed into a sound rock mass, and not in fault
zones, or where dynamic water conditions exist.
The location of instrument clusters in the 940 and 980 stoping
blocks are relative to the stoping panels in longitudinal
projection in Figure 3 (note all clusters are located in the
footwall drives 6 m behind the stope projections).
Procedure for data collection
Data acquisition is undertaken by the dataTaker® data loggers,
which are programmed to read and store all information obtained
from the instruments. The data loggers and power supplies are
FIG 2 - Layout of instruments in 930 West illustrating pre- and post-stoping damage profile.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
housed in steel enclosures. For some of the installed instruments,
steel terminal boxes are used to house intermediate connections
of the instruments to the data loggers.
Normally, the data loggers are set to six hour or 12 hour scan
rates. The data loggers are set to take readings at five minute
intervals a minimum of 30 minutes before stope firings. This five
minute scan rate is maintained for the length of the re-entry
period that is imposed to the level (generally 24 hour exclusion).
At the end of this exclusion period, all data is interpreted before
entry to the level is granted. The instrument results are compared
to the Omori analysis results from the seismic data for the same
period to ensure all key trends are not exceeding the required
limits. Any time-dependent changes (ie slow stress redistribution,
ongoing load increase, displacement from rock mass creep, etc)
can be assessed before any persons enter these areas.
Once preliminary assessments are completed and entry to the
level is granted, loggers are reset to 12 hour scan rates. This scan
rate can be changed to six hour scan rates if the stope area is
within known high seismic hazard areas. Loggers will remain in
six hour scan rates until the seismicity returns to acceptable or
background levels.
The data loggers acquire all the data in millivolts and only
display the last recorded values, therefore it is not possible to
determine whether any changes have been recorded by the
instruments at the data logger itself. All recorded data must be
converted into the appropriate units to allow analysis to be
undertaken. Conversion factors from the millivolt record to the
appropriate units for the various instruments installed in the
clusters are shown in Table 1.
OBSERVED AND MEASURED BEHAVIOUR
(DATA ANALYSIS)
The 930 West instrument cluster in the 940 West Stope Block has
been selected for discussion in this section, and its behaviour is
tracked through firing of the first 11 stoping panels since its
installation in December 2007. All other instrument clusters in
the 940 West Stope Block exhibit similar responses and trends to
the 930 West instrument cluster, and are not discussed further.
The first ten stopes were fired in the 940 West Stope Block.
167
A R PENNEY, P B HILLS and R J WALTON

TABLE 1
Data conversion factors from mV (recorded) to engineering units for instruments in use at Beaconsfield Gold Mine.

Conversion factor Units Notes

(4 × ∆mV) / Vinput × GF) Vinput = 1.99 for system used

CSIRO HI cell Mε (micro strain)
GF = 2.097 for cell and cable type used
Resistance wire (RWE1n - RWEDatum)0.8 RWEDatum is the first record in mV
mm (displacement)
extensometer RWE1n is any subsequent record in mV
N1Datum is the first record at node 1 in mV
(N1n - N1Datum) - (N2n - N2Datum) × 0.0415 N1n is any subsequent record at node 1 in mV
SMART MPBX mm (displacement)
N2Datum is the first record at node 2 in mV
N2n is any subsequent record at node 2 in mV
N1Datum is the first record at node 1 in mV
N1n is any subsequent record at node 1 in mV
SMART cable (N1n - N1Datum) - (N2n - N2Datum) × 2.55 × (0.212/D) Tonne (load) N2Datum is the first record at node 2 in mV
N2n is any subsequent record at node 2 in mV
D is the length of cable between node 1 and 2
G = 1 for E of rock
Vibrating wire (R1n - RDatum) × G × 0.00689 ∆σ MPa (pressure) RDatum is the first record
R1n is any subsequent record

Stope 3A was in the 940 East Stope Block was fired almost
concurrently with Stope 3. Stope 11 was the first stope fired in
the 980 West Stope Block in April 2008 and was accompanied
by a ML 1.9 seismic event. The location of all instrument clusters
in the 940 and 980 West Stoping Blocks and of the 11 stopes
fired during the analysis period discussed is illustrated in
Figure 3.
As was the practice throughout the 11 stope firings, each stope
was extracted in a single firing using electronic detonators. Little
change was observed on any of the instruments during the
extraction of stope panels one to four. This was primarily due to
the stope panels being 15 m west of the 930 instrumentation
cluster. Stope 5 was the closest stope firing to the 930 West
instrument cluster, located in the immediate hanging wall of the
instrument cluster, and the first significant change recorded by
the instruments was observed in response to that firing.
Following the extraction of Stope 5, minor stress changes
continue to occur throughout the extraction of the remainder of
the block but no large step changes were recorded.
The most significant change measured was that of the local
stress field as measured by the HI Cell. The use of the HI Cells in
this application does not allow determination of principle stress
directions. Orthogonal stress components σmax σint σmin refer to
the maximum, intermediate and minimum changes in
compressive stress over the measurement period. In the case of
the 930W HI Cell, this amounted to a drop in σint and σmin by
3 MPa and 10 MPa respectively, and a rotation of the stress field
of approximately 50° with associated changes in dip (Figures 4a,
4b and 4c). These changes were in line with analysis of
numerical modelling results undertaken for the design of the
drive using Map3D. This is the key outcome and requirement of
the analysis. With further iterative analysis, these observed stress
changes could be resolved into changes to the principal virgin
stress components, but this has not been done.
It was also with the firing of Stope 5 that the first real changes
were observed on the MPBX and SMART® cable (Figures 5
and 6). Displacements were recorded between nodes at 3 m to
4 m and 4 m to 5 m on the MPBX with the greatest displacement

FIG 3 - Longitudinal projection of the 940 and 980 West Stoping Blocks illustrating the location of instrument clusters and the
stoping extraction sequence.

168 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 THE USE OF GEOTECHNICAL INSTRUMENTATION TO OPTIMISE AN ENGINEERED MINE DESIGN AT BEACONSFIELD GOLD MINE

Stope 3a

Stope 11
Stope 10
Stope 1

Stope 4

Stope 6

Stope 8

Stope 9
Stope 2

Stope 3

Stope 5

Stope 7
A

4.00 Increase in compressive stress

0.00
Stress change (MPa)

∆σmax
-4.00
∆σint
∆σmin
Stope Firing
-8.00

-12.00

Decrease in compressive stress

-16.00
04/12/07 24/12/07 13/01/08 02/02/08 22/02/08 13/03/08 02/04/08
Date

Stope 11
Stope 3a

Stope 10
Stope 3

Stope 4

Stope 6

Stope 7
Stope 5

Stope 8

Stope 9
B
Stope 1

Stope 2

360

315
Bearing of stress change (degree)

270

225
∆σmax
∆σint
180 ∆σmin
Stope Firing
135

90

45

0
4/12/07 24/12/07 13/01/08 2/02/08 22/02/08 13/03/08 2/04/08
Date
Stope 10
Stope 3a

Stope 11
Stope 5

C
Stope 1

Stope 4

Stope 6
Stope 7

Stope 8

Stope 9
Stope 2

Stope 3

90

80

70
Dip of stress change (degree)

60

∆σmax
50 ∆σint
∆σmin
40 Stope Firing

30

20

10

0
4/12/07 24/12/07 13/01/08 2/02/08 22/02/08 13/03/08 2/04/08
Date

FIG 4 - Stress component magnitude change (A), bearing change (B) and dip change (C) measured by the 930 West HI Cell during
the extraction of Stopes 1 to 11.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 169

A R PENNEY, P B HILLS and R J WALTON

Stope 3a

Stope 10

Stope 11
Stope 3

Stope 4

Stope 5

Stope 6
Stope 7

Stope 8

Stope 9
Stope 1

Stope 2
2.0

1.5

6.0m and 5.0m

Displacement (mm)

1.0 5.0m and 4.0m

4.0m and 3.0m
3.0m and 2.0m
0.5
2.0m and 1.0m
Cable Damage
1.0m and head
0.0 Stope Firing

-0.5

-1.0
4/12/07 24/12/07 13/01/08 2/02/08 22/02/08 13/03/08 2/04/08 22/04/08

Date

FIG 5 - Rock mass movement measured by the 930 West MPBX during the extraction of Stopes 1 to 11.
Stope 3a

Stope 10

Stope 11
Stope 7
Stope 1

Stope 2

Stope 4

Stope 5

Stope 8

Stope 9
Stope 6
Stope 3

4.00

3.50

3.00

2.50 0.4m and 0.8m

Change in Load (t)

0.8m and 2.0m

2.00
2.0m and 4.5m
1.50 4.5m and 6.0m
6.0m and 6.4m
1.00 6.4m and 7.1m
Stope Firing
0.50
Cable Damage

Cable Damage

0.00

-0.50

-1.00
4/12/07 24/12/07 13/01/08 2/02/08 22/02/08 13/03/08 2/04/08 22/04/08

Date

®
FIG 6 - Reinforcement load change measured by the 930 West SMART cable during the extraction of Stopes 1 to 11.

occurring between the 4 m and 5 m node points. Load on the when the rock mass is subjected to high ground motions. The
SMART® cable bolt was recorded between the nodes at 2 m to debonded section of cable is from 1.0 m to 5.5 m into the rock
4.5 m and 4.5 m to 6.0 m, with the majority of load being mass and is in the region where the load and displacement
imposed between nodes at 2 m and 4.5 m. occurred. While the use of alternative reinforcing elements such
Average displacement rates for the MPBX following Stope 5 as Garford bulbed cables may help to reduce displacements, the
extraction were: resulting stiffer ground control would tend to fail when subjected
to high ground motions imposed from a large seismic event due
• ~0.5 mm/firing displacement between nodes at 4 m and 5 m, to its low energy absorption capacity. The design implications of
and the bungee cable are discussed further by Scott, Penney and
• ~0.15 mm/firing displacement between nodes at 3 m and 4 m. Fuller (2008).
Average loading rates for the SMART® cable following A key consideration in the results from the MPBX and SMART®
Stope 5 extraction were: cable are the ongoing step change increases of displacement and
load at each subsequent firing with little change of the stress
• ~400 kg/firing on cable for length between 2.0 m to 4.5 m, and field. This was interpreted to be slow rock mass degradation due
• ~100 kg/firing on cable for length between 4.5 m to 6.0 m. to reduced clamping stresses acting on the rock mass, allowing a
The displacement and load changes occurred in the section of reduction in confinement. The rock mass was able to relax and
ground coinciding with the debonded section of the SMART® displace due to the use of debonded bungee cables.
cable. This was expected as the bungee bolts do not form active The results from the RWE started showing some signs of
reinforcement in this zone. Rather, they provide more passive movement following the firing of Stope 2 and reached the
reinforcement, designed to act as energy absorption anchors maximum change following the extraction of Stope 5. There

170 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 THE USE OF GEOTECHNICAL INSTRUMENTATION TO OPTIMISE AN ENGINEERED MINE DESIGN AT BEACONSFIELD GOLD MINE

was only minor creep in the instrument records beyond that There has been little observed change to the rock mass around
point (Figure 7). Unfortunately, a significant amount of data the footwall drives. Damage induced from stope firing, has
was lost due to cable damage, limiting the analysis of these occurred around the blasthole collars in the wall. Up to 1 m of
instruments. In a gross sense, most displacement occurred on damage was observed around some rings with all support in the
the 2 m RWE which was installed at a depth range of 1 m to 3 area being totally destroyed and requiring reinstatement.
m into the nominal 6 m wide pillar between the footwall drive However the damage has been reduced in subsequent firings due
and ore drive. Observations in the area did not reveal any to the introduction of improved stemming techniques.
damage or movement in the footwall drive, so it is most
plausible that the rock mass was relaxing towards the open ALERT MECHANISMS FROM ANALYSIS
stope. The fact that displacements stopped soon after extraction
and subsequent backfilling supports that interpretation. The Communication of the occurrence of stope firings and the
backfill placement will provide a level of confinement expected rock mass behaviour following those firings
minimising any further movement. As an aside to the (ie seismicity, expected areas where rock mass change is likely
observations of RWE, raw strain change data gathered from the upon re-entry, etc) is the most effective alert mechanism for
HI Cell showed an immediate response to the backfilling of managing risk associated with stoping activity at the
Stopes 2 and 5 (Figure 8). That change has not been analysed in Beaconsfield Gold Mine. Advance notification to all site
terms of stress change. personnel and key stakeholders a minimum of one day before a
Stope 3a

Stope 10

Stope 11
Stope 9
Stope 6
Stope 7

Stope 8
Stope 1

Stope 2

Stope 3

Stope 4

Stope 5

6.0

5.0

4.0
Displacement (mm)

RWE 1m
3.0
RWE 2m

2.0 Stope
Firing

1.0

0.0

-1.0
4/12/07 24/12/07 13/01/08 2/02/08 22/02/08 13/03/08 2/04/08 22/04/08
Date

FIG 7 - Rock mass movement measured by the 930 West resistance wire extensometers during the extraction of Stopes 1 to 11.
Stope 3a

Stope 10

Stope 11
Stope 4

Stope 5
Stope 1

Stope 3

Stope 6

Stope 9
Stope 7

Stope 8
Stope 2

Backfill
Backfill

600
REF

A90
400
A45

B45
Micro-strain Change (mV)

200

B135

0 B90

C0
-200
C90

C45
-400
E90

-600 F90

Stope
Firing
-800
7/12/07 27/12/07 16/01/08 5/02/08 25/02/08 16/03/08 5/04/08 25/04/08

Date

FIG 8 - Raw strain change data recorded by the 930 West HI Cell during the extraction of Stopes 1 to 11 and illustrating the response to
backfilling of Stopes 2 and 5.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 171

A R PENNEY, P B HILLS and R J WALTON
planned stope firing is desirable. There is an understanding by all
personnel that what is predicted or expected is not guaranteed
and significant departures from this can and will occur. However,
there is also an understanding within the workforce, that where a
departure from the predicted or expected (beneficial or
otherwise) occurs, feedback will be provided once all data has
been processed and analysed.
Following each stope firing in the west zone of the mine, there
is a minimum 24 hour exclusion period for personnel access to
any footwall drive within that stope block. That period is based
on Omori analysis of seismicity from previous stope blasts.
Seismic flaring is also considered when analysing the seismic
data. The basis for this flaring analysis is:
• Moderate hazard: three times the 30 day average. Entry with
caution for specific required jobs only.
• High hazard: five times the 30 day average. Entry will be
excluded until the seismic flaring rate reduces and causes can
be fully assessed following the seismic management systems
outlined in Hills and Penney (2008).
With the additional benefit of the installed instrumentation,
additional rock mass and time dependant changes are
incorporated into the re-entry analysis. Where rock mass changes
recorded by the instrumentation exceed site guidelines, the
heading re-entry restriction will be extended. Typically, the
restriction will remain in force until the measured changes fall
below the guidelines.
In the event data records indicate a cable has been damaged,
then a geotechnical inspection is carried out before repairs are
made to the cable. Once communications have been restored, the
data is analysed and appropriate actions effected. Typically,
damage to the instrumentation cables will result in a loss of
detail, but the quantum change recoded by the instrument is not
compromised.
Ultimately, it is expected that real time alert mechanisms will
be established, reporting seismic hazard zones and observed
changes from instrumentation in excess of predetermined levels.
CONCLUSIONS
The change in mining method in the west zone of the
Beaconsfield Mine resulting from the Case for Safety review
process requires ongoing monitoring to ensure those areas
required for personnel access remain acceptably safe and
accessible for their entire designed life. Results from the first
stope block extracted under this new method (the 940 West
Block) indicated that low levels of rock mass deformation had
taken place, and that load capacities of the installed ground
control was still within acceptable limits of the design criteria.
Monitoring will continue to occur to ensure that the integrity of
the excavations is maintained. As mining progresses deeper,
continued analysis of instrumentation data may allow a
modification of ground control designs in line with observed
rock mass behaviour.
Time-dependent changes have already been identified and
ongoing analysis of this cause and effect relationship is taking
place. Ultimately, this will result in a more robust re-entry
protocol and increased understanding of the rock mass behaviour
as mining progresses. It will also permit better ground control
designs to be implemented.
Results to date confirm that the current ground control designs
are working well within the design tolerances. Once full
extraction of the 940 West Block is complete, a detailed back
analysis will be undertaken to determine the nature of the
response which the rock mass exhibited to stoping. This in turn
will be fed back into the design procedures to determine the
reliability of the engineered ground control design in future. The
process will be iterative.
172 Ballarat, Vic, 14 - 15 October 2008
ACKNOWLEDGEMENTS
The authors acknowledge the management of Beaconsfield Gold
NL for permission to publish this paper, and the assistance of
Top Rock Technologies and Coffey Mining in the design,
installation and commissioning of these instrument clusters.
REFERENCES
Bawden, W F, Tod, J, Lausch, P and Davison, G, 2007. The use of
geomechanical instrumentation in cost control underground mining,
in Challenges in Deep and High Stress Mining (eds: Y Potvin,
J Hadjigeorgiou and T Stacey), pp 317-325 (Australian Centre for
Geomechanics: Perth).
Dutta, P K, 1985. Some recent developments in vibrating wire rock
mechanics instrumentation, in Proceedings 26th US Symposium on
Rock Mechanics, pp 1043-1053 (South Dakota Schools of Mines and
Technology: Rapid City).
Hills, P B, MacDonald, G, Nero, L and Mueller, A M, 2001. Beaconsfield
Gold Mine – Tasmania Reef, in Structure and Setting of Proterozoic
and Palaeozoic Rocks in the Tamar Region, Northern Tasmania
Excursion Field Guide (ed: A R Reed), 9:76-88 (Geological Society
of Australia Specialist Group in Tectonics and Structural Geology:
Sydney).
Hills, P B, Mills, J, Penney, A R and Arthur, S, 2008. The development
and implementation of a fully remote stoping method at Beaconsfield
Gold Mine, Tasmania, in Proceedings Narrow Vein Mining
Conference, pp 199-206 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Hills, P B and Penney, A R, 2008. Management of seismicity at the
Beaconsfield Gold Mine, Tasmania in Proceedings Tenth AusIMM
Underground Operators’ Conference, pp 157-170 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Hyett, A J, Bawden, W F, Lausch, P, Moosavi, M, Ruest, M and Pahkala,
M, 1997. The S.M.A.R.T. cable bolt: An instrument for the
determination of tension in 7-wire strand cable bolts, in Proceedings
International Symposium on Rock Support – Applied Solutions for
Underground Structures, Lillehammer, June, pp 25-40.
ISRM Commission on Standardization of Laboratory and Field Tests,
1978. Suggested methods for monitoring rock movements using
borehole extensometers, Int J Rock Mech Min Sci and Geomech
Abstr, 15:305-317.
Kaiser, P K, 2008. Geomechanics and ground support review –
Beaconsfield Gold Mine, MIRARCO – Mining Innovation, Sudbury.
Kaiser, P K, McCreath, D R and Tannant, D D, 1996. Canadian
Rockburst Support Handbook, pp 324 (Geomechanics Research
Centre: Sudbury).
King, R, Thomas, S and Scott, C, 2007. Geotechnical assessment –
Resumption of ore production (east zone), Coffey Mining P/L
report no HZ00054.01.BH.
Pfitzner, M, 2006 Geotechnical assessment – Resumption of decline
development, Coffey Mining Pty Ltd report no HZ00054.01.A1.
Reeves, I, 2008. Procedure for support selection, Coffey Mining Pty
Ltd report no HZ00054.01.CC.
Scott, C, Penney, A R and Fuller, P, 2008. Competing factors in support
selection for the west zone of the Beaconsfield Gold Mine, Tasmania,
in Proceedings Narrow Vein Mining Conference, pp 173-178 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Scott, C and Reeves, I, 2007. Geotechnical assessment – Resumption of
ore production (west zone), Coffey Mining Pty Ltd report no
HZ00054.01.BY.
Sidea, D, Scott, C and Reeves, I, 2007. Geotechnical assessment –
Resumption of ore driving development, Coffey Mining Pty Ltd
report no HZ00054.01.AT.
Turner, M H, 2005. Beaconsfield site visit February 2005, AMC
Consultants Pty Ltd, 31 March 2005.
Windsor, C R and Worotnicki, G, 1986. Monitoring reinforced rock mass
performance, in Proceedings International Symposium on Large
Rock Caverns, Helsinki, pp 1087-1098 (Pergamon Press: Oxford).
Worotnicki, G and Walton, R J, 1976. Triaxial ‘hollow inclusion’ gauges
for determination of rock stresses in situ, in Proceedings ISRM
Symposium on Investigation of Stresses in Rock – Advances in
Stress Measurement, national conference publication no 76/4:1-8
(Institute of Engineers Australia: Sydney).
Narrow Vein Mining Conference
Competing Factors in Support Selection for the West Zone of the
Beaconsfield Gold Mine, Tasmania
C Scott1, A R Penney2 and P Fuller3
ABSTRACT
Re-commencement of mining in the Western Zone of Beaconsfield Gold
Mine following a seismically-induced rockfall fatality in April 2006
required a change of mining method involving drilling and blasting from
specially developed drives in the footwall of the orebody. Broken ore is
mucked remotely from ore drives on each level. Support in the footwall
drives is required to maintain stability during their development and as
stopes are retreated, which increases induced stresses and further seismic
events are triggered.
The paper outlines the challenges involved and methods used to select
the footwall drive support to handle both static and dynamic load
demand. The importance of load development in the support system, its
displacement capacity, the compatibility of support elements and energy
absorption capacity of the support system are highlighted.
It is usual in any support selection process to ensure that capacity
exceeds demand to create a factor of safety. In this case the demands on
the system relate to load, displacement and energy absorption and the
support system capacity for each must ensure that appropriate factors of
safety for each are achieved. The paper shows how these factors are not
independent and that selecting a high factor of safety for load can lead to
compatibility problems with support elements in the support system.
INTRODUCTION
Mining methods that have been employed historically at the
Beaconsfield Gold Mine have been top down methods. When
production restarted in 1999, a downward retreating, uphole
benching mining method was continued. A top down mining
method maintains an ever increasing abutment stress ahead of
(below) the current mining area as mining progressively becomes
deeper. This abutment stress led to stress related problems and
seismicity in areas of the abutment.
At Beaconsfield the mining induced abutment stresses have
increased at depth, leading to crushing in areas of the more
brittle quartz ore zone. These abutment stresses may also induce
movement along structures that may produce seismic events as
they move. Seismicity has increased with the largest recorded
event of ML = 2.3 occurring on a structure adjacent and
subparallel to the orebody in April, 2006. Following this event, a
complete re-evaluation of the mining method and support
requirements has been conducted for the entire mine, with a large
focus on the more seismically prone Western Mining Zone
(Pfitzner, 2006; Sidea, Scott and Reeves, 2007; King, Thomas
and Scott, 2007; Scott and Reeves, 2007).
Support requirements have been re-designed based on the
following criteria:
• expected depth of fracture due to stress driven spalling based
on a calibrated model between rock material properties,
modelled induced stresses and observations/measurements
made as stoping has recommenced;
1. Associate Engineer, Coffey Mining Pty Ltd, 2 Melville Street,
Hobart Tas 7000. Email: clint_scott@coffey.com
2. MAusIMM, Geotechnical Geologist, Allstate Explorations NL,
PO Box 58, Beaconsfield Tas 7270.
Email: AdrianP@allstateexp.com.au
3. FAusIMM, Senior Principal – Chief Mine Geotechnical Engineer,
Coffey Mining Pty Ltd, 1/21 Howleys Road, Notting Hill Vic 3168.
Email: peter_fuller@coffey.com
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
• yielding support elements with matching yielding capabilities;
• energy absorption capacities of support elements matched to
the energy release of potential strainburst events; and
• excess load capacity to account for potential seismic
shakedown events.
GEOLOGY AND SETTING
The Beaconsfield Gold Mine is centred on the Tasmania Reef, a
steeply dipping tabular quartz-carbonate vein hosted by a
sequence of siliciclastic sediments. The sediments, which dip
towards the east, comprise the Salisbury Hill Formation of
conglomerates and sandstones overlain by the Eaglehawk Gully
Formation of sandstones, siltstones and limestones with
occasional pebble bands. All siliciclastic rocks have been
metamorphosed to quartzites although they retain much of their
sedimentary character. The Tasmania Reef is essentially planar
over a strike length of 350 - 400 m, with an average width of
2.5 m and occupies a shear which cuts across stratigraphy
orthogonally. The west zone of the Tasmania Reef is hosted by
the Salisbury Hill Formation, and it is the brittle nature of the
conglomerate horizons within that formation in particular, which
are prone to mining induced seismicity. The geology of the
Beaconsfield Gold Mine is described in detail by Hills et al
(2001).
MINING METHOD
The re-designed mining method for the Western Zone involves
no man access into the brittle narrow ore zone once stoping
commences in a block. This has required the development of a
network of footwall drives parallel to the ore zone (for access for
drilling/blasting) on successive levels and implementation of
fully remote mining. Figure 1 illustrates the current stoping
mining method in the West Zone at the Beaconsfield Gold Mine,
stopes are extracted by a Modified Avoca style with a planned
annual production of up to 300 000 tonnes in 2009. Mining
methods are further discussed in Hills et al (2008).
Initially the ore drive is developed, followed by the footwall
drive. Minor strainburst events (spitting of small amounts of rock
from the development face) are experienced when developing the
FIG 1 - Schematic of current mining method, Beaconsfield
Gold Mine – Western Zone.
173
C SCOTT, A R PENNEY and P FULLER
footwall drive through the more competent lithologies. Mining
then retreats from both directions toward the central access drive.
Whilst this does create a closure pillar above the access
intersection where both eastern and western mining districts
exist, the ‘no man access to the ore zone’ policy is strictly
enforced, greatly reducing the risks associated with potential
pillar instability. The extraction of the access pillar is completed
from the footwall drive on that level, or from the footwall drive
above with a modified production hole pattern. Induced stresses
generally increase as stoping fronts retreat, leading to small
stress driven spalling in the footwall drive, crushing of the more
brittle quartz reef in the ore drive and increased seismicity
(particularly in the hanging wall of the orebody).
SUPPORT SELECTION CRITERIA
An outline of methods for assessing load, displacement and
energy absorption requirements for support selection at
Beaconsfield for various failure mechanisms is presented in
Figure 2.
In any given area, an assessment of support adequacy against
static failure and varying levels of load energy/displacement
requirements against seismic events is considered.
The support component and support system capacity is defined
in terms of load, displacement and energy-dissipation. In
burst-prone mines, it is likely that the yield load capacity of the
support system or individual components is temporarily
exceeded as it yields (ie SF(load) approaches 1.0). However, as
long as the displacement or energy capacity limit is not
exceeded, the support capacity is still greater than the demand
and the overall or ultimate safety factor (SF >1). Thus, the
adopted support selection approach aims to ensure that the
FIG 2 - Support selection procedure chart.
174 Ballarat, Vic, 14 - 15 October 2008
support system has sufficient load, displacement and energy
capacity. In burst-prone ground, the support must be designed to
survive large deformations and to dissipate energy. Different
safety factors may be chosen for each criteria and satisfying only
one of these three criteria is unlikely to be sufficient.
Each support component as well as the entire support system,
consisting of all support components, must provide sufficient
capacity to survive the demands placed on them. Unless some
components are designed to provide specific functions and are
sacrificed in excessive loading or deformation conditions, all
components will be deformed simultaneously (in parallel) but all
components must be deformation limit compatible (designed to
fail at the same displacement). For example, commonly installed
fully grouted rock bolts behave in a relative stiff manner and may
fail at deformations less than those required to fail less stiff deep
seated cable support. In which case, after sufficient ground
movement has occurred to fail the rock bolts, full load will be
transferred to the deep seated cable support.
The capacity of all connections (plates, nuts, etc) must be
sufficient to transfer the design load and displacement between
each support component of the overall support system.
Static assessment
Static assessments are conducted based on the following
potential failure mechanisms (Reeves, 2008):
• kinematic wedge failure, and
• release of a loosened zone with a parabolic arch shape.
Kinematic wedge assessments have been conducted using
mine wide joint set data and planned drive profiles. The loosened
parabolic arch simply assumes a loosened area defined by a
Narrow Vein Mining Conference
 COMPETING FACTORS IN SUPPORT SELECTION FOR THE WEST ZONE OF THE BEACONSFIELD GOLD MINE, TASMANIA

parabola shape in the backs of development with a mid-span 150

Volume of fractured rock (m3)

height equal to one third of the span (Lang, 1961) that is to be
supported statically by installed support.

100
Strainburst support requirement assessment
method

Self triggered strainburst 50

In a self triggered strainburst the back or wall(s) of a drive fail

violently due to high stresses and the depth of failure is suddenly
created or increases. The seismic event magnitude and the energy
released is related to the volume of rock failing, the system
stiffness and the seismic efficiency (as some seismic energy is
used up in breaking rock).
For the selection of an appropriate support system, the
dynamic depth of failure needs to be determined in order to:
• estimate the energy that could be released during
strainbursting,
• estimate the increased static load after strainbursting,
• assess the required bolt length, and
• estimate the anticipated straining of the bolts.
The assessment method for support requirement (energy
absorption) against a self triggered strainburst is detailed in the
following discussion. The process and equations used are taken
or adapted from Kaiser, McCreath and Tannant (1996).
The dynamic depth of failure (df, m) can be determined for the
anticipated σmax (maximum induced principal stress determined
from modelling, MPa) as defined by Equation 1:
df ~ w/√2 * (1.25*(σmax/UCS) – C) (1)
where:
UCS = unconfined compressive strength (MPa)
w = width of drive (m)
C = constant (calibrated by noting the modelled induced
stress at which initial stress failure/spalling occurs)
It is then assumed that this depth of failure is created instantly
(if created by gradual slabbing, the incremental released energy
is significantly less, which therefore makes this assumption very
conservative) and the anticipated sudden wall displacement that
the support will experience is the product of the depth of failure
and a representative bulking factor (BF) when the fractured rock
remains confined by the support system. For the support systems
recommended for Beaconsfield, values of BF are anticipated to
be less than 5 ± 1 per cent (from recommended bulking factors
for yielding support and major damage (Kaiser, McCreath and
Tannant, 1996)) and this needs to be accommodated as stretch in
the bolts.
The volume of failing rock per excavation face (Vfailure) is then
assumed to be approximately parabolic in cross-section and is
determined using Equation 2:
Vfailure = 2/3*df*w*l, (2)
where:
l = length of failure along the drive (to be determined based
on field evidence; generally at least equal to the width or
height of a drive) (m)
The resulting event magnitude can be calculated as proposed
by Obert and Duvall (1969) or estimated from Figure 3
(assuming a rock mass shear modulus of 20 GPa and a circular
tunnel in a uniform stress field). The higher the maximum stress
0
-0.5 0 0.5 1.0 1.5 2.0
Richter magnitude
FIG 3 - Relationship between volume of failed rock and event
magnitude (after Kaiser, McCreath and Tannant, 1996).
at the time of the strainburst, the higher the released energy and
related event magnitude.
Energy release (Es) is then calculated using the event
magnitude in Equation 3:
Es =10^(ML±0.15) (MJ) (3)
where:
ML = magnitude of seismic event (local magnitude that is
approximately equivalent to the Richter scale)
An energy efficiency, , correction is applied to determine the
energy that would transfer to the support. Seismic efficiency of
strainbursting rock is generally assumed to be ten per cent until
field evidence suggests otherwise.
The energy required to be absorbed by the support per square
metre of excavation surface is therefore calculated by dividing
the calculated releasable energy by the area over which it acts in
Equation 4:
E(support) = Es * / w*l (4)
The total energy demand on the support system is then
calculated as SF(energy) × E(support).
Remotely triggered strainburst
A remote seismic event may add dynamic stress to a drive,
increase the maximum tangential stress and thus deepen the
depth of failure. The dynamic stress increase ∆σmax (MPa) is a
function of induced peak particle velocity (PPV) (Kaiser,
McCreath and Tannant, 1996) and thus a function of the
magnitude of the expected design event and the anticipated
design distance R from the event.
For an assumed magnitude of seismic event and distance from
area of concern, the dynamic stress increase ∆σmax is estimated
from Figure 4.
The increase in depth of failure (∆df) due to the passing
seismic event is calculated using Equation 5:
∆df ~ w/ 2 * 1.25*∆σmax/UCS (5)
The anticipated sudden, incremental wall displacement that the
support will experience is equal to the product of the depth of
failure times the representative confined bulking factor (BF) as
outlined earlier.
The volume of incrementally failing rock involved (∆Vfailure) is
then determined from Equation 6:

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 175

C SCOTT, A R PENNEY and P FULLER
FIG 4 - Estimate of dynamic increase of ∆σmax due to passing
seismic wave (Kaiser, McCreath and Tannant, 1996).
∆Vfailure = 2/3* df*w*l. (6)
The resulting event magnitude can be calculated as proposed
by Obert and Duvall (1969) or estimated from Figure 3. The
energy release (Es) is then calculated using the event magnitude
determined and Equation 3.
An assumption is again made that only ten per cent of the total
energy release calculated impacts on the support systems.
The energy demand on the support system is FS (energy) × Es.
Seismic shakedown support requirement
assessment
Seismically-induced falls of ground are caused when a seismic
wave passes around a drive, causing additional stress fracturing
and adding a seismic acceleration increment to the gravitational
acceleration (static weight) of failed material.
How much ground motion a supported rock mass can survive
is dependent on the static factor of safety for survival (SFsurv)
defined in Equation 7:
1  1   The ultimate deformation (dult) for both resin bolts and cables
SFSurv = ( n. ppv )2 + 1 (7)
md  2 gd ult  
where:
dult (m) = is the remaining displacement capacity of the support
(ie after prestress or after other deformations imposed
on the support)
md = multiplier on the support element strength (ie steel
strength) when loaded dynamically, typically between
1.1 and 1.4 (suggested value of 1.25)
n = ejection velocity ratio, conservatively assumed to be
unity (1)
At this static safety factor for survival, the actual SF during the
event would be one. Hence, to achieve a real safety margin, the
design SFdesign must be greater than SFsurv, eg SFdesign (load) = SFsurv
× 1.5.
Steps involved in the assessment of support requirements
against seismic shakedown are detailed in the following:
For far-field situations, PPV can be calculated using
Equation 8:
176 Ballarat, Vic, 14 - 15 October 2008
10 0 .5 ×( ML + 1 .5 + 2 log( C *))
ppv = (8)
R
where:
C* = an empirically determined constant (0.25 m2/s)
The actual pressure on the support ps is determined from the
total (static plus dynamic) depth of fracture defined in
Equation 9:
df ~ w/ 2 * (1.25*(( max+ ∆σ max)/UCS) – C) (9)
From the calculated depths of fracture, the corresponding load
(t/m2) is:
ps = df * Y (10)
where:
Y = density of rock (t/m3)
The support load demand for the amount of failed rock, the
selected design event and the chosen displacement capacity can
now be determined by multiplying ps by SFdesign (load).
BEACONSFIELD FOOTWALL DRIVE SUPPORT
SELECTION
By following the steps outlined above and using the assumptions
and measured/modelled input parameters for the West Zone at
Beaconsfield the following support criteria have been derived:
• energy absorption requirements against self triggered
strainburst = 10 kJ/m2;
• energy absorption requirements against remote triggered
strainburst = 10 kJ/m2 (up to 20 kJ/m2 where faults are
identified);
• support load demand against seismic shakedown (dependent
on induced stress levels) = (typically) 7 - 14 t/m2 and SFsurv
ranging from 1.5 to 3.1; and
• support element displacement requirement against strainburst
events ~50 mm.
Details of elements chosen for use in the support selection are
summarised in Table 1.
has been specifically matched at 0.16 m in order to provide
complimentary total displacement behaviour.
Support standards and performance
In specifying the 980 footwall drive support, the procedures
outlined above were used with the following assumptions:
• Maximum magnitude of a seismic event = 2.5 ML
determined from Gutenberg Richter analysis for the largest
expected event using the recorded seismic history in this
domain.
• Calculated energy release from strain burst is 20 kJ/m2 as
numerous faults are either intersected or are in the near field
for the event magnitude used. While the self triggered
strainburst is still present, the intensity of faulting in the area
of the drives, and the near field relationship of these faults
justified the higher energy level selected.
• Maximum stress calculated around the drive from 3D elastic
numerical modelling at any stage of the mining sequence is
selected.
The demand requirements imposed on the standard ground
control (friction bolts and mesh) using the procedure above are
Narrow Vein Mining Conference
 COMPETING FACTORS IN SUPPORT SELECTION FOR THE WEST ZONE OF THE BEACONSFIELD GOLD MINE, TASMANIA

TABLE 1
Support element properties and behaviour.

Support element 20 mm threadbar Friction bolt Debonded single Mesh Mesh straps
properties (splitset) strand cable
(bungie bolts)
Length/ dimensions (m) 2.4 2.4 7.3 Aperture 0.1 × 0.1 m Aperture 0.1 × 0.1 m
#6 gauge 4.9 mm #0 gauge 8.0 mm
Anchorage With resin anchor and 46 mm diameter full 4.5 m debonded with n/a n/a
1 m debonded length friction contact grouted anchors each
end (refer detailed
specifications)
dult (m) 0.16 0.2 0.16 High High
Energy absorption (kJ) 26.6 9.4 (for 1 m thick slab) 31.5 n/a n/a
Energy absorption (kJ/m2) n/a n/a n/a 4-6 Negligible

loading and deformation from a dynamic depth of fracture up to
1.1 m deep and an energy dissipation shortfall of 17 kJ/m2. As
the depth of failure for these failure mechanisms do not exceed
the site limit of 1.7 m (determined by standard support patterns
and expected seismic conditions), the selection process uses resin
threadbar bolts (Posimix4) as the dynamic upgrade. To achieve
the design factor of safety and SF(surv) for both shakedown and
strainburst conditions and maintaining a SF(load) as low as
possible, an additional five resin threadbar bolts are required in
the drive backs and four resin threadbar bolts in each wall at a
ring spacing of 1.3 m. Debonded resin threadbar bolts are chosen
for the upgrade support due to the higher load capacity compared
to friction bolts, but still have a sufficient elongation at
yield/failure. This upgrade is installed over the as-installed mesh
and through additional heavy gauge mesh straps. The mesh
straps are intended to create reinforced ribs across the drive
backs and walls to enhance the capacity of the surface support
(Figure 5).
Where intersections are designed (ie drawpoints), the system is
further enhanced by the use of debonded cables (Bungee bolts).
These longer and high capacity (load, deformation and energy
absorption) bolts are used to support the larger drive spans and
increased depth of potential instability. In other localised areas
along the drive, debonded cables have also been used where poor
ground conditions have been encountered.
Ground support performance for initial mining
areas
Back analysis of the drive design from visual inspections and
data collected from geotechnical instrumentation (Penney, Hills
and Walton, 2008) suggests that the engineered design and
support selection is sound with potential for later refinement. To
date, there have been three stope firings in the 980 stope block
(all at the 980 level) extracting a total of approximately 60 m
strike length of the reef. These firings have triggered a number of

Point anchored Posimix4 bolts

Mesh Straps

Grade Line

FIG 5 - Shakedown and strainburst ground control upgrade for the 980 footwall drive.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 177

C SCOTT, A R PENNEY and P FULLER
significant and large seismic events, with the largest recorded
event being 1.9 ML immediately after the first stope firing in a
known seismic hazard zone. Minor stress fracturing on the
footwall shoulder of the drive has been observed in positions
close to where stope brows exist in the reef. Moderate stress
changes (both magnitude and direction) have been recorded and
localised increases in load and displacement have also been
recorded. To date the largest displacement change is 18 mm at a
depth of <2 m (shear displacement along stress fracturing
observed through diamond drill observation holes) with a
maximum of five tonnes load developed in the SMART® cable
instrument. Details of the instrumentation installation and results
are described by Penny, Hills and Walton (2008).
Overall, based on the performance to date the specified ground
support for the footwall drives is considered satisfactory. As
stoping fronts advance and create an expected increase in wall
closure within the stope, more deformation and support load is
anticipated in the footwall drive. Once the full range of expected
conditions in the footwall drives is experienced, further back
analysis and refinement of support parameters and support
selection will occur and be implemented for future blocks. This
experience will also be used for specifying support outside of the
western geotechnical domain.
CONCLUSIONS
A support selection method for support against several failure
mechanism; static, strainburst (self and remotely seismic event
triggered) and seismic shakedown has been developed for the
Beaconsfield Gold Mine. Support selection has been made in
order to satisfy determined load, energy absorption and
displacement requirements. In order to meet these requirements,
various support elements have been chosen with; sufficient
load-displacement characteristics to withstand expected
deformations, sufficient load capacity to support expected load
and sufficient excess energy absorption to survive anticipated
seismic events.
Important considerations identified in selecting a support
system in a seismic environment include:
• Factors of safety other than the traditional static load factor
of safety need to be considered.
• The need for displacement compatibility of support elements.
• The importance of maintaining mechanical connection
between support system components under seismic loading
conditions.
• SFdesign (load) should be selected to be as low as possible so
that loads in the high energy absorption support elements
before seismic activity is as close as possible to the yield
load. This ensures that when seismicity applies a dynamic
load, the rate of further support element loading is minimised
and the connections between the supports and the surface
restraint are not subjected to large pulse loading and the
integrity of the support system is maintained.
178 Ballarat, Vic, 14 - 15 October 2008
This support selection process has allowed the safe
re-commencement of production from the underground workings
at the Beaconsfield Gold Mine with current results being very
positive. Systems are in place to determine the ongoing support
performance and allow for future refinement of the support
system to be made using this design procedure.
ACKNOWLEDGEMENTS
The authors would like to thank Coffey Mining and the
Management of Beaconsfield Gold Mine for the opportunity to
present this data. We also would like to thank P K Kaiser for his
contribution to the understanding of this design process and
providing invaluable reviews of the work undertaken at critical
points through the support selection process.
REFERENCES
Hills, P B, MacDonald, G, Nero, L and Mueller, A M, 2001. Beaconsfield
Gold Mine – Tasmania Reef, in Structure and Setting of Proterozoic
and Palaeozoic rocks in the Tamar region, Northern Tasmania
Excursion Field Guide (ed: A R Reed), pp 76-88 (Geological Society
of Australia Specialist Group in Tectonics and Structural Geology:
Sydney).
Hills, P B, Mills, J, Penney, A R and Arthur, S, 2008. The development
and implementation of a fully remote stoping method at Beaconsfield
Gold Mine, Tasmania, in Proceedings Narrow Vein Mining
Conference, pp 199-206 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Kaiser, P K, McCreath, D R and Tannant, D D, 1996. Canadian
Rockburst Support Handbook, p 324 (Geomechanics Research
Centre: Sudbury).
King, R, Thomas, S and Scott, C, 2007. Geotechnical assessment –
Resumption of ore production (east sone), Coffey Mining Pty Ltd
report No HZ00054.01.BH.
Lang, T A, 1961. Theory and practice of rockbolting, Transactions of the
American Institute of Mining, Metallurgical and Petroleum
Engineers, 220:333-348.
Penney, A R, Hills, P B and Walton, R J, 2008. The use of geotechnical
instrumentation to optimise an engineered mine design at
Beaconsfield Gold Mine, Tasmania, in Proceedings Narrow Vein
Mining Conference, pp 165-172 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Pfitzner, M, 2006 Geotechnical assessment – Resumption of decline
developments development, Coffey Mining Pty Ltd report No
HZ00054.01.A1.
Reeves, I, 2008. Procedure for support selection – Beaconsfield Mine
Joint Venture, Coffey Mining Pty Ltd report No HZ00054.01.CC.
Scott, C and Reeves, I, 2007. Geotechnical assessment – Resumption of
ore production (west zone), Coffey Mining Pty Ltd report No
HZ00054.01.BY.
Sidea, D, Scott, C and Reeves, I, 2007. Geotechnical assessment –
Resumption of ore driving development, Coffey Mining Pty Ltd
report No HZ00054.01.AT.
Narrow Vein Mining Conference
Design Factors Leading to Prevention of Bridged Stopes at
Jundee Operations
C Miles1 and G Lind2
ABSTRACT
Bridging of production stopes in narrow vein gold deposits contributes to
significant financial losses and production delays. Bridging is caused
when underbreak in the stope production blast occurs, resulting in less
than the designed stope blast volume breaking out during blasting. Jundee
Operations in the Northern Goldfields of Western Australia extracts ore
from its expansive narrow vein orebodies utilising longhole open stoping.
In 2006 it experienced bridging in ten panels, or 8.77 per cent of all
panels, as well as multiple bridging in 30 per cent of these bridged
panels. The Jundee orebodies are west dipping with the bridged panels
dipping between 44° and 62°. This paper examines some solutions from
the drill and blast design perspective to minimise bridging occurrences. It
was found that the then current drill patterns and number of holes were
adequate from a longhole stope design perspective. However, changes to
the blasting design, specifically modifications to the blast delay initiation
sequencing, primer positioning and blasthole dump combined with timely
stope surveying and back analysis, were implemented with the aim of
preventing bridging in this narrow vein operation.
INTRODUCTION
Stoping is the main source of underground ore production at
Jundee and hence is critical to maintaining the revenue flow of
the operation. The preliminary economics of a stope, which will
determine profitability, is based on the assumption that all ore
will be recovered. A tonnage and grade report is compiled for
each panel including acceptable amounts of dilution. Where
significant underbreak occurs, which is known as bridging,
re-rising is required in an attempt to resolve the situation. With
1. GAusIMM, Graduate Mining Engineer, Newmont Asia Pacific,
Jundee Operations. Email: milesc03@hotmail.com
2. MAusIMM, WorkSafe Victoria, Level 18, 222 Exhibition Street,
Melbourne Vic 3000. Email: gavin_lind@worksafe.vic.gov.au
0.0M 250.0M 500.0M
500.0M
250.0M
0.0M 250.0M 500.0M
Scale:
JUNDEE Date : 14-Sep-2006
OPERATION Plot :
Dept : UG Mining
FIG 1 - 2006 Barton Deeps, Westside, Hughes, Hampton and Nimary declines providing access to predominant Jundee mining area.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
rise development costs to Newmont of around $1500 per metre
and with rise heights up to 20 metres (Torlach, 1998), re-rising
not only results in additional cost to the company, but also a
delay in production. The additional cost of rising in 2006 was
1.7 per cent of the total drilling and blasting costs.
Newmont’s Jundee Operations, situated in the Northern
Goldfields of Western Australia, target small narrow vein
orebodies, mined utilising longhole open stoping methods.
During 2006, 8.77 per cent (ten out of 114) of all mined panels
bridged, with 30 per cent of these panels bridging more than
once.
In search of an answer as to why stopes bridge, it is beneficial
to analyse issues which can occur throughout the stoping
process. There are, however, natural factors such as poor ground
conditions, water inflow, voids or geological structures which
can impact on the outcome of a stope (Villaescusa, 1998).
Ten panels bridged at Jundee in 2006 and these were analysed
along with one control stope to determine the cause of the
bridging.
JUNDEE OPERATIONS
There are two underground declines from within open pits on the
Jundee lease. One of these is from the Barton Pit and the other
from the Invicta Pit. The Barton decline then splits to access the
various orebodies, including Nimary, Upper North, Lower North,
South, Westside South, Westside North, Hamptons and Hughes,
as shown in Figure 1. Although some of these decline drives
access the same orebody, the access point determines the
nomenclature. The Invicta orebody is situated farther away and is
accessed by a separate, yet straightforward, decline. The newest
underground discovery has been of the Cadassian orebody, which
is accessed between the forth and seventh Invicta levels.
750.0M 1000.0M
500.0M
250.0M
750.0M 1000.0M
1:3000 Jundee Gold Mine
Barton, Westside, Hughes,
SM Hampton and Nim
181
C MILES and G LIND

GEOLOGY AND MINING METHOD

Jundee employs the unsupported mining method of sublevel
open stoping (SLOS) with pillars used as required to allow for
stable openings until work in the area has been completed. The
narrow orebody is diverse and consists of numerous veins which
are ‘chased’ by geologists, with an example of the geology
shown in Figure 2.
The sublevel stoping at Jundee uses up-holes to remove the
dirt from each individual sublevel rather than from a universal
drawpoint at the bottom of the final stope as classical SLOS
does. The numerous lodes are comprised of gold veins which are
diverse in geological position and therefore, for maximum
extraction, each level needs to be mined independently of levels
below. Jundee currently undertakes a top down approach, such
that mining of upper levels is completed before extracting levels
further down.
FIG 3 - Representation of Toe, Island and Collar bridges.
BRIDGING
Underbreak (as defined by Jundee) is where less of the designed Bridging in sublevel stopes can cause a loss of stope back
stope breaks out during blasting, resulting in ore loss when the height due to significant underbreak, which can only be remedied
ore becomes inaccessible due to the inability to access past the with specialised blasting or a re-rise.
stope brow (Mariset, 2005). This underbreak can occur as a
bridge in the form of either a toe, collar or island bridge, as DESIGN ISSUES
demonstrated by Figure 3. Collar bridges are sometimes easy to
remedy, island bridges create minimal problems and toe bridges When designing stopes, there are many drilling and blasting
cause large problems. Geological features, such as faults, may aspects which need to be considered in order to create a
cause bridging where the features are perpendicular to the successful design. These include drill hole dump, planned
hanging wall. dilution and the rise height and location. When designing

FIG 2 - Composite level plan showing some of the Barton and Westside lodes, which occur at different levels.

182 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 DESIGN FACTORS LEADING TO PREVENTION OF BRIDGED STOPES AT JUNDEE OPERATIONS
blasting patterns, void ratio, powder factor, primer placement and
detonation delay are the main factors which determine the
outcome of the stope (Dyno Nobel, 2005).
While these many design issues were investigated, along with
some uncontrollable variables (such as the dip of the orebody),
the following are the ones which were determined, through
analysis, to be the prime cause of problems relating to bridging.
Breakthrough or blind
Stopes can be designed to either breakthrough into a top drive, or
to be blind. Blind stopes are mined without a top drive, which
may occur when a crown pillar is being left or when the orebody
terminates at the top of the stope.
Dump
Drill holes can be drilled in most directions, limited only by the
capabilities of the drill rig. The holes are drilled in a series of
parallel rings, with rings either being vertical or angled forwards
or backwards. The forward and backwards angling is known as
dump. In areas of bad ground or for blind stopes, it is suggested
that drill holes are dumped forward, which will result in a better
stope formation. When dump is incorporated, stope height is
more easily sustained, break-out angles at the toe holes are
cleaner and the distance from the brow is safer for the charge-up
crew (Villaescusa, 2007).
Height gain
A stope design issue occurring at Jundee is related to height
gains, with some designs attempting to gain up to 5 m in height
throughout the panel. When rings are dumped, it is possible to
gain up to 1 m per ring due to toe break-out angles. It is also
possible to gain some height on breakthrough stopes where the
top drive provides a free face. It must be noted though that large
height gains are unachievable and need to be eliminated from the
design (Workman-Davies, 2006).
Lead/lag time
When there is a large time period between the drilling and the
charging of a stope, induced stresses in the rock mass can close
the blastholes resulting in the need for re-drilling. If the
re-drilling is not done, poor charging will occur and little or none
of the stope may break out in the blast.
Detonation delay
NONEL detonators can offer up to 94 preset delays with one per
cent accuracy which equates to ±0.5 milliseconds for every
25 milliseconds of delay. The Dyno Nobel guidelines (Dyno
Nobel, 2005) suggest that a minimum of 50 milliseconds
between holes and 100 milliseconds between rings should be
used unless fragmentation is an issue or it is a large stope with a
large free face. Where small voids are present, such as around a
rise or in a narrow stope, longer delays are required to allow the
dirt to fall, to create a free face. Equation 1 (below) is used to
determine the delay between holes for small voids, with a
20 metre hole requiring approximately two seconds delay
between initiation of the hole and the proceeding hole.
Delay (seconds) = √(Rise length/5) (1)
Primer placement
The Dyno Nobel guidelines also suggest that for optimal blasting
results, it is beneficial for holes over 10 m in length to use two
primers, with one placed at the toe (or 1.5 m to 2 m back from
the toe on breakthrough holes) and one in the centre of the hole.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
Holes less than 10 m in length only require the toe primer, unless
there is jointed ground, in which case, more primers may be
required to ensure a complete detonation of charge. Primers must
be placed in competent ground and if voids or water inflows are
known, the hole must be lined before charging. Sometimes when
a particularly hard section of ground is found, or when there have
been problems in the previous blast, two or three primers (or just
the boosters) may be incorporated in the blast design by the drill
and blast engineer.
The Dyno Nobel guidelines for primer placement are followed
at Jundee; however, this is not alleviating the bridging problem.
A Dyno Consult consultant gave the mine engineers some
suggestions on primer placement which have reduced bridging
issues at other sites. These include centre initiation on
breakthrough stopes, whilst leaving the initiation for blind stopes
at the toe (Wiggin, 2007).
Back analysis
Back analysis should always form a part of the mining process as
it is an extremely valuable tool used to improve designs
(Sen, 1995). A Cavity Monitoring System (CMS) survey must be
completed regularly. Stopes should be surveyed every 20 m when
the drives are straight, or more frequently if there is a curve or
kink in the drive. It has been noted that stopes with flat dips
provide a less accurate CMS result, as broken rock is more likely
to remain on the footwall, blocking the instrument view of the
final stope shape. The CMS survey is the only real means of
analysing stope performance, as it allows for reconciliation
reports to be completed using geological models to determine
actual versus planned mined ounces.
Geological problems
Unfortunately some problems in stoping can not be avoided due
to the influence of geological structures, including joints,
foliations, bedding and folds. A drill hole intersection of these
features can result in the cutting off of holes, the introduction of
water or may affect the manner in which the rock breaks upon
firing.
A fault cutting through a hole can cut off the charge, as ANFO
does not have the ability to pass on charge without contact, due
to a high gap sensitivity. Drillers should be able to feel a change
in ground conditions and note these on their drilling plod, while
experienced drillers may also be able to detect whether this
change is due to a fault. Engineers need to note the drillers’
comments and include a note on the charge plan to line the hole
to avoid ‘losing’ the hole and if the primer is placed within this
ground, there is a loss of confinement and hence not enough
directed force to break the required ground. Alternatively,
multiple priming, using packed explosives or charge decking can
be used to alleviate the issues of problem holes. Holes can also
be ‘squeezed shut’ or damaged in high stress areas, which may
result in requiring re-drilling of the holes before charging can
commence.
Jundee usually experiences oversize in stopes from hanging
wall failures (overbreak); however, this does not usually occur in
stopes which bridge. Similar geological features may cause both
bridging and oversize, with oversize occurring where the faults
are parallel to the hanging wall and bridges where the faults are
perpendicular to the hanging wall.
ANALYSIS
It was identified by the Jundee management team that stope
bridging was a major area of loss in the form of production
tonnes and profits and hence needed to be investigated. A
complete year, 2006, was chosen for analysis. Every panel mined
in that year was analysed, specifically from a design perspective
183
C MILES and G LIND

to see what the drill and blast team could do to minimise or
eliminate the problem.
The initial stage of analysis was to collect all geological
models, economic models, rise instructions, drilling plods,
charge plans and survey pick-ups for the stope panels which
bridged in 2006. It was also vital to collate the relevant
procedures for the mine (Nandlal, 2003).
The next step was to identify the design parameters associated
with the drill and blast process and then examine each stope and
compare the actual design with the design parameters and the
actual stope shape. Back analysis was then done to determine
actual tonnages and grades and therefore actual economics.
RESULTS
The ten problem stopes and the control stope from 2006 were
analysed from the design perspective. Table 1 is a summary of
some of the properties used to determine which stopes were
prone to bridging, with B/T indicating breakthrough stopes. Flat
dipping stopes have the orebody at less than 50° from the
horizontal while steep dips are greater than 70°.
The results indicate that narrow, breakthrough stopes with no
dump were the most likely to bridge and while there was no
major standout in the dip characteristics, steeper stopes were
more inclined to bridge.
Table 2 demonstrates the variations between design and actual
stope outcomes in terms of ounces and profit margins.
Occasionally there is ore in outlying regions, which is
uneconomical to extract through design as extensive drilling
would be required. This occurred in the 14S360S P2 which
actually resulted in more ounces and a greater profit than design
due to unplanned overbreak. The bridging standouts occurred in
19S325S P1 where the majority of the design ounces were lost
due to the abandonment of the panel and the 8S131N P1, which
bridged numerous times, resulting in losses being made on both
the stopes.
RECOMMENDATIONS
The most important recommendation is for on-site engineers to
follow procedures that are in place for drilling and blasting and
carry out the vital back analysis via prompt CMS survey
pick-ups in order to manage the problems associated with
stoping. In 2007, the implementation of bridge memorandums,
completed immediately after every bridge, as well as stope
reconciliation reports at the end of each drive has occurred.
It is suggested that initiating the centre primer first on
breakthrough stopes will alleviate the bridging problem, while
toe primers are still detonated first in blind stopes. Dumping drill
holes is also likely to prove beneficial as shown by the panels
without dump which were more liable to bridge.
The knowledge exists at Jundee to ensure that stopes do not
fail through bridging; however, the practices in place are not
always followed by the design engineers. This indicates
management failures, hence a production design manager is
being sourced for the site to monitor and control designs of
future stopes. This is expected to reduce costly design errors
from occurring in future.

TABLE 1
Property summary of analysed stopes.
B/T Blind Dump No dump Flat dip Medium dip Steep dip Narrow Bulk
- 8S124N P1
- 8S131N P1
- 14S360S P2
- 18S356S P2
- 19S125S P1
- 19S325S P1
- 19S354N P2
- 7V630S P4
- 7V640N P4
- 8Nim03N P2
13S910S Inc P1

TABLE 2
Design versus actual stope summary.
Ounces Profit
Design Actual % Lost Design Actual % Lost
8S124N P1 2179 1423 34.69 $424 804 $218 183 48.64
8S131N P1 910 726 20.22 $66 457 -$3 916 105.89
14S360S P2 3505 3881 -10.73 $754 297 $924 192 -22.52
18S356S P2 2726 1425 47.73 $761 682 $366 173 51.93
19S125S P1 400 317 20.75 $67 426 $30 750 54.39
19S325S P1 727 3 99.59 $229 762 -$52 724 122.95
19S354N P2 3327 3134 5.80 $938 563 $826 242 11.97
7V630S P4 3852 3612 6.23 $1 170 254 $1 036 915 11.39
7V640N P4 2832 2552 9.89 $727 736 $610 711 16.08
8Nim03N P2 15456 12284 20.52 $4 470 962 $3 411 776 23.69

184 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 DESIGN FACTORS LEADING TO PREVENTION OF BRIDGED STOPES AT JUNDEE OPERATIONS
With all these recommendations taken on board by the
management team at Jundee Operations it is expected that the
bridging occurrences can be greatly reduced. This will allow the
planned tonnages and grades to be achieved, while minimising
costly production downtime.
Other narrow vein mining operations need to be aware of the
scale of problems that bridging can cause and need to ensure that
they are following site-specific procedures which have been put
in place over time to enable safe and economic removal of ore.
CONCLUSIONS
The loss of ore through bridging is an extremely costly problem
at Jundee with over $2.2 million dollars of potential profit lost,
along with 45 days of lost production while re-rising took place
in these problematic panels.
After investigation it was deduced that the stopes most likely
to bridge broke through into an upper level, had no dump and
were of narrow proportions. The orebodies which were steeply
dipping were also slightly more prone to bridging, however the
dip was not a deciding factor.
The dip of the orebody cannot be altered and having
breakthrough stopes maximises ore recovery. The width of an
orebody is also fixed; however, the cost of planned dilution needs
to be weighed against the risk of bridging to determine the
optimal stoping width. Dump can be included in all designs for
improved safety and better break-out angles at the toe, and
breakthrough stopes can have the centre detonator initiated first.
These changes are likely to significantly reduce the problem of
bridged stopes at Jundee Operations.
ADDENDUM
Since this research was carried out, changes have been made on
site both in design and in underground procedures. The most
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
effective measure to prevent bridging at Jundee has actually been
determined to be the correct preparation of holes prior to
charging. There has been a significant increase in the amount of
re-drilling required; however, this is still more cost and time
effective than re-rising. It must be noted that no matter how
many measures are put in place, bridging is unfortunately still
occurring occasionally.
ACKNOWLEDGEMENTS
The authors wish to thank the management of Jundee Operations
for their support during the trial and analysis and for permission
to publish these results.
REFERENCES
Dyno Nobel, 2005. Underground products and services reference guide
[online]. Available from: <http://www.dynonobel.com> [Accessed
19 May 2007].
Mariset, S, 2005. NJO-INT-UG-011-S16 Ground Support Standard,
Newmont Australia, Perth, 04 November 2005.
Nandlal, J, 2003. Newmont Australia, Jundee Operations, drill and blast
guidelines and procedures.
Sen, G C,1995. Blasting Technology for Mining and Civil Engineers
(University of New South Wales Press Ltd: Sydney).
Torlach, J, 1998. Safety Bulletin No 39: Vertical Opening Development in
Underground Mines (Rise Development) (Department of Minerals
and Energy: East Perth).
Villaescusa, E, 1998. Geotechnical design for dilution control in
underground mining, in Proceedings Seventh International
Symposium on Mine Planning and Equipment Selection, Calgary
(Balkema: Rotterdam).
Villaescusa, E, 2007. WASM underground rock mechanics 431 lecture
materials (unpublished).
Wiggin, M, 2007. Personal communication, 18 July.
Workman-Davies, C, 2006. WASM underground mining 332 lecture
materials (unpublished).
185
Drift and Fill – The High Value, High Recovery Mining System
L Dawson1, M Yumlu2 and M English3
ABSTRACT
At times of high metal prices, mining projects that have previously been
put into the ‘too hard basket’ are now being more closely examined.
Typically these projects include the potential recovery of small mineable
tonnes with high contained metal values and often with poor ground
conditions.
For many underground mining projects, the drift and fill method is
viewed as a high-cost method and a ‘method of last resort’. As a result
the method is not well known in Australia and it is the authors’ view that
the method is under utilised.
The drift and fill method can be applied to a variety of underground
mining situations and is a preferred method for some orebody geometries
and where ore recovery is viewed as a key to project success. This paper
examines some of the critical mine planning and backfill issues involved
with the drift and fill mining method.
WHAT IS DRIFT AND FILL?
Drift and fill (DAF) is literally the mining of an underground
drift, or development drive in ore, and backfilling the completed
drift prior to mining an adjacent drift in wide orebodies or in
narrow, flat-lying orebodies. The method allows the total
extraction of the ore material with minimal exposure of
unsupported ground and minimal risk of subsidence.
When compared to mechanised cut-and-fill (MCAF) mining,
DAF is a multiple-drift arrangement used for pillarless mining in
wide orebodies or in narrow, flat-lying orebodies whereas MCAF
is used for single-drift, single-pass mining in steeply dipping,
narrow orebodies.
1. MAusIMM, Principal Mining Engineer, AMC Consultants Pty Ltd,
Level 19, 114 William Street, Melbourne Vic 3000.
Email: ldawson@amcconsultants.com.au
2. MAusIMM, Principal Mining Engineer, AMC Consultants Pty Ltd,
Level 19, 114 William Street, Melbourne Vic 3000.
Email: myumlu@amcconsultants.com.au
3. MAusIMM, Senior Mining Engineer, AMC Consultants Pty Ltd,
Level 19, 114 William Street, Melbourne Vic 3000.
Email: menglish@amcconsultants.com.au
FIG 1 - Application of drift and fill to narrow vein mining.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
Current applications of DAF mining in the narrow vein
environment, shown in Figure 1, are in areas of local orebody
widening and areas of local orebody flattening.
The three key component capabilities required for DAF
mining are:
• full face development to a given profile,
• ground support including the application of fibrecrete, and
• cemented backfill.
The main characteristics of the DAF method when compared
to bulk underground mining methods are:
• the relatively high cost per tonne of ore mined;
• the relatively low overall rate of mine production;
• the relatively low productivity of the method per person, per
machine and per working area; and
• the high number of interdependent activities required to
complete each mining cycle.
It is important to note that DAF is an entry method which
requires particular attention to safety for personnel working
beneath wide spans of rock or backfill. The method does not
readily tolerate instances of poor quality backfill or poor quality
ground support.
It is likely that the abovementioned capabilities are found to be
practiced, or are able to be readily implemented, at most
underground mining operations. The DAF mining method could
therefore be considered for all or part of the mining solution in
many underground mining situations.
WHY CHOOSE DRIFT AND FILL?
The DAF mining method may be considered where bulk mining
methods are not considered to be viable and where one or more
of the following situations exist:
• Orebodies with very high value ore material and where the
objective is for high ore extraction with high selectivity and
control.
187
L DAWSON, M YUMLU and M ENGLISH
• Orebodies with poor or variable ground conditions with
limited stand-up times. The DAF method allows for small
openings with rapid placement of permanent support.
• Situations where there is a low tolerance for large voids with
the associated risk of subsidence.
• Occurrences of single or multiple orebodies with
unfavourable or variable geometries including low angle of
dip (below angle of repose) or pinching, swelling and rolling
lenses.
• Secondary mining situations involving the recovery of pillars
left by prior mining.
DRIFT AND FILL METHOD DESCRIPTION
The DAF method involves the excavation of ore by jumbo
development techniques and requires backfilling of the
completed drift prior to excavating an adjacent drift.
The use of cemented backfill with the DAF method allows the
full recovery of a mining block without the need to leave
permanent ore pillars. An exception might be in the mining of
remnant ore where narrow permanent ore pillars may be required
to be left against uncemented backfill.
There are a number of variants of the method including:
• overhand mining: using a bottom-up sequence,
• underhand mining: using a top-down sequence,
• mining multiple parallel (strike) drifts in moderate width or
flat-lying orebodies, and
• mining a transverse layout of drifts in wide orebodies.
DAF is a progression of the MCAF method that is used for the
single-pass mining of steeply dipping, narrow orebodies. In the
past, MCAF has been extensively applied for the mining of a
variety of orebody geometries and ground conditions, mostly prior
to the common acceptance of (longhole) open stoping and where
open stoping has been rejected. However, MCAF has not always
been able to provide a satisfactory solution to difficult mining
conditions and is generally not suited to secondary mining.
In large orebodies, MCAF is a partial mining method that
requires buttress pillars to provide stability and crown pillars to
vertically separate the work areas. MCAF can be mined
overhand or underhand in narrow orebodies and generally uses
flatback stripping for overhand mining. DAF may be mined
either overhand or underhand and may be mined without pillars
and generally uses full face development. DAF involves the use
FIG 2 - Overhand mining in a narrow orebody (section view).
188 Ballarat, Vic, 14 - 15 October 2008
of cemented backfill, except for situations where the backfill is
not planned to be mined beside or beneath, where some mullock
may be co-disposed. The underhand method may be preferred
where the orebody is weak and the use of engineered cemented
backfill provides a more stable back.
In steeply dipping, narrow orebodies, typically less than 8 m
wide, the MCAF method is commonly applied overhand, in a
bottom-up sequence, as shown in Figure 2.
The typical overhand mining sequence is:
1. A central ramp access is typically established from a
footwall ramp system and a ‘T’ intersection created in the
orebody.
2. The lateral drift is developed in two directions following
the strike of the orebody. The completion of the lateral
drifts to the full extent of mining on a level is termed a
‘lift’.
3. The first lift in a series of lifts is taken as full face
development with subsequent lifts able to be taken as
flatback stripping. A full face is typically taken up to 5 m × 5
m with sidewall stripping and the subsequent flatback lifts
up to 4 m high (allowing for approximately 1 m of void).
4. The face or stripping is blasted and the broken ore material
is loaded to a stockpile located at the access. Some ore loss
and dilution occurs when bogging the broken ore material
off the backfilled floor.
5. Primary ground support including rock bolts and fibrecrete
is installed sequentially with each cut taken. Thereby only a
minimal extent of unsupported ground is exposed at any
time in the mining cycle.
6. Secondary ground support, typically cable bolts, may be
installed with the first lift so that subsequent lifts may be
presupported in poor ground. Note that the cables will be
exposed and will need to be cut off for each subsequent lift.
Secondary support is typically installed in campaigns and
may need to be repeated on subsequent lifts.
7. On completion of the drifts in both directions, the full lift is
backfilled using either mullock as rock fill or hydraulic fill
or paste backfill. Tight backfilling is generally not required
and long lengths of drive may be backfilled in a single
campaign.
8. The roof of the access is then stripped and supported ready
for the next lift to commence.
Narrow Vein Mining Conference
 DRIFT AND FILL – THE HIGH VALUE, HIGH RECOVERY MINING SYSTEM

In moderate width orebodies, greater than about 5 to 8 m wide, that are only as wide as a single drift, and leaves no permanent
the MCAF and DAF methods are applied differently. The MCAF ore pillars, except where the ground is unmineable.
method must leave permanent ore pillars to maintain stability A typical overhand DAF mining sequence in a moderate width
across the large open spans. The DAF method creates open spans orebody is shown in Figure 3.

FIG 3 - Overhand mining sequence in moderate width orebody (section view).

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 189

L DAWSON, M YUMLU and M ENGLISH
The typical overhand DAF mining sequence in a moderate
width orebody is:
1. A central access is typically established from a footwall
ramp system and a cross-cut is developed to the furthest
contact, eg hanging wall.
2. The initial drift is developed along strike in two directions,
generally following a contact.
3. The drift is developed as a full face typically up to
5 m × 5 m.
4. The face is blasted and the broken ore material is loaded to
a stockpile located at the access. Some ore loss and dilution
occurs when bogging the broken ore material off the
backfilled floor.
5. Primary ground support including rock bolts and fibrecrete
is installed sequentially with each cut taken. Thereby only a
minimal extent of unsupported ground is exposed at any
time in the mining cycle.
6. Secondary ground support, typically cable bolts, is
generally required to be installed for the planned mining
span and may be installed with the first lift so that
subsequent lifts may be presupported in poor ground. Note
that the cables will be exposed and will need to be cut off
for each subsequent lift. Secondary support is typically
installed in campaigns and may need to be repeated on
subsequent lifts.
7. On completion of the initial drift in either direction, the
drift is prepared for placement of cemented backfill with
the objective of obtaining tight backfilling where
practicable with moderate strength backfill. For paste
backfill systems, barricades are constructed so that
backfilling may occur in up to 30 m segments retreating
from the end of the drift and back to the cross-cut.
8. Following completion of backfilling of the drifts in both
directions, and following the time for the cemented backfill
to gain sufficient strength, the parallel drifts may be
developed.
9. The drift and fill sequence is repeated as for above except
the parallel drifts will have one wall of exposed cemented
backfill. Depending on the width of the orebody, the
parallel drift may adjoin or may overlap the initial drift.
10. On completion of the mining of all the drifts in a lift, the
cross-cut is barricaded and backfilled. The roof of the
access is then stripped and supported ready for the next lift
to be mined.
Alternatively, the underhand mining system may be selected
where the ground conditions in the orebody are very poor and
FIG 4 - Transverse drift and fill mining in a wide orebody (plan view).
190 Ballarat, Vic, 14 - 15 October 2008
there is insufficient stand-up time for a drift to be developed and
adequately supported. Note that the orebody contact ground
conditions need to be suitable for anchoring the beam created by
the cemented backfill.
The typical underhand DAF mining sequence in a moderate
width orebody is the same as for overhand DAF except that:
• in step 4 the floor is in ore;
• in step 5 the back will initially be in ore and, in subsequent
underhand lifts, the backfill is designed to be self-supporting
and there is effectively no exposure of unsupported ground.
• in step 6 the secondary ground support, typically cable bolts,
may be required so that the crown pillar may be
presupported;
• in step 7 the drift is prepared for placement of cemented
backfill with the objective of forming a minimum thickness
beam of high strength backfill across the mining width; and
• in step 10 the floor of the access is then stripped ready for the
next lift to be mined.
In very wide spans, typically greater than about four drifts or
about 20 m wide, a change from a strike orientation of the ore
drifts to a transverse orientation of the ore drifts may be
considered in the DAF method, as shown in Figure 4.
The typical overhand transverse DAF mining sequence in a
wide orebody is the same as for overhand longitudinal DAF
except that:
• In step 2 the initial drift may be developed along the
centreline of the orebody.
• In step 7 on completion of the central drift in either direction
to the extent of the planned mining block, the central drift is
prepared for turnout drifts to be developed to either side. The
planned mining block is to be mined in a retreat sequence
back to the central cross-cut. The endmost pair of turnout
drifts are developed concurrently and evenly until completed.
• Backfilling is planned to occur in up to 30 m segments
retreating from the end of the turnout drift and back to the
central drift. A barricade may then be constructed across the
central drift or at the entry to each turnout drift.
• Following completion of backfilling of the pair of turnout
drifts back to the intersection, and following the time for the
cemented backfill to gain sufficient strength, the next pair of
turnout drifts may be developed.
• The drift and fill sequence is repeated as for above except the
parallel turnout drifts will have one wall of exposed
cemented backfill. Variations may occur where a temporary
pillar is created to allow multiple mining areas within a lift.
Narrow Vein Mining Conference
 DRIFT AND FILL – THE HIGH VALUE, HIGH RECOVERY MINING SYSTEM

• On completion of the mining and backfilling of all the
turnout drifts and the central drift in a lift, the cross-cut is
barricaded and backfill placed. The roof of the access is then
stripped and supported ready for the next lift to be mined.
The underhand method may also be applied to the transverse
mining system.
A comparison of some key aspects of the various MCAF and
DAF mining systems is presented in Table 1.
DRIFT AND FILL PERFORMANCE ASSESSMENT
The efficiency of the DAF method is determined at two levels:
• at the face (or micro) level for the interdependency of the
development, ground support and backfilling activities in
each heading being worked; and
• at the mine (or macro) level for the interdependency of the
individual work areas.
Production is limited from each work area and a high number
of work areas are required in order to generate a reasonable level
of ore production.
Maximum efficiency for an individual work area is only
achieved through close management of the interdependent
mining activities with the objective of minimising the face
standing (or non-work) time. The components of the microcycle
for DAF mining include development, ground support and
backfill activities. The mining activities comprise variable and
fixed duration components as indicated in Table 2. As much as is
practicable, the times required for travelling to the face and for
the extensions of mine services are to be concurrent with other
face activities.
As may be seen from Table 2, only a relatively small portion
of the cycle is variable according to face size with the majority of
time being associated with the fixed duration components of the
activities. Any reduction in the time required for the fixed
duration components of the cycle, for example by reducing the
backfill set time, is likely to deliver a significant benefit to the
total face cycle time.
Figure 5 shows a time-based analysis of the underhand DAF
mining system with minimum face standing (or non-work) time
where approximately half of the DAF mining cycle is used for
development related activities and half is taken up by backfilling,
secondary support and related activities. This analysis is valid for
a range of typical face sizes and results in an average of about
40 m per month advance being achieved in a typical DAF
heading. Therefore, for an individual heading, face size is a key
determinant of productivity in terms of tonnes mined per period
of time.
The face size is also a key determinant in equipment selection
which further impacts on the variable duration activities, for
example drilling, ground support installation and mucking.
The total mining cost for the DAF method includes variable
and fixed cost components for development, ground support and
backfill. The variable cost component is related to the energy and
materials used in each activity. The fixed cost component is
related to the equipment and personnel in the mine. The efficient
deployment of principally the mining jumbos is key to driving
the total unit mining cost to as low as practicable.
Analysis of the DAF microcycle demonstrates that a
development (bore and bolt) jumbo is utilised for only about
one-third of the development time or only one-sixth of the total
DAF cycle time. Therefore, depending on maintenance regimes

TABLE 1
Comparison of mining systems.

Recovered ore value
Mining width
Face type
Face size (w × h)
Orebody dip
Orebody ground conditions
Contact conditions
No of working faces
Dilution and ore loss
Typical backfill strength
UCS
Binder requirement
No of backfill barricades
Primary support
Secondary support
Overhand single drift Underhand single drift Overhand multiple drifts Underhand multiple drifts
(MCAF) (MCAF) (DAF) (DAF)
Moderate $200 per cubic High >$250 per cubic metre High >$250 per cubic metre High >$300 per cubic metre
metre
Narrow <8 m Narrow <5 m Moderate 5 to 20 m (or Moderate 5 to 20 m (or
greater for transverse) greater for transverse)
Full face or flatback Full face Full face Full face
3.5 m × 4.0 m up to 8.0 m 3.5 m × 4.0 m up to 5.0 m 4.0 m × 4.5 m up to 5.0 m 4.0 m × 4.0 m up to 5.0 m
× 5.0 m × 5.0 m × 5.0 m × 5.0 m
Flat to steeply inclined Moderate to steeply inclined Flat to steeply inclined Flat to steeply inclined
Poor to good Very poor to fair Fair to good Very poor to fair
Poor to good Fair to good (to anchor beam) Poor to good Fair to good (to anchor beam)
1 or 2 1 or 2 1 or 2 (or more for 1 or 2
transverse)
Likely dilution and ore loss Possible dilution from weak Possible dilution for Possible dilution for
in floor for uncemented rock wall. overlapping drifts. Possible overlapping drifts and from
backfill. ore loss for areas of very poor weak rock wall.
ground.
Uncemented or low High Moderate High
<0.5 MPa 1 to 4 MPa 0.5 to 1 MPa 1 to 4 MPa
Low High Moderate High
<75 kg/m3 OPC >150 kg/m3 OPC 75 to 150 kg/m3 OPC >150 kg/m3 OPC
1 per lift 1 per 30 m of drift 1 per 30 m of drift (and at 1 per 30 m of drift
each turnout for transverse)
Variable density bolts plus Standard bolts plus standard Variable density bolts plus Standard bolts plus standard
variable fibrecrete in rock. fibrecrete in backfill. variable fibrecrete in rock. fibrecrete in backfill.
Cable bolts may be required Cable bolts may be required Cable bolts generally Cable bolts may be required
in very poor ground. to anchor backfill beams required for wide mining to anchor backfill beams
laterally. spans and for very poor laterally and vertically.
ground.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 191

L DAWSON, M YUMLU and M ENGLISH

TABLE 2
The drift and fill microcycle.

Activity Variable durations Fixed durations

Blasting (None) Mine clearance for blasting, blast fume clearance and re-entry time
Mucking Muck material from face to stockpile or truck Wash down muck pile
Surface support Hydro-scale Set-up and preparation
Spray fibrecrete Fibrecrete set time
Inspection and testing
Bolt and face drilling Drill and install rock bolts Set-up and preparation
Drill blastholes Extend services
Drill probe holes if required Survey mark up
Face charging Charge blastholes Set-up and preparation
Connect to firing system
Secondary ground support Drill and install cable bolts Set-up and preparation
(campaign basis) Plate and tension cable bolts
Backfill (campaign basis) Backfill placement Site preparation
Barricade construction (allow two days)
Minimum set time for cemented backfill (allow three to seven days)

DAF Productivity Breakdown

• the ability to achieve the same performance duty at lower
Access Stripping,
Adjacent Drives
1%
rates of cement additions and at less curing time;
Delay, 10%
• efficient production and placement of backfill with minimal
involvement of additional personnel and equipment;
Post Paste Fill
Delay, 10% • lower total backfill production, delivery and placement cost;
• continuous placement with minimal excess water drainage or
backfill segregation, thereby providing a consistent and
Paste Fill, 4%
homogenous backfill quality;
Drift
Development,
• improved local and regional ground conditions due to the
48% ability to tight backfill and with reduced dilution when
mining against the backfill; and
Barricade
Construction, • easy to mine through paste with only light drilling and
10% blasting and easy to drill into and install rock bolts where
required to do so.
Install Services, 5% It is the authors’ experience that the DAF mining system
performs at its best when used in combination with paste
Cable Bolt, 12%
backfill. To the authors’ knowledge there are in excess of 65
paste plants currently in operation throughout the world and over
FIG 5 - Drift and fill productivity chart. half of these are being used with the DAF or MCAF mining
systems. The following section describes three overseas
examples in more detail.
and travelling times, a jumbo may be cycled between at least
four and up to six headings in order to obtain maximum
efficiency. DRIFT AND FILL CASE STUDIES
Additional work faces may be planned by: opening multiple
stoping blocks concurrently, for example each 30 m vertically Case 1 – overhand drift and fill mining
within a single orebody, or; opening multiple orebodies, or;
Case 1 is a one million tonnes per annum, underground base
opening multiple faces within a single stoping block, for example
by introducing temporary pillars as shown in transverse layout in metal mine where the mining takes place from between 100 m
Figure 4, or; providing multiple entry points along an orebody. and 500 m depth below surface. The high-grade orebody is
characterised by complex geology and orebody geometries with
variable ground conditions. The hanging wall and footwall
APPLICATION OF PASTE BACKFILL sequence generally consist of weak rock types. The orebody has
Cemented backfill is an essential part of the DAF mining system a variable dip of approximately 70° in the centre and upper parts
and a number of backfill types may be considered. When of the mine and 50° at the orebody fringes. The width of
integrated into the mining cycle the backfill serves as a working mineralisation varies from a few metres in the orebody
platform, as an engineered artificial back and as a wall to work extremities and up to 50 m in the centre. The two main mining
alongside. As a local and regional ground support it provides methods used for ore extraction are open stoping and DAF
confinement to ore pillars and the host rock and prevents mining. The mine utilises paste backfill as its primary backfill
unravelling, thereby limiting convergence and subsidence into method.
the mine voids. The DAF method is applied to the fringe areas having a
In practice, the dominant backfill systems comprise paste shallow dip less than 50° and to the areas which are difficult to
backfill, hydraulic backfill or rock based backfill types mine out by longhole stoping method due to poor ground
(cemented rock fill and cemented aggregate fill). However, paste conditions. DAF mining commences with a 5 m × 5 m drift
backfill offers the following benefits when compared to the other developed longitudinally along the hanging wall contact, from
backfill types: the access through to the planned limits of the orebody. After the

192 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 DRIFT AND FILL – THE HIGH VALUE, HIGH RECOVERY MINING SYSTEM
completion of mining, the hanging wall drift is prepared for
backfilling in typically 30 m segments retreating from the end of
the drift and back to the access.
In the first segment of drift to be backfilled, a fill line and air
breather line is installed at the highest point at the stope back and
a fibrecrete barricade is constructed to retain the backfill. The
cemented paste is pumped from surface through a network of
125 mm diameter boreholes and pipes at a rate of 50 cubic
metres per hour so that the segment of drift is backfilled in a
single campaign. The preparation and backfilling is repeated for
each segment along the drift until the drift is backfilled back to
the access. When backfilling is complete and after sufficient
curing time from four to seven days, a new drift is commenced to
be mined parallel and immediately adjacent to the first drift. This
process continues until the full lift is taken at that elevation.
Access to the next lift above is obtained by slashing the roof of
the access drive and orebody mining is recommenced. This
operation continues for the full height of the planned stoping
block, approximately 15 to 20 m in height, and up to the next
sublevel.
The paste backfill is designed for optimum cement content
where the bottom 3.5 m of the lift is backfilled using five per
cent by weight (100 kg/m3) cement addition and the top 1.5 m of
the lift is backfilled using nine per cent by weight (180 kg/m3)
cement addition to provide a backfill strength (UCS) of 500 to
1500 kPa. The paste backfill is designed to provide a good
working wall to minimise dilution and a good working floor to
improve the travelling surface when mining the next lift.
The main issues experienced in Case 1 are occasional air gaps
created on top of the backfill, and barricade stability. The air
gaps are dealt with by spraying fibrecrete into the voids that are
exposed during adjacent mining. Barricade stability is ensured
through strict adherence to backfill monitoring and placement
measures.
Case 2 – underhand drift and fill mining
Case 2 is a very high-grade, underground gold mine with a
production rate of about one-half million tonnes per annum. The
orebody dip is subvertical and has variable thickness. Both the
host rock and the orebody is characterised by weak rock types.
The primary mining method is by underhand DAF with paste
backfill. In a similar drift mining and backfilling system to that
described for Case 1, mining commences with a 4.5 m × 4.5 m
drift developed longitudinally along the line of the ore, from the
access through to the planned limits of the orebody. Mostly, the
full width of the ore is mined in a single 4.5 m wide pass. After
the completion of mining the drift is prepared for backfilling in
30 m segments retreating from the end of the drift and back to
the access.
In the first segment of drift to be backfilled, a sill mat
consisting of mesh and twin plated bolts is constructed at the
base of the drift to provide pre-reinforcement for the back of the
next lift below, a fill line and air breather line is installed at the
highest point at the stope back and a fibrecrete barricade is
constructed to retain the backfill. The cemented paste is gravity
fed from surface through a network of 150 mm diameter
boreholes and pipes at a rate of 120 cubic m per hour so that the
segment of drift is backfilled in a single campaign. The
preparation and backfilling is repeated for each segment along
the drift until the drift is backfilled back to the access.
The paste backfill uses very low density natural material and is
designed for optimum cement content depending on the
application using ten per cent by weight (80 kg/m3) cement
addition and up to 20 per cent by weight (160 kg/m3) cement
addition to provide a backfill strength (UCS) of 1000 to 2000 kPa.
The paste backfill and support system is designed to provide a
safe roof when mining the next lift below.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
After completion of each lift the floor of the access is benched
to gain access for the next lift in sequence. Mining of successive
lifts takes place under prereinforced paste backfill. The exposed
rock walls are systematically sprayed with fibrecrete and are
bolted. Depending on local conditions, additional fibrecrete and
rock bolts may be installed to the exposed paste backfill in the
backs.
The main issues experienced in Case 2 are related to the
additional effort required to install the prereinforcement that may
be overdesigned and has not been required at similar operations
elsewhere, and the occasional requirement to mine wider
orebody spans with parallel drifts.
Case 3 – primary and secondary drift and fill
mining
Case 3 is a high-grade, base metal mine with a production rate of
in excess of one million tonnes per annum. The orebody is
undulating and flat lying with dip of less than 20°. The host
rocks and orebody are competent but occasionally disturbed by
faulting. The thickness of the mineralisation varies from one to
14 m and lies at an average depth of 170 m below surface. The
primary mining method is DAF with some longhole bench
stoping in thick ore zones to obtain maximum ore extraction.
In the DAF mining area the entire thickness of mineralisation
is extracted between footwall and hanging wall in a single pass.
The ore is mined in primary and secondary drifts which are up
to 150 m long, 6 m high and 10 m wide. Upon completion,
fibrecrete barricades are constructed at the entrance of the drifts
for backfilling with paste backfill. The backfilling system is
able to use the dip of the orebody to advantage so that full
length of drive may be backfilled in a single campaign over
several days.
Paste backfill is produced at rate of 90 cubic metres per hour
and delivered to underground stopes by pumping through a
network of 200 mm diameter pipes. Strict backfill monitoring
and placement measures are in place to provide tight backfill and
to maintain barricade stability. Instrumentation is installed in
each barricade to monitor and control pressures acting on the
barricades.
The backfill is designed for optimum cement content
depending on the application using four per cent by weight
(65 kg/m3) to seven per cent by weight (110 kg/m3) cement for a
backfill strength (UCS) of 400 to 700 kPa. The paste backfill is
designed to minimise wall dilution and to assist the local and
regional stability. After minimum curing of 28 days the adjacent
secondary drifts are mined.
The main issue experienced in Case 3 is associated with the
management of excessive ground water inflows that potentially
impact on the backfill quality, curing time and barricade
pressures.
FUTURE DEVELOPMENTS FOR DRIFT AND FILL
The current high metal price environment has given rise to
investigations into recovering what was previously considered to
be uneconomic or difficult orebodies and a number of important
future developments for the DAF mining method are envisaged,
such as:
• widespread application of the DAF method to secondary and
tertiary mining of remnant pillars left by prior mining;
• improved application of development and ground support
systems in very poor ground;
• wider application of the paste backfill system as a practical
and economic alternative to hydraulic backfill or rock based
backfill systems; and
193
L DAWSON, M YUMLU and M ENGLISH
• the advancement of enabling mining technologies including:
high performance, small-scale mining equipment; remote
application of fibrecrete, and; improvements to the efficient
manufacture, distribution, monitoring and quality systems for
fibrecrete and cemented backfill.
Although there is currently only a limited operating experience
base in Australia, the authors believe that there are growing
applications and a likely successful future for the DAF mining
method, particularly in applications of high-value ore requiring a
high recovery mining system.
194 Ballarat, Vic, 14 - 15 October 2008
ACKNOWLEDGEMENT
The authors would like to acknowledge the contributions of a
number of colleagues at AMC Consultants Pty Ltd in providing
subject knowledge and review assistance for this paper.
REFERENCES
AMC Consultants Pty Ltd. Various unpublished AMC internal documents
and project reports.
Yumlu, M, 2001. Paper on Cayeli underground Cu-Zn Mine, in
Proceedings 17th International Mining Congress and Exhibition (ed:
E Unal) (Kozan: Ankara).
Narrow Vein Mining Conference
Resue Firing and Dilution Control in Narrow Vein Mining
M Tuck1

ABSTRACT mucking is then undertaken, then the remainder of the material is

blasted and mucked. This is a two stage process and as such
Mining of narrow vein mineralisation, invariably depending on the mining
method selected, involves the extraction of both ore and waste rock. The greatly increases the mining cycle time and therefore the cost of
level to which mining of waste rock can be minimised depends on the operation. Other problems also are possible with this method
numerous factors, but particularly the thickness and dip of the deposit. such as the potential for safety problems and geotechnical
Resue mining is a method of mining that allows split face firing within problems particularly in seismically active mining operations.
stopes to allow waste and ore to be mined separately, allowing dilution to The second type of resue operation is where all the rock is
be minimised. This paper examines the technique of resue firing and its blasted in one firing. In this case either the ore or waste are
applicability to the mining of narrow vein deposits. Particular emphasis is blasted during the initial stages of the firing and thrown further
placed on the conditions in which resue mining presents a suitable option
for mining of narrow veins.
down the drive, then the remainder of the drive is fired with less
throw. Thus ore and waste are separated and can be mucked out
separately utilising modern mining equipment and ensuring
INTRODUCTION workers always work under fully supported ground. This method
Narrow vein mining is a challenging mining process. Mining can be termed a simultaneous method of resue firing, or a
using narrow vein mining methods in the modern industry is method of firing which fires out the face in two distinct periods
highly equipment orientated but still offers the same challenges but within a single blast.
as the old miners faced; one of these that is a critical parameter is
dilution control (Dominy et al, 1998). In order to access narrow SIMULTANEOUS RESUE FIRING
vein deposits in order to extract the mineral it is generally
The method of simultaneous resue firing is best described with
required to make the drives of larger dimensions than the vein
reference to an example. Consider the situation as illustrated in
being mined. In some cases this can be achieved using
Figure 1, with a drive in a narrow vein mining operation where
specifically designed narrow profile mining equipment, for
there is a vein of ore diagonally across the drive, at this point it
example at Waterval in South Africa (Atlas Copco, 2006). In
should be noted that the orientation of the vein is not of
many cases where standard equipment is utilised the effects of
relevance. In a typical mining situation this drive would be fired
dilution can be considerable and have an impact on the
as a single continuous blast resulting in mixing of both the waste
economics of the mining operation. Particular problems
and ore which would need to be mucked out as a single product,
associated with excessive dilution in narrow vein operations
so resulting in dilution.
include:
• difficulty in maintaining a consistent head grade to mineral
processing operations;
• increased transport costs associated with transporting highly
diluted run-of-mine ore from the mine;
• geotechnical difficulties;
• difficulty in applying modern mining equipment, leading to
enhanced labour costs; and
• low production rates.
The above list is by no means exhaustive but indicates that for
profitable production a high degree of engineering and
geological skill, good communication between all professions at
the mine site and good planning are essential.
This paper discusses the method of resue mining, in particular
simultaneous firing resue methods to narrow vein mining.

RESUE MINING
Resue mining methods can be broadly defined as:
A method of stoping in narrow vein deposits
whereby the wall rock on one side of the vein is
blasted first and then the ore (RESolutions
Publishing and Media Pty Ltd, 2007).
Unfortunately very little work has been published on the
method, particularly in terms of modern literature.
From this definition two types of resue mining can result. In
most systems either the ore or waste rock are blasted first,
1. MAusIMM, Associate Professor of Mining Engineering, School of
Science and Engineering, University of Ballarat, PO Box 663,
Ballarat Vic 3353. Email: m.tuck@ballarat.edu.au
FIG 1 - Example drive cross-section.
In the method of simultaneous resue firing the drive would be
blasted as shown in Figure 2. Figure 2a shows the drive prior to
firing. In this case the blastholes within the vein are fired as per a
normal drive blasting operation using long period detonators
allowing the material from the cut and each sequential set of
holes to be fragmented and ejected a distance down the drive,
prior to the next set of holes being fired as illustrated in Figure 2b.
Once the ore has been blasted the remaining waste is blasted

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 195

M TUCK
Ore
Ore Waste
FIG 2 - Firing of drive.
using millisecond delays to provide minimal throw but to
fragment the ore and leave it at and/or close to the freshly blasted
face, as illustrated in Figure 2c. Use of perimeter control blasting
at this latter stage also ensures minimal damage to the
surrounding rock mass, enhancing the stability of the freshly
created drive.
As can be seen from Figure 2 a blast sequenced in such a way
would result in two separate zones of fragmented material, ore
thrown further down the drive and waste close to the blasted face.
Thus the two materials could be mucked separately with ore
being transported to surface for subsequent treatment and waste
transported to be used as waste fill underground as required. This
minimises haulage costs to surface. In reality it is unlikely that
the results of such blasting would result in two distinct zones of
material. There will be a zone where ore and waste are mixed
together; this would require a decision to be made regarding the
destination of this mixed and diluted material. It should be noted
that if this material were to form part of the ore stream the
amount of dilution would be lower than in the tradition blasting
method.
PRACTICAL APPLICATION OF THE METHOD
As part of a final year student project (Place, 2007) the method
of simultaneous resue firing was to be trialled in an ore drive at
Tom’s Gully mine in the Northern Territory in early 2007. The
aim of the study was to trial the method to reduce dilution from
the underground mining operation. A number of firing options
using a variety of long period and millisecond delay periods were
outlined and being prepared to be implemented. Unfortunately
due to a significant rainy season the underground operation
flooded and the method could not be tested. In the proposed
study the waste was to be blasted first and after a period of two
seconds the ore was to be blasted, the reverse to the situation
illustrated in Figure 2. A particular issue at this mine was the
competence of the waste rock hosting the gold veins and a key
parameter of the investigation was to determine the amount of
waste to leave in place to support the ore during the two second
waiting period between the sections of the blast for the method to
work effectively.
It is hoped the method can be trialled at another mine in the
near future. As such it is very difficult to compare the resue
method with the conventional method particularly from a cost
196 Ballarat, Vic, 14 - 15 October 2008
perspective, such as bottom line benefits of using the method and
comparative costings. The method must also not be looked at in
isolation; rock breakage is a part of a whole chain of rock
processing operations, all of which need to be considered. As an
example the final destination of the waste, either carted to
surface or used as backfill underground would have different
impacts on the overall variable costs of a mining operation.
SUITABLE CONDITIONS FOR SIMULTANEOUS
RESUE FIRING
From the preceding discussion it can be seen that simultaneous
resue firing offers a potential method of mining narrow vein
deposits by, for example, a cut-and-fill mining method. It also
offers the potential to apply standard mechanised mining
equipment and to reduce dilution of the ore stream coming from
the stopes. The question arises under what conditions this mining
method could be potentially selected. It should be noted that this
mining method should only be selected if it provides a more
economic method of mining for a given geometry and set of
actual mining conditions.
Some of the conditions favouring selection of this method are
listed below:
• Well constrained vein type deposit with good definition of
both the hanging and footwall contacts.
• All dips of vein can be mined.
• Suitable for mining variable width veins.
• Good viable differentiation between ore and waste materials
would be preferred.
• Geotechnically similar and dissimilar ore and waste can be
mined. Where geotechnical competency is dissimilar account
must be taken of the need to leave supporting waste for the
ore during the rest period between the long period and
millisecond detonators if the ore is fired second.
CONCLUSIONS
Simultaneous resue firing offers the potential to reduce dilution
in narrow vein mining operations as historically standard resue
firing has. It also allows standard mechanised mining equipment
Narrow Vein Mining Conference
 RESUE FIRING AND DILUTION CONTROL IN NARROW VEIN MINING

to be utilised without the need for personnel to work under
unsupported rock. The method is to date untested and will
require further refining on a site by site basis before becoming an
accepted narrow vein mining method. It is highly suited to cut-
and-fill mechanised mining in narrow veins but could also be
used in room and pillar operations.
REFERENCES
Atlas Copco, 2006. Low seam high profit – Experience and knowledge
[CD ROM].
Dominy, S, Camm, G, Phelps, R and Sangster, C, 1998. Narrow vein
stoping practices: A United Kingdom perspective, Transactions of
the Institutions of Mining and Metallurgy, Mining Technology,
A222-A234.
Place, M D, 2007. Ore separation in narrow veined ore deposits, using/
utilising resue (split face) firing, to cut dilution, BEngSci thesis
(unpublished), University of Ballarat.
RESolutions Publishing and Media Pty Ltd, 2007. The Mining Glossary:
The Essential Handbook of Mining Terminology, 144 p (RESolutions:
Perth).

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 197

The Development and Implementation of a Fully Remote Stoping
Method at Beaconsfield Gold Mine, Tasmania
P B Hills1, J Mills2, A R Penney3 and S Arthur4
ABSTRACT
A key aspect of seismic management for narrow vein stoping of the west
zone at the Beaconsfield Gold Mine has been the development of a fully
remote extraction method that does not require personnel access to sill
drives within the orebody once stoping commences.
Initially, sill driving within the reef is undertaken in a conventional
fashion under geological control. Sill drives are developed at a vertical
backs-to-floor height ranging from 7 - 12 m. This mining stage includes
all activities prior to the commencement of stoping, after which no
personnel enter the unstoped section of the sill drive at any time for any
purpose.
The key aspect of the remote mining cycle is the use of footwall drill
drives paralleling the ore drive at a minimum separation of 6 m under
survey control. These drives provide the platform for fan drilling of the
orebody. Up to 20 m strike panels are blasted in a single shot using
electronic blasting techniques.
Ore is mucked from the extraction level sill drive using tele-remote
loaders. Aerials for tele-remote operation are installed through boreholes
from the footwall drive, negating the need for personnel access to the sill
drive. Once mucking of a stoping panel is complete, a laser scanning
survey instrument is used to survey the final void for subsequent design
and production reconciliation purposes. The survey instrument is
mounted in the bucket of the tele-remote loader, and is also operated
remotely.
Finally, tele-remote loaders are used to dump waste rock fill from the
upper sill drive. A system of flags and cameras installed in boreholes
from the footwall drive assists the operator in locating the live stope edge
during the filling cycle, again negating the need for personnel access.
BACKGROUND
The onset of mining-induced seismicity at the Beaconsfield Gold
Mine occurred during primary sill driving within the orebody at a
depth of 760 m in 2002. Initially, seismicity was manifest as
minor strainburst events with the occasional ejection of small
rocks from the hanging wall. Around that time, the first micro
seismic events were also being reported in decline development
up to 50 m in the footwall of the orebody.
Seismic management tools were put in place throughout 2003
and subsequently, as described by Hills and Penney (2008).
Seismicity peaked initially with a local magnitude (ML) 2.1 event
in the western zone of the mine in October 2005. This event
resulted in a large rockfall in a sill drive, which at the time was
closed to personnel access due to the level of seismicity being
experienced. All aspects of seismic management then in place
were re-examined and peer reviewed prior to production
resuming over a three-month period.
It was recognised that seismicity would continue once mining
resumed in early 2006, and further management strategies were
1. FAusIMM, Technical Services Manager, Allstate Explorations NL,
PO Box 58, Beaconsfield Tas 7270.
Email: PeterH@allstateexp.com.au
2. MAusIMM, Underground Operations Manager, Allstate Explorations
NL, PO Box 58, Beaconsfield Tas 7270.
Email: JasonM@allstateexp.com.au
3. MAusIMM, Geotechnical Geologist, Allstate Explorations NL, PO
Box 58, Beaconsfield Tas 7270. Email: AdrianP@allstateexp.com.au
4. Senior Surveyor, Allstate Explorations NL, PO Box 58, Beaconsfield
Tas 7270. Email: SimonA@allstateexp.com.au
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
put in place with the aim of mitigating the associated risk.
However, a fatal rockfall incident resulting from a ML 2.3 event
during the evening of Anzac Day 2006 and the ensuing two week
rescue of two trapped workers resulted in a further extensive
review being undertaken at a hitherto unprecedented level.
The requirement of a multifaceted risk review of the entire
mining operation at Beaconsfield under the umbrella of a Case
for Safety process imposed on the operation by Workplace
Standards Tasmania involved the engagement of consultants and
peer reviewers to develop an approach to mining that reduced
risk to personnel in line with the as-low-as-reasonably-practical
(ALARP) principle. Many risks were found to be adequately
dealt with; many others, though well considered and/or of
generally low tenor, were further mitigated. However, it was
ultimately determined that the residual risk of any stoping
method that required personnel access to the orebody sill drives
in the Western Zone of the mine once stoping within a given ore
block had commenced, was not acceptable.
GEOLOGY AND SETTING
The Beaconsfield Gold Mine is centred on the Tasmania Reef, a
steeply-dipping tabular quartz-carbonate vein hosted by a
sequence of siliciclastic sediments. The sediments, which dip
towards the east, comprise the Salisbury Hill Formation of
conglomerates and sandstones overlain by the Eaglehawk Gully
Formation of sandstones, siltstones and limestones with
occasional pebble bands. All siliciclastic rocks have been
metamorphosed to quartzites although they retain much of their
sedimentary character. The Tasmania Reef is essentially planar
over a strike length of 350 - 400 m, with an average width of
2.5 m and occupies a shear which cuts across stratigraphy
orthogonally. The west zone of the Tasmania Reef is hosted by
the Salisbury Hill Formation, and it is the brittle nature of the
conglomerate horizons within that formation in particular, which
are prone to mining-induced seismicity. The geology of the mine
is discussed by Hills et al (2001).
GEOTECHNICAL DESIGN
Scott, Penney and Fuller (2008) examine the complex
geotechnical design process required in the west zone (Figure 1).
That process is the cornerstone of support selection to adequately
address the requirements of the Case for Safety, which has seen
production resume in the west zone of the Beaconsfield Gold
Mine following the Anzac Day 2006 rockfall. Based on the
principles described in the Canadian Rockburst Handbook
(Kaiser, McCreath and Tannant, 1996), summarised by Reeves
(2008) and peer reviewed by Kaiser (2008), the process has been
rigorously applied in developing the mining method described in
this paper. The reader is directed to Scott, Penney and Fuller
(2008) for a detailed discussion of the geotechnical design
process.
MONITORING, MEASURING, MODELLING AND
INSTRUMENTATION
Implicit in the Case for Safety design process is the ability to
monitor and thus manage both seismicity and the response of
the rock mass to mining. This has been achieved through a
199
P B HILLS et al
FIG 1 - Long section of the Beaconsfield Gold Mine illustrating the
boundary between the west and east zones.
combination of seismic monitoring and analysis, in situ stress
measurement, rock material property measurement, numerical
modelling and rock mass instrumentation.
An extensive underground seismic array was first installed at
Beaconsfield in mid 2005 after a temporary system had been in
operation for a period of 18 months. The permanent array was
extended in 2007 and again in 2008 to comprise 12 uniaxial and
four triaxial geophones covering the full extent of the current
workings, but particularly the western area of the mine below
700 mL. A surface array was installed in 2007, primarily to
monitor the effect of seismicity on residential amenity in
Beaconsfield. However, it too can provide data to assist with
management of the underground environment. A detail synopsis
of seismic management at Beaconsfield was recently provided by
Hills and Penney (2008) and is beyond the scope of the current
paper.
Stress measurements have been undertaken in three campaigns
in 2003, 2006 and 2008 and augmented by material property
measurements undertaken in 1997, 2004, 2007 and 2008.
Numerical modelling has been used as a tool at Beaconsfield
since 2003, and has been a cornerstone step in mine planning,
sequencing, scheduling and support design since the
implementation of the Case for Safety process. This work was
discussed briefly by Hills and Penney (2008).
Monitoring of the behaviour of the rock mass in response to
mining and seismicity through instrumentation is also beyond the
scope of this paper, and the reader is directed to Penney, Hills
and Walton (2008) for a detailed discussion.
FIG 2 - Modified Avoca stoping sequence after two panels in the second horizontal lift (light grey = cemented fill, dark grey = waste rock
backfill, white = void).
200 Ballarat, Vic, 14 - 15 October 2008
MINING METHOD
The mining method at Beaconsfield Gold Mine evolved over a
number of years. Prior to April 2006 most production was
achieved by the modified Avoca method whereby uphole bench
stopes are extracted from a series of sill drives. The sills were
developed along the strike of the orebody under geological
control. Typically, stoping panels would comprise three sill
drives, and three stoping lifts would be extracted per stoping
panel. An uphole bench was extracted from the lowest sill drive
to facilitate the construction of a cemented fill pillar to allow
close to 100 per cent extraction. Uphole stoping would retreat
along strike for some arbitrary distance, and ore was extracted
via the lower sill drive using line-of-sight remote loaders. Once
the hanging wall started to fail, stoping would cease and the
stope would be backfilled from the upper sill drive. A modified
Avoca style firing against the backfill would then recommence
the stoping cycle. A full stoping lift would be completed prior to
the commencement of the second lift (Figure 2).
From late 2005, a modification to the stoping method was
implemented with the introduction of ‘checkerboarding’ (Turner,
2005). The modification was a response to increased seismicity
in successive lifts within each stoping panel, designed primarily
to minimise the impact of seismicity by limiting the strike length
of remnant crown pillars. The checkerboard method limited
the strike length of the modified Avoca stope to 20 m. After
backfilling, stoping continued vertically, including crown
extraction, prior to advancing the strike laterally. The
checkerboard stoping method is illustrated in Figure 3.
REMOTE MINE DESIGN
Introduction
Incumbent in the Case for Safety process implemented at the
Beaconsfield Gold Mine from 2006 was a requirement to
eliminate or minimise seismicity to a level that is as low as
reasonably practicable, determine the most appropriate ground
support regimes and development and production methods, and
adequately assess the risks to the health and safety of all
personnel. The approach taken was to adopt the ‘as low as
reasonably practicable’ (ALARP) principle to manage the risks
involved. Scott and Reeves (2007) used numerical modelling and
analysis of the extant seismic record to demonstrate that seismic
events up to local magnitude (ML) 2.5 could reasonably be
expected to occur as stoping proceeded. Work by Kaiser (2006)
had demonstrated that no known ground support system could be
adapted for use in the orebody, using accepted engineering
principles, which would survive such an event with an acceptable
factor of safety. Consequently, it was the overwhelming
conclusion of all parties involved in the development of the Case
Narrow Vein Mining Conference
 DEVELOPMENT AND IMPLEMENTATION OF A FULLY REMOTE STOPING METHOD AT BEACONSFIELD GOLD MINE, TASMANIA
FIG 3 - Checkerboard stoping sequence after two panels in the second vertical lift (light grey = cemented fill, dark grey = waste rock
backfill, white = void).
for Safety, that the west zone of the mine could not be stoped
from inside the sill drives. That is, no man access to the sill drive
could be permitted at any time, for any purpose, once stoping
within a given western mining block had commenced.
Studies completed during the Case for Safety process had
already shown that mining of capital and operating access
development (Pfitzner, 2006), sill driving (Sidea, Scott and
Reeves, 2007) and stoping in the generally aseismic east zone of
the mine (King, Thomas and Scott, 2007) could be undertaken
using previous methods with relatively minor changes. The
adoption of a remote stoping method in the west zone
necessitated consideration of all activities which would have
previously required access to the unstoped section of the sill
drives once stoping had commenced. Primarily, those activities
included:
• drilling of blastholes,
• charging,
• line-of-sight remote mucking,
• geological inspection and sign-off,
• stope cavity surveying, and
• placement of backfill.
The key requirement was the development of techniques to
achieve each of the listed activities remotely, and that required
the development of a suitable platform from which to conduct
those remote activities. Footwall drives were considered for the
purpose, and their use was incorporated in the geotechnical
modelling and analysis conducted for the Case for Safety (Scott
and Reeves, 2007). Ultimately, footwall drives were designed to
parallel the sill drives at a minimum separation pillar thickness of
6 m under survey control. Generally, the footwall drives were to
be designed at the horizon of the sill drive such that they were in
part shadowed from induced stress concentration in unmined
stope blocks. The location also facilitated the establishment of
accesses or drawpoints for remote equipment (Figure 4).
Footwall drives were deemed to be within the seismic hazard
envelope defined by the orebody, and consequently mining
practice, ground support and exposure protocols were also
developed within and as an adjunct to the Case for Safety. The
checkerboard extraction sequence proposed by Turner (2005)
remained the preferred sequence for ongoing stoping for the
same reasons that Turner had espoused. The key parameter of the
checkerboard sequence was the development of an essentially
vertical mining front within a given stoping panel as opposed to
the essentially horizontal mining front previously adopted. The
role of the footwall drives in the checkerboard stoping cycle is
discussed in the following sections.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
Orebody East Zone
West Zone
Stope
Sill Drive
Footwall
Drive
No entry to stope or red sill drives
once production commences.
FIG 4 - Schematic diagram illustrating the remote mining concept
for the west zone of the Beaconsfield Gold Mine.
Drill and blast
Blastholes in this new ‘western method’ were conceived as a
series of fan rings drilled from a single pivot point in the footwall
drives. Usually, the holes were drilled as upholes into the stope
panel above (Figure 5). This had the advantage that:
• holes were as short as possible,
• holes were close to perpendicular to the plane of the orebody
and parallel to the back of the sill drive, and
• holes were easily prepared for subsequent charging.
In reality, it has occasionally proven necessary (eg at the
location of access drawpoints) to drill downholes from the
footwall drive above the stope panel. This necessitates longer
holes drilled subparallel to the dip of the orebody. These holes
tend to droop and experience deviation, so consequently a
downhole survey tool has been adapted to allow hole paths to be
plotted. Based on the analysis of hole survey data some holes are
left uncharged, and occasionally additional holes are drilled.
Critical to the drilling cycle is hole accuracy and a high degree
of azimuth control is required. A longhole drill rig fitted with
column lasers to minimise set-up errors was selected to
undertake the task. The drilling pattern is virtually continuous
along strike and an entire series of rings (up to 150 rings in ten
stoping panels) is drilled consistently on a north/south orientation
to simplify set-up. This allows reference between, say, backfilling
activities on one level and drilling on another by simply referring
to a consistent ring number, which can be converted to an
easting. To cope with variation in the strike of the orebody (and
201
P B HILLS et al
FIG 5 - Detailed arrangement for remote stoping of the west zone.
hence the orientation of the footwall drive) the drill rig was also
selected for its second articulation point at the drilling table.
Thus the rings can be oriented north-south even though the strike
of the orebody may vary by up to 20°. Without this feature of the
drill rig, the ring set-ups would be quite onerous.
No slot rise is drilled. Each stope block is blasted in a single
shot using electronic detonators. The delay interval is kept low,
typically around 30 milliseconds, and a low instantaneous charge
mass is utilised to minimise p and s wave superposition. A total
blast period of around 800 milliseconds is achieved. An inverted
‘V’ pattern blast utilises the back of the sill drive as the free face
(Figure 6). To date, stope blasts of up to 70 holes in ten rings
have been successfully fired.
Typically, blasthole lengths range up to as much as 12 - 15 m.
Only that portion of the hole within the orebody is charged, and a
minimum uncharged collar length of 6 m remains. It is essential
that blast damage to the footwall drive is limited to the extent
possible, and satisfactory results have been achieved with full
column sand stemming.
A re-entry exclusion period of 24 hours is imposed after each
stope blast prior to personnel access is permitted to any footwall
drive or associated development within the stoping panel where
the blast occurred.
Mucking
Proven tele-remote loader technology is used for stope mucking.
An operator station is established in a suitable location to service
mucking from all three levels within a given stoping block.
The station is located outside the re-entry exclusion zone, and
therefore mucking is able to commence immediately after
blasting. Loader access to the sill drive occurs through
occasional drawpoints mined from the footwall drive. The
drawpoints themselves are non-entry areas for personnel, and
access is blocked by a mullock bund when they are not in use.
Ore is mucked to a stockpile or ore pass attached to the footwall
drive for later manual loading to trucks. Stockpile and drawpoint
202 Ballarat, Vic, 14 - 15 October 2008
location for both loading of ore and back loading of waste is a
key driver of productivity the loaders currently employed are
locked to first gear while remote mucking. Tramming distances
beyond 100 m severely erode productivity and are avoided where
possible.
The tele-remote operations at Beaconsfield Gold Mine have
solved the issue of ensuring radio coverage within a remote
orebody. Directional ‘Yagi’ type aerials are installed in the
footwall drive where access is possible. Typical aerial separation
of 50 - 70 m is required. Aerial coverage in the sill drives
presented a greater challenge as access for installation was not
possible. An omni-directional aerial capable of being installed
(and retrieved) through a 76 mm borehole from the footwall
drive was developed locally. These aerials have a typical range of
20 m (Figure 7).
Surveying
The key challenge for surveying was to develop a technique
whereby accurate and reliable survey data could be obtained
from the sill drives and completed stopes remotely. Accurate
survey information is vital for physical control of the stoping
operation, allowing for strategic set-up for stope backfilling and
the design of subsequent production blasts. It also fulfils a
traditional quality control in production reconciliation, grade
control and dilution monitoring. In essence, a Cavity
Measurement Survey (CMS) was required without the surveyor
having any personal access to the stope.
The system devised at Beaconsfield was to use a remotely
operated laser scanner mounted in the bucket of a tele-remote
loader. A purpose-designed frame is bolted into the bucket of the
loader on each occasion that a survey is to be undertaken
(Figure 8a). The surveyor can watch the progress of the loader
entering the stope on the VDU screen at the tele-remote operator
station. Once in position, the scan is initiated, and takes only a
few minutes to complete. The loader can be repositioned to allow
further scans to be completed if necessary.
Narrow Vein Mining Conference
 DEVELOPMENT AND IMPLEMENTATION OF A FULLY REMOTE STOPING METHOD AT BEACONSFIELD GOLD MINE, TASMANIA
FIG 6 - Typical charge plan for a west zone stope.
FIG 7 - Omni-directional borehole aerial for tele-remote mucking.
The Faro laser scanner is controlled by the surveyor at the
tele-remote station via a laptop computer. The communication
from the laptop to the scanner is established via a cat five cable
and two Rajant ‘Breadcrumb® SE’ units linked by Wi-Fi
connection (Figure 8b). One ‘Breadcrumb® SE’ is mounted in the
tele-remote loader and the other is placed in the footwall
drive adjacent to the stope An antenna is connected to the
‘Breadcrumb® SE’ in the footwall drive and installed through a
borehole to enable the Wi-Fi signal to be established (Figure 8c).
On completion of the scan, the surveyor is able to view the
results at the job to ensure that sufficient detail is captured.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
The raw scan, although completed in darkness, is equivalent to
a high-quality black and white photograph (Figure 9a). The
surveyor is able to identify survey stations and other known
landmarks established during the sill driving stage of the
production cycle (Figure 9b), and rectify the scan image to true
three-dimensional space. Using Faro Scene software a mesh of
the scans taken in the stope is generated and exported as a DXF
file for conversion to Surpac format. Slices as per a CMS pick up
are then generated (Figure 10). The result is a stope pick-up as
accurate as a conventional CMS cavity survey, which provides all
the information required.
Backfilling
Mullock backfilling in the checkerboard pattern is also achieved
using tele-remote loaders. The same aerial configuration is used
to backfill a given stope block from the upper sill horizon.
Mullock is delivered to the footwall drive via a pass, and filling
is able to proceed continuously. The greatest exposure for
tele-remote backfill is being able to consistently locate the
tipping point at the edge of the stope. This is managed by placing
a flag attached to a length of conduit via a borehole or (flag hole)
to signal the edge position to the operator, combined with a
stationary camera installed through a reamed out borehole
(camera hole) from the footwall drive and directed towards the
live edge. The camera is fitted with its own light source and the
view is available to the tele-remote operator at the tele-remote
station on a separate monitor (Figure 11). Initially the flag and
camera positions are located by analysing the CMS data, which
defines the position of the live edge. The camera and flag are
then repositioned as required to keep track of the advancing
tip point. Typically, camera holes are located in every second
blasthole ring or approximately every 3 m of strike length.
203
P B HILLS et al

A B C

®
FIG 8 - (A) The laser scanner mounted in the tele-remote loader bucket and attached to the ‘Breadcrumb SE’ for Wi-Fi communication.
(B) The second ‘Breadcrumb® SE’ mounted on the wall of the footwall drive. (C) The Cat 5 antenna installed in a borehole between the
footwall drive and the open stope.

FIG 9a - The raw black and white laser scan image. Note that this FIG 9b - Known survey locations are used to rectify the raw scan
image wraps a full 360°. image to 3D space.

FIG 10 - The rectified laser scan is used to generate a Surpac string file, which can be used for subsequent stope design and production
reconciliation.

CONCLUSIONS those which are commonplace within the modern mining

Overall, successful modifications to the checkerboard stoping
sequence have been devised to permit a totally non-entry stoping
method to be developed for the west zone of the Beaconsfield
Gold Mine. The key elements of the mining method range from
industry, to those which are innovative and in some cases unique
adaptations of available technology.
Tele-remote mucking operations are commonplace in mines
throughout the world. Drilling from remote drill galleries is
likewise commonplace. The total absence of personnel is where

204 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 DEVELOPMENT AND IMPLEMENTATION OF A FULLY REMOTE STOPING METHOD AT BEACONSFIELD GOLD MINE, TASMANIA
FIG 11 - Tele-remote set-up for backfilling illustrating the flag and
camera holes (camera inset left, tele-remote monitors inset right).
the Beaconsfield mining method diverges for other contemporary
operations. Implicit in the mining method is the tenet that there is
no point in the stoping cycle at which access to the stope
threshold, nor indeed the associated access sill drive, meets the
ALARP criteria for man access. Therefore, tele-remote
installation, all aspects of production drilling and charging,
excavation surveying and stope backfilling with waste rock must
also be conducted remotely. The latter are the innovative aspect
of the mining method adopted for the west zone of the
Beaconsfield Gold Mine in the aftermath of the Anzac Day 2006
tragedy.
ACKNOWLEDGEMENTS
The authors acknowledge the management of Beaconsfield Gold
NL for permission to publish this paper.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
REFERENCES
Hills, P B, MacDonald, G, Nero, L and Mueller, A M, 2001. Beaconsfield
Gold Mine – Tasmania Reef, in Structure and Setting of Proterozoic
and Palaeozoic Rocks in the Tamar Region, Northern Tasmania (ed:
A R Reed), Excursion field guide, No 9, pp 76-88 (Geological
Society of Australia Specialist Group in Tectonics and Structural
Geology)
Hills, P B and Penney, A R, 2008. Management of seismicity at the
Beaconsfield Gold Mine, Tasmania in Proceedings Tenth AusIMM
Underground Operators’ Conference 2008, pp 157-170 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Kaiser, P K, 2006. Strainburst analysis for ore developments –
Beaconsfield Mine, MIRARCO – Mining Innovation, Sudbury.
Kaiser, P K, 2008. Geomechanics and ground support review –
Beaconsfield Gold Mine, MIRARCO – Mining Innovation, Sudbury.
Kaiser, P K, McCreath, D R and Tannant, D D, 1996. Canadian
Rockburst Handbook, Geomechanics Research Program (CAMIRO
Mining Division: Sudbury).
King, R, Thomas, S and Scott, C, 2007. Geotechnical assessment –
Resumption of ore production (east zone), Coffey Mining Pty Ltd
report no HZ00054.01.BH.
Penney, A R, Hills, P B and Walton, R J, 2008. The use of geotechnical
instrumentation to optimise an engineered mine design at
Beaconsfield Gold Mine, Tasmania, in Proceedings Narrow Vein
Mining Conference, pp 165-172 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Pfitzner, M, 2006 Geotechnical assessment – Resumption of decline
developments, Coffey Mining Pty Ltd report no HZ00054.01.A1.
Reeves, I, 2008. Procedure for support selection, Coffey Mining Pty Ltd
report no HZ00054.01.CC.
Scott, C, Penney, A R and Fuller, P, 2008. Competing factors in support
selection for the west zone of the Beaconsfield Gold Mine,
Tasmania, in Proceedings Narrow Vein Mining Conference,
pp 173-178 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Scott, C and Reeves, I, 2007. Geotechnical assessment – Resumption of
ore production (west zone), Coffey Mining Pty Ltd report
no HZ00054.01.BY.
Sidea, D, Scott, C and Reeves, I, 2007. Geotechnical assessment –
Resumption of ore driving development, Coffey Mining Pty Ltd
report no HZ00054.01.AT.
Turner, M H, 2005. Beaconsfield site visit February 2005, AMC
Consultants Pty Ltd, 31 March 2004.
205
Wallaby – The Horizontally Challenged Opportunity
N S Rauert1

ABSTRACT
Barrick Gold Granny Smith Mine is currently undergoing conversion
from an open cut operation to an underground mine at the Wallaby
deposit. The underground project commenced in 2005 with the
development of the reserves immediately below the pit and is now
currently entering a production ramp-up phase for mining of these two
initial surfaces below the pit. With increasing gold prices the owner has
been forced to re-evaluate the underground project and has decided to
investigate the potential of a deep level mine.
This deep level underground mine presents a collection of challenges
with the most notable being exploitation of a series of well separated near
horizontal lodes ranging in thickness from less than 1 to 20 m. A variety
of room and pillar mining methods are employed for extraction with
careful consideration to resource recovery and pillar design for long-term
stability. Compounding these deep level challenges is a need to accelerate
mine level development to maintain the production profile of the mine.
This paper describes the challenges, key learnings and outcomes of the
studies undertaken to date associated with this project.

INTRODUCTION
The Granny Smith Mine, named after a soon to be grandmother,
not the apple, has been in operation since 1989 mining gold from
medium-sized satellite open pits in its northern and southern
lease extents. The mine is located in the Mount Margaret Mineral
Field some 25 km south of the town of Laverton, and about
700 km east-north-east of Perth, Western Australia (refer to
Figure 1).
The Wallaby deposit, discovered in 1998, lies approximately
11 km south-west of the Granny Smith mine processing plant on
the north-eastern shore of Lake Carey salt lake. Wallaby is the
most recent deposit mined at Granny Smith, with several older
open pit deposits mined towards the northern extent of leases.
The original owner of the lease containing Wallaby was a joint
venture between Placer Exploration Ltd and Delta Gold. Placer
Dome acquired complete ownership but in 2006 with the
acquisition of Placer Dome by Barrick Gold, Barrick Gold now FIG 1 - Barrick Gold Granny Smith Mine location.
holds full ownership. Barrick Gold operates the Granny Smith
site as part of its Yilgarn operation, which also includes the
Lawless, Darlot and eventually Plutonic mines. Much of the which is unbedded and intruded by a suite of fractionated
function of these mines including administration and mine alkaline dykes. The rocks that host the Wallaby deposit are part
technical areas is carried out at the Yilgarn shared services group of the eastern limb of the Margaret Anticline with the sediments
office in Perth to which the author is a member. The orebodies dipping moderately to the south-east.
associated with the mine were initially extracted by a series of Gold mineralisation is thought to be wholly contained within
medium-sized open pits. The last of these pits to be mined was the actinolite-magnetite alteration pipe and cross-cuts the
the Wallaby with a transition to underground mines occurring conglomerate overprinting the actinolite-magnetite alteration.
from 2004. In this respect the Wallaby underground deposit is The gold bearing lodes form a stacked series of ore zones with
unique due to its tabular nature. This paper summarises the
two types of lodes:
underground project to date, outlining the geologically and
geotechnically unique challenges in mining method selection and 1. thick subhorizontal to gently NE-dipping lodes; and
application. 2. narrower, moderately NE-dipping lodes with up to
40 degree dips (refer to Figure 2).
Geology
Four dominant gold related structural settings, termed
The Wallaby gold deposit is located in the southern portion of domains, have been identified in the underground exposures from
the Lancefield-Wallaby Basin, which is a tectonically late, work by (Miller, 2005), which have been termed as follows:
fault-bounded basin. The deposit as described by Placer Dome • Fracture mesh controlled gold lodes. These are intersections
(2005) is hosted by a thick matrix-supported mafic conglomerate, of brittle structures (faults and veins), dip and strike changes
along structures, and gold mineralisation linked to the
development of laminated and quartz breccia veins.
1. MAusIMM, Senior Long Term Planning Engineer and Granny Smith
-PFS Project Lead, Barrick Gold, Level 4, 200 Adelaide Terrace, • Horsetail domain gold lodes – fracture mesh developed
Perth WA 6000. Email: nrauert@barrick.com where the lode fault diverged into multiple faults.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 207

N S RAUERT

TABLE 1
Wallaby JORC compliant resource and reserves as at
December 2007.

Resource Ore t Au g/t Au oz

Measured 1 354 545 6.91 300 839
Indicated 4 650 678 6.31 943 154
Inferred 7 197 515 7.63 1 764 848
Total resource 13 202 738 7.09 3 008 841
Reserves
Proven 635 668 5.12 104 664
Probable 2 116 170 4.96 337 480
Total reserves 2 751 838 5.00 442 144

Accurate estimates of how much of the total gold resource

falls into the sub 2 m ‘narrow vein’ lode width are not available
at this stage but this does represent one of the major factors in the
resource to reserve extraction factor at the 3.5 g/t Au economic
break-even cut-off grade being 59 per cent for gold ounces
and 56 per cent for gold ore tonnes. Application of current
site mining methodology to this narrow ore giving a lower than
break-even economic cut-off grade due to high internal stope
dilution.

FIG 2 - Barrick Granny Smith Wallaby underground current and
planned workings.
• Ductile shear controlled gold lodes – where brittle structures
have intersected the earlier ductile shears and cleavages.
• Haematite associated gold lodes – an assemblage of quartz,
Fe-rich dolomite, haematite and Au and is over-printed by
later gold lodes (fracture mesh and ductile shear).
The unmined known gold resource for Wallaby Barrick Gold
(2007) as published December 2007 is shown together with the
December 2007 reserves in Table 1 in JORC-compliant form.
The major proportion of the unmined resource is contained
within the Wallaby Deeps area located 300 m below the surface
in Zones 80 and below. This area containing a total of some
7.135 million tonnes at 7.63 g/t Au for some 1.75 million ounces
of inferred gold resource, representing some 60 per cent of the
known remaining gold resource ounces at Wallaby. This area is
open in extent on all lode horizons.
Diamond drill coverage for resource estimation ranges from
over 100 to 50 m for inferred, below 50 m for indicated and
measured areas are only classified as such once developed.
Geotechnical condition
Extensive geotechnical information has already been gathered at
Wallaby. Table 2 summarises the rock properties based on some
11 904 structural measurements from diamond drill core and
mapping and two acoustic emission stress measurements
collected for the 2005 Wallaby underground feasibility study as
described in Placer Dome (2005).
The investigations conducted to date have shown that the rock
mass currently being mined is competent. This competency can
be generally attributed to the high, stiff rock mass that is
encountered below the open pit. However, it is anticipated that
ground conditions will deteriorate as operational depths and
extraction ratios increase. Further to this Table 2 shows that the
ore zone has a high Young’s modulus, indicating that bursting
could be possible. In this regard it is expected that ground
conditions would only deteriorate after an operation al depth
of 350 m has been achieved. This is based on experiential
observation of mines in the Yilgarn area; in this respect an area
of concern has been a syenite dyke intrusion that intersects most
lodes. This material has been prone to some microseismic
activity when development occurs in this area.
To this end if the current longhole room and pillar or variations
of that method are used at depth careful geotechnical
consideration must be given to pillar design and opening span.

TABLE 2
Geotechnical summary of Wallaby rock properties to date.

Rock type UCS Young’s Poisson’s ratio Principal stress Magnitude Azimuth (°) Dip (°)
(MPa) modulus (GPa) component (MPa)
Major 27 151 2
Conglomerate 133 86 0.27 Intermediate 15 061 3
Minor 10 279 86
Ore zone 160 83 0.27
Major 28 137 06
Alkali intrusive 136 70 0.26 Intermediate 15 228 7
Minor 11 4 81
Intermediate intrusive 137 70 0.26

208 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 WALLABY – THE HORIZONTALLY CHALLENGED OPPORTUNITY

The mining challenge

The ore lodes vary in dip from 30 degrees to horizontal.
Thickness varies from 1 to 2 m in the narrow vein areas up to
20 m in places, particularly in areas where the lodes converge in
dip and join. The narrow vein areas are generally the more
steeply-dipping areas and the thicker areas are generally more
horizontal. The major ore lode areas can extend over 700 m in
strike length and 500 m in dip. The grade of mineable areas can
vary due largely to ore lode thinning in low-grade areas and
swelling in higher grade areas. Major lodes are generally
separated by 100 to 150 vertical metres but many minor lodes
exist within 20 to 50 m of these main ore lodes and each area of
main and sublodes being referred to as zones (refer to Figure 2).
In many ways the Wallaby underground deposit can be
considered similar in nature to the Platimum chrome or gold
deposits found in South Africa. However, further similarity is
that production profiles may require more than one surface or ore
lode to be extracted at one time. In South African terminology
such a situation is known as a multireef extraction scenario. In
this respect multireef situations have unique problems, namely:
• middling collapse between the two surfaces or lodes,
• stress interaction between the two surfaces or lodes,
• the impact of non-superpositioning of pillars between the
two lodes or surfaces,
• the need to ensure simultaneous or subsequent extraction
scheduling,
• the additional complexities of different grade and
metallurgical profiles, and
• increased complexities in ventilation layout and the need to
reduce leakages.
However, in the Western Australian situation, due to the need
for mechanisation, the following situation is further compounded
by the following: FIG 3 - Barrick Wallaby study and mining process.
• mining a very flat orebody with mechanised equipment,
• maximising mining efficiency in predominantly narrow and An initial scoping study for the Wallaby Underground Mine
relatively small tonnage stoping blocks, was completed in 2001 followed by a feasibility study
• maximising footwall recovery in the flat dipping narrow vein commenced in 2003 and completed in 2005 for the Wallaby
areas, Underground Mine. Due to the depth of this operation it was
considered that pillar stress would not be a major concern. In this
• managing a potentially high stressed environment with respect the initial underground mining project determined by the
mining depth using pillar layouts, study relied more on the ability to achieve a required production
• medium level mining head grades of approximately 5 g/t capacity of one million tonnes per year from two surfaces.
gold forcing backfill options such as with ‘paste fill’ proved In this respect the planned mine would have the following
uneconomic, and characteristics as described below.
• the need for decline access and the sequencing problems of
continually sequencing decline development. Production profile
A major challenge with mining tabular deposits using Extraction from two surfaces with planned production one million
mechanised methods is that continuous open stoping using pack tonnes per year for three years at a head grade of 5 g/t Au.
support cannot be undertaken. In essence this necessitates that a Mining activity has been limited to three main production areas
mechanised room and pillar be undertaken with increasing depth (refer to Figure 2) which are:
and span. Further challenges are that such methods tend to have • Zone 250 – a largely narrow vein ore lode dipping 25 to
lower production capacities requiring high rates of block surface 30 degrees,
development. In effect high concentration mining is not possible.
• Zone 60 – a largely horizontal ore lode adjoining Z250 but
Establishing a viable mine at Wallaby requires meeting up to 20 m in thickness, and
industry-wide challenges in recruitment, training and escalating
material costs as well as Wallaby’s own unique physical • Zone 70 – a largely narrow vein ore lode dipping 25 to
challenges of mining narrow and flat dipping near horizontal 30 degrees currently being developed.
orebodies.
Underground mine infrastructure and operational
PROJECT HISTORY OF THE philosophy
UNDERGROUND MINE Access to the mine is via a decline with a portal some 40 m
Open pit mining operations at Wallaby Pit commenced in 2001, above the Wallaby Pit floor. An enclosed ladder escapeway to the
with production starting in 2002. Production from the Wallaby surface from current stoping areas being an alternate means of
Pit ceased in August 2006. Figure 3 shows the project history in egress. Figure 4 shows a plan view of the current working areas
terms of study and mining processes. in Zone 250 and 60.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 209

N S RAUERT
FIG 4 - Barrick Granny Smith Wallaby underground mining current
Zone 250 and Zone 60 working areas as at July 2008.
The primary ventilation circuit consists of a dedicated
centrally located fresh air shaft intake connecting via workings to
two exhaust shafts. Each exhaust shaft having a surface mounted
axial flow fan installed. The exhaust shafts are interconnected via
intermediate sublevel development located above the production
areas. This development known as the East Ventilation Drive
acting as an exhaust horizon connected a series of exhaust raises
down to production areas.
Secondary ventilation consists of 150 kilowatt powered fans
force ventilating development and some stoping areas. In the
Zone 250 area airflow exhausts upwards through mined stope
voids to the exhaust raises. In other areas air is directed from the
fresh air shaft and decline intake areas to the exhaust raises past
working areas.
The operation is owner operated for the majority of mining
activity with a contractor used for fibrecrete, diamond drilling
and surface haulage. Granny Smith is operated as a fly-in, fly-out
operation from Perth but some staff are residential in Laverton.
An eight days on, six days off roster is used for day shift
personnel with underground operators rotating to a night shift of
seven days on seven days off roster.
Materials handling for both ore and waste utilises Caterpillar
Elphinstone 1700 and 2900 and Sandvik LH514 load haul dump
units and two Atlas Copco MT 5010 dump trucks. Ore being
bogged from ore headings and stopes to either underground
stockpiles or direct truck loading. Ore is trucked to a surface
stockpile near the portal to be later trucked by Gulf Underground
Pty Ltd operated PowerTrain road train haulage the 15 km back
to the Granny Smith processing plant.
210 Ballarat, Vic, 14 - 15 October 2008
A fleet of three Atlas Copco M2D face jumbos is used for ore
and waste development with two Atlas Copco Simba M4C
production drill rigs in use for stope blasthole drilling. Two
Giamech 211 blasthole charging vehicles are used for stope and
development charge up.
Mining methods
Selection of the mining method was a critical part of the
feasibility study. In that high tonnages were required from a
tabular orebody. In order to achieve the desired tonnages and
combat the risks associated with orebody’s tabular nature it was
considered that:
• Different mining methods would be required to ensure that
tonnage targets were met.
• That mining methods would encompass a mixture of larger
scale open stopes and narrower mechanised reef mining
methods. This would entail that a balance of higher grade
and higher tonnage would be achieved.
• That a system of room and pillar would be used regardless of
the stoping method used. This is to prevent geotechnical
problems.
• That two surfaces would be subsequently extracted and
simultaneously developed to mitigate geotechnical problems
and to ensure that production tonnages were met.
• That room or ore drive development would be considered as
a production method.
In this respect taking note of the following characteristics the
described methods were evaluated for the first two surfaces:
The mining method selection used to date can best described
as random longhole stope room and pillar. The rooms being
mined by various uphole retreat longhole stope configurations
and the pillars being of three types which are:
1. stope pillars – predetermined to a nominal size of 5 m
between stope voids,
2. regional pillars – predetermined to at least 20 m in width
providing major access until mined by retreat at the end of
area mining, and
3. random pillars – permanent pillars left due to uneconomic
grade.
A room and pillar mining methodology being selected because
of:
• the mining of flat laying and horizontal ore lode types at
other mines,
• the ability to mine without backfill,
• the use of a mechanised relatively efficient mining method,
and
• ore loss due to pillars proving economically sustainable due
to medium level only mining head grade.
Inclined room and pillar stoping
Applied to the narrow inclined areas the inclined room and pillar
(IRP) mining method utilises very similar techniques to that used
in much of the Western Australian gold mining industry (refer to
Figure 5). The method incorporates ore drive development on a
fixed one in 50 gradient with lateral direction control given on
geological control by geologists and geological technicians. The
ore drive face is positioned truncating the ore lode hanging wall.
Stoping is carried out on a retreat basis from the ore extremity
retreating back to a central main lode access location. Uphole
retreat stoping is employed with blastholes drilled at 64 mm
diameter.
Narrow Vein Mining Conference
 WALLABY – THE HORIZONTALLY CHALLENGED OPPORTUNITY
FIG 5 - The inclined room and pillar method as used at Wallaby.
Where possible slot development is by uphole longhole raising
however with only moderate success a hand-held airleg excavated
raise has often been used. In very flat areas a near horizontal slot
area is developed by jumbo (refer to Figure 6).
Stope size is limited to 20 m along strike with a 5 m pillar left
between stopes. Ten to 15 m is left between near parallel ore
drives governing stope height and giving typical stope hydraulic
radius of five to six.
A major limitation with this method is the potential ore loss of
blasted ore left in the footwall area in the flatter dipping areas.
One method commonly used in the Western Australian gold
industry is ‘fly’ blasting (refer Figure 5). This method employs
drilling blasthole rings in planes angled to a slot to throw blast
across the flat footwall area to the slot and mucking areas within
the stope.
The IRP mining method although having some limited use at
Wallaby has been shown to have several major concerns. These
concerns are:
• the method is very efficient in production due to small stope
sizing of 3000 to 4000 tonnes giving low overall production
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
FIG 6 - The use of jumbo developed slots and ‘fly’ blasting for the
inclined room and pillar mining method as used at Wallaby.
rates of 4000 to 5000 tonnes per month from individual
areas;
• the method requires considerable technical effort from both
geological and mining engineering staff;
• considerable ore loss can occur with ore left in the footwall
area;
• considerable scheduling of stope sequence as well as
production activities such as production drilling, blasting and
mucking is required to avoid interaction with concurrent
stoping above; and
• maintaining geological control of ore development lateral
direction can be challenging in the flatter ore lodes.
Transverse longhole stoping
The transverse longhole stoping mining method (TLHS) has
been applied to the near horizontal areas where ore thickness can
be up to 20 m. In these extensive areas the ore lode is divided
into zones or mining panels approximately 100 m wide and
100 m long (refer to Figures 7 and 8). Each panel is separated by
regional pillars to be mined at later stage of mining by retreat
and are approximately 20 m wide.
Each panel is in turn divided into several transverse mining
stoping areas. Ore development for each TLHS area is achieved
by following the ore lode by gradient with a set horizontal
orientation. The ore drive is positioned approximately 2 m below
the ore lode high-grade footwall largely in a prevalent low-grade
gold alteration area.
Each TLHS area can then be mined by uphole retreat stopes
using a longhole raised slot mined back to a central access.
TLHS stoping panels are limited to 15 m in width and 20 m in
length with 5 m pillars surrounding all sides. A wider abutment
pillar between ore drives is left under a minimum 5 m × 5 m
vertical pillar in order to reduce pillar slenderness. This abutment
pillar acting as a slide for improved blasted ore recovery.
211
N S RAUERT

Legend
Mined Transverse Long hole
Stope areas

Future Transverse Long hole

Stope areas

Regional Pillars

Completed Transverse Long hole

Stope development

Future Transverse Long hole

Stope development

Random Pillars - permanent Low

grade pillars

FIG 8 - The current transverse longhole stoping areas at Wallaby

showing the regional and the random permanent pillars.

FIG 7 - The transverse longhole stope room and pillar method as
used at Wallaby.
The TLHS method has to date demonstrated much improved
stoping efficiency with stope sizes up to 30 000 tonnes and
production rates over 20 000 tonnes per month from individual
stoping areas being achieved. Considerable improvements in
production drilling efficiency were also noted.
Modified transverse stoping
An adaption of the TLHS mining method is planned for trial
during 2008. This method involves transverse longhole stoping
of thicker inclined ore lodes, many of which were previously
mined by IRP.
This method shown in Figure 9 involves ore development
similar to that in IRP but up to 30 m spaced with angled ore
drives mined under the ore lodes creating potentially larger
stopes than that in the IRP case.
The modified TLHS has the following advantages over the IRP
method:
• much larger size stopes giving much more efficient cost-
effective stoping,
• improved footwall recovery of blasted ore,
• more efficient drill and blast, and
• potential to backfill if needed from ore development in the
level above.
The only potential detraction being a slight loss of overall
resource recovery compared to IRP mining.
As part of the feasibility study a trial mine was developed from
a decline with a portal in the west wall of Wallaby Stage 2 open
pit. Mining is being carried out by Barminco Mining Contractors.
As well as establishing the portal and main mine access the
trial mine also established an underground diamond drill location
for infill drilling of the initial upper zone mining areas.
Following approval to mine the upper areas by the Placer Dome
Board in 2005 an owner operated underground mine was
commenced in December 2005.
CURRENT MINING OPERATIONS
To date it can be considered that the underground project has
been successful in terms of:
• establishing the mine to a stage of ramping up to the 1 Mt/a
target,

212 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference


FIG 9 - The modified transverse longhole stope room and pillar
method as used at Wallaby.
• accesses and main infrastructure such as the primary
ventilation circuit have been established, and
• different stoping methods have been fully tested with no
major geotechnical failures recorded.
The following key learnings have been generated:
• that project development schedules need to be tempered to
reflect the difficulties in mining a tabular orebody,
Narrow Vein Mining Conference
WALLABY – THE HORIZONTALLY CHALLENGED OPPORTUNITY
• that production ramp up may take longer than expected,
• that decline development rates have to be increased, and
• infrastructure development time is longer than originally
considered.
Due to the success of the current project and the understanding
that project development could take longer it has been decided to
fast track the development of lower surfaces. In this respect
surfaces below Zone 70 are considered as deep level geotechnical
environment and are hence known as Wallaby deeps.
THE WALLABY DEEPS FUTURE MINE
The major portion of the remaining resource at Wallaby is
contained within the Wallaby Deeps area. This area below the
current workings is the subject of a prefeasibility study
determining mining method selection and materials handling
options and in general improving the level of understanding of
this unique orebody.
A scoping study carried out by the author in 2007 on the then
inferred resource of Wallaby Deeps indicated a very different
mine to the current upper level mine would be required. This
mine having the following characteristics:
• mining methodology would be a room and pillar style mine
similar to method or adaptations of methods currently in use,
• with depth down to 1000 m it will have potential stress-
related geotechnical issues,
• for optimum site economics a production rate in excess of
one million tonnes per year is required,
• materials handling or ore and waste would most likely be by
convention 50 or 60 tonne capacity dump truck or by use of
the Gulf Underground Pty Ltd operated PowerTrain road
train haulage system,
• ventilation issues could exist with distribution of air through
the numerous work locations required, and
• Wallaby Deeps would have a life exceeding ten years.
Investigations of alternative bulk mining methods such as block
and sublevel cave were also carried out but were discounted at that
stage due to the large waste area between lodes generating very
low bulk grades. Alternative materials handling options were also
investigated including shaft haulage but potential reserves were
deemed in sufficient to justify shaft capital.
The prefeasibility study is further testing the applicability of
the room and pillar mining methodology and will include
monitored trials of suitable mining methods including the
modified TLHS and alternative materials handling with the Gulf
road train haulage. This study is being conducted by a team led
by the author, comprising of Barrick Gold Yilgarn Shared
Services personnel.
Pillar design with depth has already been identified as of
utmost importance for geotechnical stability and as such is the
focus of the geotechnical study component of the prefeasibility
study. This study is being conducted by Itasca Consulting
Canada Incorporated utilising the 3DEC distinct element
geotechnical modelling code. This code allows for rock
structural and seismic wave property as well as stress analysis.
CONCLUSION
The Wallaby deposit is a unique deposit with a perceived obstacle
in mining a horizontal orebody. Adaption to mining the Wallaby
horizontal areas has actually given the successful mining of the
Wallaby orebody a tremendous opportunity through the trial of the
transverse longhole stoping method.
Ballarat, Vic, 14 - 15 October 2008 213
N S RAUERT

Major challenges wait in applying the lessons already learn form REFERENCES
upper level mining to larger Wallaby Deeps mine. The use of a
Barrick Gold of Australia Granny Smith Mine, 2007. Internal report:
rigorous staged gate mining approval process at Wallaby, allowing
Scoping study for the Wallaby Deeps Project, October.
for full investigations including trial mining prior to the next stage
Miller, J, 2005. Report on the Structural Evolution of the Wallaby Gold
of mining is, however, leading to a successful underground mine in Deposit, Laverton WA, July (Predictive Mineral Discovery
times of industry-wide challenges and uncertainties. Cooperative Research Centre: Melbourne).
Placer Dome Australia Granny Smith Mine, 2005. Internal report:
Wallaby Underground project feasibility study, August.

214 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

Recent Experiences at the Kangaroo Flat Mine, Bendigo
R McLean1 and M Hernan2

ABSTRACT interpreted to contain significant amounts of spur/stockwork

Bendigo Mining Limited has established a modern underground gold
mine and processing plant at Bendigo in central Victoria. The Bendigo
goldfield produced around 22 000 000 oz of gold between 1851 and 1954.
Approximately 4 000 000 oz were sourced from surface alluvial workings
with the remaining 18 000 000 oz won from underground quartz reefs.
The modern mine is accessed via the Swan Decline, which commences
at the Kangaroo Flat site on the southern outskirts of Bendigo and
progresses in a north easterly direction between the Deborah and
Sheepshead anticlines, reaching a depth of almost 1000 m below the City
of Bendigo.
Development of this decline commenced in 1998. Since that time
20 km of underground development has been completed. Exploration
drilling, on-reef development and bulk sampling carried out between
1999 and 2005 resulted in the estimation of a Probable Ore Reserve of
700 000 tonnes at 9.0 g/t gold for 207 000 oz of gold. These Ore Reserves
were contained within four reefs located on the Deborah and Sheepshead
anticlines.
Reef mining commenced in late 2005 from the Shywolup and Greater
Garrard reefs. Material mined was stockpiled on the surface in
preparation for commissioning of the process plant which commenced in
mid 2006. Mining and processing were carried out between August 2006
and May 2007 during which time 176 000 tonnes were processed at an
average grade of 5.4 g/t gold to produce 26 735 oz of gold. Plant feed was
sourced from four reefs.
This paper discusses some of the issues which were encountered with
the definition and mining of these geologically variable and nuggetty
reefs. This paper covers exploration techniques, grade estimation, mine
geological control, mining methods, geotechnical issues and
reconciliations. It will also discuss some of the lessons learnt and how
these will be applied to future Mineral Resource and Ore Reserve
estimation, as well as underground operations.
mineralisation with widths amenable to extraction with large
modern mining equipment.
Reconciliation of mining from the reefs indicated a general
overestimation of the Mineral Resources in Shywolup, Garrard
and Upper S3 reefs, forcing the company into a major shift in
strategy. This change resulted in the deferral of production and a
focus on building Ore Reserves in more productive areas of the
goldfield.
This paper focuses on the resource definition and mining
experiences for the four reefs mined during 2006 and 2007 and
discusses the lessons learned and how these have been applied to
the revised company strategy.
PROJECT HISTORY
Gold was discovered at Bendigo in 1851 and over the next
century to 1954, 22 000 000 oz of gold were produced, largely
by underground mining. The majority of this gold was sourced
from five main lines of reef namely Garden Gully, New Chum,
Deborah, Sheepshead and Hustlers (Figure 1). From the 1860s
through to the early 1900s, annual production was typically in
excess of 200 000 oz and was as high as 400 000 oz around the
early to mid 1870s (Figure 2) (AMC Consultants, 2004).

INTRODUCTION
The Kangaroo Flat Mine site is located at the southern end of the
Bendigo goldfield. Bendigo Mining Limited’s tenements cover
100 per cent of the goldfield and its immediate extensions.
Bendigo is a large regional city which lies approximately 150 km
north of Melbourne and has a population of almost
100 000 people. The mining operation is located directly below
the central urban areas of Bendigo (Figure 1), which presents a
number of challenges for the operation.
Mineralisation within the Bendigo goldfield is characterised
by erratically distributed coarse gold although the overall
geological continuity of the reef systems may be quite good. This
extreme ‘nugget effect’ makes it difficult to produce a reliable
resource estimate. There are a number of published papers which
discuss this issue in detail. Relevant references have been
provided.
After an exploration and project development period of some
eight years, the operation commenced commercial production in
late 2006. Ore was sourced from four reefs including the
Shywolup and McDermott reefs located on the Deborah anticline
and the Garrard and Upper S3 reefs located on the Sheepshead
anticline. Apart from McDermott reef, the other three reefs were

1. MAusIMM, Planning Superintendent, Bendigo Mining Limited, PO

Box 2113, Bendigo Mail Centre, Bendigo Vic 3554.
Email: rmclean@bendigomining.com.au
2. AAusIMM, Mine Geologist, Bendigo Mining Limited, PO Box
2113, Bendigo Mail Centre, Bendigo Vic 3554.
Email: mhernan@bendigomining.com.au FIG 1 - Bendigo goldfield showing main lines of reef.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 217

R McLEAN and M HERNAN
BENDIGO GOLDFIELD
PRODUCTION BY YEAR
Thousands (ounces)
FIG 2 - Historical production of the Bendigo goldfield.
Production details from individual mines are incomplete, but it
is believed that there were more than 5000 shafts on the Bendigo
goldfield. At least 140 shafts exceeded 300 m in depth, 67
exceeded 600 m, and 11 were over 1000 m deep. Shafts deeper
than 1000 m occur on three separate anticlines. The two deepest
shafts are the New Chum Railway at 1312 m deep and the
Victoria Quartz at 1406 m (AMC Consultants, 2004).
Project development
Bendigo Mining NL (the predecessor of Bendigo Mining
Limited) was established in 1984 to explore a small lease area on
the Deborah anticline near the centre of the goldfield. In the late
1980s the Deborah workings were dewatered, followed by a
limited program of diamond drilling and bulk sampling. In 1993,
after 15 years of exploration in Bendigo, Western Mining
Corporation, who at that time held most of the goldfield, sold its
exploration and mining leases to Bendigo Mining. These
tenements combined with the existing leases held by the
Company gave it control of the entire goldfield and 300 km2 of
exploration licences.
After obtaining substantial funding in the early 1990s, the
Company continued exploration and developed an improved
understanding of the geological factors controlling gold
mineralisation. This analysis led to the conclusion that the
previously mined geological structures would repeat regularly at
depth and this repetition would provide significant new mineral
resources. Further funds were raised in 1997 and permits for a
new exploration decline were obtained in 1998.
Construction of the Swan Decline commenced in October
1998. By the end of 2000 it had advanced to a depth of
approximately 540 m below surface. Exploration drilling from
this platform defined an Inferred Mineral Resource of 700 000 oz
of gold beneath the historic workings. Exploration holes were
also drilled from surface down to a depth of 1600 m, 8 km north
of the decline to confirm that the repetition of mineralised
structures was not a localised phenomenon.
Further funds were raised in 2001, in order to extend the
decline to 850 m below surface and enable bulk sampling within
several of the reefs on the Deborah and Sheepshead anticlines
(Dominy, Johansen and Annels, 2001; Johansen et al, 2003). The
reefs were first accessed in September 2002. Over 1.2 km of
on-reef development was then completed to gather information
on grade and grade distribution.
Following evaluation of the exploration results, a Probable Ore
Reserve of 700 000 tonnes at 9.0 g/t gold for 207 000 oz of gold
218 Ballarat, Vic, 14 - 15 October 2008
was estimated for the Shywolup, Garrard, Upper S3 and
Christine reefs.
In 2004, a feasibility study was completed which assessed and
supported the economics of re-commencing gold production
from the southern end of the goldfield. To fund the initial
development, more equity was raised in July 2004 and
underground development recommenced in late 2004.
In early 2005, the Company announced the development plan
would consist of two stages with initial production of
200 000 oz/y gold planned commencing in June 2006 from the
Kangaroo Flat mine with a 600 000 t/yr process plant to be
constructed adjacent to the mine portal. Subsequently, a second
mine and plant was to be developed at the northern end of the
goldfield, near Eaglehawk.
Construction of the Kangaroo Flat process plant commenced
in late May 2005 and was completed in October 2006.
Commissioning occurred in the September quarter of 2006, with
commercial production commencing in the December 2006
quarter.
Reconciliation of the December 2006 quarter production
results indicated that a general overestimate of the Ore Reserves
had occurred, which caused the company to implement a major
shift in strategy. This change resulted in the deferral of
production and a focus on building Ore Reserves in more
productive areas in the south of the goldfield.
Mining and processing ceased in May 2007, with the plant
having processed 176 000 tonnes at an average grade of 5.4 g/t
gold for 26 735 oz of gold.
GEOLOGY
The following summary of the geology of the Bendigo Goldfield
is reproduced from the New Bendigo Gold Project – Feasibility
Study carried out by AMC Consultants.
Regional geological setting
Bendigo lies within the Bendigo-Ballarat zone of the Palaeozoic
Lachlan Fold Belt of eastern Australia. Lower Ordovician marine
sediments and metasediments comprise the main part of the
Palaeozoic bedrock at Bendigo. These sediments belong to a
sequence of turbiditic sandstones, siltstones and mudstones
known as the Castlemaine Supergroup. In the Late Ordovician to
Late Devonian the Bendigo-Ballarat zone underwent a period of
crustal shortening (regional compression) resulting in extensive
folding and reverse faulting, which have significantly increased
the apparent thickness of the Castlemaine Supergroup.
Narrow Vein Mining Conference
 RECENT EXPERIENCES AT THE KANGAROO FLAT MINE, BENDIGO

Local geology
The Bendigo goldfield lies in a 9 km wide NNW trending block
of Lower Ordovician turbidites bounded by the Whitelaw and
Sebastian Faults (thrust faults). The productive portion of the
goldfield lies in a zone 15 km long by 5 km wide in the centre of
the fault-bounded block.
The bedrock is composed of a very uniform sequence of
turbiditic sandstones, siltstones and mudstones interbedded with
minor ‘cone-in-cone’ limestones. The generalised structure of the
Bendigo goldfield is that of regular, close spaced folding with
extensive local reverse faulting related to the compression event.

Gold mineralisation at Bendigo

Gold mineralisation at Bendigo was synchronous with the period
of regional compression during which the Ordovician sediments
detached from the underlying Cambrian oceanic crust. The
increase in crustal thickness of the Ordovician sediments through
folding and thrust faulting is thought to have caused prograde
regional metamorphism of the underlying Cambrian stratigraphy
resulting in the formation and release of auriferous hydrothermal
fluids. Fluids migrated up fault splays associated with major
regional structures to be precipitated in nearby favourable low
pressure, dilation zones (eg faults, bedding planes, fold axes)
created when the hydraulic pressure of trapped fluids exceeded
lithostatic pressure during local seismic events.
Known gold mineralisation is exclusively associated with
quartz veining occurring as:
• free gold in quartz,
• in association with sulfides in quartz,
• in association with fragments or laminae of wall rock in
quartz, and
FIG 3 - Schematic structural setting of quartz reef types at Bendigo.
• occasionally associated with sulfides in wall rock
immediately adjacent to quartz veins.
The reefs are composed of two main quartz types, laminated Exploration techniques
and bucky quartz:
Modern exploration methodology involves the drilling of deep
• Laminated quartz consists of multiple laminae or veinlets of ‘sweeper’ holes through the limbs of the anticlines targeting the
quartz, separated by millimetre thick slivers of wall rock main prospective lithological horizons. The principle aim of
and/or sulfides typically developed in bedding parallel veins. these holes is to identify prospective laminated quartz veins
The gold may be both fine and coarse grained, and is (LQs) on the limbs of the fold, which can then be projected up to
typically restricted to only one or a few laminae. the rollover position, targeting the next round of more detailed
• Bucky quartz describes the coarsely crystalline, non- exploration. This methodology is based on the model that the
laminated and sometimes vuggy quartz in bedding discordant LQs are bedding parallel faults, which were the conduits for the
veins. The gold, when present, is typically free and coarse mineralising fluids during reef formation. It is thought that the
grained (0.25 mm to 2.5 mm), except when associated with larger the LQ, the greater the chance of a significant quartz reef
sulfide. being formed at the roll-over or other zone of dilation. An LQ
greater than 10 cm in thickness (with associated graphitic fault
Veins may contain a dolomitic carbonate (typically ankerite), gouge) on the fold limb is regarded as highly prospective.
albite and from 0.5 per cent to 2.5 per cent sulfides (pyrite, The historical exploration strategy involved sinking a vertical
arsenopyrite, lesser galena, sphalerite, pyrrhotite and rare shaft down or adjacent to an anticline axis with the aim of
chalcopyrite). The gold is typically free and coarse grained intersecting LQs (historically referred to as ‘backs’) in the shaft
(usually 100 µm to 10 mm). Historically, approximately five per and associated cross-cuts. The better LQs were then prospected
cent of reef gold production came from sulfide concentrates. by rising up the anticline limbs to the anticline axis where it was
Visible gold has historically often been reported in close spatial anticipated that a reef would be formed. The modern exploration
association with coarse grained galena and sphalerite. strategy is similar, although uses drill holes to obtain core
Individual quartz reefs occur in a wide variety of structural samples of LQs over multiple geological units. Reef targets are
settings and may display relatively complex cross-sectional then ranked by LQ width and other structural and stratigraphic
shapes, but can be classified into five broad categories (Figure 3): features.
1. saddle reefs,
Resource estimation
2. bedding parallel leg reefs,
The techniques used to define the Mineral Resources on the
3. fault and neck reefs, Sheepshead and Deborah anticlines prior to mining in 2006
included a combination of diamond drilling, on-reef development,
4. spur reefs, and
collection of rock chip samples, 50 kg samples and bulk samples
5. cross-course reefs. (Dominy et al, 2000; Johansen et al, 2003).

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 219

R McLEAN and M HERNAN
Resource volume and grade estimation
When a reef had been identified through wide spaced (120 m
section lines and 10 to 15 m vertical spacing) exploration
drilling, a preliminary mineralised reef shape was constructed
using a combination of quartz percentage, arsenic and gold
grades, structural and lithological continuity. Further infill
drilling (40 m section lines and 5 to 10 m vertical spacing),
mapping and sampling of reef development was then carried out
to enhance the original mineralised reef shape.
Generic transformation parameters, derived from the correlation
of diamond drill hole assays with bulk samples, were applied
to the raw grades within the mineralised reef shape. These
transformation parameters were applied with the aim of reducing
the highly variable nature of the assays in such an extreme nugget
environment. An unconstrained block model was created using the
transformed grade data, with further refinement of the mineralised
reef shape undertaken using the block model data to provide the
final shape and volume of the Mineral Resource.
MINING
Mine access
The mine is accessed via a 5.2 metre wide by 5.5 metre high
decline grading at one in 6.5 from the portal at Kangaroo Flat,
which proceeds in a north easterly direction to a position located
between the Sheepshead and Deborah anticlines under the City
of Bendigo. All ore and waste is hauled out of the mine via this
decline using a fleet of Atlas Copco MT5010 trucks, rated at
50 tonnes per load.
Ventilation system
The Kangaroo Flat mine is ventilated utilising three surface
ventilation shafts and the decline. The two Adam Street shafts
are return airways, one 3.5 m in diameter and the other 2.4 m in
diameter. Both shafts exhaust a total of approximately 170 m3/s,
via fans located 645 m below surface. The North Deborah shaft
is a fresh air shaft, 4.0 m in diameter and is fully supported to
provide a second means of egress via a mobile winch and cage.
FIG 4 - Long section showing the reefs on the Deborah anticline.
220 Ballarat, Vic, 14 - 15 October 2008
Due to the size and spatial separation between reefs, nearly all
working areas are force ventilated with the exhaust air returning
to the decline. The auxiliary fans used are typically twin stage
180 kW capacity with 1400 mm Protan ducting. It is not
uncommon to be working with vent ducting lengths in excess of
1000 m.
Stoping
The reefs at Bendigo vary in size, shape and wall rock type, so a
number of different stoping methods were selected to mine the
different reefs. In areas where ground conditions were poor and
reef shape complex, a mechanised flat back cut-and-fill method
was chosen. In areas where ground conditions were better and
reef shape more regular, a bench stoping method was chosen.
The original interpretations indicated a reef width of around
5 m or more. As such, all reef development and stoping was to be
mined using twin boom jumbos. A fleet of four Atlas Copco
M2D jumbos, two R1700 and two R2900 Caterpillar LHDs were
purchased to undertake reef mining as well as all capital
development.
Stoping was undertaken in the Shywolup, Upper S3, Garrard
and McDermott reefs. The Christine reef, included in the initial
Ore Reserve was not mined. Descriptions of the individual reefs
mined and experience during development and stoping are
outlined below. Figures 4 and 5 are long sections of the Deborah
and Sheepshead anticlines showing the locations of these reefs.
Shywolup reef – Deborah anticline
The reef is situated on the Deborah anticline some 600 m below
surface. The host lithology is the historically productive Big Blue
Shale.
Shywolup reef was originally interpreted with the upper and
narrower portion of the reef primarily composed of a bedding
parallel laminated vein, with the down dip extent being the wider
portion of the reef, comprising a complex zone of massive to
stockwork quartz in the fold axis. The reef was interpreted to be
between 10 to 27 m high, with the narrower upper part ranging
from 1 to 2 m wide and the basal part being typically between
6 to 10 m in width. The reef was defined over a strike length of
approximately 500 m.
Narrow Vein Mining Conference
 RECENT EXPERIENCES AT THE KANGAROO FLAT MINE, BENDIGO

FIG 5 - Long section showing the reefs on the Sheepshead anticline.

Mining of the reef was originally planned to use mechanised

cut-and-fill in the zone of complex massive to stockwork quartz
at the northern end of the reef, with small bench stopes to be
utilised for the narrower laminated quartz vein mineralisation.
The Probable Ore Reserve for this reef was estimated as
135 000 tonnes at 9.5 g/t gold for 41 000 oz of gold.
Initial development in the reef proved disappointing. The
massive to stockwork quartz at the northern end proved to be
far less consistent and more complex than expected (Figure 6).
However, cut-and-fill mining continued with two lifts mined
above the initial ore drive. Where appropriate, faces were split
fired to allow mining of the ore and waste separately. Ground
support consisted of split sets and mesh with fibrecrete applied in
some areas.
At the southern end of the reef, development showed the
laminated quartz veining to be very thin and poorly developed
with very little associated spur veining and little gold, so no
stoping was carried out in this area. Instead, mining followed a
wider and higher grade zone some 25 m above the initial
development. This zone comprised a laminated quartz leg reef
about 1.0 m wide, with a wider zone of stockwork mineralisation
up dip at the anticline rollover position (Figures 7 and 8). Much
of this ore was located outside the original Mineral Resource
shape. At the suspension of production, only the initial ore drive
had been mined through this zone. When the reef was narrow,
faces were split fired to allow mining of the ore and waste
separately. Ground support consisted of split sets and mesh.

Upper S3 reef – Sheepshead anticline

The Upper S3 reef is located on the Sheepshead anticline some
650 m below the surface. The reef is a saddle shaped zone of
massive to stockwork quartz with multiple veining events
reflecting continuous movement along the controlling fault
structure (refer to Figure 9). The reef is hosted by the Christine
Unit comprising an interbedded sequence of sandstones,
siltstones and shales.
As the reef is structurally constrained, it plunges 9° north in
parallel with the folding. The interpreted dimensions of the reef
were 15 m wide at the base and 15 m high. Leg mineralisation FIG 6 - Complex stockwork mineralisation at northern end of
was up to 4 m wide extending from 10 to 20 m below the saddle, Shywolup reef.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 221

R McLEAN and M HERNAN

FIG 7 - Laminated quartz leg at southern end of Shywolup reef.

FIG 9 - Upper S3 mineral resource interpretation.

resulting in a total height for the reef of 25 to 35 m. The reef was

defined over a strike length of approximately 550 m.
Mining was originally planned to be undertaken with mainly
small bench stopes. A large ore drive was planned at the base of
the saddle with two ore drives following the legs some 15 m
below which would be used to extract the ore. The Probable Ore
Reserve for this reef was estimated as 205 000 tonnes at 7.5 g/t
gold for 50 000 oz of gold.
Initial exposures of the reef proved disappointing, particularly
the development along the legs. As a result mining concentrated
on the cap and neck area of the reef. Mining of the cap proved
difficult with the face often being driven too high up into the
neck of the reef leading to a requirement for floor stripping. This
resulted in very high drive walls requiring significant ground
support. Support consisted of in-cycle shotcrete with split sets
and mesh. Cable bolts and gewi bolts were also used in wide
areas or areas with high wall exposures. The cap of the reef also
proved quite poorly mineralised apart from some isolated
pockets of higher grade. Figure 10 shows the poorly developed
mineralisation encountered in the reef.

Garrard reef – Sheepshead anticline

FIG 8 - Wider stockwork mineralisation at rollover of southern end
of Shywolup reef.
The Garrard reef is located on the Sheepshead anticline, 850 m
below surface. The reef is a combination of several quartz styles,
including a small saddle, west dipping laminated vein (or leg),
fault hosted reef and stockwork (refer to Figure 11). The reef is
hosted in the Railway Shale and controlled by the west dipping
Garrard Fault.
Garrard reef is bounded to the west by a west dipping
laminated vein with its base terminated by the east dipping

222 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 RECENT EXPERIENCES AT THE KANGAROO FLAT MINE, BENDIGO
FIG 10 - Upper S3 reef showing poorly developed mineralisation.
FIG 11 - Garrard reef mineral resource interpretation.
Young Fault. The reef has gradational boundaries based on gold
grade, vein density and vein thickness to the east and top. The
combination of a saddle in the axis and fault mineralisation to the
east resulted in widths near the top of the Mineral Resource of
greater than 10 m, with the centre of the reef ranging from 3 to
5 m wide. Flat dipping spurs reach out east from Garrard Fault to
give estimated widths at the base of the reef in excess of 7 m.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
The reef was estimated to range from 35 to 50 m high. The reef
was defined over a strike length of approximately 400 m.
Mining was planned using mainly small bench stopes with
some mechanised cut-and-fill at the top of the reef. The Probable
Ore Reserve for this reef was estimated as 240 000 tonnes at
10.0 g/t gold for 78 000 oz of gold.
Initial development into the reef proved disappointing, with
the quartz veining associated with the fault being narrower than
expected and the predicted spur mineralisation being much less
regular and less well developed than expected. Significant
unplanned internal dilution was incurred due to a dyke
(approximately 1 m thick) traversing the reef. Development
along the ore drives continued with approximately 45 per cent of
faces being categorised as mineralised waste. Ground conditions
were quite difficult due to the presence of faulting and the late
stage dyke, with ground support consisting of in-cycle shotcrete
with split sets. A number of the drives also required
rehabilitation and secondary support with mesh, gewi bolts and
cablebolts. Several small bench stopes were mined in selected
areas where ore development indicated better quartz
concentrations. Stope strike lengths were limited to about 12 m
due to wall stability concerns. Overall the stope ore mined
proved disappointing and a decision was made not to continue
with these stopes. Instead, efforts focused on a possible higher
grade zone at the top of the reef in the area where initial
exploration bulk sampling had taken place. Wall and back
stripping and some small retreat bench stopes were mined in
this area producing ore. Figure 12 shows the Garrard reef
mineralisation encountered.
FIG 12 - Garrard reef showing the mineralisation encountered.
McDermott reef – Deborah anticline
The McDermott reef is located on the Deborah anticline,
approximately 850 m below surface. The reef is a classic saddle
reef with reasonably well developed legs which extend for a
223
R McLEAN and M HERNAN
short distance below the rollover. The reef is located within the
shale dominant Railway Shale unit. The mineralisation is located
directly underneath and to the west of the east dipping
McDermott fault.
The saddle comprises moderately laminated and brecciated
quartz, is between 4 and 5 m high from the base of the rollover to
the top of the reef and between 3 and 5 m wide at the base of the
saddle. The reef has been defined over a strike length of
approximately 450 m. The total Mineral Inventory for this reef,
which was not included in the initial Ore Reserve, was estimated
at 51 000 tonnes at 8 ± 2 g/t gold for 12 900 oz of gold.
This reef was drill tested on 30 to 40 metre centres, a density
which would allow an Inferred Mineral Resource to be
estimated. Due to the compact size of the reef, it was planned to
mine the whole of the cap and the upper, thicker portion of the
legs in one pass. This resulted in mining a small portion of the
spur zone below the saddle. The width of the reef also meant that
an amount of waste rock outside the reef shape had to be mined
to accommodate the available large mining equipment. Initial
access into the reef took place around the centre of the defined
strike length, with mining progressing north and south from the
central access cross-cut.
Mining advanced quite successfully under geological control
for elevation as well as direction. There were a few occasions
where mining ended up too high in the reef with floor stripping
required to bring the heading back to the right position in the
reef. The presence of McDermott fault and a lamprophyre dyke
on the eastern wall of the reef combined with the weak host
shales presented some ground stability issues. Figure 13 shows
the typical McDermott reef mineralisation encountered. Ground
support consisted of in-cycle shotcrete with split sets. Secondary
support was required on a regular basis at a point about 20 m
behind the advancing face. Secondary support consisted of gewi
bolts/grouted split sets and mesh. Cable bolts were also used in
areas where there was wedge potential in the backs and where
wall heights were excessive.
FIG 13 - McDermott reef showing the mineralisation encountered.
LESSONS LEARNED
The reefs mined were interpreted to contain significant amounts
of spur/stockwork mineralisation over quite a wide zone making
224 Ballarat, Vic, 14 - 15 October 2008
them suitable for mining with large modern mining equipment. It
became clear that the methodologies used to define the size and
grade of these reefs had some significant shortcomings. The
approach was based on sound statistical methods; however, the
application of the method was not suitably constrained to honour
important factors such as local geological domains and quartz
percentage. The interpreted spur zones had significantly less
quartz mineralisation and less gold than had been estimated, so
the economic component of the reefs were generally narrower
and lower grade than expected.
The lessons learned from this:
• Geological factors such as quartz percentage, quartz texture,
sulfide abundance, structures and lithologies identified in the
drill core should be used to define the reef shape and grade.
• A combination of visual estimates and assays from drill core
should be used to estimate a grade range for each geological
domain within the reef.
• Focus exploration efforts on the historically more productive
anticlines of the field, namely Garden Gully and New Chum.
• Consider utilising alternative mining equipment and methods
more suited to narrow reefs.
• Close spaced drilling (15 m between sections) is essential
to enable accurate reef modelling and mine planning, 40 m
spaced sections do not enable mine geologists to accurately
anticipate changes in reef size, shape, orientation and grade.
• Effective communication between mine geology and
exploration geology is essential to the success of the mine.
Mine geologists need to be able to use exploration models to
anticipate changes within the reef. Exploration geologists
need to be able to access data collected by mine geologists
underground.
• Face rock chip samples are of limited use for grade control
purposes as the sample mass is orders of magnitude too small
to be representative, although they were used to identify the
limits of economic mineralisation at the ends of the reefs. In
addition, it is often difficult to sample every face due to poor
ground conditions, particularly when mining cut-and-fill lifts.
• Whilst shotcrete is very good at stabilising broken ground, it
hinders the collection of geological information.
• Split face firing is a good way to minimise dilution; however,
it effectively doubles the mining cycle time as the heading
has to be fired and bogged twice per round.
RECENT EXPLORATION
The exploration targeting methodology has been refined with a
strong focus on correlating historically well endowed rock units,
targeting based on LQs in these units and wide spaced initial
scout drilling prior to more intense drilling of the better targets.
There have also been significant changes made to the reef
volume and grade estimation techniques. These changes were
made based on the recent mining experience and from a better
understanding of how reefs were historically defined using visual
indicators. The extremely nuggetty gold distribution means that
assay grades are generally unrepresentative of the true grade of
the reef through which the drill hole passes. A better estimate of
the grade can be made by applying an understanding of the
physical characteristics of the quartz style and abundance as well
as the other minerals within the reef, as discussed in detail below.
Resource volume estimation
Once a reef has been identified through wide spaced (120 m
section lines and 7 to 12 m vertical spacing) exploration drilling,
the resource shape is constructed using a combination of quartz
Narrow Vein Mining Conference
 RECENT EXPERIENCES AT THE KANGAROO FLAT MINE, BENDIGO

percentage, quartz texture, structural and lithological continuity. TABLE 1

Further infill drilling is then carried out to generate more Summarised scorecard system.
information which is used to enhance the resource shape.
Visual scorecard
Grade estimation Parameter Weighting Score
Grade estimation is carried out using a combination of the Quartz percentage Low 0-4
physical characteristics of the quartz and screen fire assays. A Quartz texture Medium 0-8
scorecard system has been developed which combines the two
Visible gold High Uncapped
inputs, with the visual estimation carrying significantly more
weight than the analysis of the assays. The combined score will Sulfides in host rock Very low 0-3
result in each intercept being estimated to within a certain grade Sulfides in quartz – arsenopyrite Medium to high 0 - 12
range. Sulfides in quartz – sphalerite and Medium to high 0 - 15
The visual part of the scorecard is completed as part of the galena
logging process and is a way of quantifying an experienced Visual score Uncapped
geologist’s understanding of an intersection and its likely grade
range. This approach is not dissimilar to the approach used by Assay scorecard
miners on the field for 100 years from the mid 1850s. It focuses Parameter Weighting Score
on the key textures and minerals evident in high-grade reef Highest assay High 0 - 12
systems in Bendigo and ranks the intersection according to the
intensity of these attributes in the core. The components of the Upper quartile Au High 0 - 12
scorecard are described below. Median Au High 0 - 12
The percentage of quartz and the texture of the quartz are Lower quartile Au High 0 - 12
important as all the gold mineralisation in Bendigo is hosted Average as value Low 0-4
within quartz. Therefore the amount of quartz in the drill
intersection is fundamentally important. The second component Assay score 0 - 52
of quartz mineralisation is to identify the style of quartz which Total score
also provides strong clues as to the likely gold grade. Experience Total score = (visual score × 0.66) + (assay score × 0.34)
gained underground and from historical information confirms
that the more laminations/stylolites present in the quartz, the Ranking Grade Total score
greater the chance of significant amounts of gold being deposited Barren <2 gt/t Au <8
in the reef system. Therefore the intersections with a high Low grade 2 - 6 g/t Au 8 to 15
percentage of quartz veining and good textural development can
Medium grade 6 - 10 g/t Au 15 to 25
be expected to form the basis of an economic reef.
Obviously, the greater the amount of gold observed in a drill High grade 10 - 15 g/t Au 25 to 35
intersection, the greater the probability that the intersection is Bonanza >15 g/t Au >35
indicative of a high grade reef. There is some differentiation
between small pieces of gold and more substantial pieces, with a
scorecard weighting to match. The initial drill sections, spaced approximately 120 m apart,
Sulfide assemblages, such as arsenopyrite, sphalerite and into Gill reef indicate that it is the most continuous and highest
galena hosted in the quartz reef are proxies for gold grade reef system discovered so far. Subsequent infill and step
mineralisation. Arsenopyrite, after pyrite, is the most abundant out drilling continues to confirm consistent shape and grade. The
reef is interpreted as a neck style reef with leg reef mineralisation
sulfide species in Bendigo. The greater the abundance of large
below the rollover in the northern section. The reef is
bands or clumps of arsenopyrite in the quartz, the more likely it
approximately 1000 m below surface and is estimated to be
is to be high grade. The presence and relative abundance of 720 m long, 20 to 30 m high and between 3 and 9 m wide, with a
galena and sphalerite are also indicative of potentially grade range of between 10 g/t and 15 g/t gold. At the time of
well-mineralised reef systems, although these minerals are just as writing this paper, an access into the reef had been completed
variable in their distribution as the gold. Therefore the presence with infill drilling underway to allow detailed mine planning.
of these proxy sulfide species is integral to determining the likely Initial development on the reef was aimed at the rollover position
grade range of a reef system. with the mineralisation exposed consistent in both style and
The second part of the scorecard uses the screen fire assay. dimensions with the interpretations. Geological mapping
The assay scorecard results are combined with the visual identified a high percentage of laminated quartz, sulfides and
scorecard to give the final grade range for the intercept. The some visible gold, providing confidence that the reef grade is in
visual scorecard will carry between 66 per cent and 90 per cent line with expectations at around 10 to 15 g/t gold. Figure 14
weighting towards the overall grade range depending on the shows the Gill reef mineralisation as exposed in the initial
number of samples in the intercept. Table 1 shows a summary of development.
the scorecard system. It is important to note that, with an Other reefs discovered on the Garden Gully anticline include a
orebody as nuggetty as that of Bendigo, grade estimates are small (5 m high by 5 m wide) saddle reef (Grenfell), a large spur
presented as grade ranges to indicate the level of accuracy of the reef system (Grenfell East), a fault reef with footwall spur
estimates. development (Grainger) and another large neck reef (Gordon) of
similar dimensions to Gill. The first three reef systems as well as
Exploration results Gill are all hosted in the Railway Shale unit, with Gordon reef
hosted within the stratigraphically higher unit, the Big Blue
Exploration activity over the last 12 months has focused on
defining reefs on the better endowed Garden Gully and New Shale. Figure 15 shows the location of the reefs on the Garden
Chum anticlines, as well as refining knowledge of previously Gully anticline.
identified reefs on the Deborah and Sheepshead anticlines. To On the Deborah anticline exploration has focused on
date, this work has identified five reefs on the Garden Gully expanding the potential of three reefs which had been identified
anticline. previously. Dale reef is a small saddle reef hosted in the Railway

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 225

R McLEAN and M HERNAN
FIG 14 - Gill reef showing the mineralisation encountered.
Shale unit. Dean reef is a saddle style reef with legs offset by
faulting located in the same shale unit as Gordon reef (Big Blue
Shale). Upper McDermott reef is a spur style reef located some
40 m above the previously mined McDermott reef, again hosted
in the Railway Shale unit.
Trial mining and processing
The next step is a period of trial mining and processing which
will involve mining and processing discrete blocks of ore. During
this process each face will be mapped, scorecarded and
photographed by the mine geologists. The geologists will
estimate the tonnes and grade range of each block based on both
the infill drilling and the face mapping. Once the amount of gold
produced from each block is known the actual grade of the block
will be calculated and a reconciliation process will be used to
reconstruct the reef shape and determine the final grade of the
FIG 15 - Long section showing reefs on the Garden Gully anticline.
226 Ballarat, Vic, 14 - 15 October 2008
reef within that block. Over time the results of this reconciliation
process will be used to validate the scorecard system and to
modify the various scorecard components to more closely
estimate the reef grade.
THE FUTURE
Gill reef infill drilling (15 m section lines and 5 to 7 m vertical
spacing) continues to meet expectations which, combined with
encouraging results received from some of the previously
unmined reefs on the Deborah anticline, continue to rebuild
confidence in the Bendigo project. A period of trial mining and
processing commenced late in the June quarter of 2008 which,
in combination with the infill drilling, will be used to validate
the scorecard grade estimation approach. This will build
confidence in this grade estimation technique. Smaller mining
equipment has been purchased including an Atlas Copco
104 single boom jumbo and a Caterpillar R1300 loader. This
will allow the minimum drive size to be reduced from 5.0 m ×
5.0 m to 3.2 m × 3.2 m which will be ideally suited to mining
some of the small saddle and leg reefs discovered. Cut-and-fill
stoping will be the primary method utilising waste rock for fill.
Narrow leg style reefs will be mined by a combination of
cut-and -fill with split fired faces and small scale retreat bench
stopes. Mining of narrow reefs using hand-held methods remains
an option.
CONCLUSIONS
Mining of the geologically variable and nuggetty reefs at
Bendigo requires the use of unique methods for defining the reef
shape and grade. Mine geologists must use a combination of the
reef models and daily face inspections to anticipate changes in
the gold grade, direction and plunge of the reef. The mining
methods and equipment used must allow flexibility to mine reefs
of variable width, strike length and plunge. The techniques used
to discover the reefs are being refined with time as the controls
on gold deposition are better understood. The move to trial
mining and processing should further this understanding which,
combined with the significant infrastructure already in place,
puts Bendigo Mining in a strong position to capitalise on future
exploration success.
Narrow Vein Mining Conference
 RECENT EXPERIENCES AT THE KANGAROO FLAT MINE, BENDIGO

ACKNOWLEDGEMENTS Dominy, S C, Johansen, G F and Annels, A E, 2001. Bulk sampling as a

The authors wish to thank the management of Bendigo Mining
Limited for permission to publish this paper and gratefully
acknowledge the assistance of colleagues at Bendigo Mining in
its preparation.
REFERENCES
AMC Consultants, 2004. New Bendigo gold project – Feasibility study,
volume 2, Introduction, geology and mineral resources (unpublished
report), March.
tool for the grade evaluation of gold-quartz reefs, Transactions of the
Institutions of Mining and Metallurgy, Applied Earth Science,
110:B176-191.
Dominy, S C, Johansen, G F, Annels, A E and Cuffley, B W, 2000.
General considerations of sampling and assaying in a coarse gold
environment, Transactions of the Institutions of Mining and
Metallurgy, Applied Earth Science, 109:B145-167.
Johansen, G F, Raine, M D, Dominy, S C and Bartlett, J K, 2003.
Challenges of sampling extreme nugget-effect gold-quartz reefs at
the New Bendigo Project, Central Victoria, Australia, in Proceedings
Fifth International Mining Geology Conference, pp 299-310 (The
Australasian Institute of Mining and Metallurgy: Melbourne).

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 227

Historical Challenges, Modern Solutions at Ballarat East
K Williams1 and M Sykes2
ABSTRACT
Gold was initially discovered at Ballarat in 1851. After initial alluvial
mining, underground quartz mining commenced in 1853, continuing until
1918. The historic miners were considered to be capable of mining
anywhere in the world after overcoming the difficult conditions within the
Ballarat goldfields. Whilst mining has changed dramatically since the early
1900s the challenges with underground mining at Ballarat have not.
Modern pumping systems cope with the moderate water inflows. However,
modern machinery and efficient mining methods influence opening sizes
and extraction techniques. In some cases modern techniques have
accentuated the challenging ground conditions and complex orebodies that
were faced by the historic miners. As the Ballarat East project has
progressed towards production, technical challenges relating to
geomechanics and mine planning are being resolved resourcefully.
INTRODUCTION
The challenges faced within the Ballarat East goldfields have not
varied significantly since the 1800s when underground mining
first commenced. The continuing challenges include groundwater,
geological complexity, ground control and ventilation. Changes
including mechanised mining have accentuated these challenges
and created new challenges. Planning and scheduling to ensure
viability within a public company has become one of the new
challenges associated with modern mining.
1. Senior Geotechnical Engineer, LGL – Ballarat Operations, PO Box
1228, Bakery Hill Vic 3354. Email: kate.williams@lglgold.com
2. Senior Mine Planning Engineer, LGL – Ballarat Operations, PO Box
1228, Bakery Hill Vic 3354.
FIG 1 - Geological cross-section showing mineralisation types (Osborne, 2008).
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
Historically the mines in the Ballarat East field started as owner
operator mines with the owners/shareholders of the mine operating
the mine. Because of this, the production risk due to grade
variability and lode continuity was held by the operators. The
Ballarat East mines then moved towards forming companies with
contract miners and external shareholders who expected returns on
their investments. The onset of the First World War resulted in
reduced labour, equipment shortage, fuel shortage and no external
funding which stopped mining in the Ballarat East field.
GEOLOGY
The Ballarat East geology consists of tightly folded Ordovician
sandstones and siltstones deposited as a turbidite sequence
(d’Auvergne, 1990). The folding is developed along a northerly
axis and slightly overturned towards the east. Two major types of
faulting were developed throughout the field in a series of
deformation events. North striking, west dipping reverse faults
(called leatherjackets by the historical miners) were developed as
part of the folding deformation event; these are spaced
approximately 50 m apart vertically. Later subvertical strike slip
faults (colloquially known as cross-course faults) strike east west
cross-cutting both bedding and offset the orebodies, these
generally occur at 50 m horizontal spacings in the southern end
of the field and are more widely spaced out to 200 m at the
northern end of the field. The quartz orebodies are typically
located near the intersection of west dipping reverse faults and
vertical bedding parallel structures which were colloquially
known as ‘indicators’ (Baragwanath, 1923) as shown in Figure 1.
229
K WILLIAMS and M SYKES
Ballarat Goldfields, as part of the early exploration effort
compiled the historical documentation across the majority of the
Ballarat East field; from this reconstruction the styles and extents
of mineralisation were identified. Over half of the historic
production occurred along reverse west dipping faults, with the
remainder associated with flat spur veins coming off the west
dipping faults or indicators.
The ore shoots modelled at the intersection of west dipping
reverse faults and controlling vertical faults are persistent along
strike over hundreds of metres, with a typical true thickness
ranging from narrow veins through to complexes of mineralisation
up to 30 m thick. The up dip height of individual shoots is
proportional to the true thickness, with heights from 5 m to 70 m.
The geological models and interpretations are based on a
systematic drilling process as outlined in Table 1.
EXPLORATION
Historically exploration was not completed as it is defined in
today’s terms; there was little concerted effort into understanding
the geology and orebodies prior to committing capital
expenditure, ie shaft. With knowledge from the deep lead mining
and quartz outcrops the initial shafts were located close to what
was known as the ‘indicator’ bed and mining began.
In the modern mining environment capital expenditure is
difficult to acquire based on ‘we think it is there’ philosophy as
was the case in the 1800s. Ballarat Goldfields use a systematic
drilling program to identify potential orebodies and then to
ensure that resources and costs are allocated to the economically
viable orebodies.
The first level of drilling L1 (regional exploration) was
completed from accessible positions on the surface to determine
the basic geological information including fault locations and fold
locations. With the limited access to surface drill sites within the
city of Ballarat this was a very widely spaced campaign. Further
campaigns were then restricted to drilling from underground. One
of the main challenges to the operation, which is located below a
major regional city, is the limited surface area which can be used
for drilling and infrastructure as shown in Figure 2.
This initial understanding of the geology and the regeneration
of the historical ore zones has allowed the geologist to create
exploration targets where ore zones are expected. The latter
drilling is then targeted at these prospective areas.
The second level of drilling, L2, is completed at 100 m
spacings along the strike of the field. This is used to determine if
potential exploration targets are worth chasing up and to define
inferred resource targets and some indicated resources which
based on geological interpretation are located on the stronger
fault structures. Results from this level of drilling are used for
decline concepts and low level economic studies.
Systematic drilling and relative mine design stages.
Typical drill
section
Increasing level of geological
L1 Drilling 500 m
knowledge and confidence
Inferred 100 m
(L2 Drilling)
Indicated 1 (L2 Drilling) 50 m
Indicated 2 (L3 Drilling) 33 m
Indicated 3 20 m
(L4 drilling and mining)
Measured
230
CAPITAL DEVELOPMENT
Historically underground capital infrastructure consisted of a
single shaft, some larger mines developed secondary shafts for
ventilation purposes. Typically mines had a shaft that serviced a
strike length of approximately 250 m. The majority of the main
shafts sunk in the Ballarat East field were approximately 300 m
deep with the deepest extending to around 490 m. Shafts consisted
of multiple compartments for access, haulage and ventilation.
Typical shaft dimensions were 3.5 - 4.5 m × 1.5 m timber lined.
By the late 1800s mining was conducted by larger mining
companies and planning appeared to be at a higher level with
15 to 20 major mines operating over a strike length of 4.1 km.
The success of an individual mine was largely due to the location
of the mining lease that was issued to that company with some
zones producing larger amounts of gold than others. Although
individual mines had its ups and downs, gold production from
the entire field was at a relatively consistent level. This consistent
production level was achieved by having a large number of
production areas available across the field.
Shaft access allowed the historical mines to operate in close
proximity to each other. Shafts also allowed for greater flexibility
in the positioning of cross-cuts to locate suspected key
geological features and ore.
Today the underground capital infrastructure consists of four
parallel declines at an average spacing of 110 m vertically to the
north and two parallel declines in the southern end of the field.
An intake air shaft of 5.0 m diameter and an exhaust shaft of
6.0 m diameter. The number and positioning of the declines is
based on providing both horizontal and vertical access across the
3.2 km strike length that has currently been drilled. The declines
are also positioned to intersect surface infrastructure in relation
to areas in the city of Ballarat where space is available.
Infrastructure placement is managed via diamond drilling data.
To ensure infrastructure is placed correctly, drill platforms
developed on the Sulieman Decline (highest of the four parallel
declines) need to be well in advance of mining in the lower
declines. This enables exploration and resource definition
(L2 drilling) to guide the placement of infrastructure. If the
L2 definition drilling has not been completed, infrastructure can
be placed in a less than optimal position either increasing access
development to ore locations increasing total mining costs or
intersecting potential ore target hence sterilising orebodies.
Developing the capital infrastructure in this manner will allow
early access to the majority of the field so that consistent gold
production can be achieved similar to that of the collective
historical mines.
The current capital development of the mine is positioned
between the First Chance and the Scandinavian lines. The standoff
distance for the declines to the First Chance line is approximately
TABLE 1
Typical target Geological definition and mine design Planning stage
spacing
100 m Geological interpretation, targeting
20 m Determine potential targets, captial development MCA
concepts
15 m Capital development designs MDA
/MPA
10 m Resource shape definition, ore access targets, MDA
conceptual stoping stratergy
5m Stope design and mining MPA
Unlikely to reach this category
Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference
 HISTORICAL CHALLENGES, MODERN SOLUTIONS AT BALLARAT EAST

LGL-Ballarat Goldfields >> 2004 Aerial Image: Current and Future Underground Development

Sturt Street Victoria Street

Bridge Mall

Eureka Street
5

6
Grant Street

1
Gold Museum

Sovereign Hill

Magpie Street

Elsworth Street
2
8

Geelong Road
3

400 0 200 400 600 800 1000

NORTH Scale in Metres

Legend
1 Planned Underground Development (in Green) 5 Decline Development - March 2008 (in Yellow)

2 North Prince Extended Ventilation Shaft 6 Golden Point Ventilation Shaft

3 Waste Rock Bund 7 Llanberris De-watering Treatment Facility

4 Woolshed Gully Decline Portal 8 Yarrowee River

Printed on 100% Recycled Paper Ref No: 310308/Dev(a)

FIG 2 - Ballarat city showing position of Ballarat East underground mine.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 231

K WILLIAMS and M SYKES
80 m west, given this, the limited gradient of 1:6.5 for access
development and 110 m between declines, planning for access
development can be difficult and requires good geological
information on ore location as well as geological structures.
Decline positioning and the number of declines is based on
effectively using and minimising waste development to maximise
the economic extraction of the ore. If declines were positioned
further from the mineralised zones some resources may become
uneconomical due to the excessive amounts of waste development
required to access the resource. As mentioned mechanised mining
is limited by the gradients at which development can be mined.
Not only is the gradient limited by mechanised equipment, the
radiuses for development in regards to both equipment and
stability due to excavation size is limited. These limitations require
a greater knowledge of the geology so that sterilisation can be
minimised.
Historically the majority of the shafts were sunk on the
Eastern side of the First Chance Anticline; this allowed access to
the First Chance, Scandinavian and Sulieman lines from the shaft
with a cross-cut developed west; however, few mines capitalised
on this opportunity. Multiple shaft accesses to the Ballarat East
field as it was in the 1800s provided greater flexibility for ore
extraction compared to the current capital plan. This is due to the
small footprint that a shaft has, as well as the flexibility to
develop off the shaft at almost any level. This level could have
been based on the west dipping faults and the indicator beds in
upper levels, although due to a lack of knowledge, levels were
typically mined at a fixed interval of 100 feet with little
consideration to geology. The close proximity of the shafts to the
ore also reduced the amount of geological information that was
required to develop a level. Multiple shafts and non-diesel
equipment meant that the large ventilation infrastructure needed
today was not a consideration and therefore a cost. Due to the
depth of the current mine and the location of Ballarat City,
together with shaft sinking costs through the highly weathered
zone, multiple shafts to access the field is not an economical
option available for the extraction of ore. Figure 3 illustrates the
comparative coverage of the field via shafts verses to current
capital decline design.
Geological interpretation also plays a key role in geotechnical
stability of the mine. Each fault location within the mine creates
poor and deformable ground conditions. With the large amount
of exploration drilling these faults are able to be modelled in
three-dimensional space, allowing development to be designed
with preferential orientations in relation to the fault surfaces.
Development in the southern area of the mine intersects
cross-course faults at approximately 50 m intervals. As the
mining moves to the northern end of the field the spacing extend
out to 200 m. Ballarat East, with its low rock strength and poor
10400mRL Intake Raise
10500mRL
10600mRL
FIG 3 - Long section of Ballarat East capital development and the historical workings.
232 Ballarat, Vic, 14 - 15 October 2008
ground conditions, has a relatively high primary ground support
cost. Continual diligence with the fault modelling is required to
ensure that unknown faults are not intersected with development
or known faults are not intersected at inappropriate orientations
with respect to the development. Faults are typically defined by a
clay gouge from 0.01 m to 1 m thick with altered rock up to 20 -
30 m from the fault zone. The rock mass around the faults is
lower strength than the surrounding rock and development
through these zones exhibits deformation when supported with
the standard support designs. By predicting the fault positions
and planning secondary ground support around these faults
deforming ground conditions can be managed. Wall convergence
is monitored through a series of SMART cablebolts and
convergence pins to determine requirements for additional
ground support. A stiff support system is used to combat wall
convergence with cablebolts, fibrecrete and shotcrete sets being
some of the methods used to control the deformation.
Ventilation
Historically ventilation shafts were not always required; the
mining methods used only produced dust as a contaminate,
therefore only small quantities of airflow were required. One
compartment of the main shaft was set aside for ventilation with
numerous methods used to create an air circuit. Early on in the
field a vent circuit was created by using wind sails to funnel air
down the shafts. As shaft depths increased ventilation techniques
moved to lighting a fire at the base of the shaft (which in some
cases led to burning of the timber lining). Later techniques
including the use of water columns and ducks to force air down
shafts. Eventually primitive steam driven vent fans (blowers)
were used to push air into the underground environment.
Ventilation circuits were created by linking accesses between
two separate mines to provide flow though ventilation for both
mines. As individual mines had low production rates and a low
requirement for airflow, shafts were small.
With modern mechanised mining techniques development
dimensions are increased to accommodate larger equipment
for higher production rates. This equipment is typically diesel
powered, which creates the majority of contaminates
underground. The increase in contaminates, depth of the current
operation and the location of Ballarat City requires the current
mine plan to have two shafts for airflow. Diameters for the intake
shaft is at 5 m and the exhaust at 6 m. These sizes allow for over
400 m3/s airflow when the mine is in full production.
The size of both the development and shafts has amplified the
effect of the poor ground conditions. Although modern ground
support technology allows for these excavations to be sufficiently
supported, the development rates can reduce dramatically.
Exhaust Raise
Historic mines
Mined development
Planned development
Narrow Vein Mining Conference
 HISTORICAL CHALLENGES, MODERN SOLUTIONS AT BALLARAT EAST
Due to the poor ground conditions vertical shafts of greater
than 2 m must be permanently supported. Planning of suitable
shaft locations must take into account mining method, timing of
shaft and location of major faults. Major faults intersecting the
shafts cause the majority of ground control difficulties as these
faults are generally clay gouged with blocky and softer ground
conditions extending 20 - 30 m away from the fault surface.
Shorter internal shafts can be planned between major faults to
reduce the risk of failure. As the major faults are spaced every
50 vertical metres and there is some minor faulting between
these it is impossible to place life of mine raise bore shafts in
these locations. Shaft positions and sizes are balanced between
the time required to construct the shaft, ground conditions
expected and the time when ventilation is required for mining, as
permanently supported shafts use slower construction methods
than unsupported methods. Raise bore shafts are used where
ventilation is critical to the mine plan and smaller diameter shafts
will achieve the required ventilation flow. They can also be used
as temporary shafts whilst the larger, supported shafts are being
excavated.
De-watering
It is a common belief that the historical mines operating within
the Ballarat East field closed partially due to their inability to
pump the volumes of water required with the primitive
technology available. This is somewhat true in that towards the
end of mining in the Ballarat East field water inflows were an
issue. The water issue was mainly due to the closure of mines
and therefore pumping systems. The majority of mines were
connected physically, meaning that the pumping systems in a
single mine were required not just to pump their own water but
percentages or sometimes all of neighbouring mines’ water.
Perched water was also an issue with mines reopening and voids
filled with water draining into working mines, in some cases
causing multiple fatalities. Not only were mines physically
connected, mines were connected geologically with the west
dipping faults acting as a conduit for water from reservoirs such
as old workings. Overall the Ballarat East field has a moderate
inflow of water. Historically mine dewatering was conducted via
bailing buckets or later steam piston pumps within the shafts.
The historical mines did not have the ability to use ‘face pumps’
and therefore levels were mined at 1:100 up slope as a minimum
with all water reporting back to the shaft for pumping to the
surface.
It is a similar story today with water reservoirs within the old
workings being connected to the current workings via the west
dipping faults. This water cannot only interfere with mining it
can also hinder the exploration drilling when a hole intersects a
fault transporting water from the old workings as the pressure is
such that drilling cannot continue safely. This makes exploration
for deeper orebodies around areas of high water levels difficult.
As the mine moves into production along these west dipping
faults the water contained within the old workings will become
of greater concern. The historical workings are being de-watered
via either boreholes underground or borehole pumps within
historical shafts. Currently 40 per cent of the historical workings
have been dewatered. By 2010 all of the historical mines in the
Ballarat East field will be de-watered to a steady state.
General mine dewatering is completed via flygt and mono
pumps. Modern mining practices use more water and produce
more silt. Silt is produced as the low strength siltstone breaks
down beneath heavy equipment tyres. The silt eventually gets
washed into sumps and into the mines pumping system. A strict
regime of staged sumps and pumps is used to control this silt,
which would otherwise cause high pump wear and consequently
costs. The mine water typically contains between five and
30 per cent solids and at times up to 50 per cent solids; therefore
de-silting the water prior to recycling is critical. This is achieved
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
by pumping water into the historical North Woah Hawp mine to
settle out the silt allowing clean water to be then pumped to the
surface via the historical shaft. This solution has a finite life
before the voids are silted up and water can no longer be pumped
through the old mine. To overcome this plans are underway for a
surface de-silting facility to provide a long-term pumping
solution. Due to the location of the mine in relation to the city of
Ballarat the pumping strategy requires all of the mine’s water to
be pumped via underground infrastructure to the southern end of
the mine and then to surface. This provides challenges in mine
planning to be able to reduce the number of pumps being staged
underground, as well as minimising the requirement to install
high-pressure steel pipes within the haulage declines. At this
point capital development continues deeper and to the north and
does not allow for a permanent pumping station to be installed at
present. Until the position is reached to install a permanent pump
station de-watering will be completed via staging 103 mono
pumps.
At present some of the historical mines still contain water,
modern mining practices generating more silt and the expansion
of Ballarat City restricting positioning of a rising main.
Improvements in pumping technology have overcome the water
issues that faced the historical miners in Ballarat East. These
improvements have also led to the use of face pumps that allow
modern techniques to mine down gradients, which gives a greater
advantage when ‘chasing’ quartz veins to the incline limitations
that faced the original miners.
ORE DEFINITION
The third level of drilling L3 is completed at 30 m spacing in
areas where the resource is deemed economic to mine. A detailed
interpretation of geology and grade variability is undertaken
which gives a guide to the size and shape of the orebody to allow
the mine design engineers to complete the concept designs. This
level of drilling focuses on defining the upper, lower and strike
extents of the resource shape. These extents are used to
determine the positions for access development and sill drives.
Once the extent of the orebody is known, and an indicated
resources shape is completed, the sill drive can be designed and
the recommended mining method for the expected stope shape
can then be determined.
ORE ACCESS
Historically main haulage shafts were sunk close to a vertical
unit (the ‘indicator’), which when intersected by a west dipping
reverse fault, was the primary location for gold mineralisation.
Historic miners would drive out from the shaft towards the
indicator unit and then once on this unit would rise up until a
west dipping fault was intersected and EUREKA off they would
mine on the payable lode. In modern mining once access drives
have been mined and costs sunk it is too late to find the orebody
is no longer payable. Ballarat uses a systematic drilling program
to ensure that resources and costs are allocated to the economic
orebodies.
Historically there was little planning undertaken from the
miners. Exploration and mine planning were all rolled into one.
Initially a drive would be extended from the shaft until a quartz
reef was intersected and then a strike drive would be driven along
the reef. Little was known about the orebody until the strike drive
was completed with rises or winzes used to determine the
vertical extent of the orebody. This type of haphazard stoping
would create a very uneven production profile for a modern day
mining company. Ore production is high whilst the orebodies
were being mined. Once that orebody has been depleted a new
ore zone is required; the original miners endeavoured to have
multiple production areas within a mine to smooth production;
however, this was not always possible. The historic miners at
233
K WILLIAMS and M SYKES
Ballarat were regularly losing ore lodes where they were off-set
by cross-cutting faults. Without understanding the off-set it was
difficult to find the ore extension. These cross-cutting faults are
weathered to 300 m below surface with a halo of poor ground
conditions up to 20 m from the fault surface. This meant that
once the faults were intersected the miners battled to control the
ground.
Modern exploration and mining uses a different approach
where the geology and orebody shapes are defined from diamond
drilling well before the ore strike drives are mined. At Ballarat
East even with large amounts of exploration drilling the geology
and coarse gold grade variability is complex and difficult to
understand. Often the interpretation of orebodies change once
mining begins and the geologists are able to obtain a greater
understanding of the geology as the lode is exposed. This can
make planning difficult especially when ore shapes have changed
once strike drives are completed. To ensure that the production
profile remains consistent with changes in the ore shapes a larger
number of production areas are scheduled than required. The
increased geological knowledge will downgrade some resource
shapes and upgrade others. By having a large number orebodies
in production at the same time the risk is spread further over the
field and the ups and downs should even out. A range of mining
methods are considered in the planning stages and the mining
operations must remain flexible to switch methods as more
detailed information is collected on the geology, orebody and
grade distribution.
STOPE DEFINITION
Historically payable lodes were assessed by parcelling bulk
samples through the stamp battery and determining grade. If the
lode was not payable then mining ceased in the area and a new
lode was sought otherwise the lode was mined until it was no
longer payable or depleted.
Modern techniques allow payable sections of the orebody to
be defined via drilling. The final level of drilling L4 is typically
completed in stope from the sill drive. This level of drilling gives
a greater understanding of the grade variability throughout the
orebody though the high nugget effect still precludes the
upgrading of the resource from indicated to measured at this
stage of the operation. This gives the final stope shape for mining
to design a detailed stoping strategy and mine plan around.
STOPING
Historically the main stoping method at Ballarat was hand-held
cut-and-fill. Now the primary stoping methods at Ballarat East
are mechanised cut-and-fill stoping and longhole open stoping
with other methods including room and pillar and hand-held
stoping available depending on the size and shape of the orebody.
With the large range of expected stope shapes and sizes it is as
important today as historically to have a variety of stoping
methods available to cope with the ever changing geometries; by
locking into one stoping method the smaller orebodies can be
diluted or production on large orebodies can be slow.
Soft rock conditions and closely spaced orebodies also mean
that stope sequencing is very important to the long-term planning
of the underground. Whilst the stress field at Ballarat is
considered average for eastern Australia the soft ground
conditions create a high deformation environment. Time
dependant deterioration of the ore drives is observed regularly
and therefore in the planning process the timing of access
development, strike drive development and stoping is critical to
ensure that no development is left open for excessive time
periods. Planning to minimise the open time for all excavations is
critical in managing this deterioration. It also reduces the ground
support rehabilitation required over the life of the opening.
234 Ballarat, Vic, 14 - 15 October 2008
PLANNING
As already stated the original miners planned the mines
differently to today’s practices. Their planning was based on the
belief that ore would be intersected with each of the cross-cuts
that were mined. Based on mine managers’ reports it is assumed
that limited production forecasting was conducted once ore had
been intersected. It is believed that forecast production figures
were not used to attract investors; however, ‘selling’ of the mine
and its philosophy was. For example new management with new
ideas and techniques would be introduced to either increase
production or grade to attract investment.
Figure 4 outlines the planning process adopted by Ballarat
Goldfields LGL.
5 Year Strategic Plan
Updated annually
Design Tool: MCA
Rolling 18 Month Forecast
Updated quarterly
Design Tool: MCA
Rolling 3 Month Schedule
Updated monthly
Design Tool: MDA
Monthly Plan
Updated monthly
Design Tool: MPA
WeeklyPlan
Updated weekly
Design Tool: MPA
Daily/ShiftPlan
Updated daily
Design Tool: Survey Memo
FIG 4 - Planning flow chart.
The planning process starts with the life-of-mine (LOM) plan.
From the life-of-mine plan a strategic business plan is developed
which is updated annually. The next level of forecast is the
18 month forecast. The 18 month forecast is reviewed and issued
every quarter. A three month plan is developed every month that
incorporates the monthly plan. From the monthly plan weekly
and shift plans are developed.
In today’s mining environment shareholders and investors
require significantly more information and security than was the
case in the late 1800s and early 1900s. Planning within the
Ballarat East Mine integrates the geological drilling strategy into
the planning process.
At the early planning stage for the extraction of resources a
number of concepts (mine concept approval: MCA) are
developed for each resource. The intent of the MCA is to
determine the viability of the resource and the impact it may
Narrow Vein Mining Conference
 HISTORICAL CHALLENGES, MODERN SOLUTIONS AT BALLARAT EAST
have on the forecast. MCAs outline the mining method and
timing with minimal detail to allow timely evaluation of a
resource prior to additional planning resources being dedicated.
These concepts are typically based on the L2 drilling and in
some cases L3. Each of these concepts are reviewed based on
economics, expected geological and geotechnical conditions and
then risk ranked based on knowledge of ore variability, ounces
delivered, mining and geotechnical risks to determine the total
risk of each concept to the overall mine forecast. In most
instances the concept with the lowest risk to the mining forecast
is chosen as the initial mining process with the expectation that
with increased information the mining concept may vary. The
chosen concept is used to design the next program of drilling so
that access take-off points can be established and the best ore
access position can be determined. MCAs are typically
completed by the planning engineer before the generation of the
forecast.
As the level of information increases the design phase moves
across to the mine design engineers and the creation of a MDA
(mine design approval). The MDA is a full design of the resource
with detailed costs and mining methods evaluated. The MDA is
to be completed in line with the approved MCA. Once detailed
designing is being conducted and it is identified that the MCA is
not suitable the process starts again, with the recreation of the
MCA.
Implementing the design is completed via the MPA process.
An MPA is a detailed design of a component within the MDA.
It is at this point that thorough design checks, risk assessments,
etc are completed prior to issuing to survey for construction.
Scheduling is a critical component of the planning process at
the Ballarat East Mine. The scheduling package used is EPS
(Gantt chart based software published by Datamine). The use of
EPS is essential for evaluating new concepts and designs. It
allows the forecast to be updated based on the latest data to
evaluate the impact due to changing information. This allows
priorities to be set and changed as required to ensure production
targets are meet.
CHALLENGE MANAGEMENT
Due to the challenging conditions at Ballarat East
communication between the planning, design, geology,
geotechnical and operations teams are critical to success of the
operation. The mine has been split into three areas, north, central
and south. Each of these areas has a dedicated mine design
engineer, two mine geologists and moving forward a
geotechnical engineer to complete all of the geological
interpretation, mine design and stope sequencing. Every design,
before being issued to the operations team, is carefully peer
reviewed by the owners of each challenging condition. The
sign-off includes reviews by the mine geologist, geotechnical
engineer, dewatering engineer, operations engineer, ventilation
engineer, mine surveyor, mine foreman, senior mining engineer
with final sign-off from the mine manager. The mine geologist
reviews the design to assess if fault models are up-to-date from
recent drilling, ore models are satisfactory and the design does
not sterilise any potential ore zones. They also check to
determine if there are any risks associated with open drill holes.
The geotechnical engineer reviews the design to determine if the
ground support standards are adequate, if there are any major
faults intersected at unfavourable angles as this increases the
required ground support or long term rehabilitation for
deforming ground conditions. They also review the changes in
stress over time for each excavation to determine if additional
support is required, or if the development is in the most suitable
location. The mine surveyor is responsible for checking that the
development is not encroaching on exclusion zones placed
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
around historic workings and therefore increasing the risk of
water or mud inrush. The dewatering engineer reviews the mine
dewatering strategy including drainage and pumping options.
The ventilation engineer reviews the vent plan for the designed
development, then determine how this fits within the entire mine
ventilation circuit.
Historically miners at Ballarat managed their challenges
through experience with the ground conditions and local
geology. There were still many fatalities due to falls of ground,
water inrushes and inadequate ventilation. The team at Ballarat
Goldfields has a rigorous approach to planning to ensure that the
challenges faced are identified at the initial design stage and
managed through the operational mining phase.
CONCLUSION
Whilst the technology for mining has changed since the Ballarat
gold rush, 150 years ago, the challenges faced have not. The
following exert from an article in a student magazine from 1900
shows that even the historical miners knew the importance of
appropriate stope methods and designs to manage the challenges.
The stoping of a lode a few feet in width is a
comparatively simple matter, but in dealing with
bodies of quartz or ore, say 80 or 100 feet wide,
the work must be governed by some well thought
out system or principle, and this applies with
greater force where the lode is heavy and liable
to crush too often do we find stoping operations
carried out in an unsystematic and haphazard
manner, the result being danger to life and limb,
bad ventilation and high cost of production,
beside, in some cases, the loss of large quantities
of valuable ore (Lake, 1900).
Ballarat East is still in the early stages of economic production
with the geological understanding growing with each orebody
extraction. During this ‘learning’ period planning is designed to
allow for flexibility in the mining method and equipment size. It
may be argued that this can lead to a suboptimal extraction
method for the resources; however, it allows for extraction of the
ore no matter what challenge arises and provides a platform
for learning. Over time this will allow improved planning as
the confidence in the geological interpretations and the
understanding of grade variability improve.
ACKNOWLEDGEMENT
The authors wish to thank LGL and Ballarat East Mine
Management for permission to publish this paper. They would
also like to thank Peter d’Auvergne for his assistance with the
historical details and interesting stories on mining in the Ballarat
gold rush period.
REFERENCES
Baragwanath, W, 1923. The Ballarat Goldfield, Memoir 14, 257 p
(Geological Survey of Victoria).
d’Auvergne, P B, 1990. Ballarat east gold deposits, in Geology of the
Mineral Deposits of Australia and Papua New Guinea (ed: F E
Hughes), vol 2, pp 1277-1278 (The Australasian Institute of Mining
and Metallurgy: Melbourne).
d’Auvergne, P B, 2008. Personal communication, June.
Lake, J, 1900. The stoping of wide lodes, in Ballarat School of Mines
Magazine, September, pp 4-5.
Osborne, D J, 2008. The Ballarat East goldfield: New insights on an old
model, in Proceedings Narrow Vein Mining Conference, pp 59-70
(The Australasian Institute of Mining and Metallurgy: Melbourne).
235
Tarnagulla Gold Mine
P L McCarthy1 and L Faulkner2

ABSTRACT
The Tarnagulla Gold Project in Central Victoria, Australia, was
developed and operated by Reef Mining NL between 1994 and early
2000, producing 53 000 ounces of gold from the Nick O’Time Shoot, a
single narrow vein shoot structure within part of the larger Poverty Reef,
with a strike length ranging from 30 m to 60 m. The ore shoot was
developed over a vertical interval of 190 m, using a small cross-section
decline and small, mechanised equipment. Various stoping methods were
employed including mechanised cut-and-fill with a single boom jumbo,
hand-held mining using airleg drills and rope scrapers, and small uphole
retreat stopes using a line of sight remote operated LHD to recover a
crown pillar. Experience during stoping showed that airleg drilling and
rope scrapers gave much better dilution control than diesel equipment.
Innovations included a low-cost hydraulic fill system.
Due to the small scale of the deposit, the feasibility studies, planning
and permitting for the mine were based on very limited drilling
information; however, the mine was brought into production within two
years of the discovery.
While the mine made an operating profit, it failed to fully repay capital FIG 1 - Tarnagulla in the late 1850s.
at the low gold price then prevailing. Applying the lessons learned from
the operation to optimise the access and stoping method, the project is
re-examined at today’s gold price and operating costs. Central Victoria of 6000 tonnes at 7 g/t gold in the old Prince of Wales and
has the potential to host many similar small deposits. Poverty shaft workings adjacent to the Bonanza Shoot. WMC
The authors were respectively Non-Executive Director and Resident rehabilitated the small Prince of Wales shaft (Figure 2) to the
Manager for Reef Mining NL when the mine was operating. No 5 (166 m) Level where some driving and cross-cutting was
undertaken. Three diamond drill holes proved the continuity of
the Poverty Reef 330 m south of the Poverty shaft to 310 m below
INTRODUCTION surface, with variable assay results and some visible gold.
Central Victoria was famous in the 19th century for large
goldfields such as Ballarat and Bendigo and also for smaller
fields on which hundreds, possibly thousands, of narrow reefs
were mined with grades often exceeding one ounce per tonne.
The Tarnagulla Gold Project is the only such mine worked on a
smaller field in the modern era. The experience provides insights
into the economics and impact of such a project in the 21st
century.

LOCATION AND HISTORY

The historic mining town of Tarnagulla is located 45 km west of
Bendigo and approximately 160 km north-west of Melbourne.
Gold was discovered at Sandy Creek, Tarnagulla in late 1852
or early 1853. The initial mining of alluvial gold turned to
mining of reef outcrops, with the first serious attempt occurring
on the Poverty Reef in 1854 and 1855 (Wilkinson, 1996). Of the
estimated 420 000 ounces (13 063 kg) of gold produced from the
Tarnagulla field prior to modern production (which started in
1996) 360 000 ounces (11 197 kg) was produced between 1853
and 1865 from mining 122 000 tonnes of ore in the Poverty Reef
‘Bonanza Shoot’ (Figure 1). The recovered grade of 92 g/t gold
implies a head grade of around 100 g/t gold. There is limited
historical information on the size and style of the Bonanza Shoot,
but it was reported to be 2 m to 7 m wide, 110 m long and 120 m
deep below its discovery position in outcrop.
Western Mining Corporation Limited (WMC) explored the
property from 1985 to 1989, defining a Measured Resource

1. FAusIMM(CP), Managing Director, AMC Consultants, Level 19,

114 William Street, Melbourne Vic 3000.
Email: pmccarthy@amcconsultants.com.au
2. Bendigo Mining Ltd, PO Box 2113, Bendigo Mail Centre Vic 3554.
Email: lfaulkner@bendigomining.com.au FIG 2 - Prince of Wales shaft head frame Tarnagulla.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 237

238
POVERTY SOUTH
NORTH PRINCE OF VICTORIA
WALES POVERTY SYNDICATE
SHAFT SHAFT SHAFT Crystal Hill
712N SHAFT 662N
Decline Portal
RL
1200 RL1200

POVERTY TRNC 23
P L McCARTHY and L FAULKNER

5.0/0.4 RMD 26
2.0/0.93
NICK O'TIME SHOOT
TRNC 26 RMD 12
1.5/0.31 1.2/0.3 RMD 69
BONANZA RMUD39 RL 1150
2 Level TRNC 21 0.1/0.2 RMD 71
RMD 2 3.0/0.2 0.8/913 0.1/0.3
TRNC 25 RMD 57
RMUD 36
SHOOT RMD 3 TRNC 19 Block 4
2.0/0.54 TRNC 22 New Block 4 0.5/67.5
RMD 4 2.5/0.47
5.0/0.9 RMD 11
2 Level Access Ramp
TRNC 20 5.0/0.14 RMUD 9 0.3/24.0
LOW RMUD 5 0.5/165
STOPE 3.0/0.3 RMUD 7
RMD 1 GRADE RL 1100
STOPE 3 Level
Crysta 0.1/567
1.0/0.4 l Hill D
RMD 20 0.4/674, RMD 24
TRNC 24 ecline RMD 32
2.5/36.5 0.1/0.79
3.5/1.2 TRND 6 W2 0.08/0.12
PDH 2 TUD 5-2 1.8/0.47
1.5/0.7 4 Level Fault Blank RMD 8 RMD 25
TUD 5-7 RMD 31 3.0/87.9
PDH 1 5.3/2.5 RMD 7
1.0/1.1 RMD 62 5.0/0.8 RL 1050
5 Level RMD 19 7.0/19.6 RMD 65
TRND 6
Poverty Reef. Twenty-six surface diamond holes had been drilled
exploration program to identify further high-grade zones in the
been raised, Reef Mining immediately embarked on a staged
WMC ceased work on the project in December 1989. Reef

by the end of December 1994, of which holes RMD19 and

Mining NL subsequently purchased it and was floated as a
‘single project’ company in March 1994. Once funds had

TRND 1501 RMD 9 0.2/6.4

1.8/0.7 2.3/3.1 RMD 17 RMUD 10 2.0/208.5
TUD 5-8 5.5/3.9 0.8/50
PDH 3 TRND 1509 1.0/2.6
PDH 4 RMD 23
1.0/1.67 TUD 5-5 3.5/5.7 RMD 46 4.0/6.6 0.3/0.03
6 Level RMUD 24 RMUD 52
PDH 7 RMD 6 RMD 13 RMUD 46 2.5/11.7
TRND 1502 3.0/33.8 RL 1000
2.2/3.1 4.0/0.57 3.5/0.9
2.5/0.7 2E/1 RMUD 27 0.2/403
RMUD 44
7 Level 0.8/2.0 RMUD 55 RMUD 53 0.5/4.1
TRND 6 W3 RMUD26 RMUD 51 0.8/22.6
3.0/30.6
1.2/2.4 3.5/9.0
2E/2 RMUD 51 0.9/10.9
RMD51 0.9/10.9 RMD 22
TRND 1503 PDH 4 RMUD 58 0.3/0.7

Ballarat, Vic, 14 - 15 October 2008

RMD 21 1.0/0.03
1.0/0.2 No Reef RMUD 57 0.5/37.5
PDH 5 RMD 64 1.0/3.4 RMD 18
RMUD56 RU72A 3.0/0.3 RL 950
RMD 18 0.3/0.3
9 Level 1.5/4.4 1.5/3.0 0.3/0.3 1.0/4.3
2E/3 RMUD 62
RMUD 60 RMD 28 0.5/2.6
0.2/2.7RMUD 60 RU 77 0.5/6.4
RMD 15 RMD 18 0.3/7.4 1.1/58.3
Author : B. CUFFLEY
0.9/0.9 0.6/13.3 RU 61
Scale : 1:2500 Surface Percussion Drillhole TRND 1508 W1 3.0/7.6
0.4/5.9
Date : July 1999 RMUD 68 RMD 63 0.2/10.6 RL 900
Surface Diamond Drillhole RU 66
RMD 10 1.5/4.3 RMUD65 0.8/14.0
312m 0.8/16.1 TRND 1508 0.8/2.8
5.3/2.5 Estimated True Width of Reef/ 1.0/<0.01
RMD 1994-1995 Reef Mining N.L. RMD 14 RU 84 RMD 73
Average Assays (grams/tonne) 1.9/8.9 RU 78B
0.5/4.6 RU 73 RU 71 0.3/7.5

REEF MINING N.L.

RU,RMUD Underground Diamond/Reef 1.8/6.6 RU 80 0.5/35.2
Underground Diamond Drillhole TRND 1507 1.5/1.4 2.0/3.0
PDH C.V.G.M. N.L. 1.0/2.2 RU 81 RU 79 1.5/1.4
Visible Gold in Core 1.4/9.1 RU 67
TRND W.M.C. Ltd. 0.7/29.1 RU 74 1.5/1.5
High Grade Stope RL 850

MIN 4756 TARNAGULLA GOLD PROJECT

TRNC W.M.C. Ltd.
Low Grade Stope

POVERTY REEF QUARTZ SEGMENT

TUD W.M.C. Ltd.
1994 Costean Un-mined shoot / High grade(20-40g/t)
which remained open to the south (Figure 3).

(Reef Mining N.L.) Un-mined shoot / Low-medium grade(7-10g/t)

5926700N
5926100N

5926200N

5926300N

LONGITUDINAL SECTION 752950E

400N

500N
RL 800

FIG 3 - Longitudinal section through Poverty Reef showing relative position of Bonanza and Nick O’Time Shoots.

Narrow Vein Mining Conference

results showed the presence of the reef some 100 m further south
of at least 50 m strike length, up to 70 m high and 5 m wide. The
width at 87 g/t gold. These results suggested a high-grade zone
19 and 61 g/t gold respectively. RMD25 intersected 3 m true

than previously known, giving an overall strike length of 550 m,

RMD20 gave true reef widths of 7 m with average grades of
 TARNAGULLA GOLD MINE

Given the nuggetty nature of the gold, underground access was that a rapid transition to lower overall grade occurs (estimated
required to progress the evaluation. Based on the dimensions and 31 000 tonnes at 6.4 g/t gold). Only the high grade section has
a presumed grade of 30 g/t gold for the high-grade zone, been mined, to a depth of 250 m below surface. The change in
underground development was justified at the then gold price of structural setting at this depth from the east limb to the west limb
A$500 per ounce. of the Poverty Syncline is considered to be a controlling factor
on grade (Figure 4).
GEOLOGY AND MINERALISATION The shoot is a near vertical pipe-like body that is lensoidal in
plan view. The structural elements of the shoot are the massive
This geological description is derived from Krokowski et al quartz core (low grade), the laminated, high-grade mineralised
(2001). All primary gold mineralisation exploited in the hanging wall and foot wall veins and a persistent footwall fault.
Tarnagulla goldfield has been mesothermal quartz vein (reef) The laminated high-grade veins are up to 1.5 m in width and
hosted in fault structures developed in a tightly folded occur as oblique hanging wall and footwall veins in an en echelon
Ordovician turbidite sequence. No gold is known to occur in the pattern around the central core of low-grade, massive, milky
sandstone/slate wall rocks. Gold occurs in discrete high-grade quartz, which ranges up to 7 m in width.
shoots within lower grade quartz reef segments. The field is The combined strike length of the overlapping economic grade
characterised by the old miners adage, ‘nugget or nothing’, veins that make up the shoot varies from 15 m at the top of
which alludes to the fact that gold is generally highly the shoot to 75 m at the bottom (Figure 5). Tension gash, massive
concentrated in favourable structural settings within a quartz reef quartz spur veins occur in the hanging wall of the shoot and in
segment, the remainder of the segment being low grade. the reef channel between the structures. Typically, the narrow
The Nick O’Time Shoot extends from 70 m below the surface ‘tail ends’ of the veins are very high grade (up to ~1000 g/t gold).
to a depth of 370 m. The upper 180 m of the shoot is high grade A number of small subshoots extend off the northern boundary of
(57 400 tonnes average head grade 29.1 g/t gold) and below the Nick O’Time Shoot proper into the main Poverty Reef area.
752800E

752900E

753050E
POVERTY BIRTHDAY
SYNCLINE ANTICLINE

mRL 1200
WES
T PO

F
REE

F
REE
VER

Y
ERT

OCK
TY R

Block
EL
POV

4
EEF

HAV
?

ZO

'A' Fault
NE

West mRL 1100

Poverty
'Black
Wall'
?

'B' Fault Block

Structure
3
Concrete Sill
pillar mRL1070
NICK
Block
2W
'D' Fault
O'TIME
BIRTHDAY REEF
Crystal Hill
Decline Loops
Block 2E/1 SHOOT mRL 1000

Cemented sandfill
Sill pillar mRL990
F
REE

Block 2E/2
CK

mRL 965
ELO
HAV

Block 2E/3
mRL940
Footwall Vein ?
Hangingwall
Vein

mRL 900

Footwall Fault
Structure

?
?
0 50m REEF MINING N.L. Author : B. Cuffley
Date : August 1999
Scale MIN 4756 TARNAGULLA GOLD PROJECT
NICK O'TIME SHOOT - POVERTY & PARALLEL REEFS

CROSS SECTION 5926220N

FIG 4 - Cross-section through Nick O’Time Shoot.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 239

P L McCARTHY and L FAULKNER

RMD 40
RMUD 43 0.4/3.4
1.0/3.6 RMUD 42 RMD 69
0.1/0.3

RMD 39 mRL1150
0.8/913
0.8/913
LEGEND
RMD 20, Surface Diamond Drillhole
TRND 1
RMD 36
0.5/67.5 RMD 55 RMUD 3 Underground Diamond Drillhole
Pinchout RU 69
Estimated True Width of Reef/
5.3/2.5
Average Assays (grams/tonne)
RMUD 37 1129-1131 sub-level
0.4/0.9 RMD 57 6.0/3.8(inc. 0.7/28.5) Visible Gold in Core
Sandfill Hole High Grade Shoot
Foot Wall
BLOCK 4 FW
RMUD 9 3.0/24.0
RMUD 8 'Incline
4.0/0.2
RMUD 5 0.5/165 Stopes' 0 5 10 20 30 40 50

RMUD 3 Metres
RMUD 6 3.5/3.7 1102-1107 sub-level mRL1100
RMUD 7 0.1/567 1098-1101 sub-level
RMD 32 Spurry Massive
Zone
0.08/0.12 Quartz RMD 20 0.4/674
RMD 50
89 - 93
0.1/5.46
RMUD 14 e 86 - 89 RMD 20 2.5/36.5
Inclin
83 - 86 RMUD 2
Stopes 81 - 83
RMUD 13 RMD 49
77 - 81 (fault blank ? or pinch-out ?)
RMD 45 74 - 77 74 - 77
70 - 74 RMD 25 3.0/87.9
RMD 31 Concrete Sill
(fault blank) Fault
A
BLOCK 3

'5 Level' Fault

B Stopes
mRL1050
48 - 52
B2/46-49 RMD 30
Spurry Massive
Zone Quartz
0.08/0.03

RMD 17 Fault
0.8/50.4 BLOCK 2W RMD 19
Blank
7.0/19.6 5
RMD 48 RMD 29
0.04 4 Fault Blank
(pinch out) 3
RMUD 22 2 RMUD 10
2.0/208.5 Fault
2.5/3.2 1 RMUD 21
0.2/41.9
RMUD 4 RMD 46(W)
0.6/8.0 4.0/6.6 TRND 1509
RMUD 12 1.0/2.6
D ? RMUD 11 RMUD 52
5.0/0.4
3.0/2.0 RU 70
RMD 13 RMUD 46
3.5/0.90
RMUD 24 2.5/11.7 BLOCK 2E
2E/1 3.0/33.8
RMUD 23 RMUD 27 RMUD 50
mRL1000
1.5/3.6 0.4/403 6.7/890 RU 69
RMUD 19
FW RMUD 49
MEASURED RESOURCE 2.8/164

High grade RMUD 44

0.8/2.0 Cemented sandfilled
RMUD 26 crown pillar RMD 53
? 3.5/9.0 0.5/4.1
RMUD 55 RMUD 51
2.0/30.6 0.8/22.6
? RMUD 58
0.3/0.7
RMD 51 RMUD 57
2E/2 0.9/10.9 0.5/37.5
RMD 21
mRL 965 mRL 965
FW RMUD 56
1.0/3.4 1.0/4.3

RU 72A
RMUD 18 3.0/0.3 mRL 950
0.3/10.3
RMUD 59 0.6/0.02
RMD 28 RMUD 62
2E/3 0.5/6.4 0.5/2.6
RMUD 60 RU 77
0.2/2.7 1.1/58.3
MEASURED RESOURCE
Low - medium Grade
RMD 15 RU 60
RMD 18
0.2/4.0 0.3/7.4
0.6/13.3

RU 61 TRND 1508/W1
0.4/5.9 4.5/11.2
DRILLING BLOCK 2E
RMUD 63
RMUD 68 0.2/10.6
1.5/4.3 RMUD 66 RMUD 65
TRND 1508 0.8/14.0
0.8/16.1
0.8/2.8 mRL 900

RU 80 RU 78B
1.8/6.6 0.5/35.2
(0.4/26.8) RU 73 RU 71
RU 84 1.5/1.4 2.0/8.0
1.9/8.9 (0.2/10.9)
RMUD 64
RMUD 67 F.W.
0.7/29.1

RU 81
1.4/9.1
RU 79
6250N
6300N

6150N
6200N

1.8/0.2 RU 74
1.5/1.5 mRL 850

REEF MINING N.L. Author : B. Cuffley

Date : July 1999
MIN 4756 TARNAGULLA GOLD PROJECT
SOUTH POVERTY REEF
752950E LONGITUDINAL SECTION , NICK'O TIME SHOOT

FIG 5 - Longitudinal section through Nick O’Time Shoot.

240 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 TARNAGULLA GOLD MINE

PROJECT EVALUATION AND DEVELOPMENT requirements. This continued until sufficient stockpiling had
occurred to provide acoustic shielding and allow the establishment
Underground access could have been gained via the existing of a night dump location (Figure 7).
Prince of Wales shaft, which had been equipped by WMC with a Six geotechnical holes were drilled along the proposed decline
head frame and winder, or via a new surface decline. Both options route, resulting in a decision to install 65 steel arch sets for the
were evaluated assuming a two-year construction schedule, first 72 m, after which an average seven to eight bolts per metre
which included eight months of gold production. The cash with mesh support proved sufficient.
requirements were found to be similar for both cases. The
decline option was preferred because it provided access for Negotiations with potential contractors showed that the small
mechanised equipment, lower manning levels and lower ongoing scale and uncertainty of the project were likely to make contract
mine operating costs. The non-availability of qualified winder mining less attractive than owner mining. The recruitment of a
drivers was also a consideration. workforce was assisted by the closure of the Maxwell mine at
If the objective had been solely to obtain a bulk sample
through 300 m of reef driving on No 5 level, this could have been
achieved for a net cost of about $1.0 M from the Prince of Wales
shaft as against $2.5 M using decline development. Both options
required substantial further capital investment to bring the mine
into production, with the production rate for the shaft option in
the event of further success limited to a maximum of 50 000 t/a.
A significant constraint on the use of the Prince of Wales shaft
was its proximity to residences, with a history of objections
based mainly on noise during the WMC period. The decline
portal could be located well away from residences within a much
larger surface works area and there would be no need to truck
development waste and ore through town. The local Loddon
Shire advised that it had determined to issue a Planning Permit
for decline development in January 1995 after the period for
objections had expired.
On completion of the eight-month drilling program, a resource
estimate was announced comprising a Measured Resource of FIG 6 - Crystal Hill decline portal.
8000 tonnes at 7 g/t gold based on the earlier WMC work and an
Inferred Resource of 527 000 tonnes including 30 000 tonnes of
high-grade material (+30 g/t gold) referred to as the ‘south shoot’
and 40 000 tonnes of probably subeconomic material, with the
balance of 457 000 tonnes assigned an estimated grade range of
7 - 10 g/t gold.
Studies based on mining a total of 450 000 tonnes of ore at
10 g/t gold recovered grade (recovering about 145 000 ounces)
showed a net positive cash flow of $18.0 M for the shaft option
and $27.0 M for the decline option, at a gold price of A$500 per
ounce. A decision was made in March 1995 to proceed with
development of an exploration decline, to be known as the
Crystal Hill decline. Subsequent drilling in mid-1995 reduced the
estimated amount of high-grade material from 30 000 tonnes to
20 000 tonnes, but by then the decline project was well underway.

DECLINE DESIGN AND CONSTRUCTION

Tarnagulla Gold project was located on Mining Licence 4756, on
the eastern edge of the town of Tarnagulla. An initial portal
position was selected adjacent to the Tarnagulla rubbish tip,
where it was felt there could be no objections based on
environmental concerns.
A better position was then found on the eastern side of Crystal
Hill, out of the line-of-site of the town, at a time when the
company was able to purchase mining licences and freehold land
in that area. This location meant the nearest residence was an
isolated cottage located in the forest approximately 200 m north
of the portal and the next nearest residences were over 350 m
away from the portal and behind Crystal Hill.
The box cut was excavated in July 1995 and the portal
established for a decline with dimensions 3.5 m × 3.5 m at a
gradient of one in seven (Figure 6).
Development waste was stored immediately adjacent to the
portal works area and was used for construction of noise bunds
around the boxcut and other areas on site. Initially waste from
development was stored in underground stockpiles and only
trucked to the surface on day shift, to comply with the site noise FIG 7 - Aerial view Tarnagulla mine site.

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 241

P L McCARTHY and L FAULKNER
Inglewood and a reduction in activity at Ballarat Goldfield’s
Woolshed Gully decline. Locals were employed from Maldon,
Maryborough, Ballarat, Inglewood and Tarnagulla with a
significant percentage already possessing previous underground
mining experience.
Equipment purchased to commence the decline development
included a diesel generator, compressor, single-boom Secoma
electric hydraulic jumbo drill, two 3 m3 LHDs; namely an Eimco
913 and EJC 130, and a 12 tonne capacity Normet NT120 diesel
truck. The EJC130 LHD was purchased new, but the remaining
equipment was purchased second hand.
Other surface infrastructure and facilities included an access
road and car park, offices, workshop, and explosives magazine.
The original Work Plan proposed de-watering the existing
historic workings in stages, ensuring all of the workings above
5 Level were dewatered before the mine went into full production.
When WMC dewatered the Prince of Wales Shaft from June
1988, the water was discharged into the Bullabul creek system,
with weekly sampling at locations along the creek system and at
the settling pond near the Prince of Wales Shaft. In approving the
1995 Reef Mining Work Plan, the Department of Conservation
and Natural Resources (DCNR) would not approve the discharge
of saline water (ranging from 7000 - 14 000 TDS), which included
other contaminants, off-site and required that the mine water to
be retained on-site and evaporated.
An 18 ML stormwater surface run-off catchment dam below
the mine site, a 210 ML evaporation dam and a series of bypass
drains were constructed to act as closed cells from the local
groundwater system, preventing any seepage of mine/site waters
into the local water system/catchment, and allowing disposal of
the mine water pumped from the Poverty shaft. The evaporation
pond was approved for construction and required clearing State
Forest within the Reef Mining tenements, subject to Reef Mining
purchasing and transferring to DCNR a parcel of land within the
Loddon catchment to provide a net environmental gain.
The first decline cut was fired on 9 August 1995. The face had
reached a chainage of 253 m by December of that year and a
chainage of 824 m by 30 June 1996. A section of decline within
the Poverty Reef provided an 850 tonne bulk sample that
returned 1 g/t gold, although this result was in doubt due to the
inability of the bulk sample plant to recover coarse gold.
A ventilation connection with the old No 2 (89 metre) Level of
the Poverty shaft was completed and the construction of a
ventilation barricade and installation of a fan turned the Poverty
shaft into a downcast airway with exhaust air exiting the decline
at the portal. An established ladderway within the Poverty shaft
from No 2 Level formed the second means of egress from the
mine. This process was repeated after the decline reached No 5
Level with a connection driven on the reef to the north, to test a
lower grade zone, also providing a ventilation connection to the
WMC development completed in 1989. The Prince of Wales
shaft and the Poverty shaft then became fresh air intakes with
exhaust air from active working areas returning via the decline to
exit at the portal. After assessing the condition of the Prince of
Wales shaft at No 5 Level, a decision was made to rehabilitate
the Poverty shaft from No 2 Level to No 5 Level to provide the
second means of egress; this work was completed by Equator
Mining Contractors. The intake ventilation system was extended
internally with a series of three rises extending from Block 2E/2
– No 5 Level (965 - 1050 mRL).
It was found to be difficult to maintain the nominal decline
dimensions because of over-scaling in the slabby ground and
because the miners consciously drilled out larger patterns than
the design to create more working room. Typically 30 per cent
more rock had to be loaded and hauled than the design, which
had a proportionate impact on advance rates.
During development the workforce comprised a resident
manager, exploration geologist, three shifts each of three miners,
242 Ballarat, Vic, 14 - 15 October 2008
two fitters, a truck driver and an administrative officer. Work
took place on 17 × eight-hour shifts per week, three shifts per
day Monday to Friday and with two shifts on Saturday. Other
essential work such as mine planning, survey, electrical, diamond
drilling from both surface and underground and general back-up
was contracted out.
By the end of June 1996, the company had $507 000 remaining
cash with $499 000 owed on hire purchase equipment, whilst
expenses were running at $300 000 per month. The high-grade
shoot was encountered on 30 July 1996 at a chainage of 920 m. It
was named the Nick O’Time Shoot for obvious reasons at a time
of make-or-break financially for Reef Mining.
Negotiations for access to gold plants in Bendigo, Inglewood,
Maldon and Chewton resulted in an agreement to process ore
through the eight tonnes per hour Wattle Gully plant at Chewton,
near Castlemaine. Ore was stockpiled on the surface ore pad at
Tarnagulla before being loaded into highway trucks with a front
end loader and trucked by a contractor 70 km to the plant, during
daylight hours only.
The processing plant was operated by Reef Mining personnel
consisting of a metallurgical superintendent (contract metallurgist
from Metallurgy International), gold room attendant, crusher
operator, CIL operator and maintenance fitter. The plant was
operated on a day shift only basis with shift length varying from
ten to 14 hours per day depending on the ore supply. Processing
on day shift only allowed for some maintenance to be conducted
during normal plant shutdown and therefore plant availability
was generally very good.
The plant consisted of a gyratory crusher, a 8’ × 6’ ball mill, a
Knelson concentrator, three leach tanks and a half tonne carbon
stripping circuit capable of one strip a day; with a Gekko Inline
Pressure Jig added later to the gravity circuit by Reef Mining.
After refurbishing the plant to return it to operational condition,
the first gold production occurred in October 1996, realising a
price of A$478 per ounce. Cash flow was so tight that it was
necessary to shut down the ball mill and strip the liners
more than once to recover hung up gold. The initial budget
production of 10 kg (321 ounces) per week was achieved, with a
metallurgical recovery of 98 per cent.
STOPING
Development in the first few months showed that the upper part
of the Nick O’Time Shoot was 25 to 30 m in strike length and
0.4 to 3.7 m wide. The initial 3.5 m wide, 4 m high sill drive
produced 1400 tonnes of heavily diluted ore with a head grade of
19.7 g/t gold.
A concrete floor pillar was placed to ensure 100 per cent
recovery beneath the 70 - 74 (1070 mRL) sill drive. The concrete
floor pillar was constructed by initially laying poly sheeting on
the floor of the drive, which was covered with 300 mm of sand.
Then 1.5 m long rebars, spaced 2 m apart, were installed in both
walls before being covered by 500 mm of 25 MPa concrete, with
100 mm × 100 mm mesh laid in the base (Figure 8). The
concrete floor pillar was then overlain with waste rock before
commencing the next ‘flat back’ cut-and-fill lift.
Attention was then focused on minimising dilution with the
mechanised equipment in the second 3 m high ‘flat back’ cut-and
-fill lift, which had a strike length of 27 m and width of 0.8 to
3.5 m. Eventually three lifts were accessed by ramp above the sill
drive level (Figure 9).
During the December 1996 quarter 5114 tonnes of ore were
processed producing 120 kg (3857 ounces) of gold, at a recovered
grade of 23.5 g/t gold yielding revenue of $1.7 M at a realised
price of A$464 per ounce. The cash production cost was $250 per
ounce. Improvements in dilution control and a ‘sweet spot’ in
the reef boosted the December head grade to 39 g/t gold, with
98.7 per cent metallurgical recovery.
Narrow Vein Mining Conference

500 MM MINIMUM DEPTH OF
25 MPA CONCRETE
1.5 METRE REBAR ON BOTH
WALLS SPACED 2 METRES APART
300 MM SAND POLY SHEETING
LAID ON FLOOR
OF DRIVE
FIG 8 - Design of concrete floor pillar poured in Block 3 (70 - 74 L).
The original development workforce was augmented by about
ten miners as production built up to allow for concurrent
development, while an 18 tonne Wagner MT420 truck was added
to the fleet. A 2 m3 LHD was also trialled but was found to be
difficult to operate on the backfilled ramps and within the cut-
and-fill stopes.
As stoping progressed, the backs above the junction of the
stope and the access cross-cut failed on several occasions,
typically requiring two to three shifts for remediation. The Nick
O’Time Shoot was found to be faulted-off into three main blocks,
further complicating access design and requiring careful ground
control at the faults. Two stoping blocks were brought into
production, one above and one below the original sill drive level
in Block 3. Ongoing ground control problems required the
installation of cable bolts in the stope accesses, with some use of
timber sets in the stopes. Remote mucking was used to recover
ore from the fall areas, but this had the potential to further
destabilise the stopes. Where the reef width had reduced to
around 2 m, a decision was made to introduce some airleg
stoping. A rise was developed on the northern end of the block to
beneath the concrete floor pillar (1070 mRL) and a sublevel was
driven south exposing the underside of the concrete prior to
airleg extraction.
Total gold production for the 1996-97 year was 260.2 kg
(8383 ounces) from 10 199 tonnes treated at a reconciled head
grade of 27.4 g/t gold.
In July 1997 stoping beneath the concrete floor pillar reached
a point where the original access drive had collapsed. The stope
was then benched downwards in flitches approximately 1.5 m
high, with timber stulls installed as extra support. The southern
end of the stope was extracted to the top of the old timber sets
below, which also enabled ore from an earlier collapse further to
the south to be scraped back to the drawpoint. However, a
significant section of the footwall of the southern area then
collapsed, exposing the concrete floor pillar to its footwall and
hanging wall contacts over a length of about 12 m. Work in the
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
TARNAGULLA GOLD MINE
REEF
FAULT
RMD 25
REEF
100 x 100 MM MESH
LAID IN BASE OF
CONCRETE
area was stopped while a means was devised for backfilling the
fall area.
To introduce hydraulic fill, an existing surface exploration hole
(RMD25) was redrilled, for a length of approximately 140 m,
extended and cased through the concrete crown pillar at NQ size
(68 mm diameter). Minus 3 mm crushed sandstone material
(crusher dust) was delivered by tipper truck to a feed hopper at
the hole collar, where it was slurried with water and gravitated
directly into the stope void. After placement, it was free draining.
This material was so free flowing that a cone of material sitting
above the hole collar would gravitate down the drill rods into the
stope void without the addition of water.
The benefit of using this method for other filling applications
was immediately apparent and a number of other surface drill
holes, cased with HQ rods (58 mm diameter internal), were used
(RMD13 – 220 m, RMD51 – 250 m, RMD57 – 85 m in length).
An adaptor plate was manufactured to fit to the base of each
cased hole and the sand slurry was then transferred using lengths
of 110 mm poly-pipe to the areas within the stopes requiring fill.
To create a floor pillar in Block 2E/1 (990 mRL) a cement
sand slurry mix (seven per cent cement at 75 per cent solids by
weight) was delivered on surface by agitator truck (Figure 10) at
the RMD13 hole collar and transferred 220 m vertically below
the surface to the Block 2E/1 drive instead of carting 25 MPa
concrete underground via the decline. The pillar was poured to a
thickness of 1 m, over reinforcing mesh secured to the stope
walls using 2.4 m split sets and 1.8 m long rebars. When set, it
was covered with a 0.5 m layer of mullock prior to mining a
mechanised lift and the subsequent development of drawpoints,
rises for ore passes and manways for accessing the Block 2E/1
stope above 1000 mRL (Figure 11). This cemented sand slurry
was also used on several occasions to backfill areas where falls
of ground had occurred. After the cemented sand was allowed to
set, the area was easily remined to regain access to the stope
block.
243
P L McCARTHY and L FAULKNER

1211mRL

BAY 3
BAY 2
Poverty
2 Level

BLOCK 4
ACCESS 5 Level
'Plat'
BAY 1
NICK O'TIME 115
0m
RL BAY 4
Portal
SHOOT 4 Level

5 Level Prince of Wales

'Plat'
Block 4 7 Level East Cross-cut

WMC
5 Level
6 Level
5 Level Drive 'Plat'
Block 3A
Poverty
6 Level
POVERTY
BAY5 SHAFT Prince of Wales
9 Level East Cross-cut

Prince of Wales
10 Level
897mRL
Block Drill PRINCE OF WALES
Cuddy
3 SHAFT

BAY 6 Drill
Cuddy

BAY 7
Decline BAY 10
to Block 2
Block
2W
High
Grade
portion
of Drill
Cuddy
Shoot

2E/1
987mRL
RL

2E/3
0m
99

Drill
Cuddy

2E/2

REEF MINING N.L.

RL
0m

MIN 4756 TARNAGULLA GOLD PROJECT

85

Medium
to CRYSTAL HILL DECLINE,
Low
Grade PERSPECTIVE VIEW
Portion Depth Limit of SHOWING
of Resource drilling NICK O'TIME SHOOT AND PRE 1900 WORKINGS
shoot
POVERTY REEF, TARNAGULLA
Author : B. Cuffley
Date : August 1999

FIG 9 - Crystal Hill decline and Nick O’Time Shoot perspective view.

FIG 10 - Agi trucks delivering cement sand slurry to the collar of

RMD13. FIG 11 - Stope Block 2E/1 sill drive.

244 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference


Mining beneath the cement sand slurry backfill was very
successful with the ore beneath being extracted and the underside
of the pillar being exposed (Figure 12). This was carried out with
both conventional airleg and a ‘non-entry’ retreat mining methods
using uphole blastholes. Under either mining condition, the
cement sand slurry pillar remained stable.
FIG 12 - Exposed underside of Block 2E/1 cement sand slurry
floor pillar.
The increase in airleg mining required an alternate backfill
strategy to the waste rock previously used in cut-and-fill stoping.
This method of filling was later adapted for use with reclaimed
tailings from an old battery sand dump at the Prince of Wales
shaft and remained in use throughout the remaining mine life.
Within the stope, timber and hessian barricades were used to
contain the sandfill for each lift. Timber manways and steel ore
pass liners were extended upwards as mining lifts progressed
within each stope.
A mobile screening unit was used to screen out the +6 mm
material from the old battery sand, removing debris, including
tree roots and other vegetation, to create a relatively clean
product so as to eliminate blockages in the cased drill hole. The
mobile screen discharged directly into a tipper truck or the sand
was stockpiled for later transfer. At the discharge point the truck
would tilt its tray and water sprays were used for mixing the sand
into slurry (50 - 60 per cent solids by weight). Once the sand and
water were mixed, the slurry was free flowing and ran directly
into a hopper that fed directly into the HQ cased drill hole.
Although simple in design, the system was an effective means of
creating hydraulic sandfill.
A comparison of production from airleg and mechanised
stopes is provided in the October 1997 monthly report, with
1470 tonnes from the Block 2 mechanised stope and 320 tonnes
from the Block 4 airleg stope.
REGULATORY ENVIRONMENT
After many years of anti-mining sentiment in Victoria under the
Cain and Kerner governments, the election of the Kennett Liberal
government in October 1992 gave a boost to economic activity
and encouraged public servants to be more supportive of mining
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
TARNAGULLA GOLD MINE
development. The initial permits from the Loddon Shire and the
Department of Minerals and Energy were based on development
of an exploration decline to test the potential of the reef, rather
than being production permits. The required bonds were
reasonable, being $100 000 to the Department of Minerals and
Energy and $150 000 to the Department of Conservation and
Natural Resources.
After the Reef Mining NL Initial Public Offering closed in
March 1994 drilling started immediately. A planning permit
application for the decline works was lodged on 9 November
1994, six objections were lodged and resolved, and the works
were approved on 31 January 1995.
The Feasibility Study, based only on an Inferred Resource, was
completed in April 1995 and a Work Plan based on the
Feasibility Study was approved on 23 June 1995. The box cut
was excavated and then the first decline cut was fired on
9 August 1995. Thus only nine months passed between the first
planning application and the start of underground work.
PROJECT FUNDING
Funding was always a problem, with barely enough capital
available to complete each phase of the project. The falling gold
price and flat capital market made it difficult to raise funds when
required and as a result, control of the company was eventually
lost.
In many ways it seemed history was repeating itself; during
the boom years of 1869 to 1872 many smaller goldfields like
Tarnagulla declined while vast amounts of capital was attracted
to the Bendigo Goldfield as investors placed their money with
established dividend paying mines rather than in small high-risk
ventures (Wilkinson, 1996).
A proposed fund-raising in mid 1995 failed to proceed when
the incoming partner drilled six test holes in the high-grade zone,
which resulted in the estimated high-grade tonnage being
reduced by one third. Later a friendly banker was able to provide
bank finance of $0.5 M over 12 months for mining equipment,
but other banks were not interested.
Against a Feasibility Study assumption of A$500 per ounce,
the average actual gold price received, over the life of the
project, was A$451 per ounce, with the spot gold price ranging
from a high of A$500 per ounce in October 1998 to a low of
A$382 per ounce in July 1999.
SAFETY PERFORMANCE
Consistent with small mine practice in the mid 1990s a number
of informal but effective safety management strategies were
used, but there was no formal safety management system. The
loose, slabby ground and the installation of ground support using
airleg mining, together with the statistical effect of a small
workforce resulted in a high lost time injury frequency rate
(LTIFR). For example, ten LTIs in 1996-97 gave an LTIFR of
209, almost ten times the national surface and underground
mining industry average. Six were rockfall injuries and the others
were shoulder and back injuries. The injuries were typically of
low duration and this result was not inconsistent with other
underground mines using airleg methods at the time. There was a
steady and consistent improvement in subsequent years and by
1999/2000 five LTIs were reported for a LTIFR of 82. The types
of injuries continued to be mainly rockfall or strain related.
OPERATING COSTS AND ECONOMICS
Typical operating costs in mid 1997 A$ are summarised in
Table 1. The direct cost of development was $1113/m with a
further indirect cost of $503/m allocated on a tonnes moved
basis. The costs in Table 1 are net of the costs transferred to the
capital development account.
245
P L McCARTHY and L FAULKNER

TABLE 1
Operating costs in mid 1997.
Ore mined (month) tonnes 1200
Recovered grade g/t Au 28

Unit costs
Geology ($/t) 10.14
Development ($/t) 60.01
Stoping ($/t) 29.76
Underground supervision ($/t) 13.39
Underground services ($/t) 8.78
Subtotal underground ($/t) 122.08
FIG 13 - Airleg driving in Stope Block 2E/1.
Ore cartage ($/t) 8.00
Treatment ($/t) 63.64 The production history of the Crystal Hill decline detailed in
Environment and rehabilitation ($/t) 2.69 Table 2 shows the increase in airleg compared to mechanised
cut-and-fill stoping over time.
Site administration ($/t) 36.43
Underground development in Block 2E showed a dramatic
Subtotal surface ($/t) 110.76 decrease in grade from high grade to low-medium grade
occurring between Block 2E/1 (990 mRL) and Block 2E/2 (965
($/t) 232.84 mRL). The reconciliation between composite drill core assays,
Total operating cost development face samples and actual plant grades obtained for
($/oz) 258.65 the development of the Block 2E/2 level was good with the plant
grade estimated to be 10 g/t gold. Development of the Block
The cost of airleg stoping was $92 per tonne including some 2E/3 (940 mRL) level did not proceed as the estimated grade
associated vertical development. The cost of mechanised mining from the remaining material was approximately 6 g/t gold. In
of ore and waste averaged $41 per tonne. The total cost of mid 1999, with a gold price of A$386 per ounce, it was not
running the operation at that time was A$419.79 per tonne or considered worthy of further development, which meant the
A$466.32 per ounce, including capital development, sustaining remaining ore production was to come from areas already
capital and surface exploration drilling elsewhere on the leases. developed, so the Nick O’Time Shoot was mined upward to
By this time, the gold price had fallen to A$436 per ounce, so the within 70 m of the surface.
sustainability of the operation was in doubt. Production for the 12 months to June 1999 was 503.2 kg
In April 1998 a review examined alternative operating plans. A (16 177 ounces) from 17 525 tonnes at a reconciled head grade
predeveloped resource of 5200 tonnes at 39 g/t gold remained to of 29.6 g/t gold. Revenue for that year was $7.4M and the cash
be extracted by both airleg and mechanised stoping methods operating cost was A$271 per ounce.
over widths of 1.2 m and 3.5 m respectively. The main resource, In September 1998 Reef Mining NL was the subject of a
which was open at depth, could be mined by mechanised takeover offer by Sydney based Hudson Resources Limited. The
methods recovering 55 000 tonnes at 10 to 20 g/t gold or by offer was rejected but repeated in May 1999, with Hudson
airleg for 23 000 tonnes at 15 to 38 g/t gold. gaining 57 per cent of shares and effective control in September
1999.
The low gold price, the prospect of declining grade with depth
and the absence of any new discovery suggested that the fixed Under its new owners, the mining operation was shut down
costs could not be sustained at the expected production rate of and Reef Mining NL became International Concert Attractions
1800 tonnes per month. The review concluded that the Limited, which survived until 2005.
development and production rates must be increased, which Over the project life, which ended in March 2000, a total of 57
pointed to mechanised stoping. However, a suitable processing 403 tonnes of ore was mined for 53 000 ounces recovered. This
facility was not available to accept the higher mine output, so the was a time when many existing exploration and mining
companies were being re-badged as ‘dot com’ companies.
alternative of a transition to 1.5 m wide hand-held stoping was
decided upon (Figure 13). The overall cash outcome from the Tarnagulla project for the
seven financial years 1994 to 2000 is summarised in Table 3.
Once this transition was made, the remaining mechanised This cash analysis ignores the accounting treatment of capital
mining of ore can be categorised into that produced using the depreciation and amortisation, showing all costs as expended.
single boom jumbo developing sill drives and that produced by While the operation made an operating cash surplus of $8.2 M,
‘uphole’ retreat stopes from mining the pillars that remained this did not compensate for ongoing capital costs of $8.1 M,
above the sill drives. The 51 mm diameter ‘uphole’ blastholes preproduction costs of $8.5 M and ongoing (discretionary)
were drilled in a dice five pattern, with alternate rows of two exploration costs of $2.6 M. Due to the very short strike length of
holes 1.5 m apart and then one centre hole per row, the row the Nick O’Time Shoot, the cost of decline development to
spacing was 1.0 m with holes angled forward at 65°. Hole length access each stoping lift more than consumed the operating
was drilled to suit the thickness of the sill pillar; 3.5 m was the surplus at the average realised gold price of A$439 per ounce.
typical length. Cable bolts, 5 m in length, were installed on the
alternate rows to provide additional support where the stope
firing would result in a hanging wall undercut by the sill drive.
BACK-ANALYSIS
Firings were typically five or six rows at a time and the ore was At the time of writing, the price of gold is approaching
mucked from the stope by remote LHD, prior to charging and A$1000 per ounce and it is tempting to believe that the project
firing the next section. economics today would be robust. However, mine capital and

246 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 TARNAGULLA GOLD MINE

TABLE 2
Crystal Hill decline production history.

Year 1995/96 1996/97 1997/98 1998/99 1999/00 Total

Mechanised development
Capital development (m) 1013 526 583 588 0 2710
Decline depth below surface 120 175 230 270 270
(vertical metres)
Stope access development (m) 0 685 322 222 201 1430
Ore development (m) 0 408 386 199 25 1018
Total mechanised development (m) 1013 1619 1291 1009 226 5158
Vertical development (m) 0 0 27 51 0 78
Stoping
Mechanised cut-and-fill stopes (t) 9490 13 360 5315 170 2 8335
Mechanised uphole stopes (t) 2 045 2045
Hand-held stopes (t) 1 040 4 860 11 730 9 460 27 090
Total mined tonnes (t) 10 530 18 220 17 045 11 675 57 470
Percentage – mechanised C&F tonnes (%) 90 73 31 1 49
Percentage – mechanised uphole tonnes (%) 0 0 0 18 4
Percentage – hand-held tonnes (%) 10 27 69 81 47
Processing
Tonnes treated (t) 10 199 16 567 17 525 13 112 57 403
Reconciled head grade (g/t) 28.0 27.6 29.6 31.8 29.2

life must be extended to recover the initial capital (through

TABLE 3 mining a larger shoot) and also the capital development cost per
Summary of cash result 1994 - 1999. tonne must be reduced.
$ $/t $/oz Optimisation of the larger (60 m) shoot would entail:
Cash from gold sales 23 289 093 406 439 • increasing the production rate to around 40 000 t/a;
Other revenue 316 286 6 6 • more use of airleg mining equipment from the outset; and
Operating costs during 15 379 885 268 290 • careful consideration of shaft access, possibly with rail
production period equipment, instead of a diesel mechanised decline.
Operating surplus 8 225 494 143 155 The latter point may be surprising, but the geometry of the
Nick O’Time Shoot is steeply plunging, with ore extending over
Property, plant and equipment 1 058 315 18 20 190 m vertically, but only 30 m laterally, or along strike. Vertical
costs during production development adjacent to the shoot would be more efficient than a
decline, and would not be a constraint on the very low production
Capital development cost during 7 015 120 122 132 rates.
production period
Capital costs 8 073 435 141 152
CONCLUSIONS
Net ‘steady state’ cash flow 152 059 3 3
The Nick O’Time Shoot was broken up into blocks by a number
of different post ore faults and the mining of these complex fault
Preproduction exploration and 8 549 162 149 161 blocks was geotechnically difficult. The change to airleg mining
development costs in the top part of the shoot not only meant smaller openings that
Exploration costs during 2 601 933 45 49 were easier to ground control, but the complex geology through
production these faulted areas was more easily followed, allowing the Nick
Total cash outcome -10 999 036 -192 -208 O’Time Shoot to be developed to its extremities. Using airleg
mining meant that the high grades, commonly assaying in the
100s g/t gold (with the highest being 3050 g/t gold over 0.25 m)
operating costs have risen substantially over the past decade and that were concentrated at the narrow ‘tail ends’ of the shoot could
particularly over the last four years. Typical unit cost increases
be exploited.
since 1997 have been labour at 45 per cent, explosives at 30 per
cent and equipment leases at more than 50 per cent. If the results The Nick O’Time Shoot history shows that underground
in Table 3 are recalculated with a gold price of A$1000 per development progressively increased the resource as mining and
ounce and all costs increased by 50 per cent, the cash outcome is exploratory development and drilling occurred concurrently with
a surplus of A$1.5 M after all costs, including the discretionary the final tonnage mined (57 000 tonnes), being nearly three times
exploration (elsewhere on the lease). the high grade resource of 20 000 tonnes that was estimated from
However, if those costs are applied to a resource twice the size surface drilling in mid 1995.
(114 000 tonnes at the same grade, by doubling the strike length Mining of the Nick O’Time Shoot enabled the upgrading of
from 30 m to 60 m) then the cash surplus becomes $47 M, an the original geological model with new information as every
attractive outcome for a small company. It is clear that the mine block mined provided an increased understanding of the

Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008 247

P L McCARTHY and L FAULKNER
underling characteristics of the Poverty Reef. This improved
understanding is the crucial exploration tool in discovering why
high-grade shoots form, where they form and therefore how to
explore for the future Nick O’Time and Bonanza Shoots that still
remain to be found and mined profitably in Central Victoria.
ACKNOWLEDGEMENTS
The authors wish to thank the shareholders and original Directors
of Reef Mining NL, especially Dick Sandner, for being ‘true
believers’ in the potential of Central Victorian Goldfields and in
the philosophy that nuggetty gold deposits can only be fully
evaluated by underground development following the old adage
‘drill for structure, drive for grade’; Brian Cuffley for being
such a good geologist; Peter Wiseman at Media Australia for
248 Ballarat, Vic, 14 - 15 October 2008
supplying photographs and all the employees and contractors
who worked on the Tarnagulla Gold Project for their contribution
over the life of the project.
REFERENCES
Krokowski, J, Vickerod, E, Cuffley, B and Evans, T, 2001. Tarnagulla
Goldfield, Central Victoria 1:10 000 map area geological report,
Victorian Initiative for Minerals and Petroleum report 71.
Reef Mining NL, 1999. MIN 4756: Annual Technical reports to Minerals
and Petroleum Victorian Department of Natural Resources and
Environment 1995 to 1999 (Department of Natural resources and
Environment: Melbourne).
Wilkinson, H E, 1996. Research into the history of quartz reefs within
Reef Mining’s tenements at Tarnagulla.
Narrow Vein Mining Conference
Reviving the Inglewood Goldfield in Central Victoria –
Exploration and Operations
J Cahill1
ABSTRACT
The Inglewood Gold Project consists of several titles in and around the
Inglewood and Wedderburn Goldfields which were purchased from Strata
Mining Corporation (now Excalibur Mining Corporation). The Inglewood
goldfield has a historical production of approximately 160 000 tonnes at
23 grams per tonne. The company has reopened and developed the
Maxwells reef, picking up a high-grade shoot (Max Shoot) identified and
worked by Goldquest et al up until 1994. Production ceased on a low
gold price, rapidly reducing vein size and a requirement for further
development. Greater Bendigo Gold Mines Ltd (GBM) has undertaken
the development and encountered the mineralised vein in an increasing
size setting with high grades and the existing mine development has
access to a further shoot to the north (Max North Shoot). The Maxwells
reef also contains the Woolf shoot south of the Maxwells shaft and an
inferred resource of ~20 000 ounces is contained from 60 m below
surface. Best intersection 2.9 m at 50 g/t. The Maxwells reef area
produced approximately 20 per cent of the Inglewood goldfields
production and GBM is preparing to increase its exploration efforts
across the rest of the field. A processing plant has been constructed at the
Maxwells mine site, which is capable of treating tailings, hard rock and
alluvial material utilising gravity recovery. The existing Maxwells
tailings have been excavated and treatment commenced. It is intended to
produce a sand product from the tailings for resale and to reduce the
storage requirements per tonne. Several large water dams constructed by
Goldquest in the 1990s are being re-engineered to enable storage of
significant tonnages of tailings. A level of surface and underground
production is planned for 2008 and this is to be expanded into 2009 and
beyond. The long-term future at Inglewood is envisaged as structural
repeats at depth below the current workings and a project plan to gain
deep access will be prepared on successful drilling results.
INTRODUCTION
Greater Bendigo Gold Mines Ltd is a listed company on the ASX
(ASX code GBM) and floated in March 2007. The company has
several project areas and numerous titles in and around central
Victoria as well as a number of joint ventures.
The Inglewood Gold Project consists of several titles in and
around the Inglewood and Wedderburn Goldfields which were
purchased from Strata Mining Corporation (now Excalibur
Mining Corporation) in 2006.
The Inglewood goldfield has a historical production of 166 194
tonnes at 23.4 grams per tonne (Table 1). Further research is
refining this figure.
1. Managing Director, Greater Bendigo Gold Mines Pty Ltd, PO Box
1345, Bendigo Vic 3550. Email: johncahill@gbgm.com.au
TABLE 1
Inglewood goldfield production.
Mine tonnes oz
Maxwells 35 546 37 086.0
Jersey 16 150 21 055.0
Columbian 38 500 39 743.0
March Unity/ Morning Star 38 563.0 20 510.0
Other 37 435.3 30 207.1
Total 1 66 194.3 1 48 601.1
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
Gold was primarily discovered in the gullies around the
goldfield before outcropping quartz was noted to be rich in the
yellow metal.
Initial work on the field exploited shallow open cuttings to
three or five metres in depth on very high grades before shallow
shafts exploited rich shoots as high as ten ounces to the tonne
down to the water table at ~50 m where the majority of the work
on the field ceased.
Due to its location, facilities for treating the ore on the
Inglewood field in the 1800s were rare and expensive and early
work on the field had a recorded cut off grade of ~1 oz per tonne.
A systematic approach to the field during the early period
resulted in a mechanism which allows a reasonably accurate
location of the many shafts on the field today. On any new area,
the first claim registered was called the prospecting claim and all
claims north and south were named no 1 north, no 2 north, no 1
south no 2 south, etc. While there are a large number of shafts
and pits on the field, the four main producing shafts noted above
account for more than 80 per cent of the production. These
workings were significant in that production was obtained below
the water table.
These companies sank shafts to depths of 200 - 300 m depth
and were from east to west the Maxwells company, the
Columbian company, March Unity – Morning star companies
and the Jersey company.
Impressive grades were the norm for these operations but all
had ceased production by the Second World War.
The field was revived in the 1980s by Goldquest NL who
undertook an aggressive exploration program followed by the
refurbishment of the Maxwells shaft and rehabilitation of the
levels to allow a mapping and sampling program to be
undertaken.
The geological team convinced management to drive the
137 m level north as all work at that stage was directed toward
the Laidlaw’s shaft, 400 m south. The company commenced a
cross-cutting program from the north drive which located the
Max shoot mineralisation. The 156 m level was driven out to the
north as well and stoping was commenced.
A further high-grade shoot was encountered and the 137 mL
drive pushed on to the north where stoping of the Max North
shoot was undertaken.
A gravity plant was erected which was followed by a CIP
batch plant.
Some spectacular grades were encountered during stoping of
both areas, which never translated into profit from the mill due to
high costs and a low gold price.
kg Grade g/t Depth of stoping
1156.3 32.5 156
656.5 40.6 125
1239.2 32.2 200
639.5 16.6 152
199.8 5.3
3891.31 23.4
249
J CAHILL

Recovery was not good, with grades as high as 10 g/t being Additionally the rise was extended for three further cuts and
recorded in the tailings before a systematic professional three 100 kg bulk samples taken from the material and sent to
approach to the mill operation increased recovery to an Gekko Systems Pty Ltd for analysis. Table 3 details the results
acceptable level. from this work. The results were impressive with the best result
returning 77 g/t, with the majority of gold reporting to the
Goldquest collapsed in the late 1980s and the mine was run by
+50 micron size range, ideal for gravity recovery.
a series of companies until 1994 when production was halted
after the gold price dropped significantly along with reducing
grades and ventilation issues forced the mine to close. TABLE 3
Just before closure, the then mine manager, the late Mr Bob Results of bulk sample analysis.
Farrell put up a rise from the back of the last stope, which Sample Average head grade Grade from sizing analysis
revealed an increasing reef with abundant visible gold. This rise
when processed as a bulk sample returned in excess of one ounce LMAX-A 37.9 g/t 73.3 g/t
per tonne. LMAX-B 88.3 g/t 58.9 g/t
Strata Mining Corporation took control in the late 1990s and LMAX-C 39.35 g/t 36.8 g/t
undertook various drilling programs, soil sampling and
background studies before expenditure issues allowed GBM to The gold in the quartz and the clay or pug seams is visually
gain control of the licences in 2005 - 2006. spectacular taking on the appearance of sprayed gold paint at
A significant RAB program was completed from the Maxwells times.
to the Laidlaw’s area on 25 m lines. This data has been To enable access to the orepass from the existing stope below
invaluable and demonstrated the issues that existed with drilling this high-grade zone, a portion of open stope required sand
with visible gold encountered in samples returning insignificant filling. During this operation a concentrate of sorts was collected
assays. which revealed a significant quantity of free gold up to 5 mm in
size in the tailings sand being washed for sand fill underground.
This indicated that the processing undertaken by Goldquest
GREATER BENDIGO GOLD MINES – appeared to be far less efficient than estimated and that a higher
REOPENING THE PROJECT quantity of gold may exist within the tailings than first thought.
The initial aim of GBM has been to gain access to the Max shoot Central Victorian Gold NL reported a resource of ~110 000+
tonnes of tailings grading ~1 g/t and this was assessed by GBM
where the rise was reported to be to assess the claims of an
as sufficient to initiate advancing the project to a production
‘ounce left in the face’. With the mine on care and maintenance stage.
since 1994, GBM reopened the workings and commenced a
Issues facing GBM were pumping, electricals and equipment
100 m level to access the rise and verify the high-grade gold. The
quality. These have been overcome slowly to be replaced by
rise was located and chip sampling returned impressive grades
hoisting issues shaft maintenance and Worksafe compliance.
(Table 2).
GOLD PROCESSING
TABLE 2 On the basis of the tailings resource, analysis of the Goldquest
Chip sample results. data and the economic conditions at the time, the GBM board
elected to erect a mill and commence processing the tailings
Sample ID Au Au(D) Au(R) Location (Max while developing the mine and the field.
Shoot Rise)
A choice was made to utilise gravity recovery due to the mines
HD-40A 0.63 location within the Inglewood Nature Conservation Reserve and
HD-40B 7.2 a tender was accepted for a custom-made plant to be erected,
HD-40C 95.6 60.5 70.3 which has evolved somewhat from the initial scope to the plant
operating today. It became obvious that despite the rush of the
HD-40D 0.43 10 m above Level 1 gold price, available cash in the market would dry up quickly for
HD-41A 292 238 239 exploration and being a producer would be the best way forward
for GBM to maintain momentum.
HD-41B 9.03
The mill comprises a drive over feed bin, a scrubber followed
HD-41C 93.1 85.6 84.3 by a 19 mm screen with oversize being presented to a Granulator
HD-41D 1.04 8 m above Level 1 style jaw crusher and a hammer mill before being re directed into
HD-43A 13.8
the balance tank below the scrubber’s 19 mm screen.
The contents of this bin are pumped up to a water injected
HD-43B 15 sluice before crossing a trash remover and a 1.2 mm shaking
HD-43C 0.81 screen. The oversize from the screen deck is sent either to the
HD-43D 0.46 6 m above Level 1 oversize conveyor or into an impact crusher, which delivers
ground material back into the balance tank to be represented to
HD-44A 0.12 the screen.
HD-44B 15.9 The screen undersize is sent to a pair of concentrators before
HD-44C 0.77 pumping to tail. The tail passes through a cyclone to deslime a
sand product for both resale and stope fill. The plant operates at
HD-44D 0.38 4 m above Level 1 ~25 t/h on tailings and 5 - 10 t/h on hard rock.
HD-45A 68.4 64.8 63.8 An initial capacity of 20 to 30 000 tonnes of tailings is
HD-45B 6.97 available from the existing tailings dam at the mine site.
HD-45C 4.01 Goldquest constructed two large water storage facilities on the
mining licence and GBM is currently having these assessed for
HD-45D 2.66 2.9 3.1 2 m above Level 1 tailings disposal as part of a new work plan variation. This will
(893.3mRL) provide a significant volume of storage close to the plant.

250 Ballarat, Vic, 14 - 15 October 2008 Narrow Vein Mining Conference

 REVIVING THE INGLEWOOD GOLDFIELD IN CENTRAL VICTORIA – EXPLORATION AND OPERATIONS
GEOLOGY
The Inglewood inlier is predominantly composed of Ordovician
sandstones and shales which are variably folded in parts. Initial
impressions are that the units are of short strike duration and
possibly represent deltaic style deposition zones. At the
Maxwells mine, a ten degree north plunge is apparent on the
Strata. Elsewhere on the field, northerly plunge is suggested but
not confirmed.
Dykes are present throughout the sequence and don’t appear to
have any defined control other than cleavage.
The field is cut by a series of strike parallel but shallow east
dipping faults, which are on average 25 m wide dipping at
approximately 60 degrees. These fault zones manifest themselves
as the ‘lines of reef’ on the field with the hanging wall and
footwall zones worked at surface along the major structures.
Internal faulting between the main walls of the faults has also
produced significant mineralised quartz bodies.
The eastern and western most lines appear to have west
dipping structures influencing them and these may be the
bounding structures of the goldfield. Of significance is the fact
that the majority of folding appears to be associated with the
fault zones and is quite possibly fault drag.
Strata Mining Corporation undertook some initial field
mapping. This work outlined a series of repeating oblique faults
cross-cutting the strike of the field. These structures trend
north-east south-west as well as north-west and south-east and it
is currently thought that these control the high-grade shoots by
the creation of steeply dipping dilation zones which plunge
obliquely down the east dipping fault planes.
At the Maxwells mine, a major east dipping fault system some
20 m wide exists and has been worked over a distance of some
400 m from the Laidlaw’s workings in the south to the Maxwells’
workings currently accessed by GBM. Within the fault system
are a number of quartz bodies associated with various internal
and boundary faults. These quartz bodies are quite extensive
and variably mineralised. The fault system dips overall at
approximately 60 degrees to the east while the quartz bodies
appear to become steeper in dip approaching the surface,
probably influenced by cleavage or internal deviation.
The mineralised areas of the fault system are vastly different in
character from the low or non mineralised areas. The high-grade
envelopes are characterised by extensive sericite alteration and
oxidisation.
RESOURCES
Modelling of the data combined with the knowledge from the
previous operators allowed a small resource to be estimated.
Two distinct shoots with high grades have been delineated
with potential extensions down dip yet to be tested. These are the
Max and Max North shoots. Both these shoots are characterised
by strong oxidation levels, strong sericitic alteration, extensive
fault pugs bounding the quartz which is highly broken and poor
ground conditions. The fault pugs are up to 50 mm thick and
often contain significant fine gold within the clay.
An total Indicated Mineral Resource of 13 000 tonnes at a
global grade of 30 g/t has been estimated for these areas (each
with 6500 t at 30 g/t Au). Further work on the resources is
ongoing with research into the Goldquest database continuing.
A further low-grade stope was recently discovered with grades
from Goldquest sampling returning 5 - 8 g/t. A parcel has been
bulk sampled with poor results and work was discontinued. This
ground is vastly different to the Max shoot with low sericite, low
oxidation and massive quartz, etc.
Narrow Vein Mining Conference Ballarat, Vic, 14 - 15 October 2008
TARGETS
A number of exploration targets exist on the field and GBM is
currently ranking them and commencing work plan variations for
them.
Goldquest undertook a systematic diamond drilling program
early in its tenure to the south of the Laidlaw’s shaft. It located
an unworked quartz zone on successive 25 m intersection in
excess of an ounce per tonne, this zone is ~60 m below surface
and the best intersection was 2.9 m at 54 g/t.
This was designated the Wolff shoot after Gus Wolff who
worked for Goldquest during this period and this is a high
priority target for GBM.
It is intended to model this deposit and cost access options.
There is potential for an open cut to be tested for up dip
extension of the Wolff mineralisation. A basic resource of some
20 000 ounces is implied from this work, and GBM intends to
define an Indicated Mineral Resource after a suitable study.
The Daly Kentish reef is on the eastern margin of the known
goldfield and is a steep west dipping fault zone, which may
represent the eastern boundary of the field with the remaining
east dipping faults interacting with it. A series of north-east
trending oblique faults cut the structure in several places. The
reef was worked over 800 m in length with the maximum depth
reached being 50 m at the water table. At this depth the reef was
up to 10 m in width with historical reports giving grades from 3
to 5 g/t.
Two higher grade shoots were worked at the Garibaldi and the
Prospecting shafts and GBM intends to undertake bulk sampling
at these locations.
On the Western side of the goldfield is the Jersey Line with
four significant areas of production, the Jersey, Old Inglewood,
South Gordon and Leicester workings. The Jersey was the most
significant with approximately 20 000 ounces produced from
125 m depth. The fault structure is quite significant and extensive
with workings over 1.5 km long.
The workings exploited both shallow south dipping fault
intersections (South Gordon mine) as well as steeper north
plunging shoots along the Jersey, Kelverton and Ophir reefs.
South West of the township is the March Unity – Morning Star
workings, which show a different morphology to the northern
workings. While the gross characteristics are similar with a
strong east dipping fault zone, the majority of gold was
recovered from large flat tension veins which have opened up
between the bounding faults.
CONCLUSION
GBM has attacked the Inglewood field utilising proven
technology and hand-held mining methods. It has accessed the
Max shoot and erected a processing plant and is working toward
profitable production.
The Company intends to undertake low-cost exploration
utilising close spaced RAB and RC drilling to assess open cut
targets and bring these online for processing.
A diamond drilling program will be planned to target down dip
extensions of the richer portions of the known production areas
below the existing workings.
As the majority of the workings are less than 100 m, a
considerable portion of the goldfield is still available for
exploration.
The company has accessed a considerable funding package via
a convertible note and this will allow a systematic and careful
approach to the development of the goldfield over the next three
to six years.
251