F. J. BaynesP. G. FookesJ. F. Kennedy

The total engineering geology approachapplied to railways in the Pilbara, WesternAustralia

Received: 28 June 2004Accepted: 13 September 2004Published online: 9 April 2005� Springer-Verlag 2005

Abstract The application of the totalengineering geology approach to theinvestigation, design, constructionand operation of several major ironore railways in the ancient cratonicPilbara region of Western Australiais described. The approach is basedon developing a thorough under-standing of the geological and geo-morphological history of the areaand adopting strategies such asstaged investigations, an emphasison geo-mapping, establishing refer-ence conditions, the development ofmodels and applying the observa-tional method. The role of engi-neering geologists in such projects,their objectives and responsibilities,and the range of issues that have tobe dealt with are reviewed and usedto illustrate how inputs from engi-neering geologists are critical toproject implementation. Some con-clusions are drawn regarding what isgood practice in heavy civil engi-neering projects.

Keywords Total engineeringgeology Æ Investigation Æ Design ÆConstruction Æ Railways

Resume Une approche globaled’engineering geology est mise enoeuvre pour les etudes de recon-naissance, conception et construc-tion relatives a d’importants projetsde voies de chemins de fer destineesau transport de minerai de fer dansla region du craton de Pilbara (Etatd’Australie-Occidentale). La meth-ode est basee sur un examen appro-fondi de l’histoire geologique etgeomorphologique des sites con-cernes, incluant des investigationsprogressives basees sur l’observa-tion, des etudes cartographiquesspecialisees et l’etablissement demodeles geologiques. Les respons-abilites des specialistes en engineer-ing geology, les apports specifiquesde cette discipline et les differentsproblemes qui doivent etre traitesdans ces projets sont soulignes. Desconclusions sont etablies concernantles conditions d’une bonne mise enœuvre des methodes et techniques del’engineering geology dans les projetsimportants de genie civil.

Mots cles Engineeringgeology Æ Reconnaissance dessites Æ Conception des ouvrages ÆConstruction Æ Traces ferroviaires

Bull Eng Geol Environ (2005) 64: 67–94DOI 10.1007/s10064-004-0271-4 ORIGINAL PAPER

F. J. Baynes (&)Consultant Engineering Geologist,10/272 Hay St, Subiaco, WA, 6008,AustraliaE-mail: fredb@iinet.net.auTel.: +61-8-93819498Fax: +61-8-93821564

P. G. FookesConsultant Engineering Geologist,‘Lafonia’, 11a Edgar Rd,Winchester, Hampshire, SO23 9SJ, UK

J. F. KennedyConsultant Engineering Geologist,11 Southport St, Leederville,WA, 6007, Australia

Introduction

Unless engineering geologists know how to add valuetoprojects their contribution will not be sought by pro-ject managers and those projects will ultimately suffer.Thus it is important to identify strategies that ensureengineering geologists contribute to projects effectively.However, accounts of the strategies that are intrinsic toeffective engineering geology have only recently beenpublished e.g. Fookes (1997) and some are still fearfulthat the discipline remains poorly defined (Knill 2003).

The ‘‘total engineering geology approach’’ is based onthe concept that site conditions should be viewed as theresult of the complete geological and geomorphologicalhistory and that an understanding of that history has to bewell developed at the earliest possible opportunity in anyproject for it to be successfully engineered. The approachhas been used intuitively by practitioners for many yearsbut it is only recently that it has been formalised (Fookeset al. 2000, 2001). The way in which this approach differsfrom the ‘‘traditional approach’’, followed by manyengineering organizations, is illustrated in Fig. 1.

The underlying premise of the total engineering geol-ogy approach is that many of the fundamental decisionsrequired for successful project implementation can onlybe reasonably made after thoroughly understanding thetotal geological and geomorphological history of theproject area and the associated engineering implications.Recent authoritative guidelines on the management ofgeotechnical risk (e.g. Clayton 2001) now embody thisadvice. It follows that adopting this approach is the mosteffective way for engineering geologists to contribute tothe implementation of any ground engineering project.

In this paper, the practical strategies involved in thetotal engineering geology approach are identified bydescribing its application in the investigation, design,construction and operation of some major railways that

have recently been developed to haul iron ore in thePilbara region of Western Australia.

The paper summarises the engineering geologicalinformation that was acquired and the lessons that werelearnt during the authors’ involvement in a successionof these railway projects and uses this experience baseto illustrate the application of the methodology. Thepaper is divided into three parts that reflect the meth-odology. In the first part an overview of the develop-ment of the Pilbara region and a summary of thegeological and geomorphological history is provided, asthis must be the starting point for any engineeringgeological study. In the second part the relationshipbetween the engineering geology and the railway pro-jects is explored, as it is the application of engineeringgeology in real life engineering projects which providesit with relevance. In the third part the involvement ofengineering geologists in the decision making processesassociated with a range of ground engineering issues isdescribed to illustrate the effectiveness of the method-ology in contributing to project engineering.

Part 1: Overview, geology and geomorphology

Upon arrival at any project site the first task of theengineering geologist is to develop a general apprecia-tion of the development of the region and to under-stand the entire geological and geomorphologicalhistory.

Development of the Pilbara region

The Pilbara region is located in northwestern Australia(Fig. 2). The iron ore in the Pilbara was first noted whenEuropean settlers started to move into the area in thelate 1800s but systematic exploration of the resource

Fig. 1 Contrasting approachesto the application of engineer-ing geology on projects.Although greater geological ef-fort is expended using the totalengineering geology approach itdoes not usually cost as muchas the traditional approach asthe latter generally involvesmore extensive subsurfaceinvestigations using expensiveplant

68

only started after an embargo on the export of iron oreimposed by the Australian Commonwealth governmentwas lifted in 1960. The area has since become establishedas one of the world’s major iron ore provinces.

Iron ore mining

The economic deposits of iron ore may be differentiatedinto three types of ore bodies (Kneeshaw 2000):

a. Bedded ores, also known as bedrock or enrichmentores, which consist of martite, haematite and goethite

Fig. 2 Pilbara region—regional geology, stratigraphy and railwaynetworks

69

and are mined from enriched banded iron formation(colloquially referred to using the acronym ‘‘BIF’’) inthe Proterozoic Hamersley Group.

b. Channel ores (or channel iron deposits—‘‘CIDs’’),which consist of cemented goethite and haematitepisolites occurring in mid-Tertiary fluvial palaeo-channels.

c. Detrital ores, which occur as late Tertiary colluvialfans derived from adjacent bedded ores.

Since mining commenced in 1966 production of ironore from the Pilbara totalled about 4 billion tonnes up to2003. Annual production during 2003 was of the order of180 million tonnes, which represented about 9% of totalworld production (including both low and high gradeore) and about 35% of total seaborne trade (which islimited to high-grade production).

The railway systems

The iron ore mined in the Pilbara is destined for exportand requires transport to shipping ports on some of thelongest, most heavily loaded railway trains in the world.Ore trains typically consist of up to 240 wagons and canbe 2.6 km in length. Each wagon carries around130 tonnes of iron ore, which is usually loaded into thewagons at a loop near the mine where the train turnsround. The locomotives can be as powerful as 6000 hpunits and up to 4 units are used to move a large train.Each wagon imposes individual axle loads of 35 tonnesonto a railway track that has a gauge of 1435 mm and issupported on pre-cast concrete sleepers 650 mm apartplaced on a 200–300 mm thick layer of 40 mm ballast.Ruling grades for loaded trains are typically an absolutemaximum of around 0.5%. Development of railwayscommenced in the 1960s and has continued as iron oreexports grow. The system is currently estimated to con-sist of about 1400 kms of track and new railways arebeing planned. The authors have been involved in oneform or another in railway developments in the Pilbarasince the 1990s.

Geology and geomorphology of the Pilbara region

The development of a detailed understanding of thegeological and geomorphological history of an area andhow that history influences the project engineering is afundamental strategy of the total engineering geologyapproach. The understanding has to be very broad andsometimes extends to aspects of geology, such as re-gional metamorphism or palaeoclimates, that might ap-pear esoteric to engineering managers. This contrastswith the traditional approach, which is often limited tothe logging of boreholes and test pits and laboratorymeasurements of strength and index properties.

Bedrock stratigraphy

Figure 2 is a geological map of the Pilbara region, whichis a cratonic area underlain by massive Archaean gran-ites and gneisses aged around 3 Ga (i.e. 3 billion yearsold). The Archaen rocks are overlain in the south of theregion by bedded volcanics and sediments which weredeposited around 2.5 Ga onto the eroded cratonic sur-face during late Archaean and early Proterozoic times(Trendall 1990).

At the beginning of the Proterozoic the earth’satmosphere contained little oxygen and consequentlyferrous iron derived from weathering could remain insolution in oceanic waters. It is postulated that duringthe Proterozoic the evolution of photosynthesizingorganisms produced oxygen which precipitated ironfrom the oceanic waters in the ferric state as haematite,to form the iron-rich layers in the BIF (Trendall 2000).

The railways have mainly been built to service themines developed around ore bodies located in the centralPilbara where rocks of the Hamersley Group occur. TheGroup has an estimated total thickness of approximately2500 m and consists of eight formations comprising BIF,‘‘shale’’ (a term used in the Pilbara region for interbed-ded fine sandstones, laminated siltstones and mud-stones), dolomite, and thick sequences of basic and acidicvolcanic rocks. Certain formations are characterized bythick sills of basic intrusive rocks such as dolerite.

A distinctive characteristic of the Hamersley Group isthe lateral stratigraphic continuity of the BIF units,which can extend over hundreds of kilometres (Trendall1990). During regional geological mapping the strati-graphic continuity allows the recognition of individualmarker horizons and conspicuous landform patternsassociated with specific formations. The alternating for-mations, which have been gently folded, control thelandscape development, with the more resistant BIFformations forming ranges of hills with distinctive cuestaor mesa shapes. Many of the deep broad valleys areassociated with the Wittenoom Formation and BeeGorge shales, the former having been subject to karstdevelopment and the latter being more easily eroded.

The late Archaean and early Proterozoic volcanicsand sediments were deformed during the Capricornorogeny between 2.0 Ga and 1.6 Ga, intruded by thickdolerite sills and subject to medium-grade regionalmetamorphism. A significant mineralogical consequenceof the metamorphism was the growth of asbestiformminerals in certain parts of the rock mass (Trendall andBlockley 1970). The presence of such minerals creates ageohazard that can be a significant health and safetyissue in the field.

The BIF consists of alternating laminae or layers ofchert, siltstone, mudstone and haematite in varyingproportions. Within the BIF sequences there are oftenbeds of massive siliceous rock which are of very high

70

strength and abrasive. The abrasive siliceous rocks canreduce production rates for both diamond coring andblast-hole drilling, cause excessive bit wear and can alsoproduce significant wear on earthmoving and processingmachinery.

Most bedrock types in the region have high uniaxialcompressive strength values. However, the engineeringcharacteristics of the bedrock reflect both the rockmaterial properties and the presence of throughgoingdiscontinuities of all types. For instance, the BIFs cancontain siliceous beds with uniaxial compressivestrengths of 150 MPa (and values up to 450 MPa havebeen measured), but can also contain beds of ‘‘shales’’which have been sheared during folding, are weaker andmore easily weathered, and can have bedding planes withresidual shear strengths as low as /¢r=15�, c¢r=0. Thewide range of engineering characteristics is furtheraccentuated by the influence of geological structuressuch as faults and joints and, in particular, by thedevelopment of deep weathering profiles. The bedrockgeology fundamentally controls the distribution andengineering performance of many of the materials inwhich the deeper railway cuts are excavated and fromwhich the thicker fills are constructed.

Bedrock structure

The degree of development and orientation of geologicalstructures within the bedrock, particularly beddingplanes, are the key concerns in cut slope design. Thebedrock in the central Pilbara was folded during theCapricorn orogeny in a regional scale foreland fold andthrust tectonic setting (Tyler and Thorne 1990). Up tofive different fold phases have been identified by struc-tural geologists working in the Pilbara region (Tyler1991). In the bedded rocks of the central Pilbara thetypical fold styles fall into two main categories of engi-neering significance, which are described below:

a. There are open large-scale regional fold styles leadingto dips of the order of 10–30�. As these folds can resultin persistent uniform bedding planes dipping at anglesthat can be greater than the angle of friction, the dipof the strata can contribute to the development ofrock slope instability. Such instability is observed inthe steeper natural rock slopes throughout the Pilbara.

b. There are tighter localized folds that are restricted tostructural corridors possibly associated with regionalfaults or intrusions. These structures tend to provideinterlock across bedding planes which reduces thepotential for large scale and medium scale instability,but can lead to small scale local instability due toadversely dipping bedding planes occurring at out-crop or in cut slopes.

Fault styles are mainly low angle thrusts associatedwith the regional compression and are rarely encoun-tered in railway excavations. Jointing largely reflects the

structural attitude of the adjacent folding and faultingsystems. At least three distinct sub-vertical joint setsoccur in most BIF units but these sets are usually lessdeveloped in other bedrock types. Well-developed mas-ter joints on gently dipping BIF fold limbs often extendseveral hundreds of metres. Steeply dipping persistentmaster joints typically exert a fundamental control oncliff line stability, especially where the rock mass is proneto toppling failure, and characteristically form a ‘‘saw-tooth’’ pattern on cliff lines.

Landform evolution

Landform evolution governs the distribution of superfi-cial deposits, weathering profiles and slope processes andthus exerts a profound influence on the engineeringcharacteristics of the near surface materials in which therailway earthworks are constructed. Understanding thelandform evolution requires an understanding of theregional tectonic evolution of the area, as this has con-trolled palaeo-latitudes and rates of uplift and hence pastclimates, weathering processes and cycles of erosion.

The Pilbara was probably glaciated during the Perm-ian, when the region was part of the polar supercontinentGondwana, as glaciogenic sediments have been foundfurther south preserved in basins in the ancient WesternAustralian cratonic land surface (Anand and Paine 2002).The supercontinent started to rift during the Jurassic toform the continent of Australia (Fig. 3). During theMesozoic, progressive erosion produced extensive rela-tively planar land surfaces cutting across the craton. Atthe end of the Mesozoic the climate is believed to havebeen warmer and more uniform globally in comparisonwith the present, due to restricted oceanic circulation(Summerfield 1991). Thus despite being so far south atthis time, the craton was subject to several cycles of deepchemical weathering and the formation of duricrusts thateventually formed the prominent ‘‘Hamersley surface’’,remnants of which can now be seen across the Pilbara(Campana et al. 1964; Twidale 1994).

During the Cainozoic progressive uplift and minorwarping, weathering and duricrust formation and ero-sion during periods of renewed incision produced acomplex series of slopes, colluvial and alluvial depositsand duricrusts. A typical Pilbara landscape reflecting theeffects of all of these processes is shown in Fig. 4.

Cainozoic changes in climate

Since the break up of Gondwana, the Australian conti-nental plate has moved north through more than 30� oflatitude (see Fig. 3) whilst the global climatic system hasfluctuated significantly (Bowler 1982). Studies of Cai-nozoic climate change and regolith development basedon sequence stratigraphy around Australia suggest thatthere may have been at least four distinct periods of

71

more intense weathering during the Tertiary (McGo-wran and Li 1998). A simpler view is that most of thedeep weathering profiles that are preserved in the Pilbaradeveloped successively between the Late Cretaceous andthe mid-Miocene (Killick et al. 2001).

Although the Pilbara has moved from a latitude ofabout 55�S to a latitude of about 25�S, it appears thatduring most of the Tertiary the climate was ‘‘tropical’’ or‘‘sub-tropical’’ i.e. warm with a heavy rainfall. It seemsprobable that for much of the time the area was heavilyforested, there was deep weathering of the erosion sur-faces and there was a high degree of mobility of iron inthe landscape. Between late Oligocene and mid-Miocenetimes erosion of duricrusts, fluvial deposition and ironenrichment led to the formation of the CIDs (Rama-naidou et al. 2003).

During the Pliocene there was incision, erosion anddeposition of alluvial/colluvial fans. This may have re-sulted from regional tectonic uplift and increasing arid-ity, coupled with loss of vegetation cover and periodicextreme storm events. The gently sloping distal portionsof the Pliocene alluvial/colluvial fans currently provideoptimal railway route corridors and usually excellentconstruction materials.

The Quaternary climate is thought to have beengenerally semi-arid with a monsoonal hot wet summer. It

Fig. 3 Tectonic movements of the Australian Plate. The recon-structions were made by S. A. Pisarevsky (Tectonics SpecialResearch Centre, The University of Western Australia), using thespherical rotation software of the Visual Paleomagnetic Database(Pisarevsky and McElhinny 2003), PLATES software of theUniversity of Texas in Austin, and GMT software of Wessel andSmith

b

Fig. 4 Typical Pilbara landscape including rolling spinifex coveredhills, broad valleys, BIF outcropping as cliff lines and concordantsummits forming the Hamersley Surface on the skyline

72

is likely that the region was subject to considerablevariations in the amount and intensity of rainfall duringthe Quaternary associated with global climatic fluctua-tions related to glacial and interglacial periods. Theremay be similarities with the nearby Kimberley regionwhere marked climatic fluctuations during the Quater-nary have been documented (Wende et al. 1997). How-ever, research is yet to be carried out in the Pilbararegion into past periods of increased aridity or therecurrence intervals of extreme flood events associatedwith monsoonal activity.

Superficial deposits

The superficial deposits that accumulated during thedifferent climatic regimes vary in age from early Tertiaryto Recent and form deposits on or close to the surface ofthe bedrock. Four main processes have formed thesuperficial deposits:

a. Deposition of colluvial and taluvial sands, gravels andboulders on slopes.

b. Deposition of alluvial and colluvial gravels, sands,silts and clays in fans, channels or on flood plains.The finer sediments tend to be deposited in the moredistal parts of the fans and the clays tend to bedeposited in the low-lying floodplains.

d. The development of weathering profiles on bothbedrock and unconsolidated sediments involvingsolution and decomposition of mineral species.

d. The growth of mineral species in the superficialdeposits or at the top of weathering profiles devel-oped on bedrock, which in places have cementedunconsolidated materials to form rock or gravellymaterials and become duricrusts (Thomas 1994).

The dominance of duricrust development (Hockingand Cockbain 1990; Killick et al. 2001) and deepweathering has resulted in the formation of a series ofcemented surfaces within the landscape. Any surfaceprocess capable of transforming soil to rock is clearlyimportant from an engineering viewpoint. Dependingupon the cementing agent the materials may be describedas: a. Ferricrete—cemented by iron oxide, typicallybrown or red, often with a nodular appearance, some-times forming fine gravels, or irregular rock material.b.Calcrete—cemented by calcium carbonate, typically ared pink or white gravely material, sometimes formingtabular rock material.c. Silcrete—cemented by silica,typically a white cherty, gravely material sometimesforming a tabular rock material.

Where reactive i.e. moisture-sensitive clay specieshave been deposited or developed on the flatter moredistal colluvial slopes and flood plains, areas known as‘‘gilgai’’ have developed (Beckman et al. 1970; Cookeand Warren 1973; Maxwell 1994). Gilgai have acharacteristic hummocky appearance with cracks and

‘‘crab holes’’. The surface features are produced bymarked seasonal volume changes due to the wettingand drying of reactive clay species that form a sig-nificant proportion of the soils. Gilgai are a significantgeohazard.

Current topography, climate and vegetation

Currently the central Pilbara is a semi-arid upland with abase elevation some 700 m above sea level and withranges of hills that extend up to 1000 m in height. Thetopographic relief leads to the development of a varietyof active slope processes that can impact upon railways,such as debris flow and rockfall.

Summer daytime temperatures range well into the 40sand winter frosts can occur. Annual evaporation is wellin excess of annual rainfall, which on average is between180 mm and 350 mm, but it is highly irregular due toperiodic cyclonic events and thunderstorm cells thatproduce localized intense falls. There may be no rain forseveral months, or even years, in a particular area, then alarge part of the yearly rainfall will occur in a short time,leading to flooding along the water courses and move-ment of debris down slopes and along drainage lines.Rainfalls of up to 200 mm in 24 hours can occur duringthe rainy season. This type of irregular intense rainfallpattern exerts a major control over current slope processrates and the occurrence of damaging floods.

Due to the semi-arid environment the vegetation coverconsists of scrub and spinifex grass with isolated gumtrees along drainage lines. The region has been subject topastoral grazing since European settlement, but this hasnot led to significant impacts to the vegetation cover.However, the naturally sparse vegetation cover increasesthe rapidity of the runoff and does little to impede themovement of sediment downslope and thus contributesto the active slope processes and flood impacts.

The outstanding scenery of the Pilbara, which istypified by the ranges of hills and dramatic gorges andthe contrast between the green vegetation cover and thered and brown tones of the underlying geology, has re-sulted in much of the area having National Park status.

Current natural slopes

Slopes in the Pilbara reflect the underlying lithology andstructure and the overprinting of different phases ofweathering, duricrust formation, erosion and deposition(Joyce and Ollier 2003). The upper parts of slopes areoften remnants of former erosion surfaces and usuallyinclude rounded duricrusted portions or flat surfacessub-parallel to the bedding. Below the upper slopes steepcliff lines controlled by persistent sub-vertical joints haveoften developed due to rejuvenation of the landscape byuplift and incision (Fig. 5).

73

The cliff lines are usually actively degrading withaprons of rockfall debris forming transportational mid-slopes (Fig. 6).

Below the cliffs and rockfall aprons and at the exitpoints of gullies and canyons there are widespread allu-vial/colluvial fans and slopes mantled with such deposits.Both alluvial deposits and channelised debris flows formlow angle fans towards the base of the slopes. Within thebroad valleys there are alluvial flats and intermittentflood-prone river systems (Fig. 7).

Total engineering geology history

The entire geological and geomorphological history ofthe Pilbara region that is described above needs to beunderstood as part of the total engineering geology ap-proach. This history, and the consequences for railwayengineering, are summarized in Table 1.

Part 2: Project strategies for total engineering geology

A high level of geological and geomorphological under-standing was not in itself sufficient to successfully supportproject implementation for the various railways in whichthe authors were involved. Various strategies had to bedeveloped to efficiently acquire and effectively commu-nicate that knowledge. Geological information had to bedelivered to the right people in the project team at theright time in an easily understandable manner and withan explanation of the significance of the information.

Project stages

To effectively contribute to the railway projects in thePilbara using the total engineering geology approachdifferent types of information were collected and com-municated at each of the typical engineering projectstages. Figure 8 summarises the range of engineeringgeological activities that were found to be most effective

in achieving the objectives of each of the idealised projectstages.

Finance is generally limited at the pre-feasibility,feasibility or initial design stage in any project and thecollection of low-cost information that will support verybroad decisions has to be the main objective. Hence, theinvestigations during these stages usually involved deskstudies, studies of aerial photographs, site reconnais-sance, ground truthing of aerial photographic interpre-tations by site visits, general mapping and inspection.Investigations involving detailed mapping, test pitting,drilling of boreholes to obtain information for deep cuts,laboratory testing etc are far more expensive and wereonly carried out as the alignments became more fixedand greater detail was required for design and costing.

Fig. 5 Composite landscape including rounded upper slopes, clifflines and colluvium covered slopes

Fig. 6 A typical cliff line controlled by joints which form acharacteristic ‘‘saw-tooth’’ pattern. A taluvial apron has accumu-lated on the slope below the cliff line, which is about 10 m high

Fig. 7 Lower slopes formed by coalescing distal portions ofalluvial/colluvial fans, which merge into alluvial flats formingextensive floodplains. The optimal railway route corridor is on thelower slopes

74

Table

1Pilbara

totalengineeringgeology

Episodeandage

Geologicalevents

Geologicalenvironments

Engineeringconsequences

Quaternary

0–2Ma

Holocene0–0.01Ma

Currentinterglacialandhighstand,

Australiancratonmovingnorthat

10mm/year,latitude23�

Hot,semi-arid,monsoonal,misfitalluvialtracts

subject

toperiodic

floodevents,minor

colluvial,alluvial,aeolian,lacustrinedeposits

androckfallaprons.

Flash

flood,debrisflows,rock

fall,insolation

weathering,shrink/swellofclays,

low

seismic

risk

Pleistocene0.01–2Ma

Oscillationsin

temperature

leading

toworldwideglacial/interglacialcycles

Sem

i-aridmonsoonalenvironmentwithim

posed

cycles

ofboth

increasedaridityandheightened

monsoonalrainfall,substantialcolluvial,alluvial,

aeolian,lacustrinedeposits

androckfallaprons.

Troublesomesuperficialdeposits

such

aswindblownsands,gilgaiclays,

loose

alluvium

etc

Tertiary

2–65Ma

Pliocene2–5Ma

Increasingrate

ofglobalcoolingand

droppingsealevels.Palaeolatitude27�

Increasingaridity,loss

ofvegetation,extrem

erainfallsproduce

massiveerosion,large

alluvial/colluvialfans,coarseproxim

alandfine

distalfacies,upto

50m

ofclayey

sandygravels

andcobbleswithlocalferricretes.

Possible

earlyPlioceneintense

weatheringevent

Form

ationofdeepcolluvialfanswhich

provideexcellentconstructionmaterials

andrailcorridor

Miocene5–25Ma

Growth

ofpolarice,

reducedsealevels,

globalcoolingoffsetbyincreased

temperature

associatedwithreduced

latitude.

Palaeolatitude30�

Late

MiocenearidityWarm

,wet

clim

ate,mature

landscape,

deepweatheringprofiles,duricrusts,

lacustrineormeanderingmature

alluvial

deposition,upto

50m

ofclays,clayey

sands,

goethitic

clays,lacustrinelimestone,

somefine

gravels,andpisolitichaem

atite

enriched

alluvium—correlatesofCID

.Mid

Mioceneintense

weatheringeventandtectonic

uplift

EarlyMiocenesealevels200m

abovepresentlevels

Extensiveweatheringprofile

andduricrust

development,karstdevelopment,gentle

rounded

slopes

develop

Oligocene(24–34)

Openingofsouthernoceans,Antarctic

icesheetestablished.Palaeolatitude35�

Continuinghighrainfall,warm

,‘‘tropical’’

weathering

Continuingweatheringanderosion

Eocene34–55Ma

Palaeocene55–65Ma

Uniform

lywarm

phase

produced

‘‘tropical’’weatheringto

45�latitude,

restricted

oceanic

circulation,anoxic

oceanbasins.Palaeolatitude45

�

Late

Eoceneintense

weatheringeventfollowed

byaridity.Generallyhighrainfall,warm

,‘‘tropical’’

weatheringandduricrustsdeveloped

onplanation

surface,alluvialsheetwash

orbraided

alluvial

deposits,residualsoilslocallypreserved

indolomite

karst,10–100m

ofsiltysandsandfinegravels,

manganiferousgoethitic

claysdeveloped

on

weathered

dolomitesurfacesE

arlyEoceneintense

weatheringevent

Deepweathering,duricrusts,development

ofkarst,earlysurfacesandinitialslopes

Mesozoic

65–241Ma

Cretaceous65–141Ma

FaunalextinctionsatendofCretaceous

continuingbreak-upofGondwana.

Palaeolatitude60�

Late

Cretaceousclim

ate

generallyhighrainfall,

warm

,‘‘sub-tropical’’andequitable

withno

polarice,

deepweatheringprofilesand

duricrustsdevelop,longterm

sealevels250m

abovepresentlevels

Weatheringanderosionproducessurfaces

whichform

concordantsummitsafter

incision

Palaeozoic

545–251Ma

Form

ationofGondwana

Continentalcrustalplatescoalesceto

form

asupercontinent,mountain

chainsform

edbetween

theplatesare

progressiveeroded

Future

Australiancratonic

areafused

together

anderoded

Proterozoic

1.6–2.5

Ga

Capricornia

orogen

Folding,faultingandjointingDeposition

ofsedim

ents,BIF

s,volcanics,intrusives

Bedded

bedrock,dips,most

discontinuities

Archean2.5–3Ga

Form

ationofcrust

Granites

andgneisses

solidifyclose

toearth’ssurface

Massivebedrock

75

Investigation reporting objectives

From an engineering perspective most investigation re-ports prepared for the railways had a dual function:

a. To provide sufficient information to support the de-sign of the railway and the generation of an engi-neer’s estimate of the cost of the project, so thatproject finance could be sought. Engineers’ estimatestypically aimed to be accurate to ±20% at feasibilitylevel.

b. To provide ground engineering information to pro-spective tenderers for the contract works in the formof site-investigation reports. As such reports wouldultimately have contractual significance it wasimportant to control their style and content.

As a consequence of this dual function, investigationreports for the railways were prepared with the followingobjectives in mind:

Fig. 8 Idealised project stages,engineering geological activitiesand strategic objectives

76

a. The overall aim of the investigation was to cha-racterise the ground to the level where project scaleunforeseen ground conditions were unlikely to beencountered. Some balance had to be establishedbetween expenditure on the investigations and theresulting benefit of the information obtained (Stap-ledon 1982). For these railways a reasonable balancebetween cost and benefit was considered to have beenmet when the following questions were answered inthe affirmative:

i. Is the geology mapped at a scale of about 1:5000,reasonably well documented and sufficiently wellunderstood to be able to generate reliable geologicalmodels (sketch maps and sections) for all cuts?

ii. Have all embankment foundation conditions beencharacterized?

iii. Have all sources of construction materials been locatedand their performance characterized?

iv. Have all geohazards been identified?v. Is it reasonable to confidently complete the geometric

design and then request tenders for construction?

b. All of the observations were presented using a stan-dardised descriptive system to minimise confusionamongst design or construction engineers about thegeological and geotechnical descriptions of thematerials. In the Pilbara the Australian Standard forSite Investigations, AS1726 (1993), formed the basisof the descriptive system.

c. There was an emphasis on non-written communica-tion i.e. colour photographs, maps and drawings, sothat those with little or no understanding of theengineering geological descriptive terms could gainan appreciation of the ground conditions by lookingat the photographs, especially where the informationrelated to machine performance such as ripping trialsand backhoe excavations.

d. All information relevant to the project was assembledand made available either as reports provided toprospective tenderers or assembled in a room forviewing by prospective tenderers (c.f. the guidelines ofthe Construction Industry Committee 1987).

e. A clear distinction was drawn in the reports betweenobservations, interpretations and opinions to mini-mize confusion as to the nature of information beingprovided. All of the site-investigation information wascompiled into a ‘‘factual report’’, which consisted ofobservations but not interpretations, and a separate‘‘interpretation/evaluation report’’ in which the inter-pretations and opinions were documented. Both kindsof report were provided to prospective tenderers.

From the Owner’s perspective, the more that carefullypresented information could be provided to prospectivetenderers, the less was their uncertainty during the briefperiod when they prepared their bids. As uncertainty canonly be allowed for in a bid price by an increase in the

cost estimate, these strategies were specifically directed atobtaining the most competitive bids for building theproject.

Importance of geo-mapping

Mapping was not a significant component of the inves-tigations of the earlier (i.e. pre 1990s) Pilbara railways,and efforts were instead directed immediately to morecostly sub-surface investigations. This proved to beineffective and is typical of the traditional approach(Fig. 1) that is adopted in many heavy civil engineeringprojects throughout the world. It has been causticallydescribed as ‘‘drilling first and asking questions later’’ andis the antithesis of the total engineering geology approach.The traditional approach is a particularly ineffective wayof investigating long linear structures such as railwaysbecause carrying out subsurface investigations too earlyin such projects almost inevitably results in having tocarry out additional expensive subsurface investigationsdue to changes in the alignment required by ongoingdesign. In contrast, early engineering geological andgeomorphological mapping of a route corridor is far lessexpensive and far more productive because it involves noexpensive plant and is more likely to provide informationrelevant to whatever final alignment is chosen.

More importantly, mapping of the route corridorrequires careful observations of the geology and geo-morphology, interpretation of the near surface groundconditions, reporting and the generation of a ‘‘model’’into which all of the information is incorporated forfuture analysis. Thus geo-mapping can be regarded asthe essential and defining strategy of the total engineeringgeological approach.

Where this approach was adopted, considerable reli-ance was ultimately placed on detailed engineering geo-logical and geomorphological mapping of the routes atvarious scales as the primary method of investigation.Using mapping as a primary investigation tool also al-lowed the subsurface investigations that were subse-quently required, such as drilling and ripping trials, to beplanned and targeted to investigate the engineeringgeological formations in a more systematic fashion. Thealternative traditional approach often locates subsurfaceinvestigations without regard for the geology; for in-stance, planning boreholes at the highest point of thedeepest cuttings, where the existing access tracks crossthe centre line or, possibly worst of all, exactly every5 km along the centre line!

The engineering geological maps were produced withthe objective of differentiating geological units withcharacteristic engineering behaviour (Fookes 1969;Dearman 1991). In practice it was found that the totalengineering geological approach required the identifica-tion of mapping units at four different scales for whichthe following terminology was adopted:

77

a. Terrain Units, which consist of distinct assemblages ofbedrock, superficial deposits and landforms with rec-ognizable engineering characteristics and which weremapped at a scale of about 1:50000 to 1:250000.Terrain systems mapping has recently been reviewedby Phipps (2001).

b. Engineering Geological Formations, which consist ofgenetically related groups of soils and rocks andlandforms with a distinctive suite of engineeringcharacteristics and which were mapped at a scale ofbetween 1:5000 and 1:50000. These units were termed‘‘engineering formations’’ by Dearman (ibid).

c. Engineering Geological Members, which consist of asingle lithological type but which might have a rangeof engineering characteristics. Variations in engineer-ing characteristics within members often result fromthe influence of tectonics and surface processes uponlithology and thus reflect the total geological andgeomorphological history. They were usually mappedat a scale of 1:5000 to 1:1000 for the detailed design of

specific project elements. These units were termed‘‘lithological types’’ by Dearman (ibid).

d. Engineering Geological Types, which are units thathave effectively homogenous engineering character-istics and were described using standard descriptivesystems, usually in logs of boreholes, test pits andexposures.

The different mapping units were logically differenti-ated in the field, on the map legend or on logs using theprocesses of division described by Varnes (1974). Themapping units provided a framework for systematicallymaking observations and collecting information.

Definition of Reference Conditions

The definition of Reference Conditions was anotheressential strategy of the total engineering geology ap-proach. Reference Conditions (CIRIA Report 1978)consist of groups of geological materials with similar

Fig. 9 Logical relationships be-tween the mapping units andthe Reference Conditions

78

engineering characteristics, and depict the range of geo-logical conditions that can reasonably be anticipated orforeseen (Essex 1997; Knill 2003). They were used to de-scribe and communicate the geological conditions toproject engineers. In contrast to the process of divisionused to establish the mapping units, the process ofgrouping (Varnes 1974) was used to establish ReferenceConditions. The relationships between the different levelsof mapping units and the Reference Conditions areillustrated in Fig. 9

The definition of Reference Conditions was also a partof developing the models that were central to the under-standing and communication process. In most projects,the geological model is based upon observing or samplingonly a small part of the ground, and alternative interpre-tations are possible. Where a project is to be constructedunder contract and where there is uncertainty in the levelof understanding of particular features of the ground,which could have had a significant influence on the Con-tractor’s choice of method or cost, the use of ReferenceConditions indicates the assumptions that have beenmade for the basis of the contract (Muir Wood 2000).

In the case of the central Pilbara, all of the litho-stratigraphic units are essentially different successions ofBIF and shale, with dolerite intruded to varying degrees.The strata have all been subject to folding, faulting, deepweathering and duricrust formation. Thus the largenumber of different geological formations (with astratigraphic nomenclature that varies from map sheet tomap sheet and has been revised on various occasions)can be reduced to a smaller number of Reference Con-ditions, which can be communicated using as few eso-teric geological terms as possible.

The important functions of Reference Conditionsmay be summarized as follows:

a. To formally define and describe the components ofthe engineering geological formations and the geo-logical models.

b. To simplify the geology by grouping together thegeological units with similar engineering characteris-tics, which eases communication to engineers.

c. Most importantly, to document the conditions thatcan be reasonably foreseen for contract purposes.

d. To allow a reduction of overall laboratory testingschedules - only representative samples from eachReference Condition have to be tested rather thantesting all geological units encountered.

e. To allow the incorporation of knowledge from similargeological units that occur outside the project areaand that may be correlated with the Reference Con-ditions.

f. To be of practical help during construction in matterssuch as predicting equipment performance andcapability.

Examples of some typical Reference Conditionsused in the central Pilbara are provided in Table 2 and

the suite of geological and engineering characteristicsof one typical Reference Condition are provided inTable 3.

The use of models

The use of geological models in project engineering hasbeen described in detail by Fookes (1997), the strategicapplication of geological modelling on a specific pro-ject has been described by Newman et al. (2003), andthe more general use of models in ground investiga-tions has been discussed by Harding (2004). The totalengineering geology approach relies heavily on the useof engineering geological models (Fookes et al. 2000)to:

a. Assemble and collate a disparate information base;b. Interpret, present and communicate the observed or

inferred conditions;c. Indicate conditions that might be reasonably antici-

pated and might require further investigation.

In order to be useful any model must satisfy threecriteria (Moores and Twiss 1995):

a. The model must be powerful, that is capable ofexplaining a large number of disparate observations.

b. The model must be parsimonious and must use aminimum number of assumptions compared to therange of observations that it explains.

c. The model must be testable, which means that it mustanticipate conditions that at least in principle can beverified by observation.

The models that were developed for the railway pro-jects in the Pilbara took the form of simple geologicaland geomorphological maps and sections, combinationsof geological, geomorphological and geotechnical infor-mation in engineering geological maps and sections,evolutionary diagrams and 3D block models. Thesemodels presented three different kinds of information:

a. Conceptual models, which indicated the geologicalrelationships between mapping units, their likelygeometry, and anticipated distribution e.g. Fig. 10.These kinds of conceptual models were used to con-veniently represent the nature of different TerrainUnits and Engineering Geological Formations.

b. Observational models, which presented the observedand interpreted distribution of Reference Conditionsin 2D models such as maps and sections and in 3Dmodels such as block diagrams, and were constrainedby subsurface data or by extrapolation from thesurface e.g. Fig. 11.

c. Evolutionary models, which illustrated the way inwhich the Terrain units, Engineering GeologicalFormations or Reference Conditions developed intime using a series of sketch maps, sections or block

79

models e.g. Fig. 12. Block models depicting the geo-logical and geomorphological development with timebecome 4D (i.e. 3D plus the dimension of time).

The models were used to develop an understanding ofthe total engineering geology, to target areas for subsur-face investigations, or for detailed mapping investiga-tions (especially for potential borrow sources orgeohazards such as landslides, unstable soils and flood-ing), to relate sites within the railway route corridors to

the local and regional geological stratigraphy andstructure, to assess cutting excavatability, embankmentfoundation conditions and the location of constructionmaterials, and to engineer controls on the occurrence ofselected geohazards. Very importantly, the models wereused as graphic tools to communicate the interpretedand anticipated conditions to project engineers. Suchrelatively simple models helped project engineers tounderstand the engineering implications of the geology

Table 2 An outline of selected Pilbara Reference Conditions

Referencecondition

Geological origin Engineering characteristics Engineering performance

Fine grainedsoils

Alluvial sands and silts, distalalluvial/colluvial deposits, aeoliansilts, gilgai (high plasticity claydeposits), clayey residual soilsdeveloped over dolerites, basaltsand some granites, mangrove muds,supratidal flats.

Uncemented soil deposits with >50%of particle sizes <2.36 mmand a significant clay or silt content.

Unsuitable as fill,potentially troublesomeembankment foundation,erodable, low cut slope angles.

Coarse grainedsoils

Alluvial gravels, proximalalluvial/colluvial deposits,colluvium and taluvium,gravelly weathering profilesdeveloped over some granites.

Uncemented soil deposits with >50%of particle sizes >2.36 mm, low clayand silt content, significant gravelcontent, often some cobbles.

Good source of borrowfor general and select fill,generally good embankmentfoundation

Duricrusts Gravels, cobbles and rock materialsdeveloped in-situ by cementationof superficial deposits and weatheringprofiles due to: accumulationof calcium carbonate (calcrete -also developed on freshwaterand aeolian coastal limestones); ironenrichment by haematite/goethiteduring weathering (ferricrete);or accumulation of silica (silcrete).

Ranges from gravely soil depositsto irregular variable masses of rockmaterial with dimensions >1000 mm,rock material formed in-situ with lowto high strength. Characteristics varyrapidly over short distances.

May break down upon handlingor exposure, not necessarilyrock in contract terms, sourceof general fill but may requireblasting, variable embankmentfoundation conditions

Extremelyweatheredbedrock

Deep clayey or silty weatheringprofiles developed over Weeli WolliDolerite, Mt McRae Shale,Mt Sylvia Formation shales,Bee Gorge Member shales, someweathered granites and basalts.

Soils developed by weatheringof bedrock, generally with asignificant clay and silt content,corestones and relict structurescommon,

Erodable if exposed in cutsor fills, prone to shrink swelland requires encapsulationif used as fill

Shaley BIF Parts of Weeli Wolli BIF,Yandicoogina Shale, Joffre unit,Whaleback Shale, Dales GorgeMember.

Fresh or slightly weatheredhigh strength BIFwith bedding planes spacedat 5–25 mm, BIF with >30%of weak siltstones, distinctlyweathered low strengthBIF with bedding planes spacedat <100 mm,

Often excavated as common,potential source of select fillbut can be too platey, maybreak down on handling.

Blocky BIF Parts of Weeli Wolli BIF, partsof Joffre unit, parts of Dales GorgeMember, Mt Sylvia FormationBrunos Band and BIF Twins,shale poor or mineralizedBIF sequences

Fresh or slightly weathered highstrength BIF with bedding planesspaced at 50–300 mm

Requires blasting, providesgood rockfill but can be slabby,steep cut slopes

Blocky bedrock Relatively unweathered partsof Weeli Wolli Dolerite, Archaengranites and gneisses, FortesqueGroup basalts, sandstonesand quartzites.

Fresh or slightly weathered highstrength rock with jointsspaced at 50–3000 mm,equidimensional blocks,no pervasive fabric

Requires blasting, providesgood rockfill, rip-rap, steepcut slopes

CID CIDs within Tertiary palaeochannels Fresh or slightly weathered low to highstrength flat bedded rock with jointsspaced at 50–300 mm, equidimensionalblocks, no pervasive fabric.

Some rippable but mostlyrequires blasting, providesgood general fill, steepcut slopes

80

and to appreciate the importance of collecting geologicalevidence and the power of geological observation,interpretation and anticipation.

Effective subsurface investigations

Subsurface investigations were an essential part of thestudies for the railways, but wherever possible the scope

of the subsurface works was limited in order to minimizeexpenditure. The objectives of the subsurface investiga-tions were to provide additional data in areas of geo-logical uncertainty defined by the geo-mapping, and toprovide representative data to document the ReferenceConditions. Test pits dug by rubber-tyred backhoes wereoften used to characterize the mapped units within thecorridor. Boreholes were used at the location of some ofthe deeper cuts to confirm the surface mapping. In the

Table 3 Example of a Reference Condition in detail

Reference condition Blocky BIF

Stratigraphic origin Weeli Wolli BIF, Joffre unit, Dales Gorge Memberand Mt Sylvia Formation, Brunos Band and BIF Twins

Lithological description Finely bedded BIF with occasional shale beds. BIF comprisesinterbedded shale, chert and hematite and is typically red-brownto red-purple and grey in colour. Units commonly silica/chert(jaspilite) rich, red grey striped with chert pods. Duricrustdevelopment has resulted in cementation at the surfacein places. Locally haematite enriched to ore, or alteredby a duricrust. Fibrous mineralisation may be encountered

Structural geological characteristics Pervasive bedding fabric throughout, orientation variesdue to folding, bedding spacing typically 2–5 mm, beddingpartings 50–300 mm. Bedding commonly more widely spacedin silica/chert rich (jaspilitic)) units. At least two orthogonalsub-vertical joint sets typically present, spaced 0.5–4 m,discontinuous to continuous, slightly rough to rough.Up to 2 other joint sets can occur. Jointing tends to be lesswell developed as shale content increases. Many joints opennear surface and in-filled with clay gravel and rootlets.

Degree of weathering Variable. For engineering applications at relatively shallowdepths expect rock mass to be Distinctly Weathered.Weathering often concentrated preferentially along beddingplanes or joints, where wall rocks may be more weatheredthan the general rock mass.

Material strength Estimated strength ranges from Low to Very High.UCS indicates materials can have Extremely High Strength(Note an extremely high UCS value 447 MPa has been reported).UCS indicates that Very High strength materials commonin the upper 10 m, suggesting increased strength due to developmentof a duricrust. Anisotropic, weaker parallel to bedding,higher strength perpendicular/across bedding.

Rock quality/defect spacing Depends on degree of weathering. Rock quality increasesas weathering decreases. Typically expect RQD less than50% and often less than 25% due to pervasive bedding fabric.

Design slope angles 1 v:1 h for extremely weathered rock or shallow cuts to 2 v:1 hfor Distinctly Weathered to Fresh rock and deeper cuttings.If slopes undercut bedding careful mapping and monitoringis required to assess potential for failure. Slopes may be requiredto be cut back to dip of bedding if bedding under-cut.

Embankment foundation suitability Suitable.Expected excavation techniques Predominantly rock requiring blasting, minor common

(typically expect less than 5%). depending on degreeof weathering, development of bedding and secondarycementation.Abrasive.

Expected excavated material types Predominantly Type 3 coarse fill, some Type 3 graded fill.Comments Often a relatively open rock mass (along bedding planes

and joints that may be infilled with soil materials).Some caves and overhangs to 5 m deep developed in cliffs.Most likely variations:• Degree of weathering and/or shale content increasesand bedding parting spacing increases or duricrust developmentresults in materials become Shaley BIF.• Bedding parting spacing increases such that materials become blocky bedrock.

81

sparsely vegetated rolling country of the Pilbara, the useof large bulldozers (typically Caterpillar D10s) to exca-vate deep wide trenches (locally called ‘‘costeans’’) toexpose the ground conditions was a particularly suc-cessful and cost-effective subsurface investigation tech-nique. Information collected from such costeans wasused to assess excavatability based on the techniquesdescribed in Pettifer and Fookes (1994).

Attempts have been made to assess rippability usinggeophysical techniques in the form of seismic refractiontraverses. However, there were difficulties in reliablyinterpreting the information and it was not found to be auseful technique for investigating railways in the Pilbara.

A limited number of boreholes were also requiredsimply to support the letting of the construction con-tracts. It was considered important to provide prospec-tive tenderers with ‘‘traditional’’ types of information(i.e. boxes of core for viewing during the tender prepa-ration) to counter any later claim that the informationprovided was somehow unusual. Selected boreholes alsoideally documented each Reference Condition and, aspart of the project information base, formed an essentialpart of the understanding of the ground.

Established design objectives

The design of the ground engineering–related aspects ofthe railway mainly involved:

a. Optimization of the horizontal and vertical alignment,and the associated excavation volumes and earth-works volumes, as these represent the principal costvariable;

b. The definition of construction materials, both fromcuttings and borrow;

c. The design of cut and fill slopes, bridge foundationsand drainage to achieve a certain performance.

Inputs from engineering geologists into ground engi-neering design included the choice of cut batter angles,excavatability assessments and assessments of any geo-hazards that had been identified.

At the Design stage (sometimes called a DefinitiveEngineering Study in the Pilbara), both the Owner andthe Engineer desire to refine the cost estimate, perhaps to±10%. This was seldom achievable for the groundengineering components because of the large degree ofuncertainty that still existed. However, because costuncertainties associated with other non-geological as-pects of the project such as steel, concrete, sleepers andmany of the fixed costs were often less than ±10%,inaccuracies in the cost estimates of the ground engi-neering components that exceeded ±10% were balancedby the other components. This often resulted in a netcost estimate that went some way towards meeting theaim of ±10%.

Detailed design occurred towards the end of the maininvestigation phases and included the finalization of thedocumentation for the contractors. At this stage, theengineering geological information was generally sum-marized into two kinds of report:

a. An ‘‘interpretive report’’, which included the details ofthe design and which was made available to the pro-spective tenderers.

b. An internal ‘‘basis of design report’’ in which detailedinformation relative to the design process andassumptions was documented. As such informationhad no bearing upon the activities of the Contractor,and it was not made available to prospective tenderers.

The detailed design phase was not always clearly de-fined as a separate activity because it was intimately tiedto the final stages of project financing and was ofteninfluenced by the developing contractual relationshipsbetween the Owner and the Engineer.

The observational method

It was not possible to completely investigate every detailof the geology and geomorphology of any of the railwayroutes prior to construction for logistic and financialreasons and hence the design and attached Engineer’scost estimate were always necessarily based on limitedinformation. In order to overcome the inherent uncer-tainty that will always be present in any ground engi-

Fig. 10 Conceptual model used to explain the genesis anddistribution of different types of Pilbara Gilgai. The model includestwo conspicuous Terrain Units (basalt hills and pediplains) andseveral mappable Engineering Geological Formations (graniteswith weathering profiles, basalts with weathering profiles, stoneygilgai, deep gilgai, colluvial deposits and floodplains)

82

neering the Observational Method was routinely adop-ted during railway construction.

The Observational Method involves more than justobserving geological conditions during constructionand responding to failures, although it is this view of themethod that prevails in many projects. To properly applythe Observational Method the engineering implicationsof a range of conditions that can be reasonably antici-pated from the Geological Models must be overtly con-sidered before construction starts and the projectmanaged with these implications in mind. (Peck 1969;Fookes et al. 2000). The uncertainty can best be dealtwith by conceiving a range of designs and allowing forthem in the cost estimate as a contingency. During con-struction, if changes in the as-encountered ground con-ditions are observed, the designs may be modifiedaccordingly. Providing the contract contains sufficientflexibility, variations will not cause undue concern to ei-ther the Contractor or the designer.

The method was found to be particularly useful inthe railway construction in the Pilbara because of thefollowing reasons:

a. The short time frame within which many of the pro-jects had to be implemented. In some projects it wasplanned to construct several hundred kilometres ofheavy-duty railway track within 3 years from com-mencement of a feasibility study.

b. The cost benefit of carrying out investigations priorto project approval. To investigate many tens of deepcuttings with boreholes in rugged country wouldhave taken a long time and cost several millionAustralian dollars—time and money that were notnecessarily available.

c. An awareness of the relative ineffectiveness ofany subsurface investigation technique when com-pared to full-scale excavations in finding what isactually in the ground. Whilst 20 m of good quality83.1 mm diameter oriented PQ3 core from onelocation can be extremely useful, it is very expensiveto acquire and it can never provide as much geo-logical detail as a 20 m deep 100 m long cutting atthe same location.

Fig. 11 Observational model ofa steep Pilbara BIF slope thatwas investigated by boreholesto assess stability. The modelindicates the factors that con-trol stability—the geometry andcharacteristics of geologicalstructures, the distribution ofEngineering Geological Mem-bers with different proportionsof BIF and shale, and thegroundwater conditions

83

Responsibilities during construction and operation

Adoption of the Observational Method during con-struction required engineering geologists to be on siteand to form part of the construction team. The respon-sibilities of those engineering geologists included thefollowing:

a. The application of the Observational Method.b. The redesign of cut slopes if required.c. Earthworks materials management including classifi-

cation of fill materials and the nomination of adequateborrow areas.

d. The location and identification of further borrow ifrequired.

e. Geohazard identification, especially of asbestiformminerals.

f. The documentation of as-encountered conditions forcontract management purposes.

g. The confirmation of the nominated Reference Con-ditions.

Upon completion of construction an as-built reportand maintenance manual was generally produced to:

a. Document the as-built conditions, particularly exca-vation stability, fill batter slopes, the finishing anddetailing of the works and surface drainage.

b. Identify any risks that the Owner should be aware ofwhen taking over the works e.g. areas where rockfallsmay occur or where culverts may be damaged byfloods.

c. Outline any maintenance and monitoring that waslikely to be required over the operating life of therailways.

Required resources

It is always difficult to estimate how long it will take tocarry out the type of studies that have been described inthis paper. Table 4 is based on a number of differentprojects and indicates in very broad terms the typicalnumber of experienced engineering geologists and theapproximate time taken for each of the different projectstages involved in building about 50–100 km of railwaytrack across relatively flat ground. Similar resources wererequired for about 10–30 km of track across more deeplyincised hilly country. Each of the engineering geologists

Table 4 Typical resource requirements for 50–100 km of railwayacross relatively flat country or 10–30 km of railway across deeplyincised hilly country

Stage Resources Comments

Pre-feasibility Two engineeringgeologists

40 man days (2·20)

Feasibility Two engineeringgeologists

100 man days (2·50)

Design One engineeringgeologist

60 man days (1·60)

Construction One engineeringgeologist

Duration of construction—300 man days (1·300)

Operation One engineeringgeologist

20 man days for commissioningreport (1·20)

Fig. 12 Evolutionary model ofPilbara slopes underlain by BIFand dolerite used to anticipateexcavation conditions. Themodel describes landformdevelopment controlled by jux-taposed Engineering GeologicalMembers

84

concerned had a minimum of 10 years experience andhad worked in the Pilbara on a number of occasions i.e.they were equivalent to someone who had attainedChartered Geologist status and had worked in a varietyof engineering projects in remote locations for over10 years.

In addition to these resources, the periodic involve-ment of other specialists was arranged to ensure the re-quired breadth of technical knowledge that was appliedto the project. At various times the geo-team included:

a. A very experienced engineering geologist to providetechnical reviews of the engineering geology studies.

b. A specialist blasting consultant.c. A specialist earthworks consultant.d. A specialist asbestiform minerals consultant.e. A specialist regional geology consultant.

All of these resources were brought together to gen-erate high-quality geological knowledge and to presentand communicate that knowledge to the engineers in-volved in the design, construction and operation of therailway. Figure 13 is an assessment of the rate at whichgeological knowledge was accumulated in several Pilbararailway projects using the resource levels indicated inTable 4 and adopting the total engineering geologyapproach.

Independent review

Independent formal review by experienced practitionersis generally acknowledged as being one of the mosteffective ways of ensuring the quality of studies relatingto ground engineering (Fookes 1997). The use of a for-

mal review was found to be extremely beneficial on theserailway projects, with two levels of review on some, oneat a local level and the other at an international level. Byincorporating these levels of review the quality ofthe engineering geological studies was maintained. Theinternational reviewer often reported directly to theOwner even though retained by the Engineer. Thisencouraged frank, objective discussions of the issues andavoided any commercial interests influencing reviewfindings.

Part 3: Engineering geology issues

The total engineering geology approach requires theinvolvement of engineering geologists in many differentaspects of investigation, risk management, design andconstruction supervision that relate to ground engineer-ing. Tasks involving routine data collection such aslogging test pits and boreholes are generally recognizedas being necessary by civil engineers managing projects.However, certain engineering geology issues requiringspecialized studies such as detailed aerial photo inter-pretation or observations of particularly relevant geo-features that were a considerable distance from therailway line were sometimes well outside the experienceof the engineering management. In such cases approvalto undertake the necessary studies was obtained once thecost benefit or potential for risk reduction were demon-strated to the engineering management. As engineeringmanagers were exposed to the benefits of the approachthey gradually began to appreciate the contribution thatengineering geologists could make to projects.

Geohazards

Identification and assessment of geohazards that couldimpact upon the construction and/or operation of anyproject is always a key investigation activity. Geohazardsthat have been identified in the Pilbara include largerockslides, karst, asbestiform minerals, collapsing soils,reactive clays, seismicity, flooding, concrete aggregatesoundness and settlement due to dewatering.

Rockslides

The occurrence of substantial rockslides in the Pilbaramay be anticipated because of the steep slopes, incisedtopography and gently folded bedrock containing weaklayers along bedding. A large rockslide has been docu-mented (Wyrwoll 1986) and several large rockslides(estimated volumes of the order of 50 million m3) havebeen identified in aerial photographs associated with oneparticular stratigraphic unit known as the MacCraeShale. Some of the rockslides have developed where thebedding planes daylight with dip angles as low as 9�.

Fig. 13 Comparison of knowledge acquisition using the totalengineering geology approach and the typical geotechnical engi-neering traditional approach for several railway projects in thePilbara. The assessed rates of knowledge acquisition are based onthe pooled judgement of the authors, who were intimately involvedin projects that used both approaches

85

During feasibility level investigations for one railwaythe possibility of initiating large scale instability byexcavating deep cuts in certain parts of the stratigraphicsuccession containing shale was identified. The highestrisk occurred where cuts ran parallel to strike andexcavations across dip slopes could undermine dippingstrata. With these very large rockslide risks a strategy ofavoidance was adopted due to the time and likely cost ofdeveloping engineered solutions and the residual levels ofuncertainty that would remain regarding the effective-ness of the proposed stabilization measures.

Karst

Karst landforms (Waltham and Fookes 2003) are neitherwell developed nor particularly common in the Pilbarabut the presence of sinkholes may be anticipated inparticular geological settings and karst features havebeen documented (Waterhouse and Howe 1994). Hazardzones are typically associated with many of the broadvalleys that are underlain by the Wittenoom Formation(which consists of metamorphosed dolomite, dolomiticpelite, chert and volcaniclastic sandstone) and whichcontain thick infills of Tertiary detritals, includinglacustrine limestones and associated calcretes. In theseareas individual sinkholes could sometimes be seen on1:40000 aerial photographs and were identifiable on low-level 1:5000 aerial photographs and during helicopterinspections. In one karst hazard zone, where ground-water had been pumped from a karst aquifer, severalsinkholes tens of metres in diameter have formed withinthe last few decades.

Where individual sinkholes were identified duringinvestigations they were avoided by the railways duringearly route-selection studies. However, there remained arisk that sinkholes could develop in the future withinkarst-hazard zones and impact on a railway crossing thezone, although this was considered rather unlikely as therate of development of sinkholes was thought to becurrently very low. The hazard was engineered using riskmanagement techniques originally developed for land-slide risk assessment (Anon 2000). The risks werequantitatively assessed as the mathematical product ofhazard and consequence and expressed in terms of an-nual probabilities of death, loss of property and loss ofearnings. The hazard was considered as the annualprobability of the process occurring and the conse-quences were related to a series of conditional proba-bilities that the elements at risk would be impacted, thus:

RðD=PÞ ¼ P ðHÞ � P ðS:HÞ � P ðT:SÞ � V � E

whereR(D/P) is the risk, which is the annual probability of

death or the annual loss of property value, annual lossof life expressed in dollar terms, or the annual loss ofearnings due to downtime.

P(H) is the annual probability of the process occurring(expressed as a rate of sinkhole formation of a certainsize per unit area)

P(S:H) is the probability of spatial impact of the pro-cess on the elements at risk, given the event i.e. theprobability that the zone of influence of the process,which is the sinkhole development, coincides with thearea or footprint occupied by the element at risk.

P(T:S) is the temporal probability that an element atrisk is in the footprint at the time of impact e.g. theprobability of a train (that may be occupied by people)being present in the footprint.

V is the vulnerability, which is the probability of lossof life of the individual given the impact or the propor-tion of property value lost or damaged or the proportionof time lost.

E is the element at risk e.g. the value of the property,probable number of people present or profit from con-tinuous production.

This information was used to calculate the riskfor various scenarios to make informed decisions aboutthe cost benefit of engineering options. Those optionsranged from relocation of elements at risk awayfrom hazard zones to managing activities that couldtrigger karst collapse such as groundwater extraction orconcentrated infiltration. The engineering optionsincluded monitoring, to detect early signs of subsidencedeveloping.

Fibrous materials

Due to the regional metamorphism, fibrous amphibolesoccur within both the BIF and dolerite, usually asactinolite. At some stratigraphic horizons the particularchemistry of the volcaniclastic host sequences and thepresence of inter-bed shearing associated with regionalfolding has produced fibrous amphiboles such as cro-cidolite (blue asbestos). Inhalation of blue asbestos fibresduring mining and processing has caused the deaths ofmany workers in Western Australia and one of theworld’s most notorious crocidolite mines was developedin the BIF sequences at Wittenoom Gorge in the centralPilbara (Trendall and Blockley 1970; Fetherston andBrown 1990; Western Australia Mining OperationsDivision Pamphlet undated).

Due to the potential for risk to the workforce strin-gent protocols were established in order to manage thehazard. The protocols involved the identification of thehazard and the adoption of various levels of protectiveclothing together with construction procedures such aswatering to keep down dust, covering spoil and workingin sealed cabs, in order to reduce any exposure toharmful fibrous minerals to acceptable levels. Regular,open communication of the nature and occurrence of thehazard to the workforce was found to be a particularlyeffective way of managing the risk.

86

Collapsing soils

The presence of recent deposits of loose fine alluviumand loose wind-blown silts associated with flash floodsand aeolian transport processes is to be anticipated fromthe cyclonic semi-arid environment and limited vegeta-tion cover. Such materials can collapse upon inundationafter loading (Cooke 1986; Waltham 1994) and couldcause embankments to settle; they are also difficult towet and compact. The materials were recognized by acombination of mapping and test pitting and were usu-ally less than about 300 mm deep. During constructionthese fine dusty surface materials were removed frombelow the embankment footprint. Where loose silts andsands occurred as lenses within deeper soil profiles theycould not be economically removed. In such cases set-tlements were estimated to be neither catastrophic norexcessive i.e. perhaps a maximum of a few hundreds ofmillimetres. The possibility of such settlements occurringwas generally taken into account in design by planningfor re-ballasting of the track.

Reactive clays

The presence of reactive clays, i.e. moisture-sensitiveclays, is to be anticipated in soils forming part of, orderived from, deep weathering profiles developed onbasic rocks such as metabasalts and dolerites (Anon1997); the presence of reactive clays should also beanticipated in soil profiles developed in arid and semi-arid environments (Cooke and Warren 1973). Reactiveclays in the soil profile cause large volume changesduring the annual cycle of extreme wetting and drying inthe Pilbara and create characteristic landforms. Theseareas are known as gilgai and may be identified by theirdistinctive surface forms (mounds several tens of centi-metres high, deep polygonal desiccation cracks and‘‘crabholes’’) during aerial photo interpretation andsurface mapping. Figure 10 illustrates different types ofgilgai developed in certain geomorphological settings.

Deep gilgai were generally avoided by rail alignmentsas the problems associated with crossing them can bequite severe. Where older railway lines have crossed suchareas, progressive re-ballasting of formation that has‘‘disappeared’’ into soft surface materials has led to anaccumulated ballast thickness of over one metre.

Flood

The typical location of the railway corridor within thewide valleys of the central Pilbara is on the large collu-vial/alluvial fan systems. The fans have developed wheredrainage lines that cut through the strike ridges of bed-rock debouch (see Figure 14) and the distal portions of-ten coalesce to form a relatively uniform surface thatslopes gently down to the flood plain. Railways con-structed across this sloping surface can minimise earth-

works and avoid the major flood hazard and the reactiveclays within the floodplain deposits. However the semi-arid environment coupled with the propensity forcyclonic rainfall events provides potential for large-scaleflash flooding in the channels incised into the fan systems.During the time that there have been railways in thePilbara, floods have washed out many sections ofembankment and levees have been constructed to protectcertain sections of railway threatened by mobile channelswithin fan complexes. Within the design lives of theexisting railways there is the possibility of massivedestructive events and channel avulsion (a sudden changein the course of a river), when many culverts would beblocked by the movement of gravels and overwhelmed.The paucity of hydrographic data means that floodrecurrence intervals are difficult to assess and a generalappreciation of the geomorphological setting and chan-nel widths of the various drainage systems could be of usein design. However, the economics of railway construc-tion are such that culverts are generally designed simplyto cater for 25 year floods and the potential effects oflarger events and general uncertainty in the design aretreated as an acceptable risk by the Owners.

Settlement due to dewatering

The need for construction water in this semi-arid envi-ronment has led to groundwater extraction. The most

Fig. 14 Conceptual model of alluvial/colluvial fan systems in thePilbara. The model describes the relationships between differentEngineering Geological Formations in terms of their relative ageand depositional environment and may be used to understand thespatial distribution of flood hazard zones

87

prospective areas for groundwater extraction are valleyswith a thick alluvial infill containing cavernous calcrete,which are often underlain by cavernous dolomite bed-rock. Dewatering such areas can lead to sinkhole for-mation, which has been discussed earlier, but dewateringof thick sequences of superficial deposits can also lead togeneral settlement. The settlements have been estimatedto be a maximum of a few hundreds of millimetres andare probably a lot less. On one railway line the potentialimpact of such settlements was taken into account in thedesign by planning for re-ballasting of the track.

Concrete aggregate soundness

The presence of cryptocrystalline silica within the silcreteduricrusts and the BIF bedrock provides potential foralkali silica reactivity. The deep weathering profiles canresult in the presence of clay minerals and micro-crackswithin apparently sound rock at considerable depthsbelow the surface. As a consequence, the aggregate forconcrete was carefully selected to avoid incorporatingsuch materials and generally was quarried from freshdolerites, which have performed well in concrete.

Seismic hazard

The stable cratonic area of the Pilbara is subject to rel-atively low levels of seismic hazard typically with 500-year return events of 0.16 g. Seismicity is not a threat torailway formations but it is relevant to the design of highbridges where the railways cross deeply incised creeklines.

Design of cuts and fills

Railways involve many cuts and fills within linear cor-ridors and the ground engineering component of thedesign largely relates to the performance of those cutsand fills.

Cut stability

With the exception of those areas where large-scale rockslope stability had been identified, the cut batters forthese railways were all designed on the basis of ‘‘prece-dent’’ using knowledge of the geological conditions de-rived from a geological model of the excavation. Table 5indicates the precedent designs that were prescribed forsoil and rock masses of varying quality. The precedentdesigns were based on experience of the past perfor-mance of railway cuts in the Pilbara and similar cuts onother projects throughout the world.

In Table 5 the rock quality descriptors were not de-rived using rock mass classifications systems (e.g. Bie-niawski 1989) but instead were related in a more holisticmanner to the Reference Conditions e.g. fresh massive

dolerite is typically a good quality rock mass whereasweathered shaley BIF is generally a poor quality rockmass. Where adversely oriented defects were encoun-tered that had not been identified prior to construction,local modifications were made to the precedent slopedesigns using the Observational Method. A great deal ofattention was paid to the presence of specific geologicalconditions identified by mapping during excavation andthere was little reliance on the collection of discontinuitystatistics, because it is usually the presence or absence ofspecific geological structures that will determine stability.Some of the conceptual failure mechanisms associatedwith different Reference Conditions are illustrated inFig. 15.

Blasting

Blast designs were developed by a blasting specialistworking with the engineering geologists. The objectivesof the designs were to maximize safety and minimize theoverall cost to the project. Particular emphasis wasplaced on producing stable cuts and developing sufficientcomminution of the material from excavations for usewhere appropriate as fill. Precise ground conditions weredifficult to anticipate in detail owing to varying rockmaterial strength and rock mass properties. However thegeological models of each excavation developed duringthe investigations were supplemented by mapping duringconstruction and used to anticipate the conditions forindividual blasts. This directly affected the blast designwith, for instance, the adoption of pre-split blasts in themore massive dolerite, CIDs and blocky BIF rocks and

Table 5 Precedent designs for cut batters

Mass descriptor,Reference Condition

Design batter (for cutsup to 5 m highin superficial deposits and upto 20 m high in bedrock)

Loose sands, silts, clays,fine grained soils, someextremely weathered bedrock

1 v:1.5 h

Dense sands and gravels, coarsegrained soils, duricrusts,some extremely weatheredbedrock

1 v:1 h

Weathered rock with somesoil zones, jointed poorerquality rock, duricrusts,some shaley BIF,some extremely weatheredbedrock

1.5 v:1 h

Jointed medium quality rock,some shaley BIF, CIDs,some blocky BIF, someblocky bedrock

2 v:1 h

Massive good quality rock,some blocky bedrock,some blocky BIF

3 v:1 h to 4 v:1 h

88

production blasts with post-blast trimming in the shaleyBIF.

Accurate setting out, drilling, loading and detonationwere critical to the success of any blast and therefore thepreferred drill rigs were track-mounted high perfor-mance hydraulic rigs with good angled drilling capacity.Ideally, the collars of blastholes were surveyed prior tocharging in order to make sure that they were exactly inthe right position and the amount of charge introducedto each hole was carefully monitored. Experienced drilland blast crews were essential.

In addition, the early blasts on projects were generallyreviewed to ensure that good practices were being fol-lowed, as considerable damage can occur in the early

stages of a project before blasting crews gain experiencewith the local site conditions. In some of the locations,blasts were close to sensitive infrastructure and devicessuch as blast mats were employed.

Embankments

Some of the larger railway embankments were up to45 m high and were constructed of rockfill. The rockfillcomplied with the typical grading specification forrockfill as used in embankment dams (Fell et al. 1992)and were constructed in compacted lifts of 750–1000 mmthickness with about 10% by weight of water addedduring placement. A stable construction surface after

Fig. 15 Observed and antici-pated failure mechanisms forcut batters related to selectedReference Conditions

89

watering and successful trafficking by heavy trucksdemonstrated that the wheel loads were being carried bya free draining, structural skeleton of rock-to-rock con-tacts.

The design slopes of high rockfill embankmentswere also based on precedent with angles of 1.5 hori-zontal to 1 vertical or 34�. A settlement value of 1% forthe embankment height was considered appropriate(ibid). With large embankments this settlement was al-lowed for by cambering and by increasing the crest widthso that an increased thickness of ballast could be built if

it was necessary to compensate for any long-term set-tlement.

Environmental issues

Engineering geologists were generally not responsible forenvironmental matters but typically interacted withenvironmental managers providing them with informa-tion and advice, particularly with respect to borrowmanagement.

Fig. 16 Specified grading enve-lopes for the main constructionmaterial types related to thetypical Reference Conditionsdetailed in Table 2

90

Construction materials

The identification and systematic management of con-struction materials was one of the most importantactivities of engineering geologists on these rail projects.

Borrow investigations

The overall aim of the borrow investigations was to en-sure that sufficient borrow was identified to allow theContractor to carry out his works unhindered by anyrestrictions on the availability of borrowmaterial. A largeproportion of the borrow materials were located prior toconstruction and during construction additional borrowmaterials were located, usually by the engineering geol-ogists in conjunction with earthworks supervisors. Thequantities of the borrow materials were described usingcarefully chosen reporting terminology (Berkman 1989):

a. When only one or two test pits had been excavated ina borrow area of several thousand square metresprobable reserves were defined. This was consideredsufficient information for contractors to tender on.As a rule of thumb, probable reserves of about doublethe estimated fill requirements were identified prior toconstruction.

b. At the start of construction borrow areas wereinvestigated further, usually with a grid of test pits at50–100 m centres and proven reserves were defined.

Despite this cautious approach once the borrow pitswere worked it was usual for shortfalls in borrow tobe identified. This was due to variations in overburdendepth, incursions of unsuitable materials and environ-mental constraints such as a need to keep borrow pitsshallow to allow drainage and prevent water ponding.Borrow pits were especially prone to extraction diffi-culties due to the large size of much of the typicalplant e.g. Caterpillar bulldozers (D10 and D11) andCaterpillar open bowl (631) and elevating (633)scrapers, and the need to extract borrow economically.Consequently, during construction a large proportionof the engineering geologist’s time was spent in iden-tifying additional borrow. This was essential to allowthe work to proceed as quickly as possible and tominimise any opportunities for contract claims.

Material quality

Four main types of construction materials were used forrailways in the Pilbara. Figure 16 provides their specifiedgrading envelopes and indicates the typical ReferenceConditions from which the specified materials were won.

Sub-ballast capping

The embankments were capped with a thin layer of rel-atively fine-grained granular material (see Fig. 16) called

‘‘sub-ballast capping’’ that could be rolled to a smoothfinish to seal the embankment against water ingress andto allow the concrete sleepers to be placed on a surfacewith minimal protuberances, prior to ballasting, to avoiddamage. The sub-ballast was generally compacted to amodified dry density of 95%. A slight fines content wasrequired to allow satisfactory compaction and to developa cohesive non-erodable surface. Sub-ballast cappingwas carefully selected and generally came from the distalportions of alluvial and colluvial fan systems.

Type 2 fill

‘‘Type 2 fill’’ was bulk-engineered fill obtained fromborrow for embankment construction and was generallycompacted according to a performance specification, i.e.a modified dry density of 95% was achieved by the use ofcontrolled amounts of water and sufficient passes ofcompacting equipment. Type 2 fill materials were ob-tained from colluvial or alluvial gravels from nominatedborrow pits. Usually this material was a sandy gravelwith some clay and was obtained from the mid-sectionsof the alluvial and colluvial fan systems.

Type 3 fill

‘‘Type 3 fill’’ was typically a rockfill that was obtainedfrom cuttings and was placed according to a methodspecification with a maximum allowable particledimension of 750 mm. The water was added duringconstruction at a prescribed amount, not to sluice therock nor to wash the fines, but to soften the rock–rockcontacts to a degree and allow good compaction. Com-paction trials on rockfill were carried out to assess theparticular performance of different stratigraphic units.Very little of the rockfill suffered from problems ofdurability and the suitability of the materials in this re-spect was assessed by visual inspection. Some variationson type 3 fill were:

a. ‘‘Fine Type 3 fill’’ which was generally derived fromclosely bedded shaley BIF and could be placed with aperformance specification, i.e. to a modified dry den-sity of 95% at ±2% of optimum moisture content.

b. ‘‘Graded Type 3 fill’’ material which was a dirtyrockfill that was placed according to a method spec-ification.

Type 4 fill

‘‘Type 4 fill’’ was generated from cuttings in extremelyweathered shale or dolerite bedrock and was a materialthat might be prone to erosion or be expansive. Thismaterial was encapsulated so that it was not within 2 mof any of the embankment surfaces. It was generallyplaced with a performance specification.

91

Ballast

During the construction of the earlier railways granitesand gneisses were used for ballast and these rock typeshave not performed well due to excessive abrasion, andin some cases breakdown of the clasts. The poor per-formance seems to be associated with a variety of geo-logical factors such as a slight foliation, the presence ofsignificant amounts of mica in the rock material, the verycoarse grain size and resultant persistent feldspar cleav-age planes, and, importantly, penetrative weatheringwith associated clay mineral growth, micro-crack devel-opment and rock strength reduction. In contrast, ballastsourced from unweathered metabasalts and metadole-rites has performed well, as such rock types have pro-duced very high strength, equidimensional fragmentswhich are resistant to abrasion.

Contract setting

Most of the projects that the authors were involved inwere constructed using the traditional engineering con-tract i.e. involving an Owner, an Engineer and a Con-tractor. A schedule of rates contract was found to be thebest way of constructing the railways as it led to anequitable sharing of the relatively high risks associatedwith the large ground engineering component (seeFig. 17). A schedule of rates contract also provided thegreater flexibility that was required when design changesneeded to be made during construction due to theadoption of the Observational Method.

The contracts were designed to ensure that the loca-tion of borrow materials was the responsibility of theEngineer. By making the Engineer responsible forlocating the materials to be used to construct the railway,

rather than the Contractor, the opportunity for a claimagainst the Owner for unforeseen conditions by theContractor was reduced.

Payment for ‘‘rock’’ versus ‘‘common’’ excavation,which is another traditional potential source of claims,was related to production criteria for standard heavyexcavation machinery. The details of the clauses varied,but in essence common excavation was specified asmaterial that could be excavated by a nominated ma-chine at greater than a certain rate and rock excavationwas specified as any material not classed as common.This approach was found to be an effective basis forpayment and there were minimal disputes associatedwith predictions of excavatability.

Discussion and conclusions

The total engineering geology approach described in thispaper is based on a series of practical strategies that weredeveloped on real life heavy civil engineering projects tomeet demands for high quality information and to sup-port well informed, considered engineering decisions.The strategies were also developed in response to thesobering observation that many projects that sufferimplementation failure, and end up in litigation, do sobecause of a lack of investment in geological under-standing.

The approach is based on developing a thoroughunderstanding of the geology and geomorphology of theproject area at a very early stage in the project and placesgreat emphasis on strategies such as:

a. Staged investigations,b. The definition of investigation objectives,c. Investigating to answer questions,d. Geo-mapping,e. Establishing Reference Conditions,f. The development of different types of Geological

Models, andg. The application of the Observational Method.

Application of the approach requires the involvementof engineering geologists in decision making throughoutthe project life and the examples provided illustrate theimportant nature of those decisions. Observation, inter-pretation and anticipation allow information to be gen-erated that is relevant to a wide range of the fundamentalcomponents of project implementation such as feasibilityassessment, design functionality, cost estimates, contractdocumentation, construction productivity and opera-tional performance. The effectiveness with which thesecomponents of the project are engineered can all be re-lated to the adequacy of geological information, which isthus absolutely critical to the success of the project. Therequirement that engineering geologists should form partof the engineering team throughout the project lifemakes it essential that adequate numbers of experienced

Fig. 17 The relative risks associated with different types ofcontract, based on Eddleston et al. (1995) and originally presentedin CIRIA (1978)

92

individuals should be employed on any project and it isthis expenditure on people and thinking time that rep-resents the investment in geology that is advocated.

For the approach to be most successful it must beadopted at the beginning of a project. If, for instance, theObservational Method is implemented in response tosome disaster then it is unlikely that the other necessarymanagement structures, such as a flexible contract, willexist, and the situation is unlikely to be improved. Aboveall else, good communication is fundamental to the ap-proach. This requires engineering geologists who areknowledgeable of the engineering issues and engineeringmanagers who can appreciate the worth of geologicaladvice.

The advantages of the approach advocated in thispaper are that geological knowledge is acquired morequickly and more cost effectively and thus better deci-sions can be made about ground engineering issues,which are often a principal source of project risk.Additionally, there is a more equitable and transparentsharing of ground engineering risks between the partiesinvolved in the project, which leads to less conflict andeasier project implementation. A disadvantage of theapproach is that greater flexibility and management ef-fort is required as the engineering manager must takeoverall responsibility for not only implementing the

project but for ensuring that the complex geologicalimplications of the project are overtly considered.Training of engineering management in the applicationof the approach is usually necessary for it to be suc-cessful as it is often outside their experience, and someinexperienced project managers will be uncomfortablewith the greater involvement of engineering geologiststhat is required.

In general, the approach is not restricted to therailways. Other linear structures that can traverse wide-ranging geological conditions such as roads and pipe-lines, and projects involving extensive ground engineer-ing such as hydro-electric schemes, have been builtefficiently using this approach. It is suggested that thetotal engineering geology approach could be used to greateffect in all ground engineering, if only it was appreciatedthat it is a practical philosophy that aids and promotessuccessful engineering in the face of geological uncer-tainty.

Acknowledgments The authors wish to acknowledge the supportand encouragement received during the development of the ap-proach described in this paper from many colleagues who work for,or consult to various mining and engineering companies, in par-ticular Steve Brice, Tim Hagan, Doug Johannes, Dick Jupp, IanLewis, Doug McInnes, John McMeekan, Greg O’Rourke, PeterStanes, David Swifte, David Watkins and John Wood.

References

Anand RR, Paine M (2002) Regolith geol-ogy for the Yilgarn Craton, WesternAutralia: implications for exploration.Aust J Earth Sci 49:3–162

Anon (1997) Tropical residual soils. In:Fookes PG (ed) A geological societyengineering group working party re-vised report, The Geological Society ofLondon, 184 pp

Anon (2000) Landslide risk managementconcepts and guidelines. AustralianGeomechanics Society sub-committeeon landslide risk management. AustGeomech 35:49–92

Beckman GG, Hubble GD, Thompson CH(1970) Gilgai forms, distribution andsoil relationships in north-eastern Aus-tralia. In: Symposium on soils and earthstructures in arid climates, Adelaide.The Institute of Engineers, Australia,pp 88–93

Bieniawski ZT (1989) Engineering rockmass classifications. Wiley, New York,251 pp

Bowler JM (1982) Aridity in the lateTertiary and Quaternary of Australia.In: Barker WR, Greenslade PJM (eds)Evolution of the flora and fauna ofAustralia, paper 4. Peacock Publica-tions, Adelaide, pp 35–45

Berkman DA (1989) Field geologists’manual, monograph 9. AustralasianInstitute of Mining and Metallurgy,382 pp

Campana B, Hughes FE, Burns WG,Whitcher IG, Muceniekas E (1964)Discovery of the Hamersley irondeposits (Duck Creek–Mt Pyrton–MtTurner Area). Aust Inst Mining MetallProc 210:1–30

CIRIA (1978) Tunnelling—improved con-tract practices, construction industryresearch and information associationreport 79, CIRIA London.

Clayton CRI (2001) Managing geotechnicalrisk. Thomas Telford Publishing, Lon-don, 80 pp

Construction Industry Committee (1987)Guidelines for the provision of geo-technical information in constructioncontracts, The Institution of Engineers,Australia, 19 pp

Cooke RU (1986) Hot dry lands. In: Foo-kes PG, Vaughan PR (eds) A handbookof engineering geomorphology. SurreyUniversity Press, Guildford, pp 109–124

Cooke RU, Warren A (1973) Geomor-phology in deserts. B T Batsford, Lon-don, 394 pp

Dearman WR (1991) Engineering geologi-cal mapping. Butterworth–Heinemann,Oxford, 387 pp

Eddleston M, Murfin RE, Walthall S (1995)The role of the engineering geologist inconstruction. In: Eddlestson M, Walt-hall S, Cripps JC, Culshaw MG (eds)Engineering geology of construction.Geological Society Engineering Geol-ogy Special Publication 10:389–401

Essex RJ (1997) Geotechnical baseline re-ports for underground construc-tion—guidelines and practices.Prepared by the Technical Committeeon Geotechnical Reports of theUnderground Technology ResearchCouncil. American Society of CivilEngineers, 40 pp

Fell R, MacGregor P, Stapledon D (1992)Geotechnical engineering of embank-ment dams. Balkema, Rotterdam,675 pp

Fetherston JM, Brown IM (1990) Industrialminerals. In: Geology and mineral re-sources of Western Australia, Memoir3. Geological Survey of Western Aus-tralia, pp 669–679

93

Fookes PG (1969) Geotechnical mappingof soils and sedimentary rocks forengineering purposes with examples ofpractice from the Mangla Dam Project.Geotechnique 19:52–74

Fookes PG (1997) The first Glossop lecture,geology for engineers: the geologicalmodel, prediction and performance. Q JEng Geol 30:293–424

Fookes PG, Baynes FJ, Hutchinson JN(2000) Total geological history: a modelapproach to the anticipation, observa-tion and understanding of site condi-tions, Invited Paper, Geoeng 2000Conference. Technomic Publishing,Melbourne, pp 370–460

Fookes PG, Baynes FJ, Hutchinson JN(2001). Total geological history: amodel approach to understanding siteconditions. Ground Eng 34(3):42–47

Harding C (2004) Site investigation and siteconceptual models. The link betweengeology and engineering. Advances ingeotechnical engineering: the Skemptonconference. Thomas Telford, London,2:1304–1315

Hocking RM, Cockbain AE (1990). Rego-lith. In: Geology and mineral resourcesof Western Australia, Memoir 3. Geo-logical Survey of Western Australia, pp591–601

Joyce B, Ollier C (2003) The Pilbara regionof Western Australia: landscape andheritage. Aust Geol 127:27–30

Killick M, Churchward M, Anand RR(2001). Regolith—landscape evolutionand mineralisation in the Hamersleyiron province. Explor Mining Res News13:5–7

Kneeshaw M (2000) Guide to the geologyof the Hamersley and North East Pil-bara iron ore provinces. BHP Iron Oreinternal company report

Knill JL (2003) Core Values: the First HansCloos Lecture. Bull Eng Geol Environ62:1–34

Maxwell B (1994) Influence of horizontalstresses on gilgai landforms. ASCE JGeotech Eng 120:1437–1444

McGowran B, Li Q (1998) Cainozoic cli-matic change and its implications forunderstanding the Australian regolith.In: Eggleton AA (ed) State of the Reg-olith. Geological Society of AustraliaSpecial Publication 20:86–103

Mining Operations Division Pamphlet(undated) Asbestos Management inMining. Department of Minerals andEnergy, Western Australia

Moores EM, Twiss RJ (1995) Tectonics. WH Freeman & Co, New York, 415 pp

Muir Wood A (2000) Tunnelling—man-agement by design. E & FN Spon,London, 307 pp

Newman TG, Allen RG, Mortimore RN(2003) Geological modelling, theobservational method and managementof ground hazards: Ramsgate Harbourapproach road tunnel. Q J Eng GeolHydrogeol 36:35–50

Peck RB (1969) Ninth Rankine lecture:advantages and limitations of theobservational method in applied soilmechanics. Geotechnique 19:171–187

Pettifer GS, Fookes PG (1994) A revisionof the graphical method for assessingthe excavatability of rock. Q J Eng Geol27:145–164

Phipps PJ (2001) Terrain systems mapping.In: Griffiths JS (ed) Land surface eval-uation for engineering practice. Geo-logical Society Engineering GeologySpecial Publications, London 18:59–61

Pisarevsky SA, McElhinny MW (2003)Global palaeomagnetic data basedeveloped into its visual form. EOSTransact 84:192

Ramanaidou ER, Morris RC, Horwitz RC(2003) Channel iron deposits of theHamersley Province, Western Australia.Aust J Earth Sci 50:669–690

Summerfield MA (1991) Global geomor-phology. Longman, London, 537 pp

Stapledon DH (1982) Subsurface engineer-ing—in search of a rational approach.Aust Geomech News 4:26–33

Thomas MF (1994) Geomorphology in thetropics: a study of weathering anddenudation in low latitudes. Wiley, NewYork, 460 pp

Trendall AF, Blockley JG (1970). The ironformation of the Precambrian Hamers-ley Group, Western Australia, withspecial reference to the associated cro-cidolite. Geol Surv W Aust Bull119:174–254

Trendall AF (1990) Pilbara Craton—theHamersley Basin. In: Geology andmineral resources of Western Australia,Memoir 3. Western Australia Geologi-cal Survey, pp 163–189

Trendall AF (2000) The significance ofbanded iron formation (BIF) in thePrecambrian stratigraphic record, Geo-scientist 10:2–5

Twidale CR (1994) Gondwana (LateJurassic and Cretaceous) palaeosurfacesof the Australian craton. PaleogeogrPalaeoclimatol Palaeoecol 112 (1–2):157–186

Tyler IM, Thorne AM (1990) CapricornOrogen—structural evolution of thenorthern margin. In: Geology andmineral resources of Western Australia,Memoir 3, Western Australia Geologi-cal Survey, pp 223–230

Tyler IM (1991) The geology of the Sylva-nia inlier of the southeast HamersleyBasin. Geological Survey WesternAustralia, Bulletin, 138 pp

Varnes DJ (1974) The logic of geologicalmaps, with reference to their interpre-tation and use for engineering purposes.United States Geological Survey Pro-fessional Paper 837, United StatesGovernment Printing Office, Washing-ton, 48 pp

Waltham AC (1994) Foundations of engi-neering geology. Blackie Academic &Professional, London, 88 pp

Waltham AC, Fookes PG (2003) Engi-neering classification of karst groundconditions. Q J Eng Geol Hydrogeol36:101–118

Waterhouse JD, Howe PJ (1994) An envi-ronmental hydrogeological study nearMt Bruce, Pilbara region of WesternAustralia. In: Proceedings of the 25thcongress of the international associationof Hydrogeologists, Adelaide, Australia

Wende R, Nanson GC, Price DM (1997)Aeolian and fluvial evidence for LateQuaternary environmental change inthe east Kimberley of Western Austra-lia. Aust J Earth Sci 44:519–526

Wyrwoll KH (1986) Characteristics of aplanar rock slide: Hamersley Range,Western Australia. Eng Geol 22:335–348

94