Structural Controls On Orogenic Au Mineralization During Transpression: Lupa Goldfield, Southwestern Tanzania

©2013 Society of Economic Geologists, Inc.
Economic Geology, v. 108, pp. 1615–1640
Structural Controls on Orogenic Au Mineralization During Transpression:

Lupa Goldfield, Southwestern Tanzania
Christopher Lawley,1,*,† Jonathan Imber,1 and David Selby1
1 Department of Earth Sciences, Durham University, Science Labs, Durham DH1 3LE, United Kingdom
Abstract
Au mineralization in the western Lupa goldfield, southwestern Tanzania was associated with transpression
and reverse sinistral slip along a network of steeply S dipping shear zones with non-Andersonian geometries.
Slip was accommodated by: (1) frictional failure and sliding during emplacement of quartz ± Au-bearing veins;
and (2) crystal plasticity and fluid-assisted diffusive mass transfer. The Kenge mineral system is situated along
a NW-SE-trending shear zone and is characterized by ≤10-m thick, Au-bearing fault-fill veins hosted by well-
developed phyllosilicate-rich mylonites. The broadly contemporaneous Porcupine mineral system is situated
along an ENE-WSW− to E-W−trending shear zone, which is characterized by narrow, discontinuous mylonitic
shear zone within a silicified and nonfoliated granitoid protolith. Au mineralization at Porcupine occurs within
steeply dipping fault-fill and subhorizontal extension/oblique-extension veins. Three-dimensional frictional
reactivation theory provides a self-consistent explanation for the different vein styles at Kenge and Porcupine
and extends the classic fault valve model to the general case of oblique slip along multiple, arbitrarily oriented
shear zones. Analysis of the differential stress required for frictional reactivation suggests the following: (1)
the Kenge shear zone was intrinsically weaker than the Porcupine shear zone, consistent with the lack of well-
developed mylonites at Porcupine; and (2) frictional reactivation of the Kenge shear zone occurred under
suprahydrostatic but sublithostatic pore fluid pressures, whereas frictional reactivation of the Porcupine shear
zone occurred under near-lithostatic fluid pressures. We hypothesize that near-lithostatic pore fluid pressures
relieved effective normal stresses at grain-grain contacts, helping to preserve intragranular and fracture porosity
at the Porcupine orebody. As such, these pore spaces may be important microstructural sites for Au mineral-
ization. Low effective normal stresses can also explain the poorly developed phyllosilicate-rich mylonites and
limited degree of shear zone weakening at Porcupine.
Introduction in three-dimensional (e.g., transpressional) strain fields

Orogenic Au deposits are temporally associated with peri- (Krantz, 1988). Many preexisting structures will be unfavor-
ods of Earth’s history dominated by convergent tectonics, ably oriented or severely misoriented for slip during reacti-
and spatially associated with plate boundaries, terrane-scale vation, giving rise to fault valve behavior driven by abrupt
midcrustal shear zones and related subsidiary structures (e.g., temporal variations in fluid pressure (Sibson, 1990). Fault
Groves et al., 1998; Goldfarb et al., 2001). At a lithospheric
scale, the first-order control on orogenic Au deposits is related a) non-partitioned transpression
to subduction zone processes and potentially a fertile upper
mantle/lower crust (Kerrich and Wyman, 1990; Bierlein et
al., 2009; Hronsky et al., 2012). Transpressional deforma-
tion (strike-slip deformation with a component of shorten-
ing orthogonal to the deformation zone; Fig. 1; Dewey et z
al., 1998) is expected at convergent plate margins as a con- Y
sequence of oblique plate convergence and/or irregular plate X
boundaries (Harland, 1971).
Reactivation of preexisting faults, shear zones, and other
mechanical anisotropies (e.g., the margins of igneous intru- b) partitioned transpression
sions) is also to be expected (Holdsworth et al., 1997, 2001),
particularly at subduction zones where the overriding plate
contains previously deformed cratonic crust and/or a “mosaic”
of accreted arc terranes. There is no a priori reason to assume
that these reactivated structures should display Anderso- z
nian geometries, whereby the maximum principal compres- Y
sive stress (σ1) at the time of mineralization bisects the acute X
angle (typically 50°–60°) between conjugate normal, reverse

or strike-slip faults (Anderson, 1905, 1951). Moreover, new
faults with non-Andersonian geometries are likely to develop pure shear simple shear
Fig. 1. (a). Schematic block model of homogeneous or nonpartitioned
*Present address: Natural Resources Canada, 601 Booth Street, Ottawa, transpression. (b). Schematic block mode of partitioned transpression where
Ontario K1A 0E8, Canada. the wrench component of the transpressional strain is compartmentalized
† Corresponding author: email, clawley@NRCan.gc.ca into a discrete strike-slip fault (modified from Dewey et al., 1998).
Submitted: July 30, 2012

0361-0128/13/4155/1615-26 1615 Accepted: February 27, 2013
Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/108/7/1615/3469630/1615-1640.pdf

by Renanda Sevirajati
1616 LAWLEY ET AL.
valve behavior is integral to Au mineralization at midcrustal transpressional nature of the deformation; (2) the nature and
depths close to the frictional-viscous (“brittle-ductile”) transi- distribution of Au-bearing veins along the two trends; and (3)
tion (Sibson et al., 1988). the likely rheological evolution of the host faults/shear zones.
Fluid overpressure reduces the effective normal stress We use these qualitative constraints to investigate the like-
acting on a fault plane. In addition to this direct mechani- lihood of reactivation (frictional slip) and fault valve behav-
cal effect, fluid ingress under greenschist facies conditions ior along the two Au-bearing trends, adopting the modified
(typical of the frictional-viscous transition) causes widespread form of the Coulomb-Navier criterion derived by Yin and
metamorphism whereby strong, load-bearing minerals (e.g. Ranalli (1992). This approach allows us to evaluate the influ-
feldspar, hornblende; Handy, 1990) are replaced by phyllosili- ence of fault/shear zone orientation in three dimensions (i.e.,
cates (e.g. sericite, chlorite; Wintsch et al., 1995). Syntectonic accounting for variations in both strike and dip), in addition
alteration of this type is also associated with the operation to the effects of fluid pressure cycling and fault/shear zone
of fluid-assisted diffusive mass transfer deformation mecha- rheology (expressed by μ) on the style of Au mineralization.
nisms (Mares and Kronenberg, 1993; Bos and Spiers, 2000, As such, our approach extends the classic two-dimensional
2002; Bos et al., 2000a, b; Niemeijer and Spiers, 2005; Jeffer- fault valve model (Sibson et al., 1988) to the more general
ies et al., 2006b). Both processes cause weakening of faults case of oblique slip along multiple, arbitrarily oriented shear
and shear zones (Imber et al., 1997, 2001, 2008), such that zones, and explicitly considers variations in fault/shear zone
they undergo slip with apparent friction coefficients (μ) sig- rheology.
nificantly below 0.6 (Jefferies et al., 2006b; Holdsworth et al., We conclude that variations in fluid overpressure and fault/
2011, cf. Byerlee, 1978). shear zone rheology (controlled by variable degrees of syntec-
These observations suggest that the main structural con- tonic retrogression) can explain the differences in Au miner-
trols on orogenic Au mineralization in midcrustal transpres- alization along the different ore trends within this part of the
sional deformation zones are likely to be fluid pressure cycling Lupa goldfield. If correct, our results also suggest that spa-
and variable degrees of weakening (μ <0.6) along optimally tial, in addition to temporal, variations in fluid overpressure
oriented to severely misoriented non-Andersonian faults and are to be expected during Au mineralization, which implies a
shear zones. The primary aim of this study is to understand certain degree of structural compartmentalization during oro-
how differing styles of Au mineralization along the two dom- genic fluid flow and metallogenesis. Several key uncertainties
inant structural trends within the western part of the Lupa remain to be addressed, particularly in quantifying the degree
goldfield, southwestern Tanzania (Fig. 2), were influenced of fault weakening, but our approach nevertheless provides
by interaction/feedback between these factors. A second- a testable hypothesis that could assist exploration in other
ary aim is to provide a comprehensive structural description orogenic Au deposits hosted by non-Andersonian fault/shear
of the western Lupa goldfield. We combine observations of zone systems.
structures in core and outcrop with geochemical and micro-
structural analyses of the variably deformed host and ore- Methods
bearing rocks to draw qualitative conclusions about: (1) the
Oriented core
Oriented diamond drill holes from the Kenge SE, Mbenge,
30˚
6˚ and Porcupine mineral systems (Fig. 3) were surveyed using a
Reflex™ digital downhole survey tool, and diamond drill core
Ub
Tanzania
intervals were oriented using a combination of Ezy-Mark™
en
W
Ka le
(of 2iC Australia™) and Reflex Act™ (of Reflex™) tools.

de
ak
Ta
tu
o
n
Ezy-Mark™ and Reflex Act™ orientation tools are highly

m
ga
a
ny
accurate; however, the majority of orientation error can occur

ika
INDEX MAP
during the aligning and “locking” of oriented drill core sec-
Ru tions down- and updrill hole (e.g., Blenkinsop and Doyle,
kw Tanzanian Craton 2010). These errors were minimized by selecting intervals of
a
core where the orientation line at the beginning and end of
Uf
ip
the drill hole interval were within 10° of each other and avoid-
a
ing intervals of bad ground. Planar and linear measurements

Lupa
Ubendian Terranes Fig. 3 were obtained following the method of Holcombe (2008).
Up For planar measurements this involves measuring two angles
Katuma M an
bo (alpha and beta angle) relative to the core axis and the orien-
Wakole 100 km zi gw
Ubende a tation line (the core bottom). The alpha angle represents the
Ufipa Terrane boundary angle between the core axis and the long axis of the elliptical
Ma
Mbozi Ny cross section of the planar feature; the beta angle represents
law
sedimentary rocks
Lupa ika the angle between the long axis of the elliptical cross sec-
i
Meso- and Neoproterozoic

Upangwa granite tion of the planar feature extended downhole and measured
Nyika Tanzanian Craton clockwise from the orientation line. Alpha and beta measure-
Fig. 2. Regional geologic map showing the Ubendian terranes and Tan- ments were converted to strike/dip (right-hand rule; RHR)
zanian craton. The star shows the location of the study area (modified from and plunge/trend for planar and linear measurements, respec-
Daly, 1988). tively, using GeoCalculator©.

OROGENIC Au MINERALIZATION: LUPA GOLDFIELD, SW TANZANIA 1617
Lithogeochemistry
Two samples of Saza granodiorite, and two samples of
mylonite derived from the Saza Granodiorite, were analyzed
for major and trace elements using a combination of induc-
5 km
tively coupled plasma-mass spectrometry (ICP-MS) and
sanal workings visited during field seasons in 2010 and 2011 are also shown (open circles) and do not include alluvial workings, which are present in the majority of rivers
in the field area. Eastings and northings are reported as UTM coordinates (WGS84, Zone 36S). Approximate lithologic contacts are based on a series of river traverses by
the first author. Inferred lithologic contacts are based on unpublished aeromagnetic and radiometric surveys acquired from Helio Resource Corp. Shear zone locations are
based, in part, on mapping and correspond to negative magnetic anomalies, whereas dikes are buried and interpreted from linear magnetic highs. The rose diagram shows
Fig. 3. Simplified local geology map of the Helio Resource Corp. property showing the location of Kenge, Konokono, Porcupine, and Cheche Au occurrences. Arti-
the trend of magnetic lineaments picked at a 1:50,000 scale from the unpublished first vertical derivative aeromagnetic map. Contour lines are based on an unpublished
526000
instrumental neutron activation analysis (INAA) by Actlabs

4
1475 m
(Fig. 3; Table A1; Ancaster, Ontario; method 4E-Research).
sector = 5˚
3
Sample aliquants for ICP-MS analysis were first mixed with

524000
a lithium metaborate-tetraborate flux and fused in order to

n = 685
2
ensure complete digestion of refractory minerals (e.g., zir-

1
con). Detection limits for this assay package are in the low
522000
ppm and ppb range for most trace elements. These analyses
Geologic Contacts 0
were completed as part of a larger geochemical study that

approximate
5 - Konokono
8 - Porcupine included standards and duplicates as a means of quality con-

lithogeochemistry sample
7 - Dubwana
inferred
520000
6 - Tumbili
trol. The results for the two samples of unaltered Saza grano-
artisanal working
diorite were previously reported in Lawley et al. (2013a).

518000
3 - Kenge SE
4 - Mbenge
Regional Geology, Southwestern Tanzania

1 - Cheche
Magnetic Lineaments
2 - Kenge
Shear Zone
Ubendian geology
Dike
516000
The Paleoproterozoic Ubendian belt circumscribes the

western edge of the Tanzanian craton and is divided into eight
(granodiorite-diorite-gabbro)
lithotectonic Archean terranes: Ubende, Wakole, Katuma,

514000
porphyritic monzogranite
8
foliated Archean granite
Ufipa, Mbozi, Lupa, Upangwa, and Nyika (McConnell, 1950;

Ilunga Syenogranite
superficial deposits
Sutton et al., 1954; Daly, 1988; Lenoir et al., 1994; Theunissen

Saza Granodiorite
undifferentiated
et al., 1996; Boniface et al., 2009; Fig. 2). Greenschist to gran-

512000
quartz diorite
syenogranite
ulite facies igneous and sedimentary rocks are the dominant

ultramafic
gabbro
lithologies and have been intruded by voluminous Paleopro-

7
510000
terozoic granitoids (Cahen et al., 1984; Lenoir et al., 1994).

1729 m
Each Ubendian terrane is bound by steeply dipping shear

zones with subhorizontal mineral stretching lineations, which
digital elevation model by Helio Resource Corp. and are shown at 20-m intervals.
508000
ne
led to the current tectonic model invoking lateral accretion of

r zo
hea
Ubendian terranes along the western margin of the Tanzanian

za s
craton (Daly, 1988; Lenoir et al., 1994). Lateral accretion in the

Sa
506000
Ubendian belt is suspected to have occurred concomitant with

5
thrusting and accretion of the geologically equivalent Usagaran

belt, which lies to the east of the current field area (Daly, 1988).
504000
Ubendian tectonism and metamorphism occurred dia-

chronously with 2.1 to 2.0 Ga granulite facies metamorphism
Fig. 9
and is overprinted by 1.9 to 1.8 Ga amphibolite facies meta-

502000
morphism (Lenoir et al., 1994). Paleoproterozoic metamor-

3
CL1102
phism in the Ubendian belt is overprinted by Meso- and

1490 m
Neoproterozoic tectonothermal events that are correlated

1
CL0975
to the Kibaran, Irumide, and Pan-African orogenic episodes,

9080000 9078000
CL1030
respectively (Cahen et al., 1984; Boniface, 2009; Boniface

CL1101
and Schenk, 2012; Boniface et al., 2012). The current study

is located adjacent to the Rukwa rift, which is interpreted as a
strike- to oblique-slip half-graben associated with the western
branch of the East Africa rift system (Kilembe and Rosendahl,
1992). The trend of the western branch of the East African rift
system parallels the trend of the Ubendian basement and sug-
gests that reactivation of Ubendian structures has occurred
from the Paleoproterozoic until the present day (Kilembe and
Rosendahl, 1992; Theunissen et al., 1996).
Local Geology, Lupa Terrane

9076000 9074000 9072000
Lupa terrane rocks

The study area is located in the western portion of the Lupa
goldfield and corresponds to the mineral exploration licenses

1618 LAWLEY ET AL.
currently controlled by Helio Resource Corp. (Fig. 3). All rocks Greenschist facies metamorphism is also consistent with the
within the field area have undergone hydrothermal alteration chlorite ± epidote ± calcite metamorphic mineral assemblage
and greenschist facies metamorphism. Thus, all rock names (Fig. 4f) and contrasts with the amphibolites-granulite facies
are metamorphic and for the remaining discussion all igneous rocks that characterizes the other Ubendian terranes (Lenoir
rock names should have the prefix “meta-.” Archean granit- et al., 1994). Nonfoliated equigranular to porphyritic granit-
oids (ca. 2750 Ma) are the earliest magmatic phase observed oids (e.g., monzogranite, syenogranite, and granodiorite) and
in the field area and alternating quartzo-feldspathic and chlo- a dioritic-gabbroic suite of plutons and dikes intruded the
rite bands give the rock its characteristic banded appearance Archean granitoids during the Paleoproterozoic (1960–1880
(Lawley et al., 2013a; Fig. 4). This compositional banding is Ma; Lawley et al., 2013a; Fig. 3). Together, the Archean and
steeply dipping, generally trends east-west to west-north- Paleoproterozoic granitoids represent the main protoliths to
west−east-southeast, and varies in development from outcrop the Au-bearing brittle-ductile (“D2”—see below) shear zones
to outcrop. Flattened quartz crystals, undulose extinction, of the western Lupa goldfield.
and cuspate/lobate quartz crystal grain boundaries are indica-
tive of pressure solution and intracrystalline plastic deforma- Deformation events in the Lupa terrane
tion of quartz, which coupled with syntectonic metamorphic Lawley et al. (2013a) provided U-Pb ages that constrain
reactions (chlorite ± epidote ± calcite ± titanite replacement three, temporally distinct, deformation events (D1, D2, D3)
of primary Fe-Mg minerals), are the dominant deformation recognized in the field. The first deformation event (D1) is
mechanisms (Fig. 4e, f). Crystal plastic deformation micro- restricted to foliated Archean granitoids, which are cut by non-
structures in quartz and the absence of similar structures in foliated Paleoproterozoic granitoids, diorites, and gabbroic
feldspar crystals suggest that deformation occurred at green- rocks that loosely constrain the timing of D1 to between 2.75
schist facies pressure-temperature conditions (Scholz, 1988). and 1.96 Ga (Lawley et al., 2013a). Brittle-ductile mylonititc
a b feldspar rods
Archean granite
banding
Archean granite
5 cm 50 cm
c d pyrite
Archean granite
m
dike Archean ylo
granite ni
te
1 cm 1 cm
e f epidote
plagioclase
quartz
quartz
chlorite
plagioclase
1 cm 1 cm
Fig. 4. (a)–(b). Field photos of Archean granite highlighting the variability in structure and grain size of the lithologic unit.
(c)–(d). Core photos of the Archean granite cut by nonfoliated porphyritic dikes and mylonitic fault rocks (D2) cutting pre-
existing foliation (D1). (e). Crossed nicols photomicrograph of Archean granite showing undulatory extinction and cuspate-
lobate quartz crystal boundaries. (f). Crossed nicols photomicrograph of Archean granite showing alignment of chlorite and
undulatory extinction in quartz. The alignment of quartz, K feldspar, and chlorite gives the rock a banded appearance in the
field.

D2 shear zones cut all of the dated magmatic phases. Re-Os a network of narrow, discontinuous mylonitic shear zones
dating of syndeformational pyrite suggests D2 mylonitzation separated by nonfoliated but hydrothermally altered granite.
occurred at ca. 1.88 Ga (Lawley, et al., 2013b). The D2-age The highest Au grades at Porcupine are associated with shal-
shear zones, which are the main focus of our study, display low dipping quartz veins and intervals of sericitized, silicified,
evidence for contemporaneous brittle fracturing—related and nonfoliated Ilunga syenogranite (Fig. 13). These aurifer-
to emplacement of Au-bearing veins—and macroscopically ous and shallow dipping quartz veins, which occur adjacent
ductile shear accommodated by viscous deformation mecha- to (not within) mylonitic shear zones and significantly widen
nisms (e.g. crystal plasticity and fluid-assisted diffusive mass the mineralized zone at Porcupine, are notably absent at the
transfer). The main Au-bearing shear zones display north- Kenge and Mbenge orebodies. From an exploration view-
west-southeast or east northeast-west southwest to east-west point, it is important to understand why different styles of
trends, referred to here as the “Kenge” and “Saza” trends, mineralization occur along shear zones parallel to the Kenge
respectively (Fig. 3). The pyrite ± Au-bearing quartz vein sys- and Saza trends.
tems have different geometries depending on orientation of
the host shear zone (i.e. Kenge or Saza trend) but in all cases Shear zone kinematics
are locally cut by discrete D3 brittle faults. The timing of D3 D2 shear zones are characterized by a variably developed
is not constrained; however, the brittle nature of the faults is mylonitic foliation, which is considered to approximately
in contrast to the crystal plastic deformation processes charac- record the XY plane of the finite strain ellipsoid (Jones et al.,
teristic of deformation during D1 and D2 and suggests differ- 2004). As a result, subhorizontal and subvertical planes per-
ent geologic conditions during D3. The proposed temporally pendicular to the mylonitic foliation are interpreted as the
distinct deformation events are only those that are readily dis- YZ and XZ planes of the finite strain ellipsoid, respectively.
tinguished in the field and it is expected that structures have The foliation is associated with multiple kinematic indica-
been periodically reactivated from the Paleoproterozoic to tors on the subvertical plane perpendicular to the mylonitic
the present day (Kilembe and Rosendahl, 1992; Theunissen foliation (section view; XZ plane). Kinematic indicators,
et al., 1996; Boniface and Schenk, 2012). albeit less abundant, are also observed on the subhorizontal
plane perpendicular to the foliation (plan view; YZ plane).
Structural Framework During D2 Deformation The mylonitic foliation at the Kenge, Mbenge, and Porcu-
and Au Mineralization pine (Fig. 3) deposits varies in development and orientation
but is typically more steeply dipping and oriented clockwise
Shear zone geometry and spatial distribution relative to the shear zone boundary (Figs. 9, 10). This obliq-
of Au mineralization uity between the mylonitic foliation and shear zone bound-
Shear zone geometry (and kinematics) were interpreted ary implies reverse and sinistral kinematics along shear zones
from inspection of rock exposures within artisanal shafts and parallel to both the Kenge and Saza trends.
trenches, oriented drill core, and serially sliced, oriented hand Asymmetric porphyroclasts (σ type; Hanmer and Passchier,
samples (Figs. 5–10). Observations have been obtained from 1991), asymmetric pressure shadows, asymmetric boudins,
shear zones with “Kenge” (northwest-southeast) and “Saza” and S/C textures observed within the XZ plane of the finite
(east-west to east northeast-west southwest) trends, nearly strain ellipsoid are all indicative of reverse movement (south-
all of which dip steeply to the south. Inspection of the three- side up). In plan view (YZ plane of the finite strain ellipsoid),
dimensional geometry of shear zones parallel to the Kenge asymmetric porphyroclasts, asymmetric boudins, and S/C
and Saza trends shows that these structures do not comprise a textures, where observed, are all indicative of sinistral move-
conjugate (e.g. Andersonian) set, despite their apparently con- ment (Figs. 5, 6). Together the observed kinematic indicators
jugate nature in map view (Fig. 9 cf. Fig. 3). Examination of imply a finite reverse sinistral (i.e., oblique) slip vector on
diamond drill core has established that the Kenge shear zone shear zones parallel to the Kenge and Saza trends, consistent
(Fig. 3) occurs at the igneous contact between the Archean with interpretations based on the obliquity of the foliation and
granitoid (hanging wall) and Paleoproterozoic diorite-gabbro shear zone margins. This kinematic interpretation also sup-
intrusion (footwall). Therefore, some igneous contacts acted ports our earlier inference, based on shear zone geometry,
to localize the later brittle-ductile shear zones (cf. Dubé et al., that D2 deformation was associated with slip along a non-
1989; Lin and Corfu, 2002). Andersonian fault/shear zone network.
Artisanal workings and oriented core reveal a remarkable Mineral stretching lineations are variably oriented at out-
variability of Au-bearing structures in the Lupa goldfield (Figs. crop scale, despite the consistent kinematics (Figs. 9, 10).
11–14). Au occurrences along the Kenge trend (and nearby This is particularly apparent for the regionally significant
E-W−trending Mbenge mineral system) are hosted by steeply Saza shear zone, which is interpreted to have undergone
dipping quartz veins associated with an interconnected net- reverse oblique slip despite possessing generally steep, but
work of 10 to 20 m thick, anastomosing mylonitic shear zones variably plunging, mineral stretching lineations (Fig. 9). Min-
(Figs. 3, 9, 12). Hydrothermal alteration, strain, and mineral- eral stretching lineation orientations in transpressional shear
ization at these Au occurrences are largely restricted to within zones can vary continuously from horizontal to subvertical
host shear zones. Recent drilling by Helio Resource Corp. and thus generally cannot be used to infer movement direc-
has, however, identified a second, geologically distinct, min- tion (Lin and Jiang, 2001). The presence of kinematic indica-
eral system type typified by the Porcupine exploration target, tors on both planes perpendicular to the mylonitic foliation
which lies parallel to the ENE-WSW− to E-W−trending Saza implies noncoaxiality of the strike- and dip-parallel strain
shear zone (Figs. 3, 13). The Porcupine shear zone comprises components and is an observation that is typical of shear

1620 LAWLEY ET AL.
All images are section views

a looking W
b looking E c looking N
quartz vein
ite
fault-fill vein
m
ior
ylo
od
n
ite
an
mylo
Gr
za
nite
Sa
e
lo nit
elongation lineation
my
ein
ill v
lt f
fau
1m 10 cm 5 cm
d looking SE e looking SE f looking E
fault-fill vein C
pl
an
e
Cp
lan
e
S plan
S plane
Razorback mine 5m 5 cm 5 cm
e
g looking SE
h looking WSW
i looking SE
mylonite
fault-fill vein
fau
lt-fil
e
inag
l ve
boud
in
5 cm 1 cm 5 cm
Fig. 5. (a). Photo of E-W−trending shear zone from an artisanal working. The complete transition from hydrothermally
altered Saza granodiorite to mylonitic shear zone is exposed. (b). Photo of quartz veins predating E-W−trending shear zone
at artisanal working. (c). Quartz rods on fault-fill vein and mylonitic shear zone contact. (d) Photo from the Kenge mineral
system looking southeast (standing in what was formerly known as the Razorback mine). The mylonitic foliation possesses
a steeper dip relative to the shear zone boundary (ca. 70°; not shown). (e). Photo from artisanal working at the Kenge SE
zone showing S/C fabrics indicative of south-side up movement. (f). Photo from E-W−trending artisanal working showing
S/C fabrics indicative of south-side up movement. (g). Photo from artisanal working at the Kenge main zone showing S/C
fabrics indicative of south-side up movement. (h). Photo of partially oriented core from Porcupine showing asymmetric por-
phyroclasts and the mylonitic foliation wrapping quartz-feldspar microlithons. (i). Oriented sample from the Kenge SE zone
showing deformed fault-fill vein.

All images are plan view zones exhibiting triclinic, or lower, symmetry (Jiang and Wil-
N
liams, 1998; Jones and Holdsworth, 1998; Lin et al., 1998;
a) Kenge Jones et al., 2004). We therefore conclude D2-age sinistral
transpression was accommodated by oblique slip along a net-
work of non-Andersonian faults and shear zones parallel to
the Kenge and Saza trends.
Fault rock microstructures and paragenesis
D2-age mylonitic fault rocks: Au-bearing mylonitic rocks
possess a range of matrix-porphyroclast modal proportions/
grain sizes and represent the complete spectrum of mylonitic
fault rock types (Snoke and Tullis, 1998; Figs. 5–8). Mylonite
and protomylonite fault rock types are dominant and subor-
1 cm dinate, respectively, with rare ultramylonite observed as nar-
row <10-cm bands within mylonitic shear zones. The main
mylonitic foliation is defined by millimeter-scale (locally
b) Kenge centimeter-scale) bands of alternating phyllosilicate-rich and
N quartzo-feldspathic microlithons. Shear band cleavages (C
and C' types; Berthé et al., 1979) are locally observed and are
fault-fill vein characterized by a finer grain size and increased abundance
mylonite of phyllosilicates compared to the dominant mylonitic folia-
tion (Figs. 5e, 6b, c, 7f, i). Multiple C' planes observed at out-
crop scale suggest either overprinting structural deformation
mylonite produced temporally distinct shear band cleavages; and/or
ineffective flow partitioning during a single, but progressive,
5 cm
deformation event (Jiang and White, 1995). The latter is sup-
c) Saza shear zone ported by a lack of clear crosscutting relationships between
N differently oriented mylonitic shear zones (Fig. 3). The grain
size and phyllosilicate-rich nature of the Au-bearing mylonitic
fault rocks are comparable to fault rocks previously described
as phyllonites (e.g., Jefferies et al., 2006a).
Quartz and plagioclase porphyroclasts are wrapped by phyl-
S plane losilicates, possess oblate shapes (as determined by qualita-
tive shape comparison on serially sliced oriented samples on
C plane the interpreted XZ and YZ finite strain planes; Figs. 5–8),
and imply flattening of what are interpreted to be originally
equant quartz crystal shapes in the Paleoproterozoic granite
5 cm protolith. Flattening strains are consistent with the transpres-
sional nature of the deformation inferred from outcrop and
d) Cheche shear zone core observations (previous section). Foliated Archean gran-
N
itoid protoliths possessed flattened quartz crystals (Fig. 4e)
C plane
prior to mylonitization and thus were not used to infer flatten-
ing strain. Flattened feldspar porphyroclasts typically display
a “speckled” appearance under crossed nicols (e.g., Fig. 8c),
S plane
indicative of incipient alteration to fine-grained aggregates of
sericite.
Linear-shape fabrics are variably developed but are sub-
ordinate to planar-shape fabrics in all observed Au-bearing
mylonitic shear zones (i.e., S > L fabrics). This observation
5 cm is consistent with the dominant flattening strains expected to
develop within a transpressional deformation environment
(Tikoff and Fossen, 1999). Lineations, where present, con-
Fig. 6. (a). Oriented sample from the Kenge mineral system showing
sist of quartz rods on fault-fill veins (discussed further below)
asymmetric porphyroclasts in plan view suggesting sinistral movement. (b). and shear zone contacts, as ridges/grooves on the mylonitic
Oriented sample from the Saza shear zone showing asymmetric porphyro- foliation plane, and as chlorite slickensides on chlorite vein
clasts and S/C fabrics in plan view, suggesting sinistral movement. (c). Ori- surfaces and mylonitic foliation planes (Fig. 5c). The tim-
ented sample from the Kenge shear zone showing asymmetric porphyroclasts ing of lineation development is unclear and could represent
in mylonitic foliation and fault-fill vein. Note slivers of mylonite and complex
fracture patterns within the fault-fill vein. (d). Oriented sample from the reactivation of foliation surfaces during subsequent defor-
Cheche shear zone showing asymmetric porphyroclasts and discontinuous mation and/or may have developed contemporaneously with
S/C fabrics in plan view again consistent with sinistral shear. the planar shape fabrics (Theunissen et al., 1996). The latter

1622 LAWLEY ET AL.
All images are section view

a looking WSW b looking WSW c looking WSW
quartz
mylonite
feldspar
muscovite
veins
veins quartz
1 mm 1 mm 1 mm
d looking W e looking W f looking W
C plane
pyrite
pyrite
S plane
chlorite
pyrite
muscovite
mylonite chlorite
quartz
1 mm 1 mm 5 mm
g looking NW h looking NW i pyrite looking NW
C plane
S plane
pyrite
pyrite
quartz
quartz
1 mm 1 mm 5 mm
Fig. 7. Section view photomicrographs of Porcupine (a)-(c), Mbenge (d)-(f), and Kenge (g)-(i). (a). Crossed nicols photo-
micrograph of sharp and undulating mylonite-granite contact cutting muscovite ± calcite ± chlorite veins. (b). Crossed nicols
photomicrographs of sharp and undulating fault-fill vein-mylonite contact cutting muscovite ± calcite ± chlorite veins. (c).
Crossed nicols photomicrograph of quartz porphyroclasts. (d). Pyrite with quartz strain fringes. (e). Crossed nicols photomi-
crograph of muscovite included within a pyrite crystal that is concordant with mylonitic foliation. (f). Plane-polarized light
photomicrograph of S/C fabrics consistent with reverse movement (south-side up). (g)-(h). Crossed nicols photomicrographs
of pyrite with quartz strain fringes orthogonal to mylonitic foliation. (i). Plane-polarized light photomicrograph of S/C fabrics
consistent with reverse movement (south-side up).

All images are plan view

a quartz
b
quartz
oblique foliation
N N
1 mm 1 mm
c d
quartz quartz
feldspar pseudomorphs
N N brittle fractures
1 mm 1 mm
e f
pyrite chlorite
pyrite
quartz strain fringe
N N
1 mm 500 µm
g h
quartz pressure solution seams
muscovite
quartz brittle fractures
plagioclase
N N
500 µm 1 mm
Fig. 8. Plan view photomicrographs: (a). Crossed nicols photomicrograph of oblique foliation development in plan view.
(b). Crossed nicols photomicrograph showing undulatory extinction, dynamic recrystallization, and subgrain rotation in quartz.
(c). Crossed nicols photomicrograph of feldspar pseudomorphs and strained quartz. (d). Crossed nicols photomicrograph of
quartz ribbons, undulatory extinction, and dynamic recrystallization in quartz. (e). Crossed nicols photomicrograph of pyrite
with quartz strain fringe. (f). Plane-polarized light photomicrograph of asymmetric pyrite porphyroclast. (g). Crossed nicols
photomicrograph of plagioclase porphyroclasts wrapped by mylonitic fabric. Note dynamic recrystallization of quartz and
neocrystallization of feldspar at porphyroclasts margin. (h). Crossed nicols photomicrograph of brittle fractured feldspars and
quartz within the mylonitic matrix. Note dark brown, undulating pressure solution seams concordant with mylonitic foliation.

1624 LAWLEY ET AL.
9076000
9075000
Razorback mine 50 m
Ke
ng Mbenge
e
9074500
one
hear z
Keng
e SE Saza s
9074000
Cheche
500000
501500
502000
502500
500500
100 m
mylonitic foliation artisanal shaft artisanal working / exploration trench

shear zone margin
elongation lineation
Fig. 9. Map of the Kenge and Mbenge shear zones showing the orientation of individual mylonitic foliation measure-
ments, shear zone boundaries, and individual mineral stretching lineation measurements. The locations of artisanal trenches/
shafts are shown and were used to interpret the continuity of the mineralized shear zones (red polygons). The mylonitic
foliation is generally more steeply inclined and oriented clockwise relative to the shear zone boundaries and suggests reverse
(south-side up) and sinistral kinematics. Interpreted kinematics and shear zone orientations are summarized in the schematic
block model.
interpretation is supported by a common mineral assemblage The disparate microstructures observed between quartz,
between linear and planar fabrics. muscovite, and feldspar provide further constraints on the
All observed shear zones possess a similar fault rock mineral P-T conditions of D2 deformation (Tullis and Yund, 1985).
assemblage (muscovite ± quartz ± chlorite ± calcite ± epidote The onset of quartz plasticity, as observed in Au-bearing shear
± pyrite ± gold) that is consistent with metamorphism under zones, is expected to occur at ca. 300°C, whereas feldspar
greenschist facies pressure-temperature conditions (Figs. 7, plasticity is expected to initiate at 450°C (Scholz, 1988). These
8). Muscovite crystals (<0.01–5 mm) locally exhibit sweeping temperatures are consistent with the Au-bearing shear zone
extinction, comprise the mylonite matrix, and occur as fine- mineral assemblage (muscovite ± quartz ± chlorite ± calcite
grained crystals (<10 μm) derived from alteration of feldspar ± epidote ± pyrite ± gold) that is typical of other greenschist
(plagioclase and K feldspar) porphyroclasts. Chlorite over- facies mylonitic shear zones (Goldfarb et al., 2001; Groves
prints muscovite, wraps quartz and feldspar porphyroclasts, et al., 2003; Elmer et al., 2006). Fluid inclusion microther-
and gives a large proportion of mylonitic fault rocks a green mometry from Au-bearing quartz veins implies a comparable
appearance (Fig. 7f, i). Quartz exhibits sweeping extinction temperature range for hydrothermal fluids. For example,
and occurs as mantled porphyroclasts (σ type; Passchier and H2O-CO2-NaCl inclusions from the Kenge deposit, inter-
Simpson, 1986), aggregates of polygonal grains, quartz rib- preted to best reflect the mineralizing fluid, possess a range
bons, pressures shadows and/or fringes, and veins (Fig. 8b). of homogenization temperatures from 259° to 419°C (Shaw,
Feldspars (plagioclase and K feldspar) exhibit bent twin 2009).
planes, are present as mantled porphyroclasts (σ type) associ- Our microstructural observations suggest that macroscropi-
ated with quartz within microlithons, and are typically absent cally ductile shear was accommodated by variable replacement
in high-strain shear zones where they are completely replaced of feldspar by aligned aggregates of phyllosilicate minerals,
by phyllosilicates (Fig. 8d). and by the operation of both fluid-assisted diffusive mass

Kenge transfer (veins and pressure shadows and fringes) and crystal
a) plastic deformation mechanisms (quartz ribbons and aggre-
pole to foliation plane gates). Comparison with similar textures observed in natural
Z
mineral stretching lineation and experimentally deformed phyllosilicate-rich mylonites
(e.g., Mares and Kronenberg, 1993; Imber et al., 1997, 2001,
2008; Bos and Spiers, 2000, 2002; Bos et al., 2000a, b; Nie-
X 95% confidence cone
meijer and Spiers, 2005; Collettini et al., 2009) suggests that
mean vector = 036° / 15° the development of well-foliated, phyllosilicate-rich fault
(K = 44; n = 31) rocks under retrograde, greenschist facies metamorphic con-
ditions may have been associated with significant weakening
of the host shear zones. This inferred weakening has impor-
Y tant implications for Au mineralization along the Kenge and
Saza trends, discussed below.
Nonfoliated fault rocks: Nonfoliated fault rocks in the west-
ern Lupa goldfield represent the complete spectrum of fault
Kenge SE rocks classified by Snoke and Tullis (1998). Unaltered cata-
b) clasites are locally observed becoming increasingly altered
Z toward mylonite contacts and imply an early phase of brittle
deformation that predates hydrothermal alteration related to
mylonitization (D2). Pre-D2 brittle deformation is also sug-
X
020° / 9° (K = 234; n = 6) gested by vein-mylonite crosscutting relationships and Re-Os
sulfide dating, which suggests that quartz veins locally predate
mylonitization by ca. 70 m.y. (Fig. 15b; Lawley et al., 2013b).
Y In places, there is evidence to suggest foliation development
along the margins of cataclasite veins (e.g., Fig. 15e), which
is again consistent with earlier brittle structures being over-
printed and reworked during ductile shearing. Quartz vein
and mylonite clasts observed within cataclasites cutting Au-
bearing shear zones imply that cataclasis also occurred after
the D2 event (Fig. 15d) and is consistent with discrete brittle
Mbenge faults offsetting mineralized veins. Therefore, the tempo-
c) Z ral relationship between brittle and viscous deformation is
complex, locally interpreted as contemporaneous (see also
discussion on Au-bearing veins, below), and is indicative of
progressive deformation at the level of the frictional-viscous
003° / 12° (K = 112; n = 9)
transition zone.
Pseudotachylite fault rocks are also observed cutting
Archean granites (Fig. 15a). These fault rocks possess a black
X Y fine-grained matrix, angular granite clasts, offset quartz veins
and are associated with injection veins. Devitrification struc-
tures and spherulites were not observed in pseudotachylite
veins and mutually crosscutting relationships between pseu-
dotachylite fault rocks and quartz (±calcite) may indicate that
these fault rocks are in fact veins or cataclasites filled with
Porcupine
d) Z
Fig. 10. (a)-(d). Lower hemisphere equal-area stereonets of the Kenge,
Kenge SE, Mbenge, and Porcupine mineral system geometry, respectively.
The orientations of the finite strain ellipsoid axes (X ≥ Y ≥ Z; closed squares)
are based on the finite strain approach (Robert and Poulsen, 2001) where
341° / 12° (K = 327; n = 4) the intersection of the average mylonitic foliation (black solid line) and the
average shear zone boundary (black dotted line) represents the Y axis. The
X axis of the finite strain ellipsoid is assumed to lie 90° away from the Y axis
Y and along the mylonitic foliation, whereas the Z axis is fixed by the location
of the other finite strain axes (see text for further discussion). Structural mea-
X surements are from surface exposures (closed circles = mineral stretching
lineations; open circles = mylonitic foliation poles to planes; mean mylonitic
foliation pole to plane = red diamond; 95% confidence interval of mean = red
ellipse). K values are a measure of the tendency of linear data to fit a girdle
or a cluster (K values approaching zero represent girdles, whereas K values
approaching infinity represent clusters) following the approach of Woodcock
and Naylor (1983).

1626 LAWLEY ET AL.
a b
fault-fill vein
mylonite
mylonite
fault-fill vein
30 cm 1m
c d
quartz vein
molybdenite
fault-fill vein
pyrite
mylonite
1 cm 1 cm
e Ilunga Syenogranite f
pyrite
Ilunga Syenogranite
pyrite
oblique-extension vein
3 cm 1 cm
g h Ilunga Syenogranite
fault-fill vein
fluid inclusion trails

quartz
pyrite oblique-extension vein
1 mm 1 mm
Fig. 11. (a). Thick auriferous quartz vein within mylonitic shear zone at Kenge (section view looking northwest). Thick
auriferous quartz veins such as these are notably absent from the hanging/footwall, which suggests that these veins represent
fault-fill veins (sensu Robert and Poulsen, 2001). (b). Fault-fill vein from artisanal working along the Saza shear zone (section
view looking approximately northeast). (c). Boudinaged and potentially transposed quartz veins hosted by mylonitic shear
zone at Porcupine. (d). Close-up photo of thick fault-fill vein at the Saza shear zone showing stylolite-like sulfide stringer
veins. Note stylolite-like veins are notably absent from veins outside of mylonitic shear zones but are typical of the fault-fill
vein type. (e). Oblique extension veins at Porcupine. (f). Oblique extension vein at Porcupine. (g). Crossed nicols photomi-
crograph of fault-fill quartz vein. Note flattening and subgrain rotation. (h). Oblique extension vein at Porcupine. Note less-
deformed nature of quartz compared to fault-fill veins.

9074650
9074550
9074600
e
on
rz
ea
sh
ce
surfa
12
0 S
ZD
1
57
SZ
D0
46
1050 z
0.
02
0
SZ
D0
20
52
0
>20
re
v
er
se
SZD
fa
ul
153
t?
10
±5˚
0
SZ
D0
20
1000 z
23
0
Kenge Section
view looking NW
weathering profile
damage zone
fault
SZ
D0
inferred shear zone

12
43
0
Au (ppm)
Rock types 10 m
oxidized
bad ground/missing core
mylonite
quartz vein
e
on
altered diorite-gabbro
rz
ne
ea
altered foliated granite

zo
sh
granodiorite
ge
red
SZ
ma
diorite-gabbro
fer
D0
4
da
foliated granite
iin
950 z
2
0
4
Fig. 12. Cross section of Kenge with lithology logs parallel to diamond drill hole traces. Note inferred reverse fault offset-
ting the interpreted mylonitic shear zone.
dark minerals (suspected pseudotachylite veins are locally reactivation of structures from the Paleoproterozoic to the
magnetic and could be filled with fine-grained magnetite). present day (Boniface and Schenk, 2012; Boniface et al.,
Unconsolidated gouge has also been observed at Kenge (Fig. 2012). The overprinting relationship between the mylonitic
3) and is suspected to be related to near-surface faulting in foliation and earlier cataclasites implies that pre- and syn-D2
response to the Cenozoic-Recent East African rift system brittle structures may have allowed fluid ingress during D2
(Kilembe and Rosendahl, 1992; Theunissen et al., 1996). deformation. In particular, we suggest that fluid influx along
The temporal relationships between disparate nonfoli- networks of pre- and syn-D2 brittle structures led to green-
ated fault rock types remain unclear. Nevertheless, the field schist facies hydrous alteration (e.g., replacement of feldspars
evidence presented above and age dating supports periodic by phyllosilicates) and facilitated the onset of fluid-assisted

1628 LAWLEY ET AL.
9076900
9076950
9077000
9077050
9077100
surface
1200 Z
>10
>10
>20
>20
>10
>10
>
>10
1150 Z
20 0
10
GP
D0
>20
1
>20
>10
1100 Z
>10
>20
>10
GPD
10
10
10
02
0
GP
20
D1
0
5
1050 Z
20
GP
Porcupine Section
10
D1
>10
0
view looking WSW

6
ne
>10 weathering profile

ne
zo
>10
zo
damage zone
ge
ge
ma
inferred shear zone

ma
da
quartz veins / meter

da
1000 Z Au (ppm)
Rock types
oxidized
bad ground/missing core
mylonite
quartz vein
GPD
altered granite
8 10
diorite-gabbro
17
0
granite 10 m
950 Z
Fig. 13. Cross section of Porcupine with lithology logs parallel to diamond drill hole traces. Note correlation between
quartz vein intensity (measured as quartz veins/meter) and Au grade. The extent of the damage zone is inferred from the
increase in quartz vein density near the interpreted shear zone boundaries.
diffusive mass transfer mechanisms (cf. Stewart et al., 2000; unaltered Saza granodiorite (previously reported in Lawley
Imber et al., 2001) within D2 shear zones. et al., 2013a) akin to the isocon analysis approach (Grant,
1986). Isocon analysis involves plotting the composition of
Shear zone geochemistry altered samples against unaltered equivalent samples in order
Two samples of mylonite derived from Saza Granodiorite to distinguish mobile and immobile elements during hydro-
were analyzed for major and trace elements (see Table A1; thermal alteration related to mylonitization (Grant, 2005, and
Fig. 16) to compare against the chemical composition of references therein). The approach is a graphical solution to

Kenge SE
hangingwall/footwall
Archean granite foliation lithology contacts
quartz (± pyrite) veins
fault-fill vein
a) b) c) d)
mineral stretching
lineation shea
r zo
ne b
mylonitic foliation oun
dary
diorite-gabbro
N = 165 N = 71 N = 59
hangingwall/footwall
chlorite veins chlorite slickensides
e) calcite (± quartz) veins f) g)
foliated Archean granite
S/C fabric
Ken
ge
asymmetric
porphyroclast
looking north N = 149 N = 168 N = 161
Mbenge
hangingwall/footwall hangingwall/footwall
Archean granite foliation lithology contacts chlorite veins chlorite slickensides
h) i) j) quartz (± pyrite) veins k)calcite (± quartz) veins l) m)
N = 163 N = 193 N = 70 N = 126 N = 90 N = 31
Porcupine
lithology contacts chlorite veins chlorite slickensides
o) p) q)
n)
e
zon
age
dam
on
in eati
ry ng l
o u nda t re tchi N = 123 N = 303 N = 89
b l s
one minera
ar z
she oblique-extension
calcite (± quartz) veins
quartz (± pyrite) veins
r) s)
fault-fill vein e
ranit
asymmetric nog
ga sye
porphyroclast Ilun
mylonitic foliation
oblique-extension vein
ine
cup N = 714 N = 514
looking north Por
Fig. 14. (a). Schematic block model of Kenge summarizing structural fabrics. (b)-(g). Lower hemisphere equal area con-
toured stereonets of planar and linear fabrics measured from oriented core at Kenge SE. (h)-(m). Lower hemisphere equal
area contoured stereonets of planar and linear fabrics measured from oriented core at the Mbenge. (n). Schematic block
model of Porcupine summarizing structural fabrics. (o)-(s). Lower hemisphere equal area contoured stereonets of planar and
linear fabrics measured from oriented core at Porcupine. All contours are plotted following the 1% area method using OSX
Stereonet (v 1.4; contour intervals set at 1).

1630 LAWLEY ET AL.
a plan view b
quartz vein mylonite
N
Archean granite
quartz vein
pyrite
5 cm pseudotachylite 1 cm
c d
mylonite
calcite vein
mylonite pyrite
1 cm 1 cm quartz vein clast mylonite clast

e
calcite vein
Ilunga Syenogranite granite clast
chloritized Ilunga Syenogranite

1 cm
Fig. 15. (a). Plan view photo of suspected pseudotachylite fault rock offsetting quartz veins and cutting Archean foliated
granite. (b). Quartz veins cutting Ilunga syenogranite and becoming transposed subparallel to the mylonitic foliation. (c).
Calcite veins cutting Ilunga Syenogranite and mylonite. Calcite veins are in turn offset by later movement parallel to mylonitic
foliation. (d). Cataclasites cutting mineralized mylonite. (e). Cataclasites cutting and containing angular to subrounded clasts
of Ilunga Syenogranite. Note complex calcite veining.
Gresens’ (1967) zero concentration change equations and Several elements plot significantly below the unit slope, for
predicts that immobile elements will plot as a straight line example MnO and Sr, and imply that these elements were lost
extending from the origin, whereas mobile elements will plot during hydrothermal alteration. The loss of Sr is consistent
off this line. Our results show that many of the elements con- with plagioclase replacement during mylonitization and is
sidered to be immobile (e.g., TiO2, Al2O3, light rare earth ele- supported by slight depletion and enrichment of Na2O (plots
ments, high field strength elements, and transition elements) slightly below the unit slope) and K2O (plots slightly above
plot on or near the unit slope, whereas mobile elements (e.g., the unit slope), respectively (Fig. 16b). Other elements, for
MnO and Sr) plot significantly off the unit slope (Fig. 16b). example Sb and Cr, plot considerably above the others and
The significance of those elements that plot slightly above suggest that these elements were gained during hydrothermal
or below the unit slope is unclear since compositional het- alteration related to mylonitization. The dominant source of
erogeneity between samples, especially for our small sample Cr in the Lupa goldfield is present within gabbroic intrusions/
set, likely causes some immobile elements to plot off the unit dikes and we propose that fluids migrating through shear
slope (Fig. 16b). zones may have scavenged Cr from intermediate mafic rocks

a) 10 a comparable example of phyllonitic shear zones that are geo-

chemically indistinguishable from their respective granitic
wall rocks (Jefferies, 2006). If correct, the isochemical signa-
Rock / CL1030
ture of the auriferous mylonites may record homogenization

of the geochemical signature and may mask any volumetric
1 changes related to mylonitization.
Mineral System and Vein Characteristics

mylonite (CL1101)
mylonite (CL1102)
Saza Granodiorite (CL0975) Kenge mineral system
0.1
Cs Rb Ba Th K U Nb La Ce Pr Nd Zr Sm Ti Gd Tb Dy Y Ho Er Tm Yb Lu The Kenge mineral system contains a measured and indi-
b) 106 cated resource of 370,000 oz (8.7 Mt at 1.33 g/t Au using
Si a 0.5 g/t Au cutoff; Pratt and Simpson, 2012) and includes
Al what was known as the Razorback mine (Fig. 9; Van Straaten,
105 1984; Sango, 1988; Kuehn et al., 1990). The Kenge orebody is
Fe
K
Na
associated with NW-striking (ca. 120°), SW-dipping (ca. 70°),
Ca and left stepping shear zones. Each NW-SE−trending shear
104 Mg
Mylonitized Saza Granodiorite
zone segment is connected to a more WNW trending shear

Ti
Ba
zone segment and gives the ca. 2-km-long orebody en echelon
P geometry (Fig. 9). The mineralization at Kenge is discontinu-
103
ous downdip and Helio Resource Corp. has hypothesized a
reference slope = 1
Ce V Mn NW-SE−trending reverse fault offsetting the main miner-
Rb
102
Cr
Zr
Sr
alized zone (Fig. 12). The dominant NW-SE trend of the
Pr
Sc Ga
Nd Cu
S
Kenge orebody changes toward the southeast and becomes
Zn
Ni more WNW trending (ca. 103°) and this segment is known as
Ce Dy La
10 Y
Co the “Kenge SE” zone (Harrison, 2011). Structural geometries
Nb Th
Cs
Ho Sb Hf
observed at Kenge are summarized in Figure 14a.
Tb Be Sm Au is found as free grains within shear zones and is also asso-
1 U Gd As
ciated with pyrite hosted by the mylonitic fault rocks and what
Tl
Ta Yb we interpret to be syn-D2 fault-fill veins (i.e., elongate lens-like
Tm Er
0.01
Lu Eu structures within mylonitic shear zones, which contain wall-
0.01 1 10 102 103 104 105 106 rock slivers and are characterized by slickensides and slickenfi-
Unaltered Saza Granodiorite bers; Robert and Poulsen, 2001, and references therein). Pyrite
Fig. 16. (a). Trace element plot normalized to unaltered Saza Granodio- (±Au) crystals are locally observed concentrated in low-strain
rite sample CL1030. Note trace element homogeneity between all four sam- sites within the Kenge and Mbenge shear zones (e.g., quartz
ples except for large ion lithophile elements (e.g., Cs, Rb, Ba). (b). Diagram vein boudin necks). In some cases, pyrite crystals are associ-
comparing the concentrations of elements in mylonitized Saza granodiorite
and unaltered Saza granodiorite. Elements concentrations are average values
ated with quartz and/or chlorite strain fringes, whose asymme-
based on two samples of mylonitized and unaltered Saza granodiorite. Major try is consistent with the overall reverse sinistral sense of shear
element oxide analyses were converted to elemental parts per million. Solid (Fig. 8f). Apparently unmodified quartz strain fringe fibers are
line represents the unit slope (slope = 1) and defines constant composition. commonly observed in the same hand samples as transposed
strain fringe fibers (Fig. 7d, g, h). These observations provide
and deposited them within mylonitized Saza granodiorite. evidence for the operation of fluid-assisted diffusive mass
The enrichment of Sb and As is typical of many epigenetic Au transfer processes (i.e., deposition of strain fringes) contem-
deposits, however in this case Sb is not accompanied with any poraneous with sulfidation, greenschist facies metamorphism,
noticeable enrichment of As and suggests these two elements and progressive reverse sinistral deformation. In some cases,
were decoupled during hydrothermal alteration associated apparently unmodified, straight quartz strain fringe fibers on
with mylonitization (Craw, 2002). pyrite are oriented perpendicular to the mylonitic foliation
The geochemical evidence presented as part of this study (i.e., approximately parallel to the Z axis of the finite strain
records the breakdown of plagioclase to phyllosilicates by ellipsoid; Fig. 7g, h). Previous studies show that the long axis
the depletion of Sr and minor depletion and enrichment of of straight strain fringe fibers records the X axis of the instan-
Na2O and K2O, respectively. However, the general trend of taneous strain ellipsoid (e.g., Durney and Ramsay, 1973; Köhn
most elements plotting on or near the unit slope suggests et al., 2003). These quartz fibers may provide constraints on
that mylonitization was relatively isochemical. The latter is at the instantaneous extension direction during emplacement of
odds with the geochemical signature of mylonites reported in the economically important fault-fill veins, discussed below.
other studies, which typically exhibit evidence for significant Structures interpreted to be steeply dipping fault-fill veins
volume change associated with mylonitization. The unusual are present throughout the Kenge and Mbenge shear zones
geochemical signature of mylonites in the Lupa goldfield may and range in thickness from a few cm to 10 m (Fig. 12). These
reflect, at least in part, high fluid rock ratios that effectively veins intersect the mylonitic foliation and the shear zone
homogenized the geochemical composition of the mylonites boundaries at low angles and possess steeply dipping and
and the granitic wall rocks. The Miyamae shear zone provides NW-SE−trending orientations (Figs. 5d, g, 6b). Stylolite

1632 LAWLEY ET AL.
“seams” comprising molybdenite and pyrite are oriented paral- Qualitative Conclusions Based on Outcrop, Core,
lel to fault-fill vein margins. The latter observation is consistent and Microstructural Observations
with the finite subhorizontal shortening direction inferred Several important conclusions can be drawn from our obser-
from the reverse sinistral kinematics and suggest that stress- vations. First, brittle-ductile (D2) shearing associated with
induced solution transfer (i.e. fluid-assisted diffusive mass sulfidation and Au mineralization took place during reverse
transfer) may have been a process operating during sulfidation sinistral slip along a network of non-Andersonian faults/shear
and mineralization (Fig. 11d). Fault-fill veins typically contain zones. Minor structures (e.g., asymmetric porphyroclasts)
slivers of mylonitized and hydrothermally altered wall rock and and lineation orientations are typical of transpressional shear
possess sheared margins that display prominent quartz rods zones exhibiting triclinic, or lower, symmetry, which suggests
with variable plunges (Fig. 5c). The highly strained nature of that transpressional strain was not completely partitioned into
these inferred fault-fill veins makes the structural significance pure- and simple shear-dominated “domains,” at least at the
of their current orientation relative to the original vein geom- ore deposit scale (cf. Holdsworth et al., 2002).
etry unclear. In particular, narrow quartz veins (less than 10 cm Second, microstructural observations consistent with syntec-
thick and which do not contain slivers of mylonitized wall rock) tonic replacement of feldspar by phyllosilicates and the opera-
with orientations at high angles to shear zone boundaries are tion of fluid-assisted diffusive mass transfer mechanisms suggest
locally observed becoming transposed subparallel to the that significant weakening of the fault/shear zone network may
mylonitic fabric within the shear zone (Figs. 5b, 15b). Although have occurred during syn-D2 greenschist facies metamorphism
vein transposition is likely to be an important feature of pro- and Au mineralization (Fig. 17). The wide, originally continu-
gressive brittle-ductile deformation, the absence of up to ous mylonitic shear zones at the Kenge and Mbenge orebodies
10-m-thick quartz veins outside the mylonitic shear zones sug- are analogous to the crustal-scale “interconnected weak lay-
gests that the main auriferous quartz veins at Kenge and ers” described by Imber et al. (1997), which implies the bulk
Mbenge are in fact fault-fill veins (sensu Robert and Poulsen, rheological behavior of the shear zones at Kenge and Mbenge
2001), which were emplaced subparallel to the mylonitic folia- was controlled by the weak, phyllosilicate-rich mylonites. The
tion (Fig. 11a, b). We further suggest that the occurrences of phyllosilicate-rich mylonites observed at Porcupine are also
foliation-perpendicular strain fringe fibers on pyrite grains are likely to be weak. However, the narrow, discontinuous nature
most simply explained as recording the instantaneous exten- of individual shear zones suggests that the weakening effect
sion direction during transient embrittlement and dilation of was more localized (cf. Handy, 1990). We therefore infer that
the shear zone associated with fault-fill vein emplacement the NW-SE−trending Kenge shear zone slipped with an
(Kerrich and Allison, 1978). apparent friction coefficient (μ) that was lower than that of
Porcupine mineral system the ESE-WNW−trending shear zone hosting Porcupine.
Third, the presence of shallowly dipping extension/oblique-
The Porcupine mineral system contains an NI 43-101 com- extension veins at Porcupine, but not Kenge, suggests that the
pliant measured and indicated resource of 650,000 oz (15.4 fluid pressures at Porcupine locally approached, or exceeded
Mt at 1.31 g/t Au using a 0.5 g/t Au cutoff; Pratt and Simpson, the overburden pressure. By contrast, the presence of fault-
2012). The majority of mineralization is hosted by an E-NE− fill veins suggests that fluid pressures at the Kenge orebody
striking (ca. 59°) and S-dipping (ca. 64°) band of hydrother- were elevated, but sublithostatic throughout the period of
mally altered, veined, and locally sheared Paleoproterozoic D2 shearing. These inferences suggest that structures paral-
Ilunga Syenogranite (ca. 1960 Ma; Lawley et al., 2013a). On lel to the Saza trend may have been misoriented or severely
surface, the Porcupine main zone is poorly exposed in two misoriented for frictional slip, with shear being maintained
artisanal workings and so the following orebody characteris- by near-lithostatic fluid pressures. This behavior is consistent
tics are largely based on observations from diamond drill core. with the fault valve mechanism (Sibson et al., 1988). By con-
Two types of quartz veins are present at the Porcupine Main trast, shear zones parallel to the Kenge trend were able to slip
zone: (1) fault-fill veins with similar characteristics to those under sublithostatic fluid pressure conditions, which is con-
observed at the Kenge and Mbenge deposits; and (2) shal- sistent with the inferred weakness of the Kenge shear zone.
lowly dipping extension/oblique-extension veins (Robert and Finally, the presence of subhorizontal Au-bearing extension/
Poulsen, 2001; Fig. 11e, f). Both vein types are Au bearing, oblique-extension veins at Porcupine suggests that the mini-
but the oblique-extension veins are economically important mum principal stress, σ3, may have been vertical to subvertical
as they act to significantly widen and extend the mineralized during D2 transpression and Au mineralization (Fig. 14r).
zone beyond the edges of the mylonitized shear zones. Exten- Based on these conclusions, we hypothesise that the differ-
sion/oblique-extension veins are shallowly dipping structures ent style of Au-bearing quartz veins at the Kenge and Porcu-
(mean vein strike/dip = 285°/02°; n = 714), possess parallel pine mineral systems (i.e., presence or absence of extension/
and planar vein margins, are composed of massive quartz (bull oblique-extension veins) results from the different shear zone
quartz), and lack textural evidence to identify their opening orientations, variable degrees of fault/shear zone weakening,
vector (Fig. 11h). The latter makes it difficult to distinguish and spatial and temporal variations in fluid pressure.
whether quartz veins at Porcupine represent “true” exten-
sional veins with opening vectors perpendicular to the vein Reactivation Analysis
margin, or oblique-extension veins. Nevertheless, the massive
and unstrained nature of the veins coupled with the shallow- Rationale and basis
dipping orientation makes quartz veins at Porcupine distinct Yin and Ranalli (1992) presented a modified form of the
from fault-fill veins at Kenge. Coulomb-Navier criterion to assess the likelihood of frictional

a b c mylonite
quartz
K feldspar
quartz
quartz
K feldspar
sericite
sericite
1 mm 1 mm 1 mm
grain size
sericitization
finite strain
Fig. 17. (a)-(c). Crossed nicols photomicrographs of three Ilunga Syenogranite samples exhibiting various degrees of
hydrothermal alteration and finite strain. Our proposed model for shear zone development would place these three samples
into a progressive sequence from left to right. (a). Coarse-grained, nonfoliated and relatively unaltered granite becomes
hydrothermally altered, potentially due to pre-D2 brittle fracturing and results in the replacement of feldspars with intrinsi-
cally weaker phyllosilicates. (b). Continued hydrothermal promotes strain localization within the altered granite and leads to
increasing finite strain and mylonitization. (c). Complete replacement of feldspars with further grain size reduction producing
a quartz-phyllosilicate mylonite.
slip (reactivation) along an arbitrarily oriented anisotropy sub- Differential stress is expressed in terms of frictional strength
jected to an arbitrarily oriented stress system (see also Alaniz- (μ or μ0), pore fluid factor (l), cohesion (S or S0), rock density
Alvarez et al., 1998). As such, this approach can be used to (ρ), depth of deformation (z), and orientations of the princi-
explain the different styles of Au-bearing quartz veins associ- pal stresses and anisotropy. Complete definitions of symbols
ated with the Kenge and Saza trends. and their units are given in Table 1. We have used the mea-
Yin and Ranalli (1992) defined a Cartesian coordinate sys- sured orientations of the Kenge and Porcupine shear zones,
tem x1, x2, x3 corresponding to the directions of the principal together with the constraints derived from the outcrop, core,
stresses σ1, σ2, σ3, respectively, m is the unit normal to a hori-
zontal plane through the origin, where mi = cos αi (i = 1, 2, Table. 1. Notation Used for Reactivation Analysis
3) and αi is the angle between the unit normal m and the xi (after Yin and Ranalli, 1992)
axis. Thus for a vertical σ3, as inferred during D2 transpression
Symbol Definition Units
in the western Lupa goldfield, m1 = m2 = 0 and m3 = 1. The
azimuths of σ1 and σ2 are unknown, but both are assumed to μ Coefficient of friction for isotropic rock
have been horizontal. The orientation of a preexisting anisot- μ0 Coefficient of sliding friction along the preexisting
ropy, such as the Kenge or Porcupine shear zones, is specified anisotropy
ρ Rock density kg/m3
by a unit normal, n, where ni = cos γi (i = 1, 2, 3) and γi is angle g Acceleration due to gravity m/s2
between n and the xi axis. λ Pore fluid factor (ratio of pore fluid pressure to
Yin and Ranalli (1992) derived equations for the differential overburden pressure)
stress (σ1-σ3) required for the formation of a new fault plane S Cohesion Pa
in isotropic rock: S0 Cohesion along the preexisting anisotropy Pa
δ δ = (σ2-σ3)/(σ1-σ3)
2mrgz(1 – l) + 2S m Unit normal to the horizontal reference plane
s1 – s3 = —————————————, (1) n Unit normal to a preexisting anisotropy
(m2 + 1)1/2 – m + 2m(m21 + dm22) mi mi = cos αi (i = 1, 2, 3) and αi is the angle between
the unit normal m and the xi axis
and reactivation of a preexisting anisotropy: ni ni = cos γi (i = 1, 2, 3) and γi is angle between n and
the xi axis
m0 rgz(1 – l) + S0
s1 – s3 = ————————————. (2) xi Cartesian axes (i = 1, 2, 3) corresponding to the
[(n21 + d2n22) – (n21 + dn22)]1/2 + direction of the principal stresses σ1, σ2, σ3.
m0 [m21 + dm22) – (n21 + dn22)]

1634 LAWLEY ET AL.
and microstructural observations (see above) to estimate the Table 2. Parameters Used to Model Frictional Reactivation Along the
parameters in equations (1) and (2). In this way, we can assess Kenge and Porcupine Shear Zones (the values of ni (i = 1, 2) for the Kenge
and Porcupine shear zones vary according to the azimuth of σ1 and σ2)
whether our hypothesis based on qualitative outcrop, core,
and microstructural observations is consistent with three- Value (units
dimensional frictional reactivation theory. This is a reasonable Parameter Notes as in Table 1)
approach given the success of two-dimensional frictional reac-
μ Coefficient of friction for isotropic granitoid; 0.75
tivation theory in predicting the development of fault-fill and value as given by Yin and Ranalli (1992)
extension veins along fault/shear zone systems subjected to μ0 Coefficient of sliding friction along the Kenge 0.3
near-orthogonal compression. shear zone
μ0 Coefficient of sliding friction along the 0.4
Assumptions and model set-up Porcupine shear zone
ρ Rock density 2600
We assume that the absolute maximum differential stress g Acceleration due to gravity 9.81
during D2 deformation along shear zones hosting the Kenge λ Maximum pore fluid factor along the Kenge 0.57
and Porcupine mineral systems was limited by that required to shear zone
initiate a new fault plane in the intact granitoid protolith. This λ Maximum pore fluid factor along the 0.57 to 0.9
Porcupine shear zone
assumption is supported by the lack of consistent crosscutting S Cohesion for isotropic granitoid 750000001
relationships between brittle-ductile shear zones with differ- S0 Cohesion along the Kenge and Porcupine 5
ent orientations, which suggests that progressive D2 deforma- shear zones
tion was accommodated by reactivation of existing structures, δ δ = (σ2-σ3)/(σ1-σ3) 0.51
rather than initiation of new faults/shear zones. Second, we m1 σ1 horizontal with azimuth between 0° and 180° 0
m2 σ2 horizontal 0
assume that emplacement of the fault-fill veins was associated m3 σ3 vertical 1
with frictional sliding (reactivation) along the host shear zone. n1 Kenge shear zone (strike 120°; dip 70°SW) Variable
Third, we assume that the misoriented to severely misoriented n2 Kenge shear zone (strike 120°; dip 70°SW) Variable
Porcupine shear zone was locked until such time that the pore n3 Kenge shear zone (strike 120°; dip 70°SW) 0.34
n1 Porcupine shear zone (strike 059°; dip 64°S) Variable
fluid pressure approached the lithostatic pressure (i.e., l → 1). n2 Porcupine shear zone (strike 059°; dip 64°S) Variable
This assumption is consistent with the presence of extension/ n3 Porcupine shear zone (strike 059°; dip 64°S) 0.44
oblique-extension veins at the Porcupine mineral system.
Slip along this Saza parallel shear zone is assumed to have 1 Value given by Yin and Ranalli (1992)
occurred at a differential stress approximately equal to that
required for continued slip on the weaker Kenge shear zone.
Finally, we take the depth of deformation and Au mineraliza- range of azimuths between ca. 42° and 75° is unlikely, given
tion as 10 km, which is a reasonable estimate for the depth of that the differential stress required for reactivation on the
the frictional-viscous transition and greenschist facies meta- Porcupine shear zone would exceed that required to form
morphism (e.g., Scholz, 1988). a new shear fracture within the intact protolith. The Kenge
The differential stress required for reactivation of a preex- and Porcupine shear zones would slip at equal differential
isting anisotropy is very sensitive to frictional strength, μ0 (e.g., stresses when the azimuth of σ1 is approximately 8°, 39°, 98°,
Alaniz-Alvarez et al., 1998). The value assigned to this param- and 135°. These stress orientations would permit simultane-
eter also represents the biggest uncertainty in our model, ous reactivation of the both the Kenge and Porcupine shear
although we believe our qualitative inference of the greatest zones. However, the uniformly low (sublithostatic) pore fluid
degree of weakening along the Kenge shear zone, and a lesser pressures (l = 0.57) cannot explain the presence of oblique/
degree weakening along the Porcupine shear zone, is robust. oblique-extension veins at the Porcupine mineral system,
Previous authors have argued that quartz-mica phyllonites— which likely formed when l → 1.
in which deformation is controlled by pressure-solution By way of comparison, Figure 18b shows the differential
accommodated frictional sliding along mica foliation planes— stresses required for reactivation under the same conditions as
may have very low frictional strengths, μ0 = 0.25 (Jefferies et above, except that μ0 = 0.3 for both the Kenge and Porcupine
al., 2006b). We adopt slightly more conservative values for the shear zones. The graph shows that the more shallowly dip-
friction coefficient, with μ0 = 0.3 for the Kenge shear zone ping Porcupine shear zone is, in fact, more favorably oriented
and μ0 = 0.4 for the Porcupine shear zone. The complete set for slip (i.e., requires lower differential stress) than the more
of parameter values used in the modeling is shown in Table 2. steeply dipping Kenge shear zone for most orientations of σ1.
Nevertheless, the Porcupine shear zone does require a higher
Results differential stress to reactivate than the Kenge shear zone for
Figure 18a shows how the differential stress required azimuths of σ1 between approximately 52° and 82°. However,
for frictional reactivation along the Kenge and Saza shear
zones varies according to the azimuth of σ1, based on the
above assumptions but taking a uniform pore fluid pressure
(l = 0.57) for both shear zones (see Table 2). Under these Fig. 18. (a)-(c). Graphs showing the results of frictional reactivation analy-
conditions, the Porcupine shear zone is more unfavorably ses following the approach of Yin and Ranalli (1992). Each graph shows the
differential stresses required for shear zone reactivation or shear failure along
oriented for slip (i.e., requires a higher differential stress for the Kenge and Porcupine shear zones according to the orientation of σ1 and
reactivation) than the Kenge shear zone for azimuths of σ1 under varying conditions of fluid pressure and friction coefficients (see Tables
between approximately 0°–8°, 39°–98°, and 135°–180°. The 1 and 2 for symbol definitions and model parameters, respectively).

1400 the maximum pore fluid factor (l) required for the Porcu-
a) Kenge (µ = 0.3; = 0.57)
pine shear zone to slip at approximately the same differential
Differential stress required for reactivation or shear failure (MPa)
Porcupine (µ = 0.4; = 0.57)

Protolith ( = 0.57)
stress as the Kenge shear zone is <0.72, which is inconsistent
1200
with the presence of extension/oblique-extension veins at the
Porcupine mineral system. This result supports our earlier
1000 inference that the Kenge shear zone has likely experienced a
greater degree of weakening than the Porcupine shear zone.
Figure 18c shows how the differential stress required for
800
frictional reactivation along the Kenge and Porcupine shear
zones varies according to the azimuth of σ1, assuming pore
600 fluid factors (l) of 0.57 (Kenge shear zone) and 0.9 (Porcupine
shear zone). In this scenario, the Kenge shear zone (μ0 = 0.3)
is weaker than the Porcupine shear zone (μ0 = 0.4). Under
400
these conditions, both shear zones will reactivate at approxi-
mately equal differential stresses when the azimuth of σ1 is ca.
200 60°. This model can explain contemporaneous slip on the
Kenge and Porcupine shear zones and the presence of exten-
sion/oblique-extension veins at Porcupine (but not Kenge).
0 We envisage a sequence of events similar to the classic fault
0 20 40 60 80 100 120 140 160 180
1
azimuth (º) valve model, assuming that emplacement of fault-fill veins: (1)
1400 occurred during frictional slip events along the host shear
b) Kenge (µ = 0.3; = 0.57) zones; and (2) allowed rapid fluid discharge and drop in pore
Porcupine (µ = 0.3; = 0.57)

Protolith ( = 0.57) fluid pressure (i.e., decrease in l). During the prefailure
1200 period, ongoing viscous creep accommodated by crystal plas-
tic deformation and fluid-assisted diffusive mass transfer was
1000
associated with a gradual increase in both differential stress
and pore fluid pressure. When l = 0.57, frictional slip took
place along the Kenge shear zone, with emplacement of fault-
800 fill veins, resealing of the shear zone, and reduction in differ-
ential stress. During this period, the Porcupine shear zone
600
remained locked, with the local fluid pressure still rising.
Extension/oblique-extension veins were emplaced when l → 1,
with frictional slip and fault-fill vein emplacement occurring
400 on both the Kenge (l = 0.57) and Porcupine (l = 0.9) shear
zones when the differential stress had recovered sufficiently
(to ca. 300 MPa in our model; Fig. 18c). An alternative
200
sequence of events is that fluid recharge occurred at a faster
rate at Porcupine than at Kenge, with frictional slip, fault-fill
0 vein emplacement, and fluid discharge along both shear zone
0 20 40 60 80 100 120 140 160 180
azimuth (º) trends taking place simultaneously. The latter may be consis-
1
tent with the location of Porcupine close to an intersection
c) 1400
between two shear zones (Pratt and Simpson, 2012), given the
Kenge (µ = 0.3; = 0.57)

Porcupine (µ = 0.4; = 0.9) common interpretation that fluid flow is likely to be focussed
Protolith ( = 0.57)
1200 Protolith ( = 0.9) along structural intersections (e.g., Bonson et al., 2007).
Clearly there are large uncertainties in the modeling param-
eters, and the result shown in Figure 18c is nonunique. Never-
1000
theless, the quantitative results suggest that three-dimensional
frictional reactivation theory could explain the observed dif-
800 ferences in style and distribution of Au-bearing quartz veins at
Kenge and Porcupine. Importantly, the modeling provides
600
testable predictions about the frictional strength and pore
pressure evolution of the host shear zones. These predictions
have implications for orogenic Au mineralization along other
400 non-Andersonian fault systems, which are discussed below.
200 Discussion and Conclusions

Frictional strength of host shear zones
0
0 20 40 60 80 100 120 140 160 180 Our results highlight the likely importance of fault/shear
1
azimuth (º) zone rheology, in addition to shear zone orientation and

1636 LAWLEY ET AL.
fluid pressure, in controlling the style and distribution of Au- and hydrogeologic compartmentalization of the shear zone
bearing quartz veins associated with non-Andersonian faults/ network (e.g., Faulkner and Rutter, 2001). For this mecha-
shear zones. Nevertheless, fault/shear zone strength remains nism to have been effective, rupture propagation during the
the biggest uncertainty in our model. Our hypothesis suggests emplacement of individual fault-fill veins is likely to have
that the weakest shear zones, such as those containing thick, been restricted to specific fault/shear zone segments. Rup-
interconnected networks of phyllosilicate-rich mylonite, will tures along the Kenge shear zone, for example, are unlikely
reactivate at lower pore fluid pressures than stronger shear to have propagated along the Saza shear zone to intersect the
zones, i.e., those containing thin, discontinuous mylonites. region now occupied by the Porcupine mineral system. This
Drilling other exploration targets in the western Lupa gold- interpretation appears to be consistent with the (typically)
field may provide qualitative support to this hypothesis, but strain- and velocity-hardening characteristics of phyllosilicate-
direct testing requires experimental studies of rock strength. rich fault rocks (e.g., Faulkner and Rutter, 2001; Collettini et
Collettini et al. (2009) described a series of experiments in al., 2009), but further work is required to map the extent of
which they sheared both intact, foliation-parallel wafers and individual fault-fill veins.
powdered samples of phyllosilicate-bearing mylonite from the It is also important to consider the origin of overpressured
island of Elba, Italy, in a biaxial deformation rig at a range of pore fluids at each mineral system. Suprahydrostatic pres-
sliding velocities. They obtained lower values for μ (ca. 0.25) sures will develop if fluid input at the base of the seismogenic
from the wafers compared with the powered samples. These crust exceeds fluid loss at the Earth’s surface plus lateral fluid
results unequivocally demonstrate the importance of phyllo- loss perpendicular to the fault/shear zone margins (Faulkner
silicate-rich foliae in weakening the host shear zones. A key and Rutter, 2001). As noted previously, the location of Porcu-
test of our model would be to perform similar deformation pine close to the intersection of two shear zone trends may
experiments on wafers of hydrothermally altered wall rock have facilitated sustained and focused fluid flow. Compaction
and mylonite from Kenge and Porcupine. These results would of the fault rocks can also lead to increased fluid pressure.
add confidence to the use of three-dimensional reactivation Wintsch et al. (1995) showed that some greenschist facies
theory to predict the style and spatial distribution of Au-bear- hydration reactions, such as alteration of feldspar to sericite,
ing quartz veins within the Lupa goldfield and non-Anderso- are associated with up to 50% volume loss. Progressive alter-
nan fault/shear zone networks elsewhere. ation of feldspar porphyroclasts to sericite is likely to initiate
along intragranular fractures and/or cleavage planes and
Evolution and distribution of pore fluid pressures could lead to an increase in fracture porosity of the host
Our model predicts different pore fluid pressures along grain. When the proportion of weak phases (sericite) exceeds
different structural trends and implies a certain degree of a critical percentage (probably on the order of 20−30%;
structural and hydrogeologic compartmentalization between Handy, 1990), the host feldspar grain may collapse (i.e., com-
the Kenge and Porcupine mineral systems. Fluid compart- pact), giving rise to locally continuous sericite foliae. Our
mentalization is expected to be a dynamic process inextricably geochemical data suggest that mylonitization and alteration
linked with progressive deformation and fluid-rock interac- did not cause significant finite volume changes. One possibil-
tion within the host shear zone (Sibson, 1990). In principle, ity is that the high fluid/rock ratios effectively mask evidence
fluid inclusion studies could provide direct evidence of the for volume change (see above). Alternatively (or addition-
fluid pressure conditions at each deposit, given sufficient con- ally), finite volume loss caused by feldspar dissolution may
straints on the composition of the fluid inclusions, thermo- have been compensated by precipitation of silica released
metric data on the phases in the inclusion, and availability of during greenschist facies hydration reactions within pore
appropriate experimental phase data (e.g., Roedder and Bod- spaces, and by vein emplacement into the host shear zones.
nar, 1980). However, a fundamental question remains about Very high pore fluid pressures (l → 1), such as those inferred
the mechanism of compartmentalization, given the apparent at Porcupine, may have locally reduced the effective normal
interconnectivity of the host shear zones, at least at the con- stresses at grain-grain contacts within the partially altered
cession scale (e.g., Fig. 3). The implication of our model is granitoid protolith, thus inhibiting compaction and develop-
that structural compartments within the shear zone network ment of continuous phyllosilicate foliae. Silica precipitation
were able to maintain significant (but transient?) lateral pore within pore spaces during periodic drops in fluid pressure
pressure gradients (e.g., l = 0.9 at Porcupine, contempo- (e.g., during fault-fill vein emplacement) may also have sup-
raneous with l = 0.57 at Kenge). One possibility is that the pressed compaction and hence the development of a large-
Kenge and Porcupine shear zones are, in fact, not laterally scale “interconnected weak layer” structure. By contrast, the
connected. However, this suggestion seems unlikely given suprahydrostatic, but sublithostatic, fluid pressures inferred
that both structures form part of a more extensive and broadly at Kenge may have been insufficient to relieve the effective
contemporaneous shear zone network (Fig. 3; Lawley et al., normal stresses at grain-grain contacts, thus promoting com-
2013b), and both shear zones appear to be connected at depth paction of partially altered, sericitized feldspar porphyro-
to a similar fluid source. Moreover, well-accepted models of clasts. Fault-fill vein emplacement in this environment may
fault growth emphasize that structures apparently isolated in also have enhanced compaction due to the transient increase
map view are, in fact, typically connected in three dimensions in mean stress in the vicinity of veins. This simple conceptual
(e.g., Childs et al., 1995; Walsh et al., 2003; Kristensen et al., model links variations in pore fluid pressure to fault rock
2008). paragenesis and provides a rationale for the variable degrees
It is possible that the intrinsically low permeability of of fault/shear zone weakening inferred at Kenge and
the phyllosilicate-rich rocks facilitated effective structural Porcupine.

In conclusion, application of a modified form of the Cou- ——2002, Fluid-assisted healing processes in gouge-bearing faults: Insights
lomb-Navier failure criterion can explain the different styles from experiments on a rock analogue system: Pure and Applied Geophys-
ics, v. 159, p. 2537–2566.
of Au mineralization along the Kenge and Saza trends in terms Bos, B., Peach, C.J., and Spiers, C.J., 2000a, Frictional-viscous flow of simu-
of variable pore fluid pressures and different amounts of fault/ lated fault gouge caused by the combined effects of phyllosilicates and pres-
shear zone weakening. We hypothesize that these first-order, sure solution: Tectonophysics, v. 327, p. 173–194.
concession-scale effects are, in turn, related to the grain-scale ——2000b, Slip behaviour of simulated gouge-bearing faults under condi-
(microstructural) evolution of the Au-bearing mylonitic shear tions favouring pressure solution: Journal of Geophysical Research, v. 105,
p. 16,699–16,717.
zones. Additional studies, including (1) laboratory deforma- Byerlee, J.D., 1978, Friction of rocks: Pure and Applied Geophysics, v. 116,
tion experiments to quantify fault rock strength, and (2) fluid p. 615–626.
inclusion studies to quantify pore fluid pressures, are needed Cahen, L., Snelling, N.J., Delhal, J., and Vail, J., 1984, The geochronology
to test our conclusions. Nevertheless, our approach provides and evolution of Africa: Oxford, Clarendon Press, 512 p.
a conceptual link between concession- and grain-scale struc- Childs, C., Watterson, J., and Walsh, J.J., 1995, Fault overlap zones within
developing normal fault systems: Journal of the Geological Society, London,
tural processes. A key area of future research should be to v. 152, p. 535–549.
understand in detail the controls on Au occurrence (e.g., free Collettini, C., Niemeijer, A., Viti, C., and Marone, C., 2009, Fault zone fabric
Au grains vs. pyrite-hosted Au) along different ore trends and fault weakness: Nature, v. 462, p. 907–910.
hosted by non-Andersonian fault/shear zone networks. Craw, D., 2002, Geochemistry of late metamorphic hydrothermal alteration
and graphitisation of host rock, Macres gold mine, Otago Schist, New Zea-
Acknowledgments land: Chemical Geology, v. 191, p. 257–275.
Daly, M.C., 1988, Crustal shear zones in central Africa: A kinematic approach
CJML would like to acknowledge funding from a Durham to Proterozoic tectonics: Episodes, v. 11, p. 5–11.
Doctoral Fellowship at Durham University and a student Dewey, J.F., Holdsworth, R.E., and Strachan, R.A., 1998, Transpression and
research grant from the Society of Economic Geologists. transtension zones: Geological Society, London, Special Publication 135,
Helio Resource Corp. provided funding and access to min- p. 1–14.
Dubé, B., Poulsen, K., and Guha, J., 1989, The effects of layer anisotropy on
eral exploration licenses and diamond drill core. The ideas auriferous shear zones: the Norbeau mine, Quebec: Economic Geology,
presented in this paper also benefited from discussions with v. 84, p. 871–878.
Bob Holdsworth, Richard Jones, Mark Pearce, David Hol- Durney, D.W., and Ramsay, J.G., 1973, Incremental strains measured by syn-
well, Eddie Dempsey, and an anonymous reviewer. Sébastien tectonic crystal growths, in DeJong, K.A., and Sholten, R., eds., Gravity and
Castonguay, David Craw, Brett Davies, and the Economic tectonics: New York, Wiley, p. 67–95.
Elmer, F.L., White, R.W. and Powell, R., 2006, Devolatilization of metaba-
Geology Editors also improved the manuscript by providing sic rocks during greenschist-amphibolite facies metamorphism: Journal of
constructive reviews. This is Earth Science Sector contribu- Metamorphic Geology, v. 24, p. 497–513.
tion 20120448. Faulkner, D.R., and Rutter, E.H., 2001, Can the maintenance of overpres-
sured fluids in large strike-slip fault zones explain their apparent weakness:
REFERENCES Geology, v. 29, p. 503–506.
Goldfarb, R.J., Groves, D.I., and Gardoll, S., 2001, Orogenic gold and geo-
Alaniz-Alvarez, S.A., Nieto-Samaniego, A.F., and Tolson, G., 1998, A graphi-
cal technique to predict slip along a pre-existing plane of weakness: Engi- logic time: A global synthesis: Ore Geology Reviews, v. 18, p. 1–75.
neering Geology, v. 49, p. 53–60. Grant, J.A., 1986, The Isocon diagram: A simple solution to Gresens’ equa-
Anderson, E.M., 1905, The dynamics of faulting: Edinburgh Geological Soci- tion for metasomatic alteration: Economic Geology, v. 81, p. 1976–1982.
ety Transactions, v. 8, p. 393–402. ——2005, Isocon analysis: A brief review of the method and applications:
Anderson, E.M., 1951, The dynamics of faulting and dyke formation with Physics and Chemistry of the Earth, v. 30, p. 997–1004.
application to Britain, 2nd ed.: Edinburgh, Oliver and Boyd, 206 p. Gresens, R.L., 1967, Composition-volume relationships of metasomatism:
Berthé, D., Choukroune, P., and Jegouzo, P., 1979, Orthogneiss, mylonite and Chemical Geology, v. 2, p. 47–55.
non-coxial deformation of granites: The example of the South Armoricain Groves, D.I., Goldfarb, R.J., Gebre-Mariam, M., Hagemann, S.G., and
shear zone: Journal of Structural Geology, v. 1, p. 31–42. Robert, F., 1998, Orogenic gold deposits: A proposed classification in the
Bierlein, F.P., Groves, D.I., and Cawood, P.A., 2009, Metallogeny of accre- context of their crustal distribution and relationships to other gold deposit
tionary orogens—the connection between lithospheric processes and metal types: Ore Geology Reviews, v. 13, p. 7–27.
endowment: Ore Geology Reviews, v. 36, p. 282–292. Groves, D.I., Goldfarb, R.J., Robert, F., and Hart, C.J.R., 2003, Gold depos-
Blenkinsop, T.G., and Doyle, M.G., 2010, A method for measuring the ori- its in metamorphic belts: Overview of current understanding, outstanding
entations of planar structures in cut core: Journal of Structural Geology, v. problems, future research, and exploration significance: Economic Geol-
32, p. 741–745. ogy, v. 98, p. 1–29.
Boniface, N., 2009, Eburnian, Kibaran, and Pan-African metamorphic events Handy, M.R., 1990, The solid-state flow of polymineralic rocks: Journal of
in the Ubendian belt of Tanzania: Petrology, zircon and monazite geochro- Geophysical Research, v. 96 (B6), p. 8647–8661.
nology: Unpublished Ph.D. thesis, Kiel, Germany, Kiel University, 110 p. Hanmer, S., and Passchier, C., 1991, Shear sense indictors: A review: Geo-
Boniface, N., and Schenk, V., 2012, Neoproterozoic eclogites in the Paleo- logical Survey of Canada Paper 90-17, 72 p.
proterozoic Ubendian belt of Tanzania: Evidence for a Pan-African suture Harland, W.B., 1971, Tectonic transpression in Caledonian Spitzbergen:
between the Bangweulu block and the Tanzania craton: Precambrian Geological Magazine, v. 108, p. 27–42.
Research, v. 208–211, p. 72–89. Harrison, A., 2011, Mineral resource estimate of the Saza Makongolosi proj-
Boniface, N., Schenk, V., and Appel, P., 2012, Paleoproterozoic eclogites of ect for Helio Resource Corporation: Unpublished Golder Associates NI
MORB-type chemistry and three Proterozoic orogenic cycles in the Uben- 43-101 Technical Report, 201 p.
dian belt (Tanzania): Evidence from monazite and zircon geochronology, Holcombe, R., 2008, Oriented drillcore: Measurement and calculation pro-
and geochemistry: Precambrian Research, v. 192–195, p. 16–33. cedures for structural and exploration geologists: (http://www.holcombe-
Bonson, C.G., Childs, C., Walsh, J.J., Schöpfer, and Carboni, V., 2007, Geo- coughlin.com/HCA_downloads.htm).
metric and kinematic controls on the internal structure of a large normal Holdsworth, R.E., Butler, C.A., and Roberts, A.M., 1997, The recognition
fault in massive limestones: The Maghlaq fault, Malta: Journal of Structural of reactivation during continental deformation: Journal of the Geological
Geology, v. 29, p. 336–354. Society of London, v. 154, p. 73–78.
Bos, B., and Spiers, C.J., 2000, Effect of phyllosilicates on fluid-assisted heal- Holdsworth, R.E., Handa, M., Miller, J.A., and Buick, I.S., 2001, Continental
ing of gouge-bearing faults: Earth and Planetary Science Letters, v. 184, p. reactivation and re-working: An introduction: Geological Society, London,
199–210. Special Publication 184, p. 1–12.

1638 LAWLEY ET AL.
Holdsworth, R.E., Tavarnelli, E., Clegg, P., Pinheiro, R.V.L., Jones, R.R., Tracing metallogenic time scales at midcrustal shear zones hosting orogenic
and McCaffrey, K.J.W., 2002, Domainal deformation patterns and strain Au deposits: Economic Geology, v. 108, p. 1591–1613.
partitioning during transpression: An example from the Southern Uplands Lenoir, J.L., Liégeois, J.P., Theunissen, K., and Klerkx, J., 1994, The Palaeo-
terrane, Scotland: Journal of the Geological Society, London, v. 159, p. proterozoic Ubendian shear belt in Tanzania: Geochronology and structure:
401–415. Journal of African Earth Sciences, v. 19, p. 169–184.
Holdsworth, R.E., van Diggelen, E.W.E., Spiers, C.J., de Bresser, J.H.P., Lin, S., and Corfu, F., 2002, Structural setting and geochronology of aurifer-
Walker, R.J., and Bowen, L., 2011, Fault rocks from the SAFOD core sam- ous quartz veins at the High Rock Island gold deposit, northwestern Supe-
ples: Implications for weakening at shallow depths along the San Andreas rior province, Manitoba, Canada: Economic Geology, v. 97, p. 43–57.
fault, California: Journal of Structural Geology, v. 33, p. 132–144. Lin, S., and Jiang, D., 2001, Using along-strike variation in strain and kine-
Hronsky, J.M.A., Groves, D.I., Loucks, R.R., and Begg, G.C., 2012, A unified matics to define the movement direction of curved transpressional shear
model for gold mineralisation in accretionary orogens and implications for zones: An example from northwestern Superior Province, Manitoba: Geol-
regional-scale exploration targeting methods: Mineralium Deposita, v. 47, ogy, v. 29, p. 767–770.
p. 339–358. Lin, S., Jiang, D., and Williams, P.F., 1998, Transpression (or transtension)
Imber, J., Holdsworth, R.E., Butler, C.A., and Lloyd, G.E., 1997, Fault-zone zones of triclinic symmetry: Natural example and theoretical modelling:
weakening processes along the reactivated Outer Hebrides fault zone, Scot- Geological Society, London, Special Publication 135, p. 41–57.
land: Journal of the Geological Society, London, v. 154, p. 105–109. Mares, V.M., and Kronenberg, A.K., 1993, Experimental deformation of mus-
Imber, J., Holdsworth, R.E., Butler, C.A., and Strachan, R.A., 2001, A reap- covite: Journal of Structural Geology, v. 15, p. 1061–1075.
praisal of the Sibson-Scholz fault zone model: the nature of the frictional McConnell, R.B., 1950, Outline of the geology of Ufipa and Ubende: Geo-
to viscous (“brittle-ductile”) transition along a long-lived, crustal-scale fault, logical Survey of Tanganyika Bulletin, v. 19, 62 p.
Outer Hebrides, Scotland: Tectonics, v. 20, p. 601–624. Niemeijer, A.R., and Spiers, C.J., 2005, Influence of phyllosilicates on fault
Imber, J., Holdsworth, R.E., Smith, S.A.F., Jefferies, S.P., and Collettini, C., strength in the brittle-ductile transition: Insights from rock analogue
2008, Frictional-viscous flow, seismicity and the geology of weak faults: A experiments: Geological Society, London, Special Publication 245, p.
review and future directions: Geological Society, London, Special Publica- 303–327.
tion 299, p. 151–173. Passchier, C.W., and Simpson, C., 1986, Porphyroclast systems as kinematic
Jefferies, S.P., 2006, Microstructural and geochemical processes in long-lived indicators: Journal of Structural Geology, v. 8, p. 831–843.
reactivated crustal-scale fault zones: A case study from the Median Tec- Pratt, D., and Simpson, R., 2012, NI 43-101 Preliminary economic assess-
tonic Line, SW Japan: Unpublished Ph.D. thesis, Durham, UK Durham ment for the Saza-Makongolosi gold project: Unpublished company report,
University, 232 p. SRK Consulting, 313 p.
Jefferies, S.P., Holdsworth, R.E., Wibberley, C.A.J., Shimamoto, T., Spiers, Robert, F., and Poulsen, K.H., 2001, Vein formation and deformation in
C.J., Niemeijer, A.R., and Lloyd, G.E., 2006a, The nature and importance greenstone gold deposits: Society of Economic Geology Reviews, v. 14, p.
of phyllonite development in crustal-scale fault cores: An example from 111–155.
the Median Tectonic Line, Japan: Journal of Structural Geology, v. 28, p. Roedder, E., and Bodnar, R.J., 1980, Geologic pressure determinations from
220–235. fluid inclusion studies: Annual Review of Earth and Planetary Sciences, v.
Jefferies, S.P., Holdsworth, R.E., Shimamoto, T., Takagi, H., Lloyd, G.E., and 8, p. 263–301.
Spiers, J., 2006b, Origin and mechanical significance of foliated cataclastic Sango, P.M., 1988, Structural and lithological controls of gold mineraliza-
rocks in the cores of crustal-scale faults: Examples from the Median Tec- tion in the Lupa goldfield, Tanzania: Geology Department and University
tonic Line, Japan: Journal of Geophysical Research, v. 111, p. 1978–2012. Extension, University of Western Australia, Publication 12, p. 99–109.
Jiang, D., and White, J.C., 1995, Kinematics of rock flow and the interpreta- Scholz, C.H., 1988, The brittle-plastic transition and the depth of seismic
tion of geological structures, with particular reference to shear zones: Jour- faulting: Geologische Rundschau, v. 77, p. 319–328.
nal of Structural Geology, v. 17, p. 1249–1265. Shaw, E., 2009, Fluid evolution in the Lupa goldfield, southwest Tanzania:
Jiang, D., and Williams, P.F., 1998, High-strain zones: A unified model: Jour- Unpublished M.Sc., thesis, Leicester, University of Leicester, 54 p.
nal of Structural Geology, v. 20, p. 1105–1120. Sibson, R.H., 1990, Conditions for fault-valve behaviour: Geological Society,
Jones, R.R., and Holdsworth, R.E., 1998, Oblique simple shear in transpres- London, Special Publication 54, p. 15–28.
sion zones: Geological Society, London, Special Publication 135, p. 35–40. Sibson, R.H., Robert, F., and Poulsen, H., 1988, High-angle reverse faults,
Jones, R.R., Holdsworth, R.E., Clegg, P., McCaffrey, K., and Tavarnelli, fluid-pressure cycling, and mesothermal gold-quartz deposits: Geology, v.
E., 2004, Inclined transpression: Journal of Structural Geology, v. 26, p. 16, p. 551–555.
1531–1548. Snoke, A.W., and Tullis, J., 1998, An overview of fault rocks, in Snoke, A.W.,
Kilembe, E.A., and Rosendahl, B.R., 1992, Structure and stratigraphy of the Tullis, J., and Todd, V.R., eds., Fault-related rocks: A photographic atlas:
Rukwa rift: Tectonophysics, v. 209, p. 143–158. Princeton, New Jersey, Princeton University Press, p. 3–18.
Kerrich, R., and Allison, I., 1978, Vein geometry and hydrostatics during Stewart, M., Holdsworth, R.E., and Strachan, R.E., 2000, Deformation pro-
Yellowknife mineralization: Canadian Journal of Earth Sciences, v. 15, p. cesses and weakening mechanisms within the frictional-viscous transition
1653–1660. zone of major crustal-scale faults: Insights from the Great Glen fault zone,
Kerrich, R., and Wyman, D., 1990, Geodynamic setting of mesothermal gold Scotland: Journal of Structural Geology, v. 22, p. 543–560.
deposits: An association with accretionary tectonic regimes: Geology, v. 18, Sutton, J., Watson, J., and James, T.C., 1954, A study of the metamorphic
p. 882–885. rocks of Karema and Kungwe bay, western Tanganyika: Bulletin of the Geo-
Köhn, D., Bons, P.D., Hilgers, C., and Passchier, C.W., 2003, Development logical Survey of Tanganyika, v. 22, 70 p.
of antitaxial strain fringes during non-coaxial deformation: An experimental Theunissen, K., Klerkx, J., Melnikov, A., and Mruma, A., 1996, Mechanisms
study: Journal of Structural Geology, v. 25, p. 263–275. of inheritance of rift faulting in the western branch of the east African rift,
Krantz, W.R., 1988, Multiple fault sets and three-dimensional strain: Theory Tanzania: Tectonics, v. 15, p. 776–790.
and application: Journal of Structural Geology, v. 10, p. 225–237. Tikoff, B., and Fossen, H., 1999, Three dimensional reference deformations
Kristensen, M.B., Childs, C.J., and Korstgård, J.A., 2008, The 3D geometry and strain facies: Journal of Structural Geology, v. 21, p. 1497–1512.
of small-scale relay zones between normal faults in soft sediments: Journal Tullis, J., and Yund, R.A., 1980, Hydrolytic weakening of experimentally
of Structural Geology, v. 30, p. 257–272. deformed Westerly granite and Hale albite rock: Journal of Structural Geol-
Kuehn, S., Ogola, J., and Sango, P., 1990, Regional setting and nature of gold ogy, v. 2, p. 439–451.
mineralization in Tanzania and southwest Kenya: Precambrian Research, ——1985, Dynamic recrystallization of feldspar: A mechanism for ductile
v. 46, p. 71–82. shear zone formation: Geology, v. 13, p. 238–241.
Lawley, C.J.M., Selby, D., Condon, D.J., Horstwood, M., Millar, I., Crowley, Van Straaten, V.P., 1984, Gold mineralization in Tanzania—a review, in Fos-
Q., and Imber, J., 2013a, Lithogeochemistry, geochronology, and geody- ter, R.P., eds., Gold ’82: The geology, geochemistry and genesis of gold
namic setting of the Lupa terrane, Tanzania: Implications for the extent of deposits: Rotterdam, A.A. Balkema, p. 673–685.
the Archean Tanzanian craton: Precambrian Research, v. 231, p. 174–193. Walsh, J.J., Bailey, W.R., Childs, C., Nicol, A., and Bonson, C.G., Formation
Lawley, C.J.M., Selby, D., and Imber, J., 2013b, Re-Os molybdenite, pyrite, of segmented normal faults: A 3-D perspective: Journal of Structural Geol-
and chalcopyrite geochronology, Lupa goldfield, southwestern Tanzania: ogy, v. 25, p. 1251–1262.

Wintsch, R.P., Christofferson, R., and Kronenberg, A.K., 1995, Fluid-rock Yin, Z.M., and Ranalli, G., 1992, Critical stress difference, fault orientation
reaction weakening of fault zones: Journal of Geophysical Research, v. 100, and slip direction in anisotropic rocks under non-Andersonian stress sys-
p. 13,021–13,032. tems: Journal of Structural Geology, v. 14, p. 237–244.
Woodcock, N.H., and Naylor, M.A., 1983, Randomness testing in three-
dimensional orientation data: Journal of Structural Geology, v. 5, p. 539–548.

1640 LAWLEY ET AL.
Table A1. Lithogeochemistry Results
CL1101 CL1102 CL1101 CL1102
Analysis 499282 / 501137 / Analysis 499282 / 501137 /

Analyte Method1 Unit2 D.L. 90732583 90729693 Analyte Method1 Unit2 D.L. 90732583 90729693
SiO2 FUS-ICP % 0.01 67.44 62.77 Ni TD-ICP ppm 1 23 13

Al2O3 FUS-ICP % 0.01 15.94 17.19 Pb TD-ICP ppm 5 < 5 11
Fe2O3(T) FUS-ICP % 0.01 3.7 2.83 Rb FUS-ICP ppm 1 120 86
MnO FUS-ICP % 0.001 0.032 0.028 S TD-ICP % 0.001 0.003 0.004
MgO FUS-ICP % 0.01 1.38 0.73 Sb INAA ppm 0.1 2.7 2.6
CaO FUS-ICP % 0.01 0.46 3.09 Sc INAA ppm 0.01 5.78 6.37
Na2O FUS-ICP % 0.01 1.83 4.06 Se INAA ppm 0.5 < 0.5 < 0.5
K2O FUS-ICP % 0.01 4.32 3.73 Sn FUS-ICP ppm 1 1 <1
TiO2 FUS-ICP % 0.001 0.445 0.46 Sr FUS-ICP ppm 2 94 178
P2O5 FUS-ICP % 0.01 0.13 0.14 Ta FUS-ICP ppm 0.01 0.36 0.39
LOI FUS-ICP % 3.06 4.13 Th FUS-ICP ppm 0.05 7.52 6.84
Total FUS-ICP % 0.01 98.74 99.14 U FUS-ICP ppm 0.01 1.56 1.37
Au INAA ppb 1 < 1 40 V FUS-ICP ppm 5 60 74
Ag TD-ICP ppm 0.5 < 0.5 < 0.5 W INAA ppm 1 3 9
As INAA ppm 1 2 3 Y FUS-ICP ppm 1 12 9
Ba FUS-ICP ppm 1 1553 1399 Zn TD-ICP ppm 1 34 30
Be FUS-ICP ppm 1 1 2 Zr FUS-ICP ppm 1 139 140
Bi FUS-ICP ppm 0.1 0.1 0.3 La FUS-ICP ppm 0.05 35.2 31.7
Br INAA ppm 0.5 < 0.5 < 0.5 Ce FUS-ICP ppm 0.05 68.1 63.1
Cd TD-ICP ppm 0.5 < 0.5 < 0.5 Pr FUS-ICP ppm 0.01 6.15 6.81
Co INAA ppm 0.1 7.7 9.4 Nd FUS-ICP ppm 0.05 21.8 23.8
Cr INAA ppm 0.5 128 118 Sm FUS-ICP ppm 0.01 3.73 3.94
Cs FUS-ICP ppm 0.1 4.9 2.1 Eu FUS-ICP ppm 0.005 0.994 1.01
Cu TD-ICP ppm 1 32 47 Gd FUS-ICP ppm 0.01 2.73 2.65
Ga FUS-ICP ppm 1 19 20 Tb FUS-ICP ppm 0.01 0.4 0.38
Ge FUS-ICP ppm 0.5 1.7 1.7 Dy FUS-ICP ppm 0.01 2.18 1.89
Hf FUS-ICP ppm 0.1 3.5 3.4 Ho FUS-ICP ppm 0.01 0.4 0.35
Hg INAA ppm 1 < 1 <1 Er FUS-ICP ppm 0.01 1.08 0.98
In FUS-ICP ppm 0.1 < 0.1 < 0.1 Tl FUS-ICP ppm 0.05 0.42 0.32
Ir INAA ppb 1 < 1 <1 Tm FUS-ICP ppm 0.005 0.158 0.142
Mo FUS-ICP ppm 2 < 2 <2 Yb FUS-ICP ppm 0.01 1 0.91
Nb FUS-ICP ppm 0.2 4.7 5 Lu FUS-ICP ppm 0.002 0.148 0.142
1 Analysis method abbreviations: FUS-ICP = fusion inductively coupled plasma mass spectrometry, INAA = instrumental neutron activation analysis, TD-
ICP = total dissolution inductively coupled plasma mass spectrometry, MULT INAA = multi-element instrumental neutron activation analysis
2 Detection limit
3 Eastings and northings are reported as UTM coordinates (WGS84, zone 36S)


Structural Controls On Orogenic Au Mineralization During Transpression: Lupa Goldfield, Southwestern Tanzania

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Structural Controls On Orogenic Au Mineralization During Transpression: Lupa Goldfield, Southwestern Tanzania

Uploaded by

Copyright:

Available Formats

©2013 Society of Economic Geologists, Inc.

Economic Geology, v. 108, pp. 1615–1640

Structural Controls on Orogenic Au Mineralization During Transpression:

Introduction in three-dimensional (e.g., transpressional) strain fields

nian geometries, whereby the maximum principal compres- Y

sive stress (σ1) at the time of mineralization bisects the acute X

angle (typically 50°–60°) between conjugate normal, reverse

Submitted: July 30, 2012

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/108/7/1615/3469630/1615-1640.pdf

(of 2iC Australia™) and Reflex Act™ (of Reflex™) tools.

Ezy-Mark™ and Reflex Act™ orientation tools are highly

accurate; however, the majority of orientation error can occur

ing intervals of bad ground. Planar and linear measurements

Meso- and Neoproterozoic

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/108/7/1615/3469630/1615-1640.pdf

instrumental neutron activation analysis (INAA) by Actlabs

(Fig. 3; Table A1; Ancaster, Ontario; method 4E-Research).

Sample aliquants for ICP-MS analysis were first mixed with

a lithium metaborate-tetraborate flux and fused in order to

ensure complete digestion of refractory minerals (e.g., zir-

were completed as part of a larger geochemical study that

8 - Porcupine included standards and duplicates as a means of quality con-

diorite were previously reported in Lawley et al. (2013a).

Regional Geology, Southwestern Tanzania

The Paleoproterozoic Ubendian belt circumscribes the

lithotectonic Archean terranes: Ubende, Wakole, Katuma,

foliated Archean granite

Ufipa, Mbozi, Lupa, Upangwa, and Nyika (McConnell, 1950;

Sutton et al., 1954; Daly, 1988; Lenoir et al., 1994; Theunissen

et al., 1996; Boniface et al., 2009; Fig. 2). Greenschist to gran-

ulite facies igneous and sedimentary rocks are the dominant

lithologies and have been intruded by voluminous Paleopro-

terozoic granitoids (Cahen et al., 1984; Lenoir et al., 1994).

Each Ubendian terrane is bound by steeply dipping shear

led to the current tectonic model invoking lateral accretion of

Ubendian terranes along the western margin of the Tanzanian

craton (Daly, 1988; Lenoir et al., 1994). Lateral accretion in the

Ubendian belt is suspected to have occurred concomitant with

thrusting and accretion of the geologically equivalent Usagaran

Ubendian tectonism and metamorphism occurred dia-

and is overprinted by 1.9 to 1.8 Ga amphibolite facies meta-

morphism (Lenoir et al., 1994). Paleoproterozoic metamor-

phism in the Ubendian belt is overprinted by Meso- and

Neoproterozoic tectonothermal events that are correlated

to the Kibaran, Irumide, and Pan-African orogenic episodes,

respectively (Cahen et al., 1984; Boniface, 2009; Boniface

and Schenk, 2012; Boniface et al., 2012). The current study

Local Geology, Lupa Terrane

Lupa terrane rocks

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/108/7/1615/3469630/1615-1640.pdf

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/108/7/1615/3469630/1615-1640.pdf

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/108/7/1615/3469630/1615-1640.pdf

All images are section views

d looking SE e looking SE f looking E

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/108/7/1615/3469630/1615-1640.pdf

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/108/7/1615/3469630/1615-1640.pdf

All images are section view

d looking W e looking W f looking W

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/108/7/1615/3469630/1615-1640.pdf

All images are plan view

quartz brittle fractures

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/108/7/1615/3469630/1615-1640.pdf

mylonitic foliation artisanal shaft artisanal working / exploration trench