See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/223509882

Rock quality, durability and service life prediction of armourstone

Article  in  Engineering Geology · September 2006

DOI: 10.1016/j.enggeo.2006.06.004

CITATIONS READS
33 1,709

3 authors:

John-Paul Latham David Lienhart

Imperial College London DAL Engineering Geologic & Petrographic Services
143 PUBLICATIONS   3,428 CITATIONS    43 PUBLICATIONS   110 CITATIONS   

SEE PROFILE SEE PROFILE

Sébastien Dupray
Centre d’études et d’expertise sur les risques, l’environnement, la mobilité et l’amén…
26 PUBLICATIONS   165 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Numerical and experimental investigations of particle stress and fracture for complex-shaped pellets View project

Sandspit Boat Harbor and Marina, Queen Charlotte Islands, British Columbia View project

All content following this page was uploaded by David Lienhart on 27 May 2022.

The user has requested enhancement of the downloaded file.

 Engineering Geology 87 (2006) 122 – 140
www.elsevier.com/locate/enggeo

Rock quality, durability and service life prediction of armourstone

John-Paul Latham a,⁎, David Lienhart b , Sebastien Dupray c
a
Department of Earth Science and Engineering, Imperial College London, London SW7 2AZ, UK
b
DAL Engineering Geologic and Petrographic Services, 7229 Longfield drive, Cincinnati, OH 45243-2209, USA
c
CETE de Lyon-LRPC-Groupe Mécanique des roches, 69674 Bron Cedex 01, France
Received 20 April 2006; received in revised form 20 June 2006; accepted 22 June 2006

Abstract

Rock armourstone is unusual amongst geomaterials because of the extremely wide range of intrinsic rock strengths that often have to
be used and the range of end use conditions in which it must serve. The authors' experience has been distilled to formulate suitable
guidance as demanded by practicing coastal and hydraulic engineers for the forthcoming update to the 1991 Rock Manual (CIRIA, CUR,
CETMEF. in press. The Rock Manual. The use of rock in hydraulic engineering (second edition). C683, CIRIA, London.). The
armourstone evaluation theme of earlier work by Lienhart, and abrasion testing and degradation modelling of armourstone by Latham are
revisited. A general in-service degradation model for wear of armourstone, based on a progressive fractional mass loss with time, is
redeveloped. The innovation is to introduce two new methods that are both practical and simple to implement. One method uses abrasion
resistance, water absorption and block integrity in a system that couples the rock susceptibilities to various common site conditions. The
other treats intrinsic rock properties from the quarry all together to achieve a global armourstone quality designation, AQD, of the rock,
that is independent of the actual conditions of the proposed site of application. Depending on the assessment of site aggressiveness, the
degradation with time on site is generated by retarding or accelerating the standard mass loss plot invoked from the AQD value, but
adjusted by semi-qualitative engineering and/or scientific judgement of the rating geologist. Examples are used to illustrate and compare
the two methods. Discussion is included on the need for more rigorous future research approaches based on breakage prediction and on
the need for further case histories.
© 2006 Elsevier B.V. All rights reserved.

Keywords: Armourstone; Riprap; Breakwater; Degradation model; Performance prediction; Laboratory evaluation; Quarry evaluation; Abrasion
resistance; Geomaterials

1. Engineering and scientific context
Many engineers with the task of specifying geological
construction materials (geomaterials) revert to precedent
⁎ Corresponding author. Tel.: +44 20 7594 7327.
E-mail addresses: j.p.latham@imperial.ac.uk (J.-P. Latham),
dlienhart@armorstone.com (D. Lienhart),
sebastien.dupray@equipement.gouv.fr (S. Dupray).
0013-7952/$ - see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.enggeo.2006.06.004
practice and standard specifications for given end uses
when preparing their tender documents. To specify more
optimally, e.g. using a whole life cycle analysis, what is
really sought is the prediction of the performance of the
geomaterial in the civil engineering structure. In this
respect, rock used in coastal structures, termed armour-
stone, presents a somewhat unique, perhaps greater
challenge than for geomaterials use in bridges, highways,
buildings and other infrastructural projects. The key
 J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140
difference is the larger variety of potentially acceptable
rocks and the wide range in the possible severity of
loading forces, such as waves or abrasive agents.
In some more ideal locations of the world where
coastal engineering projects are planned, quarries have
been or can be opened in favourable geological settings.
Palaeozoic age rocks within rock masses stripped by
glaciation of their weathered overburdens are likely to
yield excellent quality armourstone that will be in great
demand for coast protection. Designers are generally
confident that when this armourstone is placed in the
front line of direct resistance of coastal structures and
sea defences to take the annual battering from storms
and weather cycles, it will faithfully perform its tradi-
tional design function for a century or more.
Taking a more global view, the normal situation is
very different. The only quarries available to service
coastal engineers' designs often lie within the weaker
rock formations of Mesozoic and Tertiary age that gen-
erally are found along coastlines of most wave-active
seas. Furthermore, construction projects may be re-
quired in potentially very stormy and aggressive cli-
matic regimes. Designers working in these parts of the
world are acutely aware of the high costs of materials
transport and the many advantages of using local mate-
rials. They therefore often seek advice from engineering
geologists to help decide on suitability of the alternative
potentially available rock sources. It is common practice
for designers to progress rapidly to develop a site-spe-
cific materials specification, but without the assistance
of an engineering geologist skilled in armourstone eval-
uation. In these cases only minor thought is given to
optimization of rock quality. A client commissioning a
coastal structure, whether it is a breakwater to provide
sheltered harbour waters, a rock revetment protecting
several kilometres of seawall, or a series of rock groynes
trapping valuable beach sand and shingle at a holiday
resort, is quite entitled to ask: when will my hydraulic
structure require maintenance work and when will it
become unserviceable if we perform no maintenance?
To answer these questions, the design team needs access
to a range of expertise and tools to help predict the likely
consequences of using the most readily available quarry.
Matching a design to the unique properties of a certain
quarry's armourstone, at a site where the rock blocks are
required to retain their original design mass and inter-
locking properties for the design life brings engineering
and geology together. It requires an integrated approach
that begins with a detailed materials evaluation. Any
geologist will be quick to point out that a detailed
geological investigation of the source rock is essential
because the material is not an off-the-shelf building
123
product and is often inherently heterogeneous. Further-
more, predicting full-scale performance is difficult be-
cause full-scale block testing over long i.e. service life
test periods is not an option.
The field and laboratory examination of an existing
quarry or a new rock source dedicated to a construction
project has a block size objective and a rock quality
objective. These are:
1. to establish that the required tonnages of the necessary
sizes can be produced and delivered at the desired rate.
This subject is explored in Latham et al. (2006a,b)
2. to make a prediction of the service life of the rock
with the aid of field observations and or rock quality
test results. This is the subject of the present paper.
In general, sources that yield large blocks will have
satisfactory physical and weathering resistance proper-
ties, but this trend is of limited practical value as it is not
always the case and gives no indication of the degree of
longevity. With such a range of possible rock types and
rock strengths, each with inherent variability and deg-
radation susceptibilities, an evaluation of rock quality
using laboratory and field tests is always necessary.
Once both of these objectives have been achieved, it is
possible to fully consider alternative designs and only
then is it most appropriate to write a material speci-
fication. It is common practice for designers to progress
rapidly to a materials specification with only minor
thought being given to optimisation of rock quality. The
incentives for making the extra effort needed with this
type of optimisation will be in proportion to the project
scale and risk. For the designer and contractor to make
optimum choices, it is important that the levels of in-
trinsic test results eventually set down in the materials
specification are compatible, i.e. in tune with, the deg-
radation conditions and the design intended on site. To
help practitioners achieve such a “tuning”, we summar-
ize our research findings in the form of degradation
models that attempt to predict the effect of degradation
at the structure. These take account of:
• key factors in the quarry
• key variables at the project site
Our interest in armourstone and riprap source eval-
uation stems from seminal papers by Fookes and Poole
(1981) and Lienhart and Stransky (1981) which empha-
sised the need for both field and laboratory evaluations.
Poole et al. (1983) laid the foundations for combining
intrinsic (i.e. geological) factors with site (i.e. site aggres-
siveness) factors to give an indication of deterioration
124 J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140
expectancy. Encouraged by designers who repeatedly
faced problems working with less than ideal sources of
armourstone, a stone degradation model for armourstone
in association with abrasion resistance testing was tenta-
tively proposed for the evaluation of service life (CIRIA/
CUR, 1991; Latham, 1991, 1992a, 1993). The work was
based on earlier research on abrasion testing (Latham and
Poole, 1988). This first degradation model was an early
attempt to couple specific intrinsic attributes such as
breakage resistance or water absorption, to particular site
conditions, such as wave energy or severity of freeze–
thaw action. Such coupled effects would then retard or
accelerate a typical baseline curve for mass loss history
suggested from abrasion resistance test results for the rock
type in question.
Inspired by Hudson's (1992) rock engineering systems
approach, and starting work in 1994, Lienhart sought first
to systematize all factors affecting intrinsic properties of
armourstone that could affect service life (Lienhart, 1998).
To do so, he devised both a rating system and a rock quality
index with which to compare suitability of alternative
armourstone quarries. His next contribution was to adapt
Latham's “site aggressiveness” calibration of the degrada-
tion model by stripping away any reference to intrinsic
factors relating to the quarry. This was necessary because
the full spectrum of intrinsic properties were already
incorporated within the rock quality index (see Lienhart et
al., 2002) and their influence should not be included twice.
The implication was that this service life prediction would
therefore give less bias towards laboratory abrasion
resistance test results. More recently, and with a definite
focus on US experience and test methods, Lienhart (2003)
presented an update of the systems approach and its
associated guide tables, showing the interplay between
intrinsic rock quality factors, site factors, design choices
and the role of specification. Especially useful was a new
calibration of the climatological effect on weathering
intensity. The important distinction between Lienhart's
and Latham's implementation of the model is Lienhart's
total separation of intrinsic properties from site degradation
accelerators when calculating service life outcomes.
A very different and more rigorous approach to predict
losses during loading and service life was proposed by
Dupray (2005) who introduced a new test for resistance to
major breakage. The test used is a newly devised full-scale
splitting test for block integrity (Dupray et al., 2003)
where an impactor is dropped onto blocks and the
cumulative mass curve of the sample of armourstone
pieces before the impacts is compared to that after any
fracture damage upon impact has occurred. Analysis of
the curves provides an objective measure of resistance to
major breakage of a consignment. His in-service breakage
model is applicable for those cases where the most do-
minant cause of deterioration can be identified as being
block breakages e.g. due to impact loads caused by
movement either in the construction phase or during
storms, rather than for example, by spalling and abrasion
of outer layers due to intensive climatic cycling effects.
Given that the energy level of impacts on site is known or
can sometimes be estimated relative to that of the impactor
collision in the splitting test, the model indicates how the
cumulative curve may be expected to transform as a
consequence of progressive impacts on site, provided full-
scale splitting test results are available for the armourstone
source in question. The field calibration for the model was
based on controlled repeated dumping events in a quarry
using a front wheeled loader and is further elaborated in
CIRIA/CUR/CETMEF (in press).
The general aim of the present paper is to stimulate
discussion of the problem of performance prediction in
geomaterials, using the example of armourstone. A further
aim is to provide additional coastal engineering back-
ground to the engineering geology guidance on stone
degradation prediction presented in CIRIA/CUR/CET-
MEF (in press). The guidance recommends two different
methods to implement a general degradation prediction
model that may be of interest to a broad group researching
and practicing in engineering geology. Through the pre-
sentation of tables, procedures and examples, we provide
sufficient detail to assess rock sources and facilitate a
service life prediction of rock armoured structures, making
reference to recognized European and American test
methods.
We begin by briefly discussing the concept of dur-
ability, modes of degradation in different parts of a
structure and the most reliable symptoms of a rock
source that will go on to exhibit poor durability. Follow-
ing the systems approach to armourstone evaluation,
introduced by Lienhart (1998, 2003), a system for as-
signing Armourstone Quality Designation (termed
AQD), in numerical terms, using quarry evaluation cri-
teria as well as laboratory test results is presented. The
principles common to all degradation models are then
introduced before full details of two methods to im-
plement a general empirical degradation model are pre-
sented. We conclude with a list of mitigating strategies
for the designer when low durability scenarios are im-
plicated by the degradation model predictions.
2. Armourstone durability and degradation
The durability of a material component or system is
defined here as its ability to continue performing
adequately in a specific working environment.
 J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140
It follows that durability may be quantified by the rate
of loss of performance in engineering time for the
working environment in question. Durability is therefore
a balance between the intrinsic resistance of the material
and the aggressiveness of the forces acting in service.
This means that poor quality materials exhibiting low
resistance to weathering may stand up well in mild
energy environments where there is low exposure to
climatic conditions and where degradation forces are
virtually non-existent. For example, degradation will be
exceptionally mild in the permanently submerged part of
the core of an impermeable breakwater, or in the deep toe
regions of the outer cover layer of armourstone i.e. below
the significant water velocities associated with wave
action. By contrast, tough igneous rocks will fatigue and
crack if frequently moved by storm action, e.g. on a berm
breakwater, or if used in a revetment for shoreline
protection, will wear down if persistently pounded by
shingle.
The durability of armourstone on a given project is
quantified by the rate of loss of material performance
(such as reduction of mean mass or degree of
interlock), attributable to rock property changes in
engineering time. This should be distinguished from
erosion damage to a breakwater caused by under-
design where a storm or series of storms deliver
impacts and significantly exceed design expectation.
Degradation processes can be broadly classified as
wear (embracing spalling or flaking or granular loss from
the stone surfaces by abrasion), fracture (rapidly from
block-to-block contact impacts or more slowly from
cyclic weathering effects, including major throughgoing
cracking, often but not always exploiting pre-existing
weaknesses) and pervasive disintegration (where a net-
work of numerous small cracks exploit a rock fabric with
poor resistance to weathering), see Fookes and Poole
(1981), Lutton and Erikson (1992). These depend on the
Fig. 1. Layout of a small breakwater showing armourstone in cover layer and under layer with quarry run as core, details based on Ostia breakwater, Italy.
125
position of armourstone in the structure and their relative
degree of protection by other blocks (see Fig. 1).
2.1. Significance of degradation within zones of a
typical breakwater structure
• In the armour layer, armourstone may be exposed to
weathering such as freeze and thaw, salt crystallisation,
wetting and drying, and temperature cycling when
above water. Dissolution, usually driven by acidic
water, can be significant in freshwater lakes but is
rarely significant below salted seawater. In perma-
nently submerged armour layers, effects within sea-
water are negligible and the low water velocities will
further reduce mechanical assistance of degradation
mechanisms. For coastal structures, the inter-tidal zone
is the most vulnerable to physical–chemical effects.
For structures designed to adjust dynamically, impacts
between armourstone pieces may lead to breakage,
attrition and generally greater wear. For these struc-
tures, the strength of blocks to resist the forces induced
by rocking or rolling is important. The main perfor-
mance requirement is to minimise any service life
decrease in the median block mass (M50). Prediction of
M50 reduction is therefore sought by “demand-led”
designers wishing to set quality requirements high
enough to exclude a certain mass loss occurring within
the design life, however far the armourstone may need
to be transported. “Supply-directed” designers wishing
to add degradation-compensating measures within the
design philosophy, having set their sights on locally
sourced materials that are not of high quality, also seek
it. The authors wish to emphasise that there are gene-
rally several combinations of adjustments such as
dimensioning the armourstone, selecting specific strata
within sources for specific properties and character-
istics that will provide basic physical protection, but
126 J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140
that the performance life becomes a difficult
prediction.
• In the underlayer, armourstone is exposed to weath-
ering effects to a lesser degree than in the armour
layer. The main concern is that degradation may lead
to breach of the geotechnical filter rules that safe-
guard against geotechnical failure of the mass itself.
• For core material, exposure to weathering agents is
less and the consequences likely to be less critical.
Dissolution may take place below water or be in-
duced by intermittent water circulation above water
level. This may lead to voids and decrease of the
bearing capacity. The solubility is very dependent on
the water chemistry as well as mineral content and
rock porosity. For example, most types of limestone
will not dissolve in the chemical environment of most
sea waters but will dissolve slowly in acidic fresh-
water lakes and rivers.
2.2. Indicators typical of poor armourstone durability
Whole symposia and workshops have been devoted
to this subject, for example: Magoon and Baird (1991),
McElroy and Lienhart (1993) and Latham (1992b). It is
not our purpose to look into the details of specific
mechanisms and test methods but more, to devise prac-
tical quantitative tools, building upon Lienhart's (1998)
analysis of the most pertinent factors noting recent
developments in the literature e.g. Topal and Acir
(2004). However, it is worth summarizing key findings
in the above references by stating that petrography, mass
density, porosity, mineral fabric strength, armourstone
production method and discontinuity pattern in the
quarry, can be expected to provide a good indication of
relative durability when comparing sources.
First, a word of warning about water absorption, WA,
which is often considered the best single indicator of in-
service durability. Armourstone with a low WA, less
than 0.5% to 1%, will display a good resistance to severe
cycling stresses such as salt crystallisation or freeze and
thaw. For significantly higher values of WA, appropriate
accelerated weathering tests should be performed to
evaluate concerns about resistance to weathering of the
rock. However, if taken in isolation, WA, can be mis-
leading since rocks with WA greater than 4% may have
a free-draining pore structure and thus perform well in
service for certain applications. A high proportion of
microporosity as a percentage of the absolute porosity is
detrimental. Rock that is porous but free draining is less
prone to breakdown by freeze–thaw action and salt
crystallisation. As an example, in the USA the Missis-
sippian-age Salem Limestone (Indiana Limestone) and
Berea Sandstone both have absorptions exceeding 6%.
Both of these sources have been used in breakwaters on
the Great Lakes where weathering intensities due to
freezing are extreme. The Salem Limestone in the Chi-
cago Breakwater is still performing satisfactorily after
80 years. The Berea Sandstone in the Cleveland Break-
water is still in good condition after almost 100 years.
Furthermore, rock with mass density lower than 2.3 t/m3
is often considered to have unacceptable durability for
normal use. In practice, when the circumstances force
the contractor to use rock of such low density and high
water absorption, the low resistance to weathering is
sometimes less of a problem than the low mineral fabric
resistance to breakages associated with such mineral
textures. The breakages become immediately apparent
during construction.
3. Source evaluation of armourstone quality
In a study of durability indicators for aggregates,
Fookes et al. (1988) devised two intrinsic composite
indices based on test results. Because of the different
loading environments for different aggregates uses, one
index was devised for dynamic applications, another for
pseudo-static applications. As armourstone indices, their
usefulness is limited as armourstone evaluation has only a
tenuous link with the standard tests applied to evaluation
of mineral aggregates and they lack the rigour of indices
composed of dimensionless parameters. However, the
index value was a powerful way of dividing the overall
quality of aggregates into unacceptable, marginal and
acceptable classes. Since the CIRIA/CUR (1991) manual
was published, the use of four indicative classes for given
attributes has proven popular amongst users. Table 1 is an
updated selection of factors and classification criteria
from CIRIA/CUR (1991), from Lienhart (1998) and from
recent work involving authors of the European standard
for armourstone EN 13383 Part 1 (BSI, 2002a).
3.1. Introduction to source evaluation and Table 1
Each test property is classified into bands based on
geological criteria and judgement indicative of excellent,
good, marginal or poor armourstone durability consider-
ing all conceivable rock types and quarry sources. Such a
classification is independent of design and site conditions,
and therefore cannot give categorical assurances for
suitability or service life outcomes. Each attribute (each
row in Table 1) has a direct or indirect consequence in
terms of resistance to a particular type of loading, even
though that type of loading may or may not actually be
present for the application and site. For the same site
 J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140
application, relative outcomes for different rock sources
can be inferred from the classes. In this relative sense, the
four classes given for each specific attribute or property
have the following tentative interpretation:
• Excellent — ideal. This material, with reference to
this specific attribute, can be used without any risk of
degradation with time over a typical design life.
• Good—better than average. In normal situations, no
specific attention need be paid to this attribute which
will generally not lead to any significant degradation,
although it may show progressive signs of degradation
over a typical design life in certain circumstances.
• Marginal—lower than average. Without specific at-
tention, the attribute may lead to significant degrada-
tion. It should be studied. If necessary, production,
construction or design should be adapted by using
appropriate blasting techniques, increased quality con-
trol or by over-sizing rock block sizes using appro-
priate prediction of the degradation, e.g. as outlined
later in the paper. This may be associated with short
periods where loss of performance is more severe, and
if they are cumulative, could shorten the performance
life of the project.
• Poor—much lower than average. If possible, the use of
the material should be avoided in areas where exposure
may exploit the attribute, maybe lead to rapid
degradation. If not, special attention should be paid,
as for ‘marginal' attributes. However, specific perfor-
mance evaluation of the structure (e.g. proportion of
fractured, sub-sized and rocking blocks and exces-
sively large voids as well as distorted overall profiles),
will generally be required, not just after storm damage,
and heavy maintenance may be necessary.
Used in one way, Table 1 provides the engineer with
a multi-purpose and generalised look-up table of in-
trinsic properties to compare the relative merits of alter-
nate supply sources. Used in another way, it can provide
the basic scoring template for an overall quality rating to
compare armourstone rock sources. This table provides
the input data for Lienhart's overall rating system to
evaluate all, or a subset, of attributes of an armourstone
source. Furthermore, the table provides data needed for
the application of a degradation model described later.
3.2. Armourstone quality designation, AQD, an overall
quality rating of an armourstone source
Table 2 shows an example of a completed quality
rating assessment worksheet using integer ratings with a
maximum value of 4 for Excellent and a minimum value
127
of 1 for Poor. The objective is to derive an overall rating
for the armourstone source. In this paper, this numerical
overall rating applicable to quarried rock is given the
name Armourstone Quality Designation — AQD (some-
times termed Overall Rock Quality Rating — ORQR, or
the Lienhart Quality Rating Index, LQRI).
The AQD measure is an attractive application of
Hudson's (1992) rock engineering systems approach to
developing a weighted multifactor index. It has advan-
tages over traditional multiple regression methods when
there is only a meagre set of case history records with
which to calibrate data on intrinsic quarry factors. Ideal-
ly, the profession needs access to a system of perfor-
mance indicators for different armourstone candidates
for a range of design and site conditions, which is rarely
available for armourstone.
In this discussion it is of interest to digress to the
means of performance evaluation of yet another geo-
material. The performance index for road surfacing ag-
gregates would be much more amenable to calibration
through the use of mass loss from test patches on iden-
tically trafficked road surfaces. In such cases, multiple
regression methods would normally be considered more
rigorous than formulae with weightings based on
Hudson's interaction matrices methods.
Returning to the worked example in Table 2, column
(d) provides six precise weighting values (58, 73 etc).
These values arise from Lienhart's (1998) detailed re-
search into factor significance following Hudson's ap-
proach. Designers working with experienced engineering
geologists may prefer to assign alternative weightings for
each of the separately listed attributes. For example,
keeping to four alternative weighting values for the attri-
bute importance as follows: essential attribute to
assess = more than 90; important = 80–90; may be use-
ful = 70–80; minor = 50–70, rounded weighting values
could be assigned. In this way, AQD values pertinent to a
specific application could be developed, but care should
be exercised when introducing weightings different to
those recommended i.e. customised weightings, espe-
cially if the AQD value is to be later used as the basis for
degradation modelling. For example, in application to
river engineering in fluvial systems at high altitude with
significant freezing, a higher weighting for freeze–thaw
resistance could be considered.
Many different implementations of Lienhart's systems
approach are possible depending on the ease of acquiring
data. For example, one based entirely on laboratory results
using six factors is presented in Lienhart (2003). Using
both quarry site and laboratory attributes, another des-
cribed in Lienhart (1998), was used to compare six quarries
with AQD of from 2.45 to 3.41. The determination of

Table 1
Guide to durability of armourstone from a quarry source using field and laboratory criteria
Criteria
Field evaluation Lithological
in the quarry classification
(Important Note: it
must not be assumed
that these rock types
have the results shown
in the columns below
where mentioned)
Regional in-situ stress
Weathering grade
Discontinuity analysis
(IBSD)
Groundwater condition
Production method
Rock block quality
Set-aside
Block integrity
(Visual)
128
Reference Durability guide (not intended for specification purposes)
Excellent Good Marginal Poor
BS EN 932-3: 1997 Unfoliated igneous Crystalline dolomites, Argillaceous Shaly limestones, reef breccia,
and metamorphic rocks, crystalline limestone limestones, poorly shale, siltstone, slate, schist,
quartzites and high silica and moderately well cemented sandstones, chalk, gypsiferous carbonates
cement sandstones, compact cemented sandstones dolomite reef rock
crystalline limestones with void cavities
J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140
Lienhart (1998) Low stress, no folds Medium stress, High stress. Release Very high stress. Faults may be
no faults unloading features fractures parallel to present in quarry face. Rock
may be present face may be present bursts may be present in floor
BS EN 5930:1999a IA—fresh, IB—faintly weathered II—slightly weathered III—moderately weathered
un-weathered (staining on major (staining persists (less than half the rock
surfaces) through a greater mass is decomposed)
part of the rock mass)
Wang et al. (1990), Di80 > 2 m Di80 = 1.5–2.0 m Di80 = 1.0–1.5 m Di80 < 1.0 m
Latham et al. (2006a) (Di80 = 80% passing
in-situ block
sieve size)
Lienhart (1998) Dry Moist Seepage from Water flowing from walls and
quarry walls pooling on floor
Lienhart (1998) Non-blasting methods: e.g. Tailored single row Conventional blasting Aggregate blasting with large
dimension stone quarry blast, low shock with ANFO, specific stone as by-product, specific
production methods energy, specific charge charge 0.2–0.4 charge > 0.4 kg/m3
< 0.2 kg/m3, blast kg/m3, blast hole
hole diameter ∼75 mm diameter ∼ 100 mm
Lienhart (1998) < 5% of blocks have length 5–10% of blocks have 10–15% of blocks have > 15% of blocks have LT > 3,
to thickness ratio LT > 3, LT > 3, 95% of blocks are LT > 3, 95% of blocks 95% of blocks are at least
95% of blocks are weathering weathering grade IB or are at least weathering weathering grade III,
grade IA, free of unfilled cavities better, dense or free grade II, either micro- argillaceous or micaceous
and are extremely high strength draining, very high porous or with unfilled
strength cavities, high strength
Lienhart (1998) Quarried rock is stockpiled for Quarried rock is Quarried rock Freshly quarried rock is
3 months for curing and release stockpiled for is stockpiled for transported directly to project
of stress 2 months 1 month site for placement
Lienhart (1998) > 95% of blocks are free of 90–95% of blocks are 85–90% of blocks are < 85% of blocks are
incipient fractures, flaws or fracture free after fracture-free after 2 fracture-free after 2 months
cracks due to stress relief, rough 2 months set-aside months set-aside set-aside in stockpile
handling, overblasting or other in stockpile in stockpile
causes after 2 months set-aside
 in the stockpile
Block Integrity CIRIA/CUR (1991) IM50 < 2% IM50 = 2–5% IM50 = 5–15%, IM50 > 15%
(drop test)
Block Integrity Dupray NOTE : FSST is a design tool to
(FSST) (2005) assess the effect of the actual
integrity of a source. As such,
classification of CFSST values is
not appropriate
Block Integrity Tourenq Ic > 80 Df < 20 Ic = 80–70 Df < 20 Ic = 70–60 Df < 20 Ic < 60 Df > 20
(by sonic velocity et al. (1971)
indices: Ic and Df)⁎⁎⁎
Laboratory tests Petrographic Trained ⁎ ⁎ ⁎ ⁎
evaluation petrographer
Mass density (t/m3) BS EN >2.9 2.6–2.9 2.5–2.6 <2.5

J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140

13383-2:2002
Water absorption (%) BS EN <0.5 0.5–2.0 2.0–6.0 >6.0
13383-2:2002
Microporosity / total Lienhart <2 2 to 6 6 to 20 >20
porosity (%) (2003)
Methylene blue Verhoef (1992) <0.4 0.4–0.7 0.7–1.0 >1.0
adsorption (g/100g) BS EN 933-9:1999c
Compressive strength BS EN >120 60–120 20–60 <20
(MPa) 1926:1999b
Schmidt impact index ISRM >60 50–60 40–50 <40
(% rebound) (1978a)
Sonic velocity (km/s) BS EN >6 4.5–6 3–4.5 <3
14579:2004
Point load strength (MPa) ISRM (1985) >8 4–8 1.5–4 <1.5
Fracture toughness ISRM (1988) >1.7 1.0–1.7 0.6–1.0 <0.6
(MPa m1 / 2)
Indirect tensile ASTM >10 5–10 2–5 <2
(Brazilian) strength (MPa) D3967-95a
(2004) ISRM (1978b)
Los Angeles abrasion (%) BS EN 1097-2:1998a <15 15–25 25–35 >35
Micro-Deval (% loss) BS EN 1097-1:1996 <10 10–20 20–30 >30
MgSO4 soundness BS EN 1367:1998b <2 2–10 10–30 >30
(% loss)
Freeze–thaw loss (% loss) BS EN 13383-2:2002 <0.5 0.5–1 1.0–2 >2
Sonic velocity reduced by <5 5–15 15–30 >30
freeze thaw (% change) ⁎⁎
Wet–dry (% loss) ASTM D5313-04:2002 <0.5 0.5–1 1.0–2 >2
⁎) Judgemental; no established criteria exist. ⁎⁎) provisional criteria needing confirmation from further research. ⁎⁎⁎) if terms and symbols such as Ic (index of continuity) are unfamiliar, explanations
can be found in the references given and in CIRIA/CUR/CETMEF (in press).
Explanatory note: The guidance offered in this table is intended for a detailed semi-quantitative general durability evaluation with an option to determine the AQD value of a source using the system in
Table 2. It is therefore important to provide an appropriately sensitive rating, four categories being considered reasonable.

129
130 J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140

Table 2
Example of a completed quality rating assessment worksheet (after Lienhart, 1998)
(a) (b) (c) (d) (e)
Criteria Quality rating Rating Weighting Weighted rating
value
Excellent Good Marginal Poor
(=4) (=3) (= 2) (=1) Average % {(c) × (d)} / mean of (d)
Lithological classification ✓ 3 58 2.12
Regional in-situ stress ✓ 2 73 1.78
Weathering grade ✓ 3 73 2.67
Discontinuity analysis ✓ 3 95 3.48
Groundwater condition ✓ 2 73 1.78
Production method ✓ 1 95 1.16
Rock block quality ✓ 2 80 1.95
Set-aside ✓ 3 73 2.67
Petrographic evaluation ✓ 2 95 2.32
1 Block integrity test ✓ 1.5 90 1.65
Block Integrity visual ✓
2 Mass density ✓ 3 80 2.93
Water absorption ✓
Microporosity / total porosity ✓
Methylene blue absorption ✓
3 Compressive strength ✓ 1.67 88 1.79
Schmidt impact index ✓
Sonic velocity ✓
4 Point Load strength ✓ 2.67 88 2.87
Fracture toughness ✓
L A Abrasion ✓
5 Micro-Deval ✓ 2 88 2.15
6 Freeze–thaw loss ✓ 3.67 80 3.58
MgSO4 soundness ✓
Wet–dry loss ✓
Sum 1229 34.9
n 15 15
Mean 81.9 2.33
Note 1: This sheet includes 15 factors (9 field, 6 laboratory) hence overall rating or Armourstone Quality Designation, AQD is mean of column (e)
based on all 15 factors. If no data are available for one or more factors, AQD should be based on the number of included factors. A complete and
balanced set of data is ideal.
Note 2: In addition to engineering geology indicators, each boxed grouping of tests 1 to 6, generates one average rating value in column (c) from one
or more suggested tests. They refer to 1: resistance to Major Breakage, 2: mineral fabric physical quality, 3: resistance to Minor Breakage
(compressive), 4: resistance to Minor Breakage (tensile, dynamic), 5: resistance to wear (shear and attrition), 6: resistance to in-service weathering.
Note 3: test results and field assessments can be used to generate continuously varying ratings from 0.5 to 4.5 rather than using integer values.
Similarly AQD results can vary from 0.5 to 4.5.

AQD is recommended for comparing the overall intrinsic
quality of a source of armourstone that will be indicative of
durability given similar site applications. Needless to say,
the user who understands the bases of scientific quality
criteria, the principles of generating an importance
weighting, and the use of parameter rating systems for
rock engineering may use this approach to derive AQD
and adapt it to good effect in much the same way rock
engineers have customized Bieniewski's (1984) rock mass
rating system, once the principles were understood. To
suggest one simple improvement; substituting continu-
ously varying ratings instead of integers within each
category would allow greater accuracy and for the range of
AQD to extend below a relative value of 1 for a poor
source and above 4 for an excellent source. In fact, in a
comparison study of quarries used for the Great Lakes
shore protection projects, (Lienhart's consultancy reports
from December 2004 to April 2005) utilised a rating of 5
for a Superior quality stone with a unit weight exceeding
3.3 t/m3 and other property values far exceeding the
requirements for an “Excellent” rating.
Prediction of the future performance response of
armourstone with given characteristics in the quarry and
under the wide range of possible life-time loadings, con-
tinues to be a subject in need of further research. Several
tables and modelling tools that the authors' research and
 J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140
experience suggests may be useful in the process of eva-
luating the site aggressiveness and the service life of the
armourstone are introduced below.
4. Degradation modelling
4.1. Principles
In simple terms, a degradation model is the application
of mechanical principles with the following format:
• material properties + loadings (and boundary condi-
tions) = deformation or damage response or:
• evolution of material properties + history of loadings
applied = history of damage response or:
• average material properties + average loading inten-
sity = average rate of degradation.
A typical application of a degradation model uses
armourstone properties representative of the rock
consignment at the point of leaving the quarry. This
may be measured by a specific material property such as
block integrity or abrasion resistance, or an overall
quality index such as AQD as derived above.
The model then predicts the response to future loading
intensity of the rock armour with such properties. These
may be short-term loads or long-term “in service” loads.
The model output gives the change in the performance
parameter (such as M50, or the complete mass distribution)
for any number of handling events or storm/flood events,
or alternatively, for any number of years in service e.g. up
to and including the design life of the structure.
In a study of concrete armour block integrity, Bur-
charth et al. (2000) considered in detail the static and
dynamic loads affecting armour layers, strictly focusing
on the mechanics of breakage. For a more general treat-
ment of rock armour degradation, these would need to be
combined with other deterioration mechanisms. Latham
(1991), referred to an average loading intensity as project
site aggressiveness which can be subdivided into:
• attrition loading intensity: a function of waterborne
attrition agents, rocking, sliding and rolling loads
(affected by stone size, wave energy, mobility allowed
by the design wave condition, degree of interlock due
to shape and grading);
• breakage loading intensity: a function of rocking and
rolling loads (affected by stone size, wave energy,
mobility in design, interlock due to shape and grading)
• physical–chemical climatic loading intensity: a
function of zone on structure (Fig. 1), meteorological
climate, slope angle.
131
For static armour designs, (i.e. those characterised by
smaller ratios of significant wave height to stone mass)
the mass loss from blocks of armour is by both catas-
trophic cracking and by imperceptibly slow or sub-
critical opening and propagation of cracks, spalling and
rounding aided by physical–chemical climatic effects
and accompanied by accelerated loss of interlock from
wear. A model with a comprehensively assessed index
of intrinsic properties might therefore be considered
most appropriate in such cases.
For a dynamic design, (i.e. larger ratios of significant
wave height to stone mass, as in traditional berm-type
breakwaters) attrition intensity and breakage level
loading intensity will be considerably higher than cli-
matic loading intensity. A breakage model calibrated
using block integrity and mineral fabric strength, ideally
combined with one for abrasion including resistance to
wear properties, would be more suitable.
Degradation models, focusing specifically on wear
mechanisms (Tomasicchio et al., 2003) and on breakage
mechanisms (Tørum and Krogh, 2000; Dupray et al.,
2003) have been proposed for potentially dynamic
berm-type breakwaters. These more sophisticated mod-
els consider progressive mass reduction of individual
rock blocks associated with repeated storm events. The
storm loading taken into account is usually set to be that
which exceeds a threshold level of energy for initiation
of damage, or when armour movement velocity is above
a threshold value, accumulated over the life-time of the
structure. Such models attempt to deal with mass loss by
specific wear or breakage mechanisms that ignore climatic
weathering intensity effects. The fewer the degradation
mechanisms considered in a model, the more simply and
accurately the physics can be captured by measurable
parameters and the more rigorous the model calibration
approaches can be, but the less widely applicable is the
model to long-term service life prediction.
It must be remembered that damage to a structure is not
just caused by breakdown of the construction materials.
Wave action itself is most critical in causing structural
damage to the protected engineered works, as caused by
break-up displacements of the armourstone. Performance
and damage analysis methods (Kobayashi et al., 2003;
Takahashi et al., 2003) have been proposed to assess
accumulated structural damage caused by block displace-
ments. In this hydraulic sense, the accumulated damage is
calculated in terms of number of blocks missing from the
eroded area of the armour layer profile resulting from the
predicted storm and history of storms. A summation of the
structural damage predicted by stability equations, (e.g.
Van der Meer, 1988), together with the statistics for the
probability of exceedance of the design condition, are
132 J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140
used to build a model for lifetime armour layer damage
accumulation. It is now becoming more widely recog-
nised (e.g. Takahashi et al., 2003) that such whole-life
design approaches, based on hydraulic storm erosion
damage by block movements, also require an estimate of
the reduction in M50 of the armourstone due to rock
material degradation with time. The general in-service
degradation model described below may be tentatively
applied to estimate changes in M50 for such purposes.
Little research has been done applying practical methods
of measuring current representative values of M50 on a
structure, that could reveal the rate of degradation in-
service since construction. Short of automated image
analysis sizing methods using photographic surveys of the
same sample region, direct mass measurement by remo-
ving rocks for weighing and replacing them, or measure-
ment by assessing dimensions and shape factors to deduce
masses, seem to be the main approaches, all having
disadvantages.
4.2. General in-service degradation model for wear of
armourstone
A brief introduction to the model and possible sug-
gested improvements based on the work of Lienhart
(1998) since its first presentation (Latham, 1991) are given
here before presenting the revised proposed methods.
The general model for wear is designed to introduce a
systematic approach to the forecasting of progressive
degradation of rock with time. It is however based on the
notional assumption that, over the long-term, deteriora-
tion of armourstone during service life follows the same
progressive form as would the mass loss of pieces of rock
aggregate during a laboratory wear test. In practice, mass
loss locally or even over a large part of a structure may be
temporally episodic and relatively rapid (e.g. bursts by
freeze–thaw, salt crystallisation, or oxidation of pyrite etc)
following long dormant incubation periods. Deterioration
measured by mass loss may also begin to accelerate after a
critical number of years, whereas the laboratory degrada-
tion shows a continuously decelerating or steady degra-
dation rate trend. Engineering experience and judgement,
such as may be obtained following an examination of
degradation and block-rounding processes near the pro-
posed site (e.g. see Fookes and Thomas, 1986), should
always be taken as a primary indicator of service life
performance. It should be stressed that the models are no
substitute for the expertise of a qualified and experienced
engineering geologist's judgement, they are more applic-
able as an aid to engineering judgement.
With these limitations and provisos, the user is then in a
position to incorporate these predictions into the design.
The original model (Latham, 1991) has been considered
widely in projects in Cartagena in Colombia (Assen,
2000), Brindisi Italy (Tomasicchio et al., 2003), western
Canada (Lienhart et al., 2002; Lienhart, 2003), Iceland
(Tørum, 2003), and the Middle East, to help specify stone
quality requirements or to improve understanding of
maintenance needs in these coastal structures.
4.3. Mill abrasion tests and mass loss versus time plots
The original model (Latham, 1991) was based on test
results using a wet laboratory mill abrasion test with
constant abrasive environment applied, the result of which
was presented as a plot of fractional mass remaining M /M0
versus laboratory mill time t measured in units of one
thousand abrasion revolutions. The test method required a
duration sufficiently long to obtain the long-term beha-
viour. Once the shape effects of asperities, corners and
edges were largely removed by wear, the gradient govern-
ing the smoothing process generates a linear extendable
index called the abrasion resistance index, ks, measured as
a fraction of mass loss per 1000 revolutions. From this
index, the whole mass loss with revolutions plot can be
approximated by Eq. (2) (see later discussion).
The micro-Deval test, termed here the MDE test, is
another mill test for wear resistance by a wet surface
grinding action (EN 1097-1:1996). The test runs for a
fixed number of revolutions. The test value, termed the
micro-Deval value and written MDE, is highly correlated
with ks. This test is now the standard wear test for
aggregates and is specified for resistance to wear in the
armourstone standard EN 13383. Eq. (1), which is valid
for MDE < 70 (Latham, 1998), is used to relate the
abrasion resistance index, ks, and the micro-Deval test
value, MDE. Test results needed to apply the model are
therefore now widely available for rock types typically
used as armourstone.
The general in-service degradation model can be
implemented in two ways, the MDE method (essentially
very similar to the 1991 model) and the new AQD method
following Lienhart's development. Both assume the long-
term pattern of mass loss in service to be progressive and
of the same form as a steady abrasion test which as
pointed out above is a questionable approximation.
4.4. Micro-Deval (MDE) method
The MDE method used in the model appears at first to
be only relevant to armourstone that degrades by attrition
and surface grinding mechanisms. To extend the model to
other modes of degradation, the practical and research
experience of Latham and Lienhart was employed by the
 J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140 133

authors as a means of semi-quantitatively scoring the 4.4.3. Step 3

strength of interactions between factors perceived to be
most important for all the principal degradation mechan-
isms including spalling and fracturing. For further
explanation of the coupled terms X4, X6 and X9 where
aggressive site factors preferentially weaken armourstone
that has certain intrinsic properties, see Latham (1991).
4.4.1. Step 1
Use Eq. (1) to convert micro-Deval MDE test results to
ks, or use previously published data for the abrasion
resistance index, ks, which describes the intrinsic resist-
ance to mass loss by abrasion.
ks ¼ 4:12d10−5 MDE
1:485
4.4.2. Step 2
Plot fractional mass remaining M / M0 versus time t
(in units of 103 mill revolutions) using Eq. (2) and
comzpare results with the abrasion resistance classifica-
tion chart, Fig. 2:
M =M0 ¼ 0:05exp ð−30ks d tÞ þ 0:95exp ð−ks d tÞ
where M = mass remaining after time, t, M0 = initial mass.
It has been suggested through the work of Dupray
(2005) that this double exponential decay plot may be
made more representative of the mass loss profile of
typical armourstone shapes if the pre-factor 30 in Eq. (2),
governing the initial fast decay, is raised to about 200.
To determine the site aggressiveness, calculate X, the
‘equivalent wear time factor’ using Eq. (3):
X ¼ X1 dX2 dX3 dX4 dX5 dX6 dX7 dX8 dX9 ð3Þ
where X1, X2, …X9 are a series of site aggressiveness
loading parameters obtained from Table 3.
We wish to point out the slight revisions to the former
ratings for the site loading parameters given in Latham
(1991) and CIRIA/CUR (1991). Differences are introduced
in the light of further research and case history experience.
To obtain X6, we now substitute the project site weathering
intensity factor, MCWI, given in Table 4, based on work of
ð1Þ
Lienhart (2003). We also substitute water absorption test
results to couple the weathering susceptibility with intense
weather. The drop test breakage index IM50, is given as the
test method to quantify the block integrity. If there is a high
wave energy or mobility in the design, low integrity results
are made to have a bigger impact on degradation rates than
high integrity results as shown by the integrity coupling in
ð2Þ the wave energy term (X4) and in the mobility in the design
term (X9). The accuracy of the suggested ratings for these
coupling effects is uncertain and could be low. This is
where research by interested parties would be a helpful
advancement in the state of the art. Recent work on
breakages (Dupray, 2005), extended to consider wave
action loads (see Tørum and Krogh, 2000; Trmal, 2004)
has begun to tackle such problems.

Fig. 2. Abrasion resistance classification chart based on MDE results or their equivalent based on ks results. MDE values have been converted using
Eqs. (1) and (2) to mass loss plots indicative of the equivalent mill abrasion test results. The original quality boundaries for abrasion resistance are
shown in unbroken lines. Alternative revised boundaries compatible with the more demanding MDE European standard specification categories
(MDE = 10, 20 and 30), (BSI, 2002a) are also shown in broken lines.
134 J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140

Table 3
Ratings estimates for parameters in armourstone degradation model, (revised from Latham, 1991)
Parameter Rating estimates Parameter influence Calibration
Xmax / Xmin reliability⁎
ks Rock fabric strength ∼ 500 Excellent
Use MDE test and relation: ks = 4.12 × 10–5 MDE 1.485
or AQD value and relation: ks = 0.032 AQD–2.0
X1 Size
Effect given by 0.5(M50)1 / 3 (M50 in tonnes) ∼ 10 Good
M50 15.0 8.0 1 0.1 0.01
Rating 1.23 1.00 0.50 0.23 0.11
X2 Grading ∼ 2.5 Fair
(M85 / 1.1–1.4 1.5–2.4 2.5–2.4
M15)1 / 3
Rating 1.2 1.0 0.5
X3 Initial shape ∼2 Fair
Angular/ Blocky/equant Semi- Rounded
irregular rounded
Rating 1.0 1.1 1.5 2.0
X4 Incident wave or current energy (treat as independent of size of stone) ∼ 10 Fair
Significant wave >8.0 4.0–8.0 <4.0
height, Hs (m)
Rating If IM50 > 15% 0.3 1.0 2.0
If IM50 = 5.0–15.0% 0.5 1.3 2.3
If IM50 = 2.0–5.0% 0.7 1.6 2.6
If IM50<2% 1.0 2.0 3.0
Rating If using AQD 0.7 1.6 2.6
method
X5 Zone of structure ∼ 10 Good
Inter- Supra-tidal/ Supra-tidal/ Always
tidal hot temperate submerged
Rating 1.0 2.5 8 10
X6 Meteorological climate weathering intensity ∼7 Good
(Use MCWI Index of Lienhart, 2003—see Table 4)
MCWI index <100 100–300 300–600 >600
Rating If WA > 2.0% 0.8 0.6 0.4 0.2
If WA = 0.5–2.0% 1.0 0.8 0.6 0.4
If WA < 0.5% 1.4 1.2 1.0 0.8
Rating If using AQD method 1.0 0.8 0.6 0.4
X7 Water-borne ∼ 7.5 Poor
attrition agents
Sediment Shingle Gravel Sand Silt None
type
Rating 0.2 0.5 1.0 1.2 1.50
X8 Concentration of wave attack ∼2 Fair
Tidal range (m): <2.0 2.0–6.0 >6.0
Rating for slope angle 1.0 1.2 1.5
of 1:2.5 or steeper
Rating for slope angle 1.5 1.8 2.0
of 1:3.0 or gentler
X9 Mobility of armour in design concept ∼ 20 Fair
Hs / ΔDn50: 1–2.4 2.5–3.9 4–6.9 7–20
Rating If IM50 > 15% 1.5 0.6 0.3 0.1
If IM50 = 5.0–15.0% 2.0 1.0 0.5 0.2
If IM50 = 2.0–5.0% 2.0 1.5 1.0 0.5
If IM50 < 2% 2.0 1.8 1.6 1.4
Rating If using AQD method 1.5 1.3 1.1 0.7
⁎Calibration reliability used to set the ratings estimates is variable, ranging from a simple reasoning using qualitative field observations of factor
influence; (Poor), to extensive confirmatory data; (Excellent).
 J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140
Table 4
Meteorological Climate Weathering Intensity (MCWI) for rock
(Lienhart, 2003)
Rock weathering intensity analysis
a. Mean max − mean min temperature range
b. Mean annual temperature
c. Mean number of days max temp > freezing
d. Mean number of days max temp ≤ freezing
e. Extreme max and min temperature range
f. Mean number of days with precipitation > 0.25 mm
g. Annual precipitation, cm
h. Total normal degree − days, base 18 °C⁎
MCWI = (a / b) × (d / 365) × (e / c) × ((g / f) × h) Note: units are
degrees2 × cm / number of days2 as the second term 'd / 365' is
dimensionless (365 = no. days in a year) Care is required in
calculating MCWI owing to the unusual units of the various weather
parameters
⁎Explanatory note: Total normal degree − days, base 18 °C= sum of
heating degree − days and cooling degree− days, and is measured in
degrees. For each day where the outside temperature is such that a
building may require heating, the heating degree − day (in °C) = 18 minus
the mean outside temperature for a 24-hour period. For each day with
average temperatures above 18°C, the heating degree–day is recorded as
zero. The weather offices total these heating degree–day values for a
specific weather station for a year and provide the “normal” total heating
(in °C) for the average year. The cooling degree − day (in °C) = the mean
outside temperature for a 24 hour period minus 18°. The “normal” total
cooling (in °C) is reported like the heating degree − days.
4.4.4. Step 4
Finally, to determine the service life prediction,
convert the M / M0 curve in Step 2 by multiplying the
time scale (t in thousands of revolutions) by X to give
the correct time in terms of in-service engineering time
on the structure, T in years. Alternatively, plotting the
change in mass M / M0 against in-service engineering
time, T in years, using Eq. (4) will give the same result.
This plot gives the calculated estimate of the fraction of
original mass remaining for any period of service life or
design life in years.
M =M0 ¼ 0:05expð−30ks ðT =KÞÞ
þ 0:95expð−ks ðT =KÞÞ ð4Þ
4.5. Armourstone quality designation (AQD) method
The AQD method (suggested here using parameter
ratings given in Table 3) aims to scale the rate of loss of
performance using a site aggressiveness and design appli-
cation factor that is totally independent of the global
intrinsic property assessed by the AQD value. In this way,
it differs from the MDE method.
Calibration for the AQD was based on Canadian
work where a basalt-like pyroclastic material was in
place. A conservative approach in the projection of
135
potential service life was chosen. This same approach
was used for the evaluation of several quarries for pro-
jects on the Great Lakes and the Atlantic coast.
4.5.1. Step 1
Perform a systematic quarry evaluation with a com-
bination of field and laboratory assessments and obtain
the AQD value using Tables 1 and 2. Use Eq. (5) to
convert AQD to ks, where ks now describes the intrinsic
resistance to mass loss of the armourstone quarry source
in question. Consideration of all potential susceptibil-
ities to possible degradation mechanisms that may be
active in any structure over the long-term are repre-
sented by ks through AQD.
ks ¼ 0:032 AQD−2:0 ð5Þ
Eq. (5) was derived empirically by Lienhart as a simple
means to convert AQD values into plots with fraction of
original mass remaining versus time T in years on the
structure, assuming the equivalent wear time factor X = 1.
To derive the relationship in Eq. (5), AQD values of 1.5,
2.5 and 3.5 were set by curve fitting to correspond with ks
values previously used to separate resistance to progres-
sive degradation at three quality boundaries — bound-
aries that are essentially similar to those given by abrasion
resistance classes shown in Fig. 2. It may therefore be
inferred that the calibration in use that links intrinsic data
with service life has changed very little from the original
one based on the limited work reported in Latham (1991).
For the AQD method, at this point in the procedure, ks has
the units of change in fraction of mass remaining per year
simply because it is assumed the equivalent wear time
factor on the structure (X) is equal to 1.
4.5.2. Step 2
Use Eq. (4) to plot reference curves for intrinsic
armourstone performance with AQD = 0.5 to AQD = 4.5.
These are given in Fig. 3 where performance is plotted
in terms of mass fraction remaining against time T on
the structure in years, assuming the equivalent wear time
factor X = 1.
4.5.3. Step 3
Obtain X using Eq. (3) and Table 3, taking care with
X4, X6 and X9 to use ratings that specifically apply to the
AQD method (rather than the MDE method).
4.5.4. Step 4
Use Fig. 3 to interpolate for the actual AQD value and
change the time scale according to the actual X value from
136 J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140
Fig. 3. Intrinsic armourstone performance classification chart based on AQD values, assuming equivalent wear time factor, X = 1.
Step 3, or, plot the change in mass M / M0 against service
life time, T in years, using Eq. (4). This plot gives the
calculated estimate of the fraction of original mass re-
maining for any period of service life or design life.
4.6. Example comparison between MDE and AQD
methods
Both methods have been illustrated using two exam-
ples of data. To demonstrate the utility of the general
degradation model for estimating the predicted wear of
armourstone and to compare the two methods, four curves
are shown together in Fig. 4. Example 1 is largely theo-
retical and the calculation of AQD and X is explained in
detail (Tables 2 and 5) whereas Example 2 is an appli-
cation of data from Assen (2000). In Table 5, the column
for the MDE method indicates three test results. The MDE
test result of 20% is used with Eq. (1) to provide the ks
value of 0.00352 (boundary between Marginal and Good
in terms of abrasion resistance). The water absorption and
integrity drop test results are also required to more
accurately select the coupling of the different attributes of
the source rock with the site aggressiveness parameters
X4, X6 and X9, as identified in Table 4. All nine parameters
are scored according to Table 4 and the product X = 1.12 is
determined using Eq. (3). The column for the AQD
method takes a more in-depth assessment of the intrinsic
properties of the source material using the assessment
scheme in Tables 1 and 2 and in this example gives an
AQD = 2.33, (top of Marginal range) which from Eq. (5)
yields the ks value of 0.00589. The parameters in Table 4
this time yield the product X = 1.80, using Eq. (3). The
resulting service life predictions plotted using Eq. (4) and
shown in Fig. 4 are similar for both methods in Example 1.
They suggest that M50, originally of 8 tonnes, will have
fallen to 7.2 tonnes (90%) in about 20 years.
The remarkable similarity of the result in Example 1
is fortuitous. Example 2, uses laboratory test criteria
from Assen's (2000) field and laboratory investigation
of armourstone degradation in rock groynes constructed
with Quaternary coral reef and Miocene age limestones.
The (ks, X) values for the MDE and AQD methods
respectively were (0.0084, 0.91) and (0.0106, 1.77). The
notable differences in X values combined with similar ks
values suggest two slightly different responses. In
summary, they predict that a mass reduction of 30% in
the armour (initially 2000 kg) would occur within about
35–50 years. Field observations and measurements of
degradation from groynes of known ages dating back
some 25 years provide valuable data that are highly
variable and not conclusive for validation purposes.
However, they do suggest that degradation mass losses
have been at about this rate and significantly faster than
predicted in places. One possible explanation for the
faster average rates of degradation discussed by Assen
(2000) is that the intrinsic properties of the limestone
and coral armourstone reduce in quality during service
life. This may not be adequately accounted for in the
model's representation of progressive degradation.
The AQD method is attractive as it draws upon many
more intrinsic factors of relevance. The MDE method on
the other hand, is probably better calibrated as it was based
on case history calibrations where abrasion test results
were related to site case histories of degradation with time.
It is therefore suggested that in practice, results of both
methods be examined along the lines given in the
 J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140 137

Fig. 4. Service life predictions for the general in-service degradation model, using both MDE and AQD methods. EX 1 is a theoretical scenario based on
quarry evaluation and AQD estimation according to the data in Table 2 (as taken from the original paper by Lienhart, 1998) combined with a theoretical
example of a set of site variables as given in Table 5, with guidance from Tables 1, 3 and 4) to implement the models. The very close agreement between
methods shown in Example 1 (EX1), is somewhat fortuitous and not likely to be reproduced in practice. EX2 uses data extracted for illustrative purposes
from the study of degradation of armourstone on rock groynes built from Miocene age porous bioclastic limestone and coral reef limestone (Assen, 2000).
Assen's field observations of mass loss from dated groynes suggested that the degradation prediction model (she used an approach similar to the MDE
method) was over optimistic, one possible explanation being that while the rock was in service, the intrinsic fabric strength and quality of the rock
decreased below that assessed at the quarry. The authors would like to point out that the results reported by Assen, upon which the AQD for EX2 were
calculated, lead to higher AQD values than an experienced petrographer would expect for Miocene age porous bioclastic limestone and coral reef
limestone. Unrepresentative samples may have been used for testing. Furthermore, the AQD value estimated here from Assen’s study was based entirely
on laboratory results. The inclusion of ratings based on field criteria such as set aside and lithological classification, as well as petrographic evaluation
would have probably resulted in a significantly lower AQD than that deduced from the laboratory tests, giving a service life prediction more in line with
reported observations of degradation over 25 years.

examples above and illustrated in Fig. 4. It would be
advisable to perform a sensitivity analysis considering the
range of parameter values that may be applicable. At
present, there is too little case history data to give gui-
dance on which of the two methods is to be preferred.
As mentioned above, both methods assume the long-
term pattern of mass loss in service to be progressive and
of the same form as a steady laboratory abrasion test.
However, the authors agree that after a certain value that
might be anywhere between 10% and 50% mass loss, an
accelerated deterioration may be more realistic than the
assumed continuously slowing rate given by the form of
the abrasion test plot. Increasingly inaccurate predic-
tions are likely for M / M0 below say, 0.7.

Table 5
Application of armourstone degradation model for in-service mass loss to cover layers, see Example 1 in Fig. 4
Parameter Site loading information relating MDE method e.g. MDE = 20%, AQD method e.g. AQD = 2.33 based upon AQD for
2to degradation WA = 1.2%, IM50 = 12% quarry source data given in Table 2
ks 0.00352 0.00589
X1 M50 = 8 tonnes 1 1
X2 M85 / M15 = 1.3 1.2 1.2
X3 Angular irregular shape 1 1
X4 Significant wave height for design 1.3 1.6
storm Hs = 5.0 m
X5 Zone of concern is intertidal 1 1
X6 Meteorological climate weathering 0.4 0.4
intensity, MCWI = 700
X7 Attrition by silt 1.2 1.2
X8 Steep (1:2) slope with high (7 m) tidal range 1.5 1.5
X9 Mobility in design expressed by Hs / ?Dn50 = 3.0 1 1.3
X Equivalent wear time factor 1.12 1.80
138 J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140
5. Discussion
5.1. Design implications
If available, good service record observations can be a
guide for calibration of degradation models and validation
of predictions. Service records are valuable if you can
match the in-place stone properties with the properties of
the stone that will come out of the same source when you
construct the new project. However, Mother Nature is
rarely consistent in either the vertical or the horizontal
direction. Also the new project will not be subject to the
same exact service conditions as the old one.
Damage prediction modelling is more complex where
the armour is mobile during storms. Aggregated
distances travelled by storm-displaced blocks, possibly
producing severe degradation by corner breakage and
rounding, as well as by splitting, need to be considered.
Both integrity and abrasion resistance are therefore ex-
tremely important considerations for design input. For
dynamic structures such as berm breakwaters and other
novel low cost designs with expected movements, deg-
radation models have advanced sufficiently to warrant
application of calibrated block integrity testing (Full-
Scale Splitting Test results) and mineral fabric strength
tests (e.g. Point Load strength index and MDE) to assess
mass losses during block splitting and rolling. To
understand the more rigorous models applicable to
berm breakwater designs with mass losses, breakages
and rounding from armour layer reshaping during storm
events, key research papers by Tørum and Krogh (2000)
and Tomasicchio et al. (2003) should be studied. For
static designs, especially where marginal and poor rock
is to be considered, the general armourstone degradation
models that attempt a comprehensive average degrada-
tion assessment over the service life are recommended.
Needless to say, for the foreseeable future, armourstone
degradation models will remain in a state of research and
development as new data becomes available. They
should only be applied with considerable caution and
backed up by sensitivity analysis that will make clear the
degree of uncertainty in the service life predictions.
Sometimes, the degradation prognosis concluded from
the modelling will be severe and of concern to the
designer, notwithstanding the need to proceed with the
project, so mitigation strategies will be required.
5.2. Mitigation strategies for low durability scenarios
of rock armour
The greater the movement of armour and material
components in service that is expected in a chosen
design, the greater is the need for more highly durable
materials. Similarly, to reduce the rate of degradation,
low durability material should not be placed in a manner
susceptible to rocking movements.
Experience imparted to the authors from practicing
coastal engineers suggests that the decision to go ahead
with lower quality material for armouring and/or filter-
ing may be justified in the following circumstances:
• satisfactory performance records for the same
lithological type of rock in similar uses are available
• the assessment of the aggressiveness of the site con-
ditions is extremely mild
• high frequency of maintenance and repairs are con-
sidered acceptable because they can be done relative-
ly cheaply and with little disturbance
• the armouring has been deliberately over-dimen-
sioned in order to reduce mobility in storms
• the slopes have been flattened and greater material
volumes used to reduce mobility in storms
• the armour has been placed with a higher degree of
interlock to reduce mobility in storms
• there is no alternative: to find more durable materials
or to use concrete units is much too costly — a
relatively short design life is therefore acceptable
Some innovative low cost options, referred to in
Crossman et al. (2003), highlight the use of marginal
quality materials in addition to less conventional design
and construction practice. The innovation of grouting
to achieve the necessary stability in the cover layer is
also an option but it should be noted that the cemen-
titious or bituminous grout's reduction in binding per-
formance is more rapid when in sites affected by severe
climate.
6. Concluding remarks
It is important to stress that stone degradation pre-
diction is by its nature inexact and burdened with dif-
ficult scientific judgements. The variables involved and
the complexity of modelling such a process have been
reviewed for the purposes of providing the best guid-
ance for the forthcoming update to the Rock Manual
(CIRIA/CUR/CETMEF, in press). Systematic meth-
odologies for degradation prediction have been pre-
sented. Further validation and refinement of these
recently proposed degradation modelling approaches,
e.g. by back analysis of case histories, is required. The
tools described will reduce the amount of guesswork
that would otherwise be involved, especially in cases
where less than ideal rock has to be used.
 J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140
Acknowledgements
This paper extends the content of work presented in the
Rock Manual (CIRIA/CUR/CETMEF, in press) and is
printed with kind permission of CIRIA/CUR/CETMEF.
The authors are grateful for comments provided during
review and the motivation provided by the Rock Manual
team.
References
Assen, R.L., 2000. Durability assessment of armourstones along the
coast of Cartagena, Columbia. Memoirs of the centre of engineering
geology in the Netherlands, vol. 191. Delft, 84 pp.
American Society For Testing Materials (ASTM), 2002. Standard Test
Method for Evaluation of Durability of Rock for Erosion Control under
Wetting and Drying Conditions. D5313-04. ASTM, Philadelphia.
American Society For Testing Materials (ASTM), 2004. Standard Test
Method for Splitting Tensile Strength of Intact Rock Core Specimens.
D3967-95a. ASTM, Philadelphia.
Bieniawski, Z.T., 1984. Rock mechanics design in mining and tunnelling.
A.A. Balkema, Rotterdam, p. 272.
British Standards Institution (BSI), 1996. Tests for mechanical and
physical properties of aggregates. Determination of the resistance to
wear (micro-Deval) (AMD 14864) BS EN 1097-1. BSI, London.
British Standards Institution (BSI), 1998a. Tests for mechanical and
physical properties of aggregates. Methods for the determination of
resistance to fragmentation. BS EN 1097-2. BSI, London.
British Standards Institution (BSI), 1998b. Tests for thermal and
weathering properties of aggregates. Magnesium sulfate test. BS EN
1367-2. BSI, London.
British Standards Institution (BSI), 1999a. Code of Practice for site
investigation. BS 5930. BSI, London.
British Standards Institution (BSI), 1999b. Natural stone test methods.
Determination of compressive strength. BS EN 1926. BSI, London.
British Standards Institution (BSI), 1999c. Tests for geometrical
properties of aggregates. Assessment of fines, Methylene blue test.
BS EN 933-9. BSI, London.
British Standards Institution (BSI), 2002a. Armourstone-Part 1:
Specification. BS EN 13383-1. BSI, London.
British Standards Institution (BSI), 2002b. Armourstone — Part2: Test
Methods. BS EN 13383-2. BSI, London.
British Standards Institution (BSI), 2004. Natural stone test methods.
Determination of sound speed propagation. BS EN 14579. BSI, London.
Burcharth, H.F., d'Angremond, K., van der Meer, J.W., Liu, Z., 2000.
Empirical formula for breakage of Dolosse and Tetrapods. Coastal
Engineering 40 (3), 183–206.
CIRIA/CUR 1991. Manual on the use of rock in coastal and shoreline
engineering. CIRIA special publication 83, CUR report 154, 607 pp.
CIRIA, CUR, CETMEF, in press. The Rock Manual, second
edition. The use of rock in hydraulic engineering. CIRIA,
London, p. C683.
Crossman, M., Sequra-Dominquez, S., Allsop, N.W.H., 2003. Low
cost structures for beach control and coast protection: practical
design guidance. DEFRA/EA Technical Report, FD 2409.
Dupray, S., Latham, J.-P., Durville, J.L., 2003. Proposal of a full-scale
destructive test method to assess integrity of natural armourstone. In:
Melby, J. (Ed.), Proceeding of the Conference Coastal Structures.
ASCE, Portland, USA, pp. 150–162.
139
Dupray, S. 2005. Integrity of rock blocks for armouring of hydraulic
works — towards full-scale test to predict mass distribution changes,
PhD thesis, Imperial College Royal School of Mines, Earth Sciences
and Engineering Department, London, 325 pp.
Fookes, P.G., Thomas, R.D., 1986. Rapid site appraisal of potential
breakwater rock at Qeshm, Iran. Proceedings of the Institution of
Civil Engineers 80, 1297–1325.
Fookes, P.G., Poole, A., 1981. Some preliminary considerations on the
selection and durability of rock and concrete materials for break-
waters and coastal protection. Quarterly Journal of Engineering
Geology 14, 97–128.
Fookes, P.G., Gourley, C.S., Ohikere, C., 1988. Rock weathering in
engineering time. Quarterly Journal of Engineering Geology 21, 33–57.
Hudson, J.A., 1992. Rock Engineering, Theory and Practice. Ellis
Horwood, Chichester.
ISRM, 1978a. Suggested method for determining hardness and
abrasivity of rocks. International Journal of Rock Mechanics and
Mining Sciences & Geomechanics Abstracts 15, 89–97.
ISRM, 1978b. Suggested method for determining tensile strength of
rock materials. International Journal of Rock Mechanics and
Mining Sciences & Geomechanics Abstracts 15, 99–103.
ISRM, 1985. Commission on testing methods. Suggested method for
determining Point Load Strength (revised version). International
Journal of Rock Mechanics and Mining Sciences & Geomechanics
Abstracts 22, 51–60.
ISRM, 1988. Commission on testing methods. Suggested method for
determining the fracture toughness of rock. International Journal of
Rock Mechanics and Mining Sciences & Geomechanics Abstracts
25, 71–96.
Kobayashi, N., Zhao, H., Pozueta, B., Melby, J.A., 2003. Virtual
performance of rubble mound structures. In: Melby, J. (Ed.),
Proceeding of the Conference Coastal Structures. ASCE, Portland,
USA, pp. 1–13.
Latham, J.-P., 1991. Degradation model for rock armour in coastal
engineering. Quarterly Journal of Engineering Geology 24 (1),
101–118.
Latham, J.-P., 1992a. In-service durability evaluation of armourstone.
In: Magoon, O.T., Baird, W.F. (Eds.), Durability of Stone for
Rubble Mound Breakwaters. ASCE, pp. 6–18.
Latham, J-P. 1992b. Seminar on armourstone: Production, evaluation,
specification and use. Un-published, Queen Mary and Westfield
College–University London, 194 pp.
Latham, J.-P., 1993. A mill abrasion test for wear resistance of
armourstone. In: McElroy, C.H., Lienhart, D.A. (Eds.), Rock for
Erosion Control, ASTM Special Technical Publication, vol. 1177,
pp. 46–61.
Latham, J.-P., 1998. Assessment and specification of armourstone
quality — from CIRIA/CUR (1991) to CEN (2000). In: Latham, J.P.
(Ed.), Advances in aggregates and armourstone evaluation. Engi-
neering Geology Special Publication, vol. 13. Geological Society,
London, pp. 65–85.
Latham, J.-P., Poole, A.B., 1988. Abrasion testing and armourstone
degradation. Coastal Engineering 12, 233–255.
Latham, J-P., Van Meulen, J.A., Dupray, S., 2006a. Prediction of in-
situ block size distributions with reference to armourstone for
breakwaters. Engineering Geology 86, 18–36.
Latham, J-P., Van Meulen, J.A., Dupray, S., 2006b. Prediction of
fragmentation and yield curves with reference to armourstone
production. Engineering Geology 87, 60–74.
Lienhart, D.A., 1998. Rock engineering rating system for assessing the
suitability of armourstone sources. In: Latham, J.P. (Ed.),
Advances in aggregates and armourstone evaluation. Engineering
140 J.-P. Latham et al. / Engineering Geology 87 (2006) 122–140
Geology Special Publication, vol. 13. Geological Society, London,
pp. 91–106.
Lienhart, D.A., 2003. A systems approach to evaluation of riprap and
armour stone sources. Environmental and Engineering Geoscience
9 (2), 131–149 (May).
Lienhart, D.A., Stransky, T., 1981. Evaluation of potential sources of
riprap and armor stone-methods and considerations. Bulletin of the
International Association of Engineering Geology XVIII (3),
325–332.
Lienhart, D.A., Sayao, O.J., Gerdsen, A.W., 2002. Predicted service
life of armour stone: a case history. In: Magoon, O.D., Prehn, J.D.
(Eds.), Proceedings of breakwaters '99: First International
Conference on Monitoring of Breakwaters: American Society of
Civil Engineers, Reston, VA, pp. 145–159.
Lutton, R.J., Erikson, R.L., 1992. Problems with armour-stone quality on
Lakes Michigan, Huron, Erie. In: McElroy, C.H., Lienhart, D.A.
(Eds.), Rock for Erosion Control. ASTM Special Technical
Publication, vol. 1177.
Magoon, T., Baird, W., 1991. Durability of stone for rubble mound
breakwaters. ASCE, Cleveland, Ohio, p. 277.
McElroy, C.H., Lienhart, D.A., 1993. Rock for Erosion Control.
ASTM Special Technical Publication, vol. 1177. Philadelphia.
Poole, A.B., Fookes, P.G., Dibb, T.E., Hughes, D.W., 1983. Durability
of rock in breakwaters. Breakwaters-design and construction.
Thomas Telford Ltd, London, pp. 19–30.
Takahashi, S., Hanzawa, M., Sugiura, S., Shimosako, K., Van der
Meer, J., 2003. Performance design of maritime structures and its
application to armour stones and block of breakwaters. Proceeding
of the Conference Coastal Structures 2003. ASCE, pp. 14–26.
View publication stats
Tomasicchio, G.R., Lamberti, A., Archetti, R., 2003. Armour stone
abrasion due to displacements in sea storms. Journal of Waterway,
Port, Coastal and Ocean Engineering, ASCE 129 (5), 229–232
Sept/Oct.
Topal, A., Acir, O., 2004. Quality assessment of armourstone for a
rubble mound breakwater (Sinop, Turkey). Environmental Geol-
ogy, vol. 46, no. 6. Springer, Berlin Heidelberg, pp. 905–913.
Tørum, A. 2003. Abrasion mill test results in relation to stone
movements on berm breakwaters NTNU Norwegian University of
Science and Technology, Report No IBAT/MB R2/2003.
Tørum, A., Krogh, S., 2000. Berm breakwater-stone quality, STF22
A00207. SINTEF Civil and Environmental Engineering, Trondheim,
Norway.
Tourenq, C., Fourmaintraux, D., Denis, A., 1971. Wave propagation and
discontinuities in rock. Symposium of the International Society for
rock mechanics. Nancy, France.
Trmal, C. 2004. Study of mechanical loadings applied to natural
armourstone in cover layer of breakwater under wave action, MSc
Thesis, Faculty of Technology, Plymouth (UK), 88 pp.
Van der Meer, J.W., 1988. Deterministic and probabilistic design of
breakwater armour layers. Journal of Waterway, Port, Coastal, and
Ocean Engineering, ASCE 114 (1), 66–80.
Verhoef, P.N.W., 1992. The methylene blue test applied to geomater-
ials. GEOMAT, vol. 02. Delft. 85 pp.
Wang, H., Latham, J.-P., Poole, A.B., 1990. In-situ block size assessment
from discontinuity spacing data. Proceedings of the 6th Congress of
the IAEG, Amsterdam, August 6–10, pp. 117–127.