research news (jan) no.33 15/2/01 11:17 am page 1 newsletter · research news (jan) no.33 15/2/01...

In this ISSUE:

➤ Nutrients from sediments: Implications for algal blooms in Myall Lakes 2

➤ More sources for gas and oil in Perth Basin: Study highlights potential for multiple petroleum systems 5

➤ The importance of the ‘backend’ to online delivery of geoscience information 10

➤ Minerals laboratory staf f develops new ICP—MS preparation method 12

➤ Bonaparte Basin:Geochemical characteristics of hydrocarbon families andpetroleum systems 14

➤ Regolith maps incorporatinghydrologic modelled attributescustomised for geochemicalexploration 21

ALSO INCLUDED

List of recent publications involving AGSO authors

RESEARCHGeoscience Australia

newsletter DECEMBER 2000 • Number 33

PP255003/0

0266

Research News (JAN) No.33 15/2/01 11:17 AM

Nutrients fromsedimentsImplications for algal blooms in Myall Lakes

D Palmer, DJ Fredericks, C Smith & DT Heggie

During April and October 1999, the Myall Lakes experiencedblue green algae blooms that persisted until August 2000.1

The bloom conditions had considerable impact on the localtourism and fishing industry. In response to communityoutcry and political pressure to find solutions to theproblem, the managing authorities developed monitoringand assessment programs in an attempt to discover thecause of the algal blooms. An important component of theassessment was to understand the nutrient dynamics andsediment–water interactions within the Bombah Broadwater.

AGSO conducted an 11-day survey in Bombah Broadwaterwithin the Myall Lakes system, measuring fluxes of nutrientsfrom the sediments using benthic chamber instrumentation.The results showed that denitrification, a natural microbialprocess which removes nitrogen from estuaries as nitrogengas, was operating inefficiently. As a consequence, a highproportion of the nitrogen recycled in the sediments wasbeing returned to the water column as biologically availableammonia, potentially enhancing algal growth in the watercolumn.

The Myall Lakes region is on the central coast of New SouthWales approximately 280 kilometres north of Sydney. TheMyall Lakes drain a catchment approximately 780 square

kilometres in area, of which about 25 per cent is cleared andunder agricultural production; the remainder is relativelyundisturbed vegetation within state forest, national park oruncleared private land holdings.2

Bombah Broadwater (figure 1) is the southern most lakewithin the Myall Lake system. It is a relatively shallow, flat-bottomed lagoon, approximately 22 square kilometres in size.Bombah Broadwater receives the main freshwater input to thelake system (Myall and Crawford Rivers and Boolambayte Creekvia Boolambayte Lake) and drains out via the lower Myall Riverwhich flows south, approximately 20 kilometres into PortStephens. There is very little tidal flushing of the Myall Lakes asmarine water only moves up the lower Myall River into BombahBroadwater during extended periods of low rainfall.2

The distribution of sediment facies within the lake correspondsclosely with water depth. The outer margins and a broad shoal inthe centre of the lake consist of medium to coarse sand,dominantly quartzose in composition. Areas greater than twometres in water depth are dominated by mud with a total organiccarbon content of around five to seven per cent.3

Benthic flux measurementsThe flux of nutrients and metabolites between the sediments andthe overlying water was measured using benthic chambers.4 Thechambers were deployed on the sediments of the lake andcaptured approximately nine litres of seawater. Data loggersrecorded dissolved oxygen concentrations both within theconfined chamber waters and in bottom waters outside the

AGSO Research Newsletter

December 2000, no. 33

Editor: Julie Wissmann

Graphic Designer: Karin Weiss

This publication is issued free ofcharge. It is published twice a yearby the Australian Geological SurveyOrganisation. Apart from any usepermitted under the Copyright Act1968, no part of this newsletter is tobe reproduced by any processwithout written permission. Requestsand enquiries can be directed toAGSO’s Chief Executive Officer at theaddress shown below.

Every care is taken to reproducearticles as accurately as possible, butAGSO accepts no responsibility forerrors, omissions or inaccuracies.Readers are advised not to rely solelyon this information when making acommercial decision.

© Commonwealth of Australia 2000

ISSN 1039-091X

Printed in Canberra by National Capital Printing

Australian Geological Survey OrganisationGPO Box 378, Canberra ACT 2601cnr Jerrabomberra Ave & Hindmarsh DrSymonston ACT 2609 AUSTRALIA

Internet: www.agso.gov.au

Chief Executive OfficerDr Neil Williams

SubscriptionsDave HarrisPhone +61 2 6249 9333Fax +61 2 6249 9982E-mail [email protected]

Editorial enquiriesJulie WissmannPhone +61 2 6249 9249Fax +61 2 6249 9984E-mail [email protected]

AGSO Research Newsletter isavailable on the web atwww.agso.gov.au/information/publications/resnews/

2 AGSO Research Newsletter DECEMBER 2000


DECEMBER 2000 AGSO Research Newsletter 3

chamber. Samples were drawn from within the chamber at predeterminedintervals and analysed for dissolved inorganic nutrients (NOx

-, NH4+, HPO4

2-,SiO4

2-), pH, TCO2, alkalinity, N2 and Cs concentrations. The flux of nutrientsand metabolites across the sediment water interface was determined from therate of change in concentration within the chamber, during the course of eachincubation.

Benthic chambers were deployed at three sites as selected by the NewSouth Wales Department of Land and Water Conservation (figure 1). One sitewas located on the sand facies (site 3); the other two were on the mud facies (sites 1 and 2).

Sediment denitrificationNitrogen is delivered to coastal lake and estuarine environments in dissolvedand particulate forms. Nitrogen added to the estuary is either captured byprimary producers—including phytoplankton, various seagrass species andmangroves—or is flushed out to the sea. In most Australian barrier estuaries(those separated from the ocean by a sand barrier) most nitrogen is trappedor/and recycled within the coastal lake or estuary. The dominant, naturallyoccurring, self-cleansing mechanism for these lakes and estuaries isdenitrification. Denitrification is a bacterially-mediated process that occurswithin sediments. It converts nitrates, and nitrites generated from thebreakdown of organic matter, into nitrogen gas, which is subsequently lost tothe atmosphere. The identification of this denitrification process and theefficiency to which it is occurring are key sedimentary indicators ofenvironmental condition.

Denitrifying bacteria are ubiquitous in nature and require an organicsubstrate, a supply of nitrate, and a sub-oxic to anoxic environment or nichewithin the sediments for metabolism. When denitrifying bacteria are operatingefficiently, the majority of dissolved inorganic nitrogen (DIN) generated viathe breakdown of organic matter is converted to gaseous N2. However, whendenitrification is operating inefficiently, most DIN is returned to the watercolumn, thus remaining available for plant growth.

Denitrification efficiency, expressed as a percentage, was calculated using

Figure 1. Bombah Broadwatersite map

N Predicted is the calculated flux of DIN from the measured Total CO2

flux assuming a Redfield stoichiometry of 106C:16N.

Figure 2. Calculated denitrificationefficiency

0 1km

ML-1 ML-2

ML-3

Denitrification Efficiency = (N Predicted – DIN Measured) * 100

N Predicted

Calculated denitrificationefficiencies were greatest at site ML-3(average 79% ± 4%) (figure 2). Thissuggests that the sand facies at ML-3is very efficient at converting nitrogenfrom degrading organic matter intoN2. Calculated denitrificationefficiencies for mud facies sites, ML-1and 2 (figure 2) varied over a widerrange (30% to 78%), yet had aconsiderably lower average (38% ±9%). These efficiencies are low bycomparison to muddy sites withinother Australian estuaries such as thecentral portion of Port Philip Bay(~60% to 100%) and Wilson Inlet(~50% to 80%).

In Myall Lakes, the majority ofnitrogen recycled from organic matteris being returned to the water columnas ammonia at sites 1 and 2. Incontrast, the small amount of organicmatter being recycled at site 3 is

Map: AUSLIG



efficiently converting organic nitrogen into nitrogen gas (figure 3).It is important to consider the significance of low denitrification efficiency.

Recycling of nitrogen from sediments is known to have a non-linear impact onproductivity5— that is, a small decrease in denitrification efficiency may have adisproportionately large impact on primary production.

The impact of denitrification efficiency on the residence time of nitrogenin an estuary with limited flushing is illustrated in figure 4. It shows that theresidence time of nitrogen in the estuary increases rapidly when thedenitrification efficiency decreases below about 40 per cent. A similar effectwas found in the Port Phillip Bay Environmental Study where primaryproduction was predicted to increase rapidly when denitrification effecienciesfell below about 40% (equivalent to a doubling of the N load).5,6

ConclusionIt is difficult to assess the system-wide denitrification rate from the dataavailable for Myall Lakes. However, the limited measurements of sedimentdenitrification in Myall Lakes indicate that denitrification efficiency is low, atleast during the winter month of June at the mud sites. Furthermore, the datasuggest that Bombah Broadwater may be close to a state in which feedbackof labile nitrogen from the sediments may fuel plant growth. It is possible thatthe extended cyanobacteria bloom experienced in Myall Lakes over thesummer of 1999–2000 was sustained by poor sediment denitrification. Anyfurther decline in sediment denitrification is likely to result in more extensivephytoplankton production, though not necessarily cyanobacteria.

Figure 3. Schematic of nitrogen cycling in Bombah water

Figure 4. Relationship between residence time of nitrogen in an ideal estuaryand sediment denitrification. In this figure, residence time is defined as thenumber of times added nitrogen is recycled between sediments and theoverlying water column before 50% of it is lost as nitrogen gas.

References1. Carter G. 2000. NSW Dept Land &

Water Conservation, personalcommunication, Jun 12.

2. Atkinson G, Hutchings P, JohnsonM, Johnson WD & Melville MD.1981. An ecological investigationof the Myall Lakes region.Australian Journal of Ecology; 6: 299–327.

3. Thom BG, Shepherd M, Ly CK,Roy PS, Bowman GM & Hesp PA.1992. Coastal geomorphology andquaternary geology of the PortStephens–Myall Lakes area.Canberra: Australian NationalUniversity.

4. Berelson WM, Heggie DT,Longmore A, Kilgore T, NicholsonG & Skyring G. 1998. Benthicnutrient recycling in Port PhillipBay, Australia. Estuarine, Coastal& Shelf Science; 46: 917–34.

5. Murray AG & Parslow JS. 1999.Modelling of nutrient impacts inPort Phillip Bay: A semi-enclosedmarine Australian ecosystem.Marine & Freshwater Research; 50:597–611.

6. Heggie DT, Skyring GW, OrchardoJ, Longmore AR, Nicholson GJ &Berelson WM. 1999.Denitrification and denitrifyingefficiencies in sediments of PortPhillip Bay: Direct determinationsof biogenic N2 and N-metabolitefluxes with implications for waterquality. Marine & FreshwaterResearch; 50: 589–96.

Acknowledgment: This project wasjointly funded by the New SouthWales Department of Land and WaterConservation, New South WalesNational Parks and Wildlife Serviceand the Australian Geological SurveyOrganisation.

➤ Duncan Palmer, Petroleum &Marine Division, AGSO, phone+61 2 6249 9710 or [email protected]

➤ Dr David Fredericks, Petroleum &Marine Division, AGSO, phone+61 2 6249 9434 or [email protected]

➤ Craig Smith, Petroleum & MarineDivision, AGSO, phone +61 2 6249 9560 or [email protected]

➤ Dr David Heggie, Petroleum &Marine Division, AGSO, phone+61 2 6249 9589 or [email protected] "



Shale.1,7,8 Organic matter in thesesediments is considered to be thesource of gas, condensate and oil inthe Beharra Springs, Mondarra,Woodada, Dongara, Mount Hornerand Whicher Range fields of theonshore Perth Basin.8–11 In contrast,oil in offshore Gage Roads-1 isthought to originate from LateJurassic rift-related sediments of theYarragadee and/or ParmeliaFormations.8,12–14 Liquids from theGingin and Walyering Gas fields ofthe onshore Dandaragan Troughwere probably sourced from theCattamarra Coal Measures.8

ReservoirsLithostratigraphic units with reservoirpotential are widespread throughoutthe entire sedimentary succession.15,16

The Lower Permian sandstones of theIrwin River Coal Measures producegas of economic significance in theDongara field. Discontinuous thinsandstones in the CarynginiaFormation reservoir gas in theDongara field and on Beagle Ridge,while the Woodada gas field is foundin thick carbonates of the CarynginiaFormation. The Wagina Sandstoneproduces gas in the Dongara andMondarra fields. The best reservoirpotential is present in the UpperPermian Dongara Sandstone andBeekeeper Formation. Thesereservoirs, together with the high-grade reservoirs in the basal Triassicsandstone of the Kockatea Shale,contain the bulk of hydrocarbonsdiscovered in the basin. There areminor accumulations in thesandstones of the Lower TriassicArranoo Member (gas and oil inDongara and oil in Mount Horner).Several thin sandstone horizons ofthe Lower to Middle JurassicCattamarra Coal Measures produce oil from the Mount Horner field.

SealsRegional seals are provided by theCadda Formation and by someintervals within the Cattamarra CoalMeasures, but mainly by the thickand laterally extensive KockateaShale.15,16 Shales in the CarynginiaFormation may provide a seal to theIrwin River Coal Measures, orjuxtaposition of Kockatea Shale andintra-formational seals of theCarynginia Formation across faultboundaries can providecompartmentalisation ofhydrocarbon.16,17 Seals for thesandstone reservoirs within theCattamarra Coal Measures/EneabbaFormation are either intra-formationalor provided by the regional CaddaFormation, while the YarragadeeFormation is sealed by the ParmeliaFormation in the Dandaragan Trough.

More sources for gasand oil in Perth Basin Study highlights potential for multiplepetroleum systems

CJ Boreham, JM Hope, B Hartung-Kagi & BJK van Aarssen

Perth Basin has been intermittently explored for the last few decades,resulting in the production of gas and oil from several onshore fields.The bulk of known hydrocarbon reserves has been produced, however,and new ideas are needed for Perth Basin to contribute to Australia’spetroleum stock in the future. Notwithstanding this long explorationhistory, the accepted sources for gas have been based on minimalgeochemical data; even the generally accepted major Early TriassicKockatea Shale source for oil has been questioned recently.1 Toimprove understanding, carbon isotopic and biomarker analyses ofgases, condensates and oils have been analysed as part of AGSO’s Southand South-west Regional Project. The study has documented numerousoil families from Permian, Triassic and Jurassic sources and positivelyidentified, for the first time, both Permian and Triassic sources for gasin the Perth Basin.

Perth Basin is a deep, linear north–south trending trough extending morethan 1000 kilometres from Geraldton in the north to the south coast ofWestern Australia (figure 1). The basin covers an area of approximately

45 000 square kilometres onshore and 98 000 square kilometres offshore andcontains sediments of Permian to Cainozoic age. A generalised stratigraphy forPerth Basin is shown in figure 1.

The basin is bounded to the east by the north–south trending DarlingFault and this has been downthrown on its western side.2,3 The maindepocentre is the Dandaragan Trough, where up to 15 kilometres of Permianand Mesozoic sediments were deposited. The succession shallows to thenorth and west, where it is bounded by Beagle Ridge. To the south,Dandaragan Trough is separated from Bunbury Trough by Harvey Ridge.Offshore and to the north, the Abrolhos Sub-basin contains sediments of EarlyPermian to Late Cretaceous age. Offshore and to the south and west of thecity of Perth, the Vlaming Sub-basin contains about 10 kilometres ofCretaceous and Tertiary sediments.

The structural history of the basin is recognised as being very complex withnone of the existing models giving completely satisfactory explanations for alltectonic elements.4 There are considerable problems in accurately dating thePermian sections and this adds to the difficulties in reconstructing the basinhistory. Mory and Iasky recognise two major phases in the structural evolution ofPerth Basin related to the breakup of Australia and India.5 The first of these beganwith north–south extension in the Early Permian resulting in east–west trendingnormal faults and probable sinistral strike-slip faults along the Darling Fault. TheLate Jurassic extension and subsequent Early Cretaceous separation of GreaterIndia from Australia caused reactivation of these faults and major uplift anderosion. This second event was probably associated with increased heatflow.

Petroleum systemsThe onshore Perth Basin has yielded volumes of 4.2 billion barrels, 1.4 billionbarrels and 0.7 trillion cubic feet of oil, condensate and gas, respectively.6 Thebulk of these reserves have already been produced. The Dongara fieldcontains more than half the oil and gas reserves, while approximately 85 percent of the condensate is found in the Beharra Springs field.

SourcesPetroleum accumulations in the Perth Basin are believed to originate fromsources within the terrestrial source rocks of the Early Permian Irwin RiverCoal Measures and some marine mudstone source rocks of the Early PermianCarynginia Formation, Permian Wagina Sandstone and Early Triassic Kockatea



Present investigationGas compositionThe highest CO2 content is seven per cent in gas from Houtman-1 in theoffshore Northern Perth Basin, which is similar to the average composition ofAustralian natural gas,18 whereas Whicher Range-1 and Yardarino-3 have onlytrace amounts (figure 2a). A much better appreciation of the origin of CO2 isseen using the relationship between CO2 content and carbon isotopiccomposition of the CO2 (figure 2b). The strong relationship betweenincreasing CO2 content and enrichment in 13C is governed by the degree ofmixing of isotopically light thermogenic, organic-derived and isotopicallyheavy inorganic (mantle and/or igneous)-derived CO2. Very high N2 content isfound in Yardarino-3 while all other wells (figure 2a) are below the averageof 3.2 per cent for Australian natural gases.18

Maturity is the principal control on the composition of the gaseoushydrocarbons and is reflected in the strong relationship between the ratio%C1/%C1–C5 and the ratio %C2/%C3 (figure 3a). The bulk of the gases havehigh ratios suggesting relatively high maturities. This is supported by thecarbon isotopic composition of individual gaseous hydrocarbons (figure 3b),which suggests gas generation at vitrinite reflectance (VR) >1.1 per cent. Theconsiderable ‘scatter’ around the predicted maturity trend indicates that theisotopic composition is also governed by source effects.18 Figure 3c shows thecarbon isotopic composition of individual C1–C5 gaseous hydrocarbons inrelation to the range in carbon isotopes of Australian natural gases.18

Oil geochemistryOf the 10 oils analysed for C33-alkylcyclohexane (C33ACH; table 1), acharacteristic biomarker of the organic matter in the Kockatea Shale, relativelyhigh amounts were found in North Erregulla-1, Woodada-3 and Mt Horner-1,and in somewhat lower abundance in Erregulla-1. Although its presenceconfirms a major contribution from the Early Triassic Kockatea Shale, itsabsence is equivocal. On the other hand, carbon isotopes are one of the mostdiagnostic indicators of source in the Perth Basin (see below).

The results of the analysis of aromatic hydrocarbons for methylatednaphthalenes, methylated benzenes and higher plant-derived biomarkers arelisted in table 1. The naphthalene parameters TMNr, TeMNr and PMNr for alloils fall in or close to the centre when plotted in a ternary diagram (figure4).19 Since none of the samples deviates appreciably from the ‘maturitycentre’ there is no positive indication, within the limitations of the technique,of any significant in-reservoir mixing of oils of different maturities,biodegradation or migration contamination.

The oils from Dongara-4, East Lake Logue-1, Erregulla-1 and MountHorner-1, sourced predominantly from the Kockatea Shale, have relativelylow abundances of the land-plant markers retene and ip-iHMN.20,21 The highHPI for the oil from Woodada-3 suggests an additional input from

Figure 1. Tectonic elements, selected well locations and generalisedstratigraphy in the Perth Basin. Note: stratigraphy modified after Owad-Jonesand Ellis15 and Crostella25.

Reservoir unit Regional seal unit Source rock unit Oil Gas Oil and gas

140

205

250

300

435

490

160

180

200

220

240

260

280

420

0 100 km

114

116

PERTH

Bunbury

LeeuwinBlock

VasseS

helf

YilgarnBlock

NorthamptonBlock

TUR

TLE

DO

VE R

IDG

E

Harvey

Ridge

Bunbury T

rough

28

34

Batavia Arch

Edward’sIslandBlock

DongaraSaddle

Geraldton

30

32

Houtman 1

Ocean Hill 1

Warro 1

Walyering 1

Gingin 1

Gage Roads 1

Whicher Range 1

BEA

GLE

RID

GE

TR

OU

GH

DA

ND

AR

AG

AN

E. LakeLogue 1

Woodada 1

Mount Horner Field

Yardarino / Elegans

Beharra Springs Field

Dongara Field

allochthonous terrestrial organicmatter to the extensively marinedepositional environment of theKockatea Shale.

The oils from Gage Roads-1 andGingin-1 contain relatively abundantconifer-derived retene, consistent witha source from the Late JurassicYarragadee Formation and EarlyJurassic Cattamarra Coal Measures,respectively.8

The HPF for Whicher Range-1 isunexpectedly weak for oil derivedfrom a Permian land-plant source.The HPF for Walyering-2 is similar tomarine-sourced oil from NorthErregulla-4 but its high TeMBr mayindicate an additional input from aterrestrial source biased towards thelow molecular weight components(see below).

Gas-to-oil-to-source correlationThe most reliable approach forassigning source rocks for the variousnatural gases firstly involves defininggas-to-oil correlations and secondly,by using better understood oil-to-source correlations, extrapolating tothe desired gas-to-source correlations.The critical step relies on the utility ofthe n-alkane carbon isotope profile ofoil as a good diagnostic tool in oil-oiland oil-to-source correlations.8,22 The>C7+ n-alkane carbon isotopic ratiosfor representative oils from PerthBasin are shown in figure 5. A rangein δ13C of <2 ‰ for n-alkanes of thesame carbon number but fromdifferent oils is typical for variationsin organic facies from essentially thesame source rock interval. To unravelthe source of the gas, the positionand shape of an extended n-alkanecarbon isotope profile is used,involving the combination of thecarbon isotopic data for gas (C1, C2 C3,n-C4, n-C5) components and the >C7+

n-alkane carbon isotopes fromaccompanying oils and condensates(figure 5).

For East Lake Logue-1 andDongara field gas and oil there is afairly smooth trend (i.e. continuity)across the carbon number range C4 toC9, which defines the gas to oiltransition zone (figure 5). This trendsupports the idea that the gas and oilare genetically related, both generatedfrom the Early Triassic KockateaShale. Gases from the Beharra Springsfield, Indoon-1 and Woodada-6 arealso thought to have a majorcontribution from Early Triassic rocks.Oils from Mount Horner-1 (figure 5),North Erregulla-1 and Yardarino-1 areextremely isotopically light (depletedin 13C) compared with oils from olderPermian (Whicher Range-1) andyounger Jurassic sources (GageRoads-1 and Gingin-1).8 The rather



flat (constant carbon isotopes) profilefor the C15+ n-alkanes is typical ofmarine-sourced oil.23 A flat profile isalso characteristic of the lacustrinesource (Gage Roads-1), whileincreasing isotopic lightness withincreasing carbon number is typicalof a land plant source.8,23

The oils from Erregulla-1,Woodada-3 and Walyering-2 show‘intermediate’ n-alkane carbonisotope profiles (figure 5).8 The firsttwo oils are isotopically similar andmost likely from the same source.Summons et al. suggest an EarlyTriassic Kockatea Shale source, albeitfrom slightly different organic facies.8

0

10

-20

-10

Carbonateor biogenic

Inorganic (igneous and/or mantle)

Organic (thermogenic)

1 2 3 4 5 60

4

6

0

2

East

Lak

e Lo

gue

Beha

rra S

prin

gs

Wal

yerin

g

Gin

gin

Don

gara

Eleg

ans

Mon

dara

Indo

on

Woo

dada

Hou

tman

Whi

cher

Ran

ge

Yard

arin

oTable 1. Results from biomarker analyses

HPF

Sample HPI % % % TMNr TeMNr PMNr 136/ TeMBr DBT/ C33ACHret cad iHMN 137 1367

Dongara-4 0.27 3 96 1 0.82 0.75 0.60 1.10 0.64 0.10East Lake Logue-1 0.12 0 100 0 0.91 0.86 0.59 1.29 0.78 0.45Erregulla-1 0.07 0 100 0 0.75 0.80 0.70 1.45 0.49 0.15 +Gage Roads-1 0.67 70 25 5 0.53 0.56 0.41 1.23 0.74 0.14Gingin-1 0.18 38 62 1 0.74 0.71 0.63 1.16 0.75 0.12Mt Horner-1 0.10 0 100 0 0.76 0.75 0.61 1.42 0.48 0.20 ++North Erregulla-1 0.12 30 70 0 0.72 0.70 0.52 1.49 0.47 0.05 ++Walyering-2 0.13 32 68 0 0.83 0.85 0.74 1.26 0.73 0.28Whicher Range-1 0.13 9 90 1 0.67 0.65 0.58 1.10 0.70 0.30Woodada-3 0.72 12 85 3 0.67 0.70 0.55 1.37 0.51 1.07 ++

HPI : Higher plant index = (retene + cadalene + ip-iHMN) /1,3,6,7-TeMNHPF: Higher plant fingerprint%ret = retene / (retene + cadalene + ip-iHMN); % cad = cadalene / (retene + cadalene + ip-iHMN); %iHMN = ip-iHMN / (retene + cadalene + ip-iHMN)TMNr = 1,3,7-TMN / (1,3,7-TMN + 1,2,5-TMN)TeMNr = 1,3,6,7-TeMN / (1,3,6,7-TeMN + 1,2,5,6-TeMN)PMNr = 1,2,4,6,7-PMN / (1,2,4,6,7-PMN + 1,2,3,5,6-PMN)136/137 = 1,3,6-TMN / 1,3,7-TMNTeMBr = 1,2,3,5-TeMB / (1,2,3,5-TeMB + 1,2,3,4-TeMB)DBT/1367 = DBT / 1,3,6,7-TeMNC33ACH = C33 alkylcyclohexane

Figure 2. Plots showing a. average molecular percentage of CO2 and N2 and b. carbon isotopic composition of CO2

versus molecular percentage of CO2 for natural gases from Perth Basin.

The Woodada-3 oil was described as a ‘vagrant’—that is, it stood alonecompared with the other Perth Basin oils using statistical principal componentcluster analysis based on biomarker ratios and bulk carbon iotopes.24 Theadditional biomarker data in table 1 also support the unusual composition ofthis oil. If this interpretation is correct, then the Kockatea Shale source cangive rise to a wide isotopic variability in the same n-alkane (e.g. 3.5 ‰ for n-C15).

The gas isotope data for Woodada-6 indicate that the gas is from the morecommon organic facies of the Kockatea Shale source, indicating a rathercomplex charge history.

It is apparent that the carbon isotopic composition of the wet gascomponents in Elegans-1 (a later re-entry of Yardarino-1) is heavier (enrichedin 13C) compared to the other Perth Basin gases as well as to the shallower oilfrom the original well on the same site (Yardarino-1). This enrichment in 13C isattributed to a source effect and is consistent with either a Jurassic or Permiansource. The geological setting and regional maturation profiles indicate aPermian source for the gas.15



-30

-40

-50

Ethane Propane

100

95

90

85

80

75

0

1

2

3

4

1 2 3 4 -10 -5 0

Maturity

Biodegradation

Biogenicmethane

Waterwashing

Maturity

-20Methane

Hydrocarbon gas componentButane Butane Pentanei- i- i- Pentanei-

Implications forexplorationThis study identified numerous oilfamilies and petroleum systems inPerth Basin. It also positivelyidentified, for the first time, bothPermian and Triassic sources for gasin Perth Basin.

The Early Triassic Kockatea Shaleis the principal source for oil, and itis of the highest quality in theonshore Northern Perth Basin.Carbon isotopic evidence for gasindicates that the Kockatea Shale isalso the major source for gasonshore. Gas generation should stillhave occurred offshore, even thoughthe Kockatea Shale has diminishedpotential for oil.24 The Permian, andto a lesser extent, Jurassic sedimentsare also gas sources in Perth Basin.The existence of leaky Permian sealsfor gas leads to a large scale ‘gasflush’ in the subsurface,compounding the widespread gas-stripping of oil in Perth Basin.17

However, this re-mobilisation ofhydrocarbons should result in long-range migration and a mechanism foremplacement of petroleum higher inthe section.

In summary, the identification ofPermian and Triassic sources for gasin the onshore Perth Basin, coupledwith recognised oil and gas potentialin the Mesozoic sediments offshore,25

highlight the potential for multiplepetroleum systems active in theregion and points to new explorationopportunities.

Figure 4. Triangular plot of TNMr,TeMNr and PMNr. The circle definesthe ‘maturity centre’ representing a 10% variability in the ratios.19

C1 C5-21

-23

-25

-27

-29

-31

-33

-35

13 C

(

PD

B)

C7 C10 C15 C20 C25

Figure 5. Carbon isotope profile for n-alkanes from gases and oil in the Perth Basin

Figure 3. Plots of a. Percentage methane%/(methane% +ethane% + propane% + iso- & n-butane%+ iso & n- pentane%) versusethane%/propane%; (%C1/C1–C5 vs C2/C3) b. δ13Cmethane minus δ13Cethane versus δ13Cethane

minus δ13Cpropane (the predicted evolution ofcarbon isotopic difference26) c. δ13C of individual C1–C5 gaseoushydrocarbons for natural gases from thePerth Basin (the shaded area is the rangein carbon isotopes for unalteredAustralian gases18).



References1. Gorter J. 2000. ‘Basal Triassic’

shales in the Northern PerthBasin—will the real source rockstand up? PESA News, Oct/Nov;47: 58–59.

2. Hall PB. 1989. The futureprospectivity of the Perth Basin.APPEA Journal; 29(1): 440–449.

3. Marshall JF, Lee CS, Ramsay DC &Moore AMG. 1989. Tectoniccontrols on sedimentation andmaturation in the offshore northPerth Basin. APEA Journal; 29(1):450–465.

4. Tupper NP, Phillips SE & WilliamsBPJ. 1994. Advances in theunderstanding of upper Permianreservoir distribution and quality,north Perth Basin. In: Purcell PG& Purcell RR, eds. Thesedimentary basins of WesternAustralia: Symposiumproceedings. Perth: PetroleumExploration Society of Australia;823–837.

5. Mory AJ & Iasky RP. 1994.Structural evolution of theonshore northern Perth Basin,Western Australia. In: Purcell PG& Purcell RR, eds. Thesedimentary basins of WesternAustralia: Symposiumproceedings. Perth: PetroleumExploration Society of Australia;781–789.

6. Longley IM, Bradshaw MT &Hebberger J. 2000. Australianpetroleum provinces of the 21stcentury. In: Downey M, Threet J& Morgan W, eds. Petroleumprovinces of the 21st century.AAPG Memoir 74: in press.

7. Warris BJ. 1988. The geology ofthe Mount Horner oilfield, PerthBasin, Western Australia. APEAJournal; 28(1): 88–99.

8. Summons RE, Boreham CJ, FosterCB, Murray AP & Gorter JD. 1995.Chemostratigraphy and thecomposition of oils in the PerthBasin, Western Australia. APEAJournal; 35(1): 613–32.

9. Thomas BM. 1979. Geochemicalanalysis of hydrocarbons occurrencesin the northern Perth Basin. AAPGBulletin; 63: 1092–1117.

10. Thomas BM. 1982. Land-plantsource rocks for oil and theirsignificance in Australian basins.APEA Journal; 22(1): 164–177.

11. Jefferies PJ. 1984. Petroleumgeochemistry of the northernPerth Basin. Perth: WesternAustralian Institute of Technology;unpublished postgraduatediploma report.

12. Kantsler AJ & Cook AC. 1979.Maturation patterns in the PerthBasin. APEA Journal; 19(1): 94–107.

13. Backhouse J. 1984. Revised LateJurassic and Early Cretaceousstratigraphy in the Perth Basin.Perth: Geological Survey ofWestern Australia, report 12(Professional papers for 1982); 1–6.

14. Bradshaw MT, Bradshaw J, MurrayAP, et al. 1994. Australianpetroleum systems. PESA Journal;21: 43–53.

15. Owad-Jones D & Ellis G. 2000.Western Australia atlas of petroleumfields, onshore Perth Basin— vol 1.Perth: Department of Mines &Energy Western Australia.

16. Mory AJ & Iasky RP. 1997.Stratigraphy and structure of theonshore northern Perth Basin,Western Australia. Perth:Geological Survey of WesternAustralia, report 46.

17. Ellis GK & Bruce RH. 1998.Dongara oil and gas field. In:Purcell PG & Purcell RR, eds. Thesedimentary basins of WesternAustralia 2: Symposiumproceedings. Perth: PetroleumExploration Society of Australia;625–635.

18. Boreham CJ, Hope JM & Hartung-Kagi B. 2001. Understandingsource, distribution andpreservation of Australian naturalgas: A geochemical perspective.APPEA Journal; 41(1): in press.

19. Van Aarssen BGK, Alexander R &Kagi RI. 2000. Reconstructing thegeological history of Australiancrude oils using aromatichydrocarbons. APPEA Journal;40(1): 283–92.

20. Van Aarssen BGK, Alexander R &Kagi RI. 1998. Higher plantbiomarkers on the North WestShelf: Application in stratigraphiccorrelation and palaeoclimatereconstruction. In: Purcell PG &Purcell RR, eds. The sedimentarybasins of Western Australia 2:Symposium proceedings. Perth:Petroleum Exploration Society ofAustralia; 123–128.

21. Van Aarssen BGK, Alexander R &Kagi RI. 1999. Age determination ofcrude oils in the Barrow Sub-basinusing palaeoclimate-related variationsin higher plant biomarkers.APPEA Journal; 39(1): 399–407.

22. AGSO & Geotechnical Services. 2000.Characterisation of natural gases fromwest Australian basins: Perth Basinmodule. Canberra: AGSO; unpub-lished proprietary study.

23. Murray AP, Summons RE,Boreham CJ & Dowling LM. 1994.Biomarker and n-alkane isotopeprofiles for Tertiary oils:relationship to source rockdepositional setting. OrganicGeochemistry; 22: 521–542.

24. AGSO & GeoMark Research. 1996.The oils of Western Australia.Canberra: AGSO; unpublishedproprietary study.

25. Crostella A. 2000. Geology andpetroleum potential of theAbrolhos Sub-basin, WesternAustralia. Perth: Geological Surveyof Western Australia, report 75: inpress.

26. James AT. 1983. Correlation ofnatural gas by use of carbonisotopic distribution betweenhydrocarbon components. AAPGBulletin; 67: 1176–91.

➤ Chris Boreham, Petroleum andMarine Division, AGSO, tel +61 2 6249 9488 or [email protected]

➤ Janet Hope, Petroleum and MarineDivision, AGSO, phone +61 2 6249 9487 or [email protected]

➤ Birgitta Hartung-Kagi,Geotechnical Services Pty Ltd, 41–43 Furnace Road, Welshpool,Perth WA 6106

➤ Ben van Aarssen, Centre forPetroleum and EnvironmentalOrganic Geochemistry, School ofApplied Chemistry, CurtinUniversity, GPO Box U1987, PerthWA 6001 "



The importance of the ‘backend’to online delivery of geoscienceinformationRJ Ryburn

‘Backend’ is a term commonly used for the combination of people,hardware, software and data that lies behind corporate informationsystems. In this article I apply the term particularly to the corporate

database component of the backend conglomerate. The ‘front end’ refers tothe part the user sees—the forms, client software or browser pages that areused to access the backend. The front end is like a butterfly: eye catchingbut with a short life span, soon to be replaced by better, brighter interfaces.The backend, particularly the data and their logical structure, is made ofsterner stuff. Good backends take time and effort to construct, but whendone properly should survive for decades, outlasting any number of frontends. The hardware and system software aspects of the backend evolve overtime, but data and their logical structure should be made to last.

Geological surveys such as AGSO are very much a part of the dotcom sceneas we strive to deliver more and more geoscience information via the internetand web.1 We are presently constrained by narrow communication channels, butthese limitations will be largely overcome in just a few years. Although part ofAGSO’s output will always be in the form of electronic documents, images, GISdatasets and other project-related datasets, enhanced band-widths will increasedemand for integrated, seamless national datasets that can be automaticallysubsetted and delivered online for specified areas of the Australian lithosphere(figure 1). These standardised national datasets replace the map series that weretraditionally one of the main outputs of most geological surveys. They will comewith added dimensions, greatly enhanced usefulness, and methods ofpresentation that are limited largely by the imagination.

The love affair geoscientists havehad with their PCs leads many tothink that a do-it-yourselfapproach can carry us into the‘dotcom’ era. However, the secretto the success of major onlinebusinesses is their mastery of the‘backend’—the logical, physicaland human infrastructure that isthe foundation of web sites. Thesebusinesses know that in the longrun their customers are bestserved by building a solidbackend. Attractive web pages getcustomers in, but what keepsthem returning is the quality,quantity and timeliness of thecontent behind the web site. Mostsuccessful dotcom companieshave restructured, or built fromthe ground up, to provide the bestpossible backends. Geologicalsurveys must do likewise tosurvive in the information age.

Figure 1. A recent example of integrated online delivery of geoscience information via the web—in this case petroleumexploration titles. The titles layer has been zoomed to the Tasmanian region, and the T/28P title polygon clicked toobtain a report on that title from three tables in the PTAG petroleum accumulations database. See http://www.agso.gov.au/map/national



Limitations of hierarchical directories Every PC user is familiar withhierarchical, or tree-structured,computer disk directories. They havebeen around as long as random-access storage devices—that is,almost as long as the computer itself.Every PC and workstation, and mostservers, are organised around tree-structured directory systems. Forpersonal- and project-scalemanagement of information,hierarchical directory systems are stillvery useful and entirely appropriate.

For large-scale data management,however, tree-structured filingsystems are cumbersome, as theypermit just one way of classifyingand locating files. In corporatecomputing networks, files can easilybe lost in a maze of directories, andfile searches are slow and inefficient.Duplication of data is rife, and thereis often no way of knowing whichversion of a file is the most up todate. Other problems concernfrequent changes to directorystructures, leading to isolation of filesrelying on pathnames for access.These limitations may not seemserious on a PC, but on the computernetworks of large organisations theyare major headaches. Similarly, websites organised around tree-structuredfiling systems can become difficult tomanage once they grow beyond acertain size. Applications becomedifficult to write, requiring manychanges to keep up with fast-changing directory structures.

The evolving backend Relationaldatabase management systems (andtheir forerunners) arose partly toovercome the problems encounteredwith traditional filing systems. With arelational database the user has manyways of finding data that areefficiently indexed, obviating system-wide file-by-file searches. Similarly,the problem of duplicated data canbe largely overcome in a properlydesigned corporate database systemby rejection of inadvertent attemptsto re-enter the same information, andclear version control. In reality thereare many reasons for using relationaldatabase management systems—suchas transaction management, businessrules, concurrency, security andautomatic backup.

Corporate databases have existedalmost as long as the computer itself,and as relational databases for thepast 20 years. Corporate databases,albeit in a simpler form, werecertainly to the fore in the days ofcorporate mainframe computers withdumb terminal access.

With the rise of the minicomputerin the seventies, and the PC in theeighties, databases were able tomigrate to the department, theproject and the person. At the sametime the idea of distributedcomputing gained currency, andpeople liked the relative freedom ofbeing able to do their own thing ontheir own machine (if at some cost).

The advent of the web in the earlynineties gave rise to ‘thin-client’computing in which the web browserbecame the graphical equivalent ofthe old character terminal, and thebulk of the computing power was putback into the server. Suddenly therewas the potential for thousands ofterminals anywhere in the world to beconnected to your server at little cost.

You could upgrade your systemsoftware and applications in the oneplace, without the need to upgradeall those client systems. Theefficiencies to be gained wereastounding—the backend was backwith a vengeance. The net result hasbeen the reunification of thecorporate backend into oneintegrated system, albeit spread overa number of CPUs. OracleCorporation, fed up with theinefficiencies of dispersed systems,recently consolidated thousands oftheir business servers spread aroundthe world into what is effectively onecorporate backend at theirheadquarters in Redwood Shores,California.2 In so doing they claimthey will save a billion dollars.

The rise of object-relational databasesIn the early nineties there werepredictions that object-orienteddatabases would take over completelyfrom relational databases. This hasnot happened, and pure objectdatabase systems now form only asmall, specialised part of the currentdatabase scene. What has happened,though, is that the main relationaldatabase systems have graduallytaken on some of the more usefulqualities of the object-oriented world.A variety of different object types maynow be conveniently handled in whatare known as universal, or object-relational database managementsystems. For example, Oracle hasacquired the technology for handlingimages in the database, to the extentthat it can now deal with commandslike ‘get me all pictures that look a bitlike this one’—and that is withoutrecourse to textual metadata. InAGSO’s case, the ability to handle GISdatasets in Oracle 8i (with or withoutESRI’s Spatial Data Engine) and to

store documents in the database willundoubtedly prove to be veryvaluable. Images, too, will beincreasingly placed inside thedatabase.

At the same time the database isexpanding to accommodate a greatervariety of data types, the applicationsoftware that allows one to easilyhandle these new forms of data ismaturing. With Oracle’s latestapplication development tools, onecan rapidly construct a web site thatincludes text, images, documents,video, sound and ‘XML’ extracts—allfrom within the corporate database.Once the corporate database includesall these new forms of information ina well-organised and accessiblemanner, constructing an attractive,highly functional web site becomesan easy and relatively quick exercise.Furthermore, multimedia front endssuch as these are now easily modifiedor expanded to meet clientexpectations. ‘Portals’ for use on otherweb sites can also be projected.

Burgeoning databaseadministrationA consequence of the rapidlyexpanding backend is that databaseadministration (DBA) is a morecomplex and critical job than it was afew years ago. The software neededto place corporate information on theweb has grown enormously involume and sophistication, and thevariety of information types beingstored in the backend forpresentation on the web is increasingin leaps and bounds. New products,such as Oracle’s WebDB 3.0, allowthe contents of the entire web site tobe brought within the database.Control of content can then bedistributed to the appropriate areas ofthe organisation with securityhandled by the databasemanagement system. GIS datasets arealso moving into the database, as inESRI’s Spatial Data Engine now usedin AGSO. We will soon be requiredto use a document managementsystem, in which all documents andcommunications are stored within thecorporate database system andaccessed via the intranet and web.Information can now be supplied viathe web encapsulated in the XMLdata definition language.

Efficient management of acorporate web site requires peoplewith at least three types of expertise.First you need people who arecompetent with hardware, operatingsystems, communications,

➥ Continued page 20



AGSO’s Minerals Division laboratory staff has developed a new methodof sample preparation for ICP–MS analysis that replaces the time-consuming, multi-step acid dissolution technique used to date.

The new method involves digesting pieces of the 12:22 lithiumtetraborate/lithium metaborate fusions that have been prepared andrun for XRF major element analysis.

To date sample preparation for ICP–MS analysis at AGSO has beenbased on a method outlined by Jenner et al.1 The method involves aseries of acid digestion and drying stages over a period of four days.

Comparisons between XRF Zr results and Zr results from the ICP–MSindicated that for many samples the method was not achieving total digestionof the zircon present. Subsequent replicates of problem samples also showedlarge variations in Zr results. Hot or cold spots on the hotplate surfaces,affecting digestion, may have been the reason for these variations.

Comparison of XRF and ICP–MS Cr results for those samples withsignificant Cr values suggests that similar problems are present for thedissolution of the refactory mineral chromite. Other elements present inrefactory minerals generally tend to be at levels approaching or below thedetection limit of the XRF method making comparisons meaningless.

To overcome these problems laboratory staff first experimented with Parrbomb and microwave digestions. Little or no improvement was found withthe microwave and only slight improvement with the Parr bombs. SteveEggins from the Australian National University’s School of Geology was usinglaser ablation on XRF fusions to do trace element analyses. His work ledMinerals laboratory staff to the idea of digesting pieces of XRF fusions.Digesting the fusions proved straightforward.

The only problems encountered arose from contamination of the digest bythe platinum ware used to produce the fusions. All platinum cruciblescontained some rhodium and palladium, even the recently purchased 95% Pt/

Minerals laboratory staf f developsnew ICP–MS preparation method J Pyke

5% Au crucibles. Rhodium is the mostcommonly used internal standardused in ICP–MS because of itsposition in the mass range.

Introducing it into the sampleobviously precluded its use as aninternal standard. Fortunately theelements for which Rh had been usedas an internal standard weresuccessfully divided between Ni 61and Sm 147. Lanthanum was alsopresent in many older crucibles—most probably residual from theNorrish and Hutton flux that thelaboratory had used for many years.The La was removed successfullywith a number of dummy fusions.One crucible that was highlycontaminated with zinc is no longerused.

During the development of thismethod, the cause of an intermittentBa and Pb contamination problemwith the ICP–MS analysis wasidentified. One of a number of glasspipettes that could be used in thepreparation of the internal standardwas having these elements leachedout of the glass by the 1% HF used inthe preparation. No glassware is usedin the ICP–MS laboratory now.



MethodApproximately 100 micrograms ofchips from the smashed discs areweighed accurately into Savillexteflon vessels. Five millilitres ofinternal standard, one millilitre of HFand five millilitres of HNO3 are thenadded.

The vessels are sealed and heatedfor 12 hours overnight at 200 degreescentigrade on a timed hotplate, suchthat cooled samples are ready thefollowing morning. The digests arethen transferred to volumetric flasksand made up to volume ready for theICP–MS.

Occupational health and safety There are substantial benefits from anoccupational health and safetyperspective. The digests take place ina sealed vessel. If the fume cupboardsystem failed, dangerous acid fumeswould not leak (unlike the presentsituation). The use of acids,particularly HF, is substantiallyreduced. With this method onemillilitre of HF is used instead of six,and five millilitres of HNO3 are usedinstead of 18. This greatly reducesstaff handling of acids, with less acidneeding to be distilled and stored inthe laboratory.

Sample digestionThe sample is totally digested. Any undissolved sample in the glass disccreates stress points and the disc shatters. Because the discs have survivedXRF analysis and a few days stored in a plastic bag, laboratory staff can beconfident that dissolution in the glass is complete.

Disregarding the fusion process, which is standard practice for the XRF,this method is a one-step 24-hour preparation compared to a multi-step four-day procedure. The time could be cut even further by using a microwave forsmall lots of ‘specials’.

Contamination of samples should be effectively eliminated since thedigestion takes place in a sealed container. There are no drying stages as inthe current method where samples spend many hours in open vessels.Because the sample is totally in solution, laboratory staff can now confidentlyanalyse for those elements associated with the refractory minerals—elementssuch as Zr (see table 1), Hf, Cr and the REE likely to be tied up in zircons.

Table 1. Zr results obtained on some international standards

Standard ICP–MS ICP–MS AGSO Recommended

Old New XRF value2

W-2 78 ppm 95 ppm 93 ppm 94 ppmBIR-1 15 15 15 15.5DNC-1 36 37 36 41QLO-1 171 189 188 185BHVO-1 151 176 175 179AGV-1 205 235 235 227

Figures 1–9. XRF vs ICP–MS: Result comparisons for anumber of elements from a recent suite of samples


Bonaparte BasinGeochemical characteristics ofhydrocarbon families and petroleumsystems

DS Edwards, JM Kennard, JC Preston, RE Summons, CJ Boreham & JE Zumberge

The Bonaparte Basin has been actively explored for more than 20 years, with oil production from several fields (Jabiru,Challis–Cassini, Laminaria–Corallina, Elang and the depleted Skuafield) and proposed production from giant gas/condensate fields(Bayu–Undan, Sunrise–Loxton Shoals–Troubadour and Petrel–Tern).Despite this focused exploration and appraisal, to date geoscientistshave had a relatively poor understanding of the region’s petroleumsystems.

To improve this understanding, isotopic and biomarker analysesof numerous oils, condensates and gases have been undertaken togeochemically characterise the hydrocarbon families in theBonaparte Basin, and to correlate them with likely source rocks.Preliminary results of this study show that two Palaeozoic andseven Mesozoic oil families can be identified in the BonaparteBasin. Details of the petroleum systems active in this basin werepresented at the recent AAPG International Conference in Bali(October 15–18, 2000) by Dianne Edwards and John Kennard, and continue to be investigated by AGSO’s North-north-westRegional Project.

The Bonaparte Basin lies between north-western Australia and theisland of Timor (figure 1). It has a complex tectonic historyinvolving two phases of Palaeozoic extension and Late Triassic

compression prior to the onset of Mesozoic extension. Initial rifting occurred in the Late Devonian to form the north-west-

trending Petrel Sub-basin in the south-east. The resultant thick LateDevonian–Carboniferous rift and sag succession was orthogonallyoverprinted in the Late Carboniferous to Early Permian by north-east-trending rift basins to form a proto-Malita and possible proto-Vulcan Sub-basin. Late Jurassic extension resulted in a series of linked, north-west-trending (Vulcan Sub-basin and Malita Graben) and south-east-trending(Sahul and Flamingo Synclines) intracontinental grabens. Thick marinemudstones accumulated within these grabens, and passed laterally to fandelta sandstones on the adjacent horst blocks and terraces. TheseMesozoic depocentres are surrounded by structural highs (AshmorePlatform, Londonderry High, Sahul Platform and Darwin Shelf; figure 1)which have relatively thin Jurassic–Cretaceous sediments across an upliftedand eroded Triassic–Palaeozoic section.

Most of the commercial and soon to be developed oil and gasaccumulations are reservoired in Middle and Upper Jurassic sandstones(Plover and Montara/Elang Formation, respectively; figure 2). Commercialaccumulations also occur in Upper Triassic and Upper Cretaceous sands inthe Vulcan Sub-basin. In the Petrel Sub-basin, gas and gas/condensateaccumulations occur in the Upper Permian Hyland Bay Formation (Petreland Tern Fields, Fishburn-1 and Penguin-1), and gas discoveries on theLondonderry High (Prometheus-1, Ascalon-1A) and Sahul Platform (KelpDeep-1) also occur within this unit (figure 2).

Hydrocarbon familiesOil–oil comparisons were made using cluster and principal componentanalysis—the results of which are displayed as a dendrogram in figure 3.The GeoMark protocol was adhered to which utilises 16 geochemicalparameters (two bulk carbon isotopic values, 13 source-specific terpane


The distribution of zircon, ifpresent, in a powdered sample is anexcellent indicator of samplehomogeneity. Any major discrepancyin Zr results from the XRF and theICP–MS will now suggest problemswith the sample grinding rather thananalytical problems. Results,particularly trace results, could bethen treated with appropriate cautionor the sample(s) reground andreanalysed.

The laboratory will be able toreport both results for those elementsthat can now be equally welldetermined by XRF and ICP–MS.These elements will probably be Ba,V, Cr, Ni, Cu, Zn, Rb, Sr, Y, Zr and Pbbecause they are generally present insilicates at levels significantly abovethe detection limit of the less-sensitive XRF (see figures 1 to 9).

Although the effective sampleweight in the final solution is smallerthan presently used, the weight usedto make the fused disc is approxi-mately three times larger than atpresent. The end result is thatlaboratory staff is now using whatshould be a more representativealiquot of the sample.

Because discrete chips of glassare used, there are no problems withelectrostatic charges that causesegregation in some sample powders.One chip digests as easily as a dozensmaller pieces.

Similar rock standards willroutinely be run with each batch ofboth XRF and ICP–MS. Collection andstorage of this data will continuallymonitor the performance of bothtechniques.

DisadvantageLithium and boron analyses will nolonger be available from the MineralsDivision laboratory because of the Liand B present in the flux and theirmemory effect within thespectrometer.

References1. Jenner GA, Longerich HP, Jackson

SE & Fryer BJ. 1990. ICP–MS—Apowerful tool for high-precisiontrace-element analysis in earthsciences: Evidence from analysisof selected USGS referencesamples. Chemical Geology; 83:133–148.

2. Govindaraju K, ed. 1994.Geostandards Newsletter, specialissue Jul; 18.

➤ John Pyke, Minerals Division,AGSO, phone +61 2 6249 9288 ore-mail [email protected] "



Figure 1. Location of petroleum exploration wells in the Bonaparte Basin

128

123

Darwin

10

15

100 km0

PrudhoePrudhoe

Terra

ce

Terra

ce

Berkley Platform

Berkley Platform

DarwinDarwin

ShelfShelf

LondonderryLondonderry

HighHigh

Malita

Gra

ben

Malita

Gra

ben

Timor Trough

Timor Trough

Calder

Gra

ben

Calder

Gra

ben

AshmoreAshmore

PlatformPlatform

BROWSE BASINBROWSE BASIN

Yampi

Yampi

S

helf

S

helf

Sahul Platfo

rm

Sahul Platfo

rm

Petrel Sub-basinPetrel Sub-basin

Sub

-bas

in

Sub-b

asin

Vulc.

Vulc.

Sahul Syn.

Sahul Syn.

Oil, gas show Oil Dry well Gas Condensate

Figure 2. Generalised stratigraphy of the Bonaparte Basin

AGEASHMOREPLATFORM

VULCANSUB-BASIN

DARWINSHELF

SAHULSyncline High

CARB

PERMIAN

TRIA

SSIC

E

L

M

E

M

L

JUR

ASSI

CC

RET

ACEO

US

E

L

Not penentrated

LONDON-DERRYHIGH

MALITAGRABEN

PETRELSUB-

BASIN

and sterane biomarker ratios, andpristane/phytane ratio).1 TwoPalaeozoic oil families are recognisedin the Petrel Sub-basin, interpreted tobe sourced by Carboniferous andPermian sediments. Seven Mesozoicoil families are currently recognisedfrom the Bonaparte Basin, three ofwhich are present in the Vulcan Sub-basin, and four in the northern partof the basin, in and adjacent to theTimor Gap Zone of Co-operation.The two oil families in the adjacentBrowse Basin identified by Blevin et al.plot separately to those of theBonaparte Basin.2 It is apparent fromthe number of discrete oil familiesthat there are many effective sourceunits in the Bonaparte Basin.

Palaeozoic hydrocarbonfamiliesOils interpreted to be derived fromthe Lower Carboniferous MilligansFormation in the Petrel Sub-basin(Barnett, Turtle and Waggon Creekwells) are isotopically light (δ13Cvalues for the n-alkanes between -28to -29 ‰) and have a slightlynegative trend with increasing n-alkane number (figure 4). Theirbiomarker signatures arecharacterised by low pristane/phytane


16 AGSO Research Newsletter DECEMBER 200016 AGSO Research Newsletter DECEMBER 2000

ratios (Pr/Ph = 1.2), an abundance ofrearranged hopane and steranes, andhigh sterane/hopane and tricyclicterpane/hopane ratios, which indicategeneration from marine, anoxic, clay-rich source rocks.3

In contrast, the condensates fromthe Petrel and Tern Fields have heavyisotopic signatures (δ13Csat = -24 ‰)and their n-alkane isotopic profilesexhibit a strong negative-slope. Then-alkane isotopic trends of the Tern-5gases (C1–C5) are generally continuouswith the n-alkane trends of thecondensates, suggesting that the gaseshave the same source as the liquids.These heavy 13C enriched isotopicvalues of the gas/condensates areconsistent with derivation from land-plant material, and this interpretationis supported by the high abundanceof C29 diasteranes. Geologically, themost likely source of thesehydrocarbons are the Permiansediments of the Keyling and HylandBay Formations which are rich inland-plant remains and weredeposited in coastal plain and deltaicenvironments, respectively.3

Dry gas from the Upper PermianHyland Bay Formation in Kelp Deep-1is extremely enriched in 13C, which isin keeping with generation from aland-plant-rich source rock that is nowovermature. Its stratigraphic andstructural position on the SahulPlatform also suggests a Permian origin.

The gas discoveries at Penguin-1and Fishburn-1 in the Petrel Sub-basin, as well as Ascalon-1A and,most recently, Prometheus-1 on theLondonderry High (all reservoiredwithin the Upper Permian HylandBay Formation; figure 1), are alsoattributed to this Permian system.

Analysis of oil-bearing fluidinclusions in the Torrens-1 well onthe Londonderry High indicates aninterpreted 42-metre gross palaeo-oilcolumn within the Permian FossilHead Formation.4 Isotopic andbiomarker profiles of a residual oilfrom this palaeo column (Core 2,Permian Fossil Head Formation) arecomparable to the Petrel and Terncondensates (figure 4). This oil is thusalso attributed to a Permian source.5–7

In contrast, the residual oil in Core 3at Torrens-1 (Upper CarboniferousKuriyippi Formation) has an isotopicsignature similar to the Barnett andTurtle oils, and is interpreted to bederived from a Carboniferous source.

Oil-bearing fluid inclusions havealso been interpreted to indicate apalaeo-oil column in the UpperPermian Hyland Bay Formation inOsprey-1 on the western margin ofthe Londonderry High.8 The strati-

graphic and structural setting of the Osprey-1 and Torrens-1 wells alsosupports the interpretation that these prior-oil accumulations were sourcedfrom Permian sediments. This finding is significant because it indicates that aviable Permian oil play may be present across much of the LondonderryHigh.

Additional evidence of a viable Permian oil play in the Bonaparte Basin issuggested by the occurrence of clusters of interpreted oil slicks on SyntheticAperture Radar (SAR) satellite scenes on the Ashmore Platform.9 These slicksoverlie areas where Jurassic and Cretaceous source rock are known to be absentor immature. Based on the current state of knowledge of potential Triassic andPalaeozoic source rocks, these slicks most probably indicate an active, oil-prone Permian petroleum system.

New gas isotopic data indicate that the Permian system also extends intothe eastern Browse Basin.10

Mesozoic hydrocarbon families in the Vulcan Sub-basinIn the Vulcan Sub-basin, two major oil families are recognised: one with adominant marine source signature and the other with a dominant terrestrialsource signature (figure 5). A third oil family comprises condensates withvariable geochemical composition.

The Vulcan Sub-basin marine oil family comprises oils from the Puffin,Skua, Cassini, Challis, Talbot, Jabiru and Tenacious wells (figure 5). They havelight isotopic signatures (δ13Csat = -27.5 ‰) with the n-alkanes displaying a‘lazy-S’ profile (figure 4). This type of isotopic profile is seen in many UpperJurassic sourced oils on the North West Shelf and is characteristic of mixinglow molecular weight n-alkanes of the oil fraction with the higher molecularweight components of a more mature gas.10 Whole oil chromatograms showthat these oils have a unimodal n-alkane distribution with a maximumbetween C10 and C13 and intermediate Pr/Ph ratios (mean = 2.8), which areindicative of marine organic matter deposited in a sub oxic environment.Other source-dependent characteristics include a slight predominance of theC29 homologue among the regular and rearranged steranes (C29 > C27 > C28). Asiliciclastic source lithology is indicated by the abundance of rearrangedsteranes and hopanes. Preliminary oil–source correlations indicate that theseoils are derived from the Upper Jurassic Lower Vulcan Formation.

The Vulcan Sub-basin terrestrial-influenced oil family includes oils fromMaret-1, Bilyara-1, Montara-1 and Oliver-1, all of which have relatively heavyisotopic signatures (δ13Csat = -25.5 ‰). Whole oil chromatograms show thatthese oils have moderately high Pr/Ph ratios (2.4–6.2) and high wax contents,consistent with their derivation from land-plant organic matter. Diagnosticbiomarker features include an abundance of C29 sterane in comparison to theC27 and C28 homologues; diasteranes are more abundant than steranes.Rearranged hopanes are present, but in different relative amounts (e.g.C29Ts/C29 hopane ratio is lower) compared with the aformentioned marine oilfamily. The lower correlation co-efficient between the Oliver, Montara-Bilyaraand Maret oils (figure 5) indicates derivation from several localised sourceunits. The most likely source of these oils is the Lower–Middle Jurassic PloverFormation.

Condensates in the Vulcan Sub-basin—including those at Tahbilk-1,Eclipse-2 and Swan-1—show some variation in their geochemistry, and plottogether as a separate family in figure 5. Data for these condensates were notused in the generation of the Bonaparte dendrogram (figure 3), because it isbelieved that their present composition is more reflective of reservoiralteration effects (such as leakage and gas flushing) rather than the type oforganic matter in their source rocks.

Mesozoic hydrocarbon families in the northern Bonaparte BasinThe oils and condensates from the northern Bonaparte Basin fall into fourfamilies. The condensates from the Bayu–Undan Field and oils from the Elangand Kakatua Fields have a dominant marine signature, but plot as a separatefamily to the Vulcan Sub-basin marine oil family (figure 3). This is probablydue to these oils originating from multiple marine-influenced source facieswithin the Middle Jurassic Plover Formation and Upper Jurassic Elang andFlamingo Formations.11

The terrestrial-influenced oils from the Laminaria, Corallina, Buffalo andJahal Fields make up a second oil family in the northern Bonaparte Basin.These oils probably also arise from several Jurassic source rocks, rich in land-plant remains.11


DECEMBER 2000 AGSO Research Newsletter 17DECEMBER 2000 AGSO Research Newsletter 17

Figure 4. Plot of δ13C versus carbonnumber for n-alkanes from selectedBonaparte gases, condensates and oils

1 5 9 13 17 21 25 29 33-20

-22

-24

-26

-28

-30

-32

-34

-36

C13

1.0 0.8 0.6 0.4 0.2 0

Darwin Formation), was recognised inthe northern Bonaparte Basin byPreston and Edwards.11 These oilshave a marine clastic signature andhave been correlated to the underlyingEchuca Shoals Formation. Somesimilarity is seen between these oilsand the Early Cretaceous oil family(Caswell-2 and Gwydion-1; figure 3)identified by Blevin and co-workers inthe Browse Basin.2

Regional implicationsThese studies have identifiednumerous oil families and petroleumsystems in the Bonaparte Basin, andhighlight the fact that both Palaeozoicand Mesozoic source units areeffective in the region.

The Permian system, previouslyknown only as a gas/condensatesystem in the Petrel Sub-basin, hasbeen shown to be more widespreadand extends across the LondonderryHigh to the Sahul Platform andeastern Browse Basin. Fluid inclusiondata indicates that this system hasgenerated substantial palaeo-oilaccumulations on the LondonderryHigh. Source rock data and

Figure 3. Oil familydendrogram showingthe degree ofcompositional similaritybetween theoils/condensate of theBonaparte and BrowseBasins

The third oil family comprises condensate from the Sunrise Field. Thiscondensate has a biomarker signature that is atypical of North West Shelfhydrocarbon accumulations in that it appears to have a marine carbonateorigin, as demonstrated by the abundance of 30-norhopanes, with C29 hopanebeing the dominant hopane. The only other known condensate in this regionwith similar composition is from Buffon-1 in the Browse Basin. Thesehydrocarbons are reservoired within the Lower–Middle Jurassic PloverFormation, and may originate from pre-Jurassic source rocks. Alternatively,the carbonate signature in these condensates may originate from dieselcontamination in the drilling muds; further work is required to resolve thisissue of possible contamination.

A fourth oil family (not shown in figure 3), comprising the Elang West-1,Kakatua North-1 and Layang-1 oils (all reservoired in the Lower Cretaceous



palaeogeographic facies mapssuggest that oil-prone coaly shales ofthe Lower Permian KeylingFormation probably extend acrossthe Londonderry High and aroundthe southern and north-easternmargin of the Petrel Sub-basin. ThePermian system is untested on theAshmore Platform, but oil potentialhere is suggested by the presence ofinterpreted SAR oil slicks.

Previous studies focusing onMesozoic plays in the BonaparteBasin have made the somewhatsimplistic assumption that mosthydrocarbon accumulations in thewestern and northern BonaparteBasin have been charged from UpperJurassic source rocks. Although this istrue for the producing fields in theVulcan Sub-basin, it is now apparentthat there has been a significantcontribution to hydrocarbon reservesin both the northern Bonaparte Basinand southernmost Vulcan Sub-basinfrom the Lower–Middle JurassicPlover Formation. If plays can beidentified where the timing ofhydrocarbon generation and trapformation is more favourable topreserve Early–Middle Jurassic-derived liquids, then this systemcould add significant reserves to theregion. Furthermore, non-Jurassic,oil-prone petroleum systems havenow been identifed in the BrowseBasin (Gwydion-1, Caswell-2 andCornea Field2) and northernBonaparte Basin (Elang West-1),

both of which are sourced from Lower Cretaceous marine mudstones. An additional (?)Early Mesozoic petroleum system may be indicated by

the condensates at the Sunrise Field and Buffon-1 near the outer margins ofthe Australian plate. These condensates have a distinctive, marine carbonatebiomarker signature, and may form part of the oil-prone Late Triassic–EarlyJurassic carbonate system known in the Australian–Banda boundarycomplex to the north.12 This carbonate system includes oils from the islandsof Buton, Buru, Seram and Timor, and may also be prospective along theouter margin of the Bonaparte–Browse Basins.

In summary, the Bonaparte Basin has greater hydrocarbon potentialthan is currently recognised since several other source units besides thewell-known Upper Jurassic marine mudstones are also capable ofgenerating liquid hydrocarbons.

References1. AGSO & GeoMark Research. 1996. The oils of Western Australia. Canberra

& Houston; unpublished proprietary report. 2. Blevin JE, Boreham CJ, Summons RE, Struckmeyer HIM & Loutit TS. 1998.

An effective Early Cretaceous petroleum system on the North West Shelf:Evidence from the Browse Basin. In: Purcell PG & RR, eds. Thesedimentary basins of Western Australia 2: Proceedings of PetroleumExploration Society of Australia Symposium, Perth; 397–420.

3. Edwards DS, Summons RE, Kennard JM, Nicoll RS, Bradshaw J, BradshawM, Foster CB, O’Brien GW & Zumberge JE. 1997. Geochemicalcharacterisation of Palaeozoic petroleum systems in north-western Australia.APPEA Journal; 37(1): 351–79.

4. Lisk M & Brincat MP. 1998. Oil migration history of the west Bonapartemargin. Canberra: CSIRO confidential report 98-040; unpublished.

5. Summons RE, Bradshaw M, Crowley J, Edwards DS, George SC &Zumberge JE. 1998. Vagrant oils: Geochemical signposts to unrecognisedpetroleum systems. In: Purcell PG & RR, eds. The sedimentary basins ofWestern Australia 2: Proceedings of Petroleum Exploration Society ofAustralia Symposium, Perth; 169–184.

6. Passmore VL, Edwards DS & Kennard JM. 1999. [abstract]. Enhanced Permo-Triassic prospectivity in the Timor Sea provided by a new Paleozoic petrol-eum system. AAPG annual meeting, San Antonio, Texas, Apr 11–14; A105.

Figure 5.Dendrogram ofthe Vulcan Sub-basin oils andcondensates

Marine

Terrestrial

1.0 0.8 0.6 0.4 0.2 0


Hydrocarbon seepage inCarnarvon Basin subjectof major studyAGSO and its partners, Signalworks Pty Ltd and Nigel Press &Associates are investigating hydrocarbon leakage and seepage in theCarnarvon Basin (figure 1). The study is using several independentremote-sensing technologies, namely:

• double-coverage RadarSat Synthetic Aperture Radar (SAR). SAR iseffective for mapping oil-prone leakage and seepage;

• 20 000 kilometres of reprocessed Mark III Airborne LaserFluorosensor (ALF) data. ALF effectively maps oil and condensateleakage and seepage. Interpretations from reprocessing BP regionallegacy Mark II ALF data through the area are also being used, as areBP’s original interpretations from the Mark II ALF surveys;

• more than 2000 kilometres of water column geochemical sniffer(WaSi) data, which detects oil, condensate and gas leakage andseepage; and

• Landsat data.

The interpretations derived from these technologies (i.e. SAR, WaSi,ALF and Landsat) are being been compared and contrasted, and thenintegrated with regional seismic data, isopach maps of key reservoir,source and sealing units, and fault maps.

The goals of the study are to:

1. provide a soundly based understanding of the relative responses ofthese technologies to different types and rates of hydrocarbonseepage; and

2. determine the nature of, and principal controls on, hydrocarbonseepage within the Carnarvon Basin.

Lessons learnt in the Carnarvon Basin, and from a recentlycompleted similar study in the Timor Sea, will be applied toevaluations of frontier exploration areas around Australia.

For more information about the study phone Dr Geoff O’Brien on +61 2 6249 9342 or e-mail [email protected] "

Figure 1. Study area of seepage investigation, Carnarvon Basin, North WestShelf. Area covered by BP ALF data (now reprocessed) is shown in red; areacovered by Fugro ALF survey (now reprocessed) is shown in blue. Watercolumn geochemical sniffer data covers parts of the basin.


7. Ruble TE, Edwards DS, Lisk M, etal. 2000. Geochemical appraisal ofpalaeo-oil columns: implications forpetroleum systems analysis in theBonaparte Basin, Australia. AAPGAnnual Meeting, New Orleans,Louisiana, Apr 16–19.

8. Lisk M, Brincat MP, Eadington PJ &O’Brien GOB. 1998. Hydro-carbon charge in the Vulcan Sub-basin. In: Purcell PG & RR, eds. Thesedimentary basins of WesternAustralia 2: Proceedings ofPetroleum Exploration Society ofAustralia Symposium, Perth;287–302.

9. O’Brien GW, Lawrence G, WilliamsA, Webster M, Lee J & Burns S. 2000.Hydrocarbon migration and seepagein the Timor Sea and northernBrowse Basin North West Shelf,Australia: an integrated SAR,geological and geochemical study.Canberra: AGSO, Nigel Press &Associates, RadarSat International,AUSLIG (Australian Centre forRemote Sensing).

10. Boreham CJ, Hope JM & Hartung-Kagi B. 2001. Understanding source,distribution and preservation ofAustralian natural gas: ageochemical perspective. APPEAJournal; 40 (1): in press.

11. Preston JC & Edwards DS. 2000.The petroleum geochemistry of oilsand source rocks from the northernBonaparte Basin, offshore northernAustralia. APPEA Journal; 40 (1):257–282.

12. Sykora JJ. 2000. The buried fold-thrust belt of offshore Seram.[abstract]. American Association ofPetroleum Geologists InternationalConference and Exhibition, Bali,Oct 15–18.

➤ Dianne Edwards, Petroleum andMarine Division, AGSO, phone +61 2 6249 9782 or [email protected]

➤ John Kennard, Petroleum andMarine Division, AGSO, phone +61 2 6249 9204 or [email protected]

➤ Jim Preston, BHP Petroleum Pty Ltd,600 Bourke Street, Melbourne,Victoria 3000

➤ Dr Roger Summons, Petroleum andMarine Division, AGSO, phone +61 2 6249 9515 or [email protected]

➤ Chris Boreham, Petroleum andMarine Division, AGSO, phone +61 2 6249 9488 or [email protected]

➤ John Zumberge, GeoMark Research,9748 Whithorn Drive, Houston,Texas 77095, USA "



networking, computer security andweb servers. You then need goodwebmasters, database administrators,and application developers. Finallyyou need people to supply thecontent—the information to go in thedatabases and on the web. Inorganisations of any size it isimpossible for one person to beacross all these skills; a team isgenerally needed to run web sites ofdotcom organisations. Increasingly,though, it is the databaseadministrator who occupies thepivotal position, and has the mostdemanding job. In organisations thatfloat on seas of information, thedatabase administrator is the pilot.One must look after the DBA rolecarefully, or risk disorganisation.

Web delivery ofgeoscience dataIn organisations like AGSO there is aclear distinction to be made betweenstandard national databases and themore ad-hoc datasets and documentsproduced particularly by researchprojects. However, both types ofinformation have to be adequatelymanaged as we move towards an erain which online delivery ofgeoscience information is the norm.Both information types demand thatattention be given to the backendpart of the total informationmanagement environment.

In regard to ad-hoc types ofinformation, I include all traditionaldocuments, papers and publications,most of which will soon be routinelyavailable on AGSO’s intranet and/orweb site. Some may be handled by adocument management system, suchas ‘TRIM’, that can use the Oracle 8idatabase management system as itsbackend. Alternatively, Oracle’sWebDB 3.0 may be used to placemany documents on the intranet andweb. Either way, vast amounts ofunstructured written material will findtheir way into the backend databasemanagement system in the nearfuture. It is also likely that processedimages and project GIS datasets willeventually migrate into the corporatedatabase system, rather than, as atpresent, be kept in computer files on‘corporate’ UNIX disk systems.Almost all datasets that projectscurrently publish on CD-ROM willprobably end up in the corporatedatabase system for online delivery.3

The demands on the backend will beheavy indeed. The infrastructure tohandle the metadata for ad-hocinformation is already at hand as theAGSO Catalog.4

The formally structured nationalgeoscience databases that have been

an integral part of AGSO’sinformation store for more than 20years are being made available onthe web in user-specified chunks. Inone mode, the user is able toconstruct a map of the area in whichthey are interested by selectingrequired GIS layers. They shouldthen be able to obtain all thespatially related attribute data thatAGSO has in the backend databases(as in figure 1). All laboratory datashould be made available, and in thecase of geological field data there arethoughts of capturing videos ofcritical outcrops, in which geologistsexplain the relationships.

Another mode of use is whereother web sites transparently accessAGSO data via portals, as if it weretheir own. AGSO is a participant inthe Australian Spatial Data Initiativewhich has already demonstrateddistributed web mapping, with datadrawn online from dispersedagencies. The web enables manydifferent data types to be integratedfrom many sources, but thecomponents are best kept in well-controlled backend databases.

Restructuring for onlinedeliveryOne of the main lessons to belearned as we enter the dotcom erais that we must organise properly foronline delivery of geoscience data.The task should not be left toindividuals scattered among disparateprojects. Individual projects can, andsometimes do, set up web interfacesfor online delivery of certain types ofdata, and some of these interfacesare effective. However, such practices‘reinvent the wheel’, and time andmoney are wasted on systems thatcannot be readily integrated withother online systems. The web sitecan easily become an expensivemishmash of disparate systems. Wehave already experienced instancesof this in AGSO.

Information management anddelivery is too important a corefunction in today’s geological surveysto relegate to dispersed groupswithin the organisation. In thecurrent climate it is axiomatic for adotcom organisation that onlineinformation delivery be managed bya well-integrated, multidisciplinaryteam with good communicationskills. Furthermore, this team shouldact as an ongoing service group,rather than a project. The team mustinclude a strong backend group,specifically those with databaseadministration and developmentskills. In time, when the systems andprocedures for online delivery

become established, responsibility canbe more devolved. As proceduresbecome routine, the team size can bereduced and many membersredeployed into other areas.

There is little doubt in my mindthat geological surveys shouldconsider setting up formalinformation divisions (if they havenot already done so) to properlyhandle the difficult new paradigm ofonline geoscience informationdelivery.

Conclusions➤ Successful online delivery of

geoscience information requiresgood backend databases.

➤ Web sites that rely exclusively onhierarchical directory systems donot scale well.

➤ Object-relational databases, suchas Oracle 8i, are built to handleall sorts of information.

➤ Database administration hasrecently become a much morecomplex and critical job.

➤ Online delivery requires a focusedapproach to data acquisition anddatabase design.

➤ Both structured and unstructuredtypes of information must becatered for on the web.

➤ Geological surveys may need torestructure to better supportonline information delivery.

References1. AGSO. 2000. Australian

Geological Survey OrganisationOnline Action Plan.http://www.agso.gov.au/information/online_action_plan.html.

2. Karpain G & Myers L. 2000. Keepit simple: How Oracleconsolidated its global infra-structure into a centralised e-business architecture. OracleCorporation white paper, Sept2000. http://www.oracle.com/applications/story.html.

3. AGSO. 2000. AustralianGeological Survey Organisation2000–2001 work program.Canberra: AGSO; unpublished.

4. Ryburn RJ, Ross SI &Wijatkowska-Asfaw BS. 2000.Guide to the AGSO Catalog.Canberra: AGSO, record 2000/13.

➤ Dr Roderick Ryburn was Leader of AGSO’s CorporateDatabase Group when this articlewas written. He can be contacted by phoning +61 2 6291 7362 or e-mailing [email protected] "

➥ Continued from page 11



Regolith-landform maps provide an important geomorphic andlandscape evolution framework for developing more effectivegeochemical exploration strategies, and in the interpretation ofgeochemical datasets in highly weathered terrains. Recently,customised and more focused regolith-landform maps have beencompiled to provide direct information for the exploration industry.These new maps show specific information for geochemical

Customised regolith maps incorporatehydrologic modelled attributes forgeochemical explorationJR Wilford & D Butrovski

Regolith is particularly well developed in Australia with more than 80per cent of the continent’s surface characterised by highly weatheredbedrock and/or transported materials. The formation of an extensive

blanket of transported regolith and weathered bedrock in Australia is largelydue to:• long exposure of most of the land surface to sub-aerial weathering; • preservation due to the overall low relief and recent arid climate with

associated low rates of geomorphic processes; and• tectonic stability of the Australian landmass.

Mapping the regolith, understanding past and presentgeomorphic/geochemical processes and developing models of landscapeevolution are key factors that underpin geochemical exploration in deeplyweathered landscapes.1

Mapping regolith in Australia is largely based on a lands system approach.A lands system is defined by Christian and Stewart as an area of landthroughout which recurring patterns of topography, soils and vegetation canbe recognised.2 The key to the land system approach is the recognition of theinterrelationships between landforms, soils and vegetation. Soil scientists havelong recognised these relationships and have used the term ‘catena’ todescribe a repeating sequence of soils that are spatially associated withchanges in topography. These relationships form the basis for regolith-landform mapping. Regolith-landform maps use landforms as the principalsurrogate for mapping regolith materials and this is justified by the closespatial and genetic associations between regolith and topography. Landformattributes are derived from aerial photography and, where available atappropriate detail, digital elevation models. In areas of poor landformexpression, other mapping surrogates such as gamma-ray spectrometry andenhanced Landsat TM imagery become the principal datasets from whichregolith boundaries are derived.3

The first 1:250 000 regolith-landform map over the Ebagoola map sheetwas compiled by AGSO in 1992.4 The map was compiled from existinggeological mapping, aerial photographs and gamma-ray spectrometricimagery. The use of gamma-ray imagery for mapping regolith including soils5

and in understanding landscape processes was relatively new at the time, butthe imagery is now used widely in regolith map compilation and morerecently in soil/landscape modelling.6

The Ebagoola map and other similar maps7,8 are generic products withequal application for mineral exploration and land-use assessment. Since themid-90s a new generation of specialised and tailored regolith and thematicmaps for the exploration industry have been developed—through theactivities of the Cooperative Research Centre for Landscape Evolution andMineral Exploration (CRC LEME). Examples from Leonora (WA), Tanami (NT),and Selwyn (Qld) are used to illustrate these new thematic maps (seelocations in figure 1).

Figure 1. Location diagram of thestudy areas—Leonora, Tanami andSelwyn

exploration including, forexample, recommended samplingstrategies, estimated thickness oftransported cover, and modelledhydromorphic attributes to assistin metal dispersion studies.

Geochemical samplingstrategy mapsGeochemical sampling strategy mapsare hybrid regolith-landform maps(figure 2) that have been specificallydesigned to aid interpretation ofgeochemical datasets and fordeveloping geochemical samplingstrategies. Geochemical samplingstrategy maps build on the standardregolith maps that show in-situ andtransported materials. They do this byincorporating additional informationfrom detailed regolith-geochemicalorientation studies.9 These studies areusually undertaken at local districtscales (e.g. a 5 x 5 km area or minesites). They provide detaileddescriptions on the composition anddistribution of the regolith, andinformation on the physical andchemical dispersion processes thathave occurred or are currently activewithin the regolith.

Knowledge gained from thesestudies, in particular metal dispersionprocesses and recommended sample



types, are then incorporated and extrapolated to the wider area viageochemical sampling strategy maps. Units on a geochemical samplingstrategy map are grouped by their regolith-geochemical characteristics andsampling approach—for example, lag (lithic, ferruginous, calcareous), soil ordrilling. Although the units on a geochemical sampling map are largely basedon the physical and chemical characteristics of the regolith, genetic landscapemodels are involved in the classification processes. The geochemical samplingstrategy map is therefore an interpretive product that has more focus than thestandard descriptive-based regolith-landform map. Primary descriptive regolithand landform attributes are also provided in the GIS from which the maps aregenerated, allowing the user to modify or develop new geochemical strategyclassification schemes where necessary. The maps are further enhanced byincorporating hydrological attributes (see figure 3) derived from digitalelevation models (DEMs).

Figure 3 (left). Surface flow vectorsdraped over a digital elevation model.The length of the arrows generated bythe algorithm can be proportionallyscaled to the flow accumulation orslope of the DEM. Surface flowdirection including areas of flowconvergence and divergence is readilyinterpreted from the model.

Integrating hydrologicattributesFrom a mineral explorationperspective, hydro-geomorphologicalprocesses are critical inunderstanding landscapegeochemistry—particularly inelement dispersion studies. Chemicalweathering through hydrolysis,oxidation and reduction reactions,and movement of sediment andsolute materials (both vertically andlaterally in the landscape) are mainlycontrolled by surface and near-surface hydrologic pathways.

Water will move downslope intwo ways: as overland flow and asthrough-flow. Overland flow orrunoff consists of surface flow thatoccurs when rainfall exceeds theinfiltration rate of the upper part ofthe regolith. Sheetflow is a commonform of overland flow. Concentratedor channelled overland flow typicallyresults in stream flow. Through-flowin contrast is the movement of waterthrough the regolith. Water

Figure 2 (above). Major geochemical sampling groups that are used in a geochemical sampling strategy compiled overthe Selwyn area 140 kilometres south-east of Mount Isa.9 Bedrock age is used first to separate the landscape intoprospective and non-prospective terrains. Then major in-situ and transported regolith materials are identified andgrouped into geochemical exploration domains. These major geochemical groups are specific to the Selwyn region;other areas are likely to show different associations reflecting their specific geologic and landscape histories.

Duricrusts and saprolite

Major bedrock type(Proterozoic–Post-Proterozoic)

Duricrust—Fe and SiSaprolite Saprolite covered by lags

Bleached–Ferruginous and thin residual soils

Transported

Alluvial sediments Colluvial sediments

Old silicified–Recent Thin (<1.5m)–Thick (>1.5m)



Figure 4 (left). Flow vectorssuperimposed on a regolith map overpart of the Tanami region in theNorthern Territory. Regolith units areclassified according to thickness ofsediments (orange-brown <1 m, green1–10 m). The Tanami area has poorlydefined drainage patterns and verylow relief. Most sediment is beingshed into lower parts of the landscapein the form of colluvial sheet floodfans, rather than in discrete channels.The surface flow vectors predict thelikely movement of colluvial sedimentsin the landscape. As a consequence,flow vectors are important tounderstanding geomorphic dispersionprocesses—particularly in relation tosoil or lag geochemical surveying.

Figure 5 (above). Flow vectorssuperimposed on a DEM over part ofthe Leonora map sheet in the EasternGoldfields of Western Australia.Sediments derived from granite andgreenstone lithologies have beencoloured red and yellow, respectively.Catchment boundaries are shown in blue.

Figure 6 (left). An airborne magneticimage with a first vertical gradient(courtesy of Mount Isa Mines) hasbeen embedded into the map to adda bedrock/structure dimension to theinterpretative process. Flow vectorsoverlayed on a geochemical samplingstrategy map over part of the Selwynregion 140 kilometres south-east ofMount Isa. This combination allowsgeochemical datasets to beinterpreted in terms of the regolithmaterials, areas of active erosion,potential mechanical and chemicaldispersion patterns, and provenanceof sediments.

movement as through-flow has an important control on the intensity anddepth of weathering and the movement of solute materials in the regolith.

Although regolith map units imply past palaeo and present hydro-geomorphic processes (e.g. colluvial sheet flood fans, alluvial sediments),the direction of transport and provenance of sediments are generally notwell shown. This is particularly true in areas with gentle slopes andpoorly defined surface drainage, which typify large parts of centralAustralia (figure 4).



DEM-derived surfaces can complement regolith maps by providinginformation about hydrological and geomorphological processes in thelandscape.10 Holyland discusses the use of DEM for generating slope vectorsmaps for soil sampling and hydromorphic dispersion studies, and ingenerating drainage divides and stream intersections points for streamsediment surveys.11 Building on Holyland’s work, an algorithm has beendeveloped that uses DEMs to generate overland (surface) flow vectors andthen integrates the vectors with regolith and geological maps using streamcatchment boundaries.

The first part of the algorithm generates surface flow vectors in the formof arrows that indicate the likely movement of sediments and solutes in thelandscape.

Overlaying the flow vectors on regolith and geochemical maps allowssurface flow directions, including areas of flow convergence anddivergence, to be analysed with respect to regolith/lithological materials.

After the arrows have been generated, the second part of the algorithmuses drainage catchments to analyse these flow characteristics in associationwith regolith or geological units. Catchment boundaries are generatedautomatically from the DEM or imported as a polygon layer. The algorithmis capable of automated delineation of catchment boundaries at variousscales. Catchment scaling is controlled by factors such as the minimumallowed catchment size (area), flow accumulation threshold at whichdrainage networks derived from the DEM begin to form, and the minimumstream order at which catchments may form. To prevent discontinuity increation of drainage networks, and to ensure correct catchment delineation,all sinks that may be present in the DEM are filled before applying thealgorithm.

Regolith types or lithologies likely to contribute to the greatest supply ofsediment within each catchment are automatically identified. The arrowsare then coloured to indicate sediments derived from different sources(figures 5 & 6). This is done by firstly identifying the active eroding parts ofthe catchment using a slope derived from the DEM. High slopescorrespond to zones of active erosion. Intersecting the zones of high slopewith the underlying geological or regolith map units is used to determinethe predominant lithology or regolith material that is likely to becontributing the most sediment within each catchment.

Combining the datasets in this manner allows potential mechanical andchemical dispersion patterns to be linked directly to regolith andlithological materials. These DEM-derived surfaces are then overlayed onthe geochemical sampling strategy map allowing geochemical datasets tobe interpreted in a regolith, hydromorphic and geomorphic context. Furtherintegration and customisation are achieved by combining these themes withother datasets such as airborne magnetics (figure 6). This integration isachieved using colour–space transformations where the intensity, hue andsaturation components of colour are used to combine different datasets.

Limitations in interpretationsThere are some important assumptions and limitations to consider wheninterpreting surface flow maps. Surface water flow maps generated by thistechnique are based exclusively on information presented in a DEM. Theydo not take into account variation in infiltration rates caused by factorssuch as soil depth, texture and vegetation cover. Flow patterns are likely torelate to present-day geomorphic processes or the last geomorphic event. Ifthe present-day topography differs markedly from the palaeo-topography,the hydrological regime predicted by this technique is unlikely to correlatewith the distribution of older regolith materials. In such cases, the surfaceflow maps will show the likely re-distribution of these older materials in thelandscape by present-day processes. Also the flow grids are based onsurface flow only; movement of deeper groundwater may have littlecorrelation with the topography-derived water flow model, particularly inlow-relief landforms.

References1. Butt CRM & Zeegers H. 1989. Classification of geochemical exploration

models for tropically weathered terrains. Journal of GeochemicalExploration; 32: 65–74.

2. Christian CS & Stewart GA. 1952. General report on survey ofKatherine–Darwin region, 1946. Canberra: CSIRO; land research series 1. In:

Cooke RU & Doornkamp JC, eds.1990. Geomorphology inenvironmental management. Edn 2.Oxford: Clarendon Press.

3. Wilford JR, Craig MA, Tapley IJ &Mauger AJ. 1998. Regolith–landformmapping and its implications forexploration over the Half MoonLake region, Gawler Craton, SouthAustralia. Canberra: CRC LEME,report 92R.

4. Pain CF, Wilford JR & DohrenwendJC. 1992. Ebagoola 1:250 000 mapsheet. Canberra: AustralianGeological Survey Organisation.

5. Wilford JR, Pain CF & DohrenwendJC. 1992. Enhancement andintegration of airborne gamma-rayspectrometric and Landsat imageryfor regolith mapping—Cape YorkPeninsula. Proceedings of the 9thASEG Geophysical Conference &Exhibition, Gold Coast, Oct 5–8.

6. Cook SE, Corner RJ, Groves PR &Grealish GJ. 1996. Use of airbornegamma radiometic data for soilmapping. Australian Journal of SoilResearch; 34: 183–94.

7. Craig MA & Churchward HM. 1995.Leonora regolith landforms 1:250 000 scale map. Canberra:Australian Geological SurveyOrganisation.

8. Chan RA. 1995. Bathurst regolithlandforms 1:250 000 scale map.Canberra: Australian GeologicalSurvey Organisation.

9. Wilford JR. 1997. Regolith-landformcharacteristics, evolution andimplications for exploration over theSelwyn region, Mount Isa. Canberra:CRC LEME, report 44R/ E&M 372R.

10. Moore ID, Lewis A & Ladson AR.1991. Digital terrain modelling: Areview of hydrological,geomorphological and biologicalapplications. Hydrologic Processes;5(1): 3–30.

11. Holyland V. 1995. High-resolutiondigital elevation models: A new datasource. Exploration Geophysics; 26:302–306.

Acknowledgment: Dr PhilMcFadden is thanked for hisassistance with developing the flowvector code in ArcView GIS.

➤ John Wilford, CRC for LandscapeEvolution and MineralExploration, phone +61 2 6249 9455 or [email protected]

➤ Dmitar Butrovski, InformationManagement Branch, AGSO,phone +61 2 6249 9825 or [email protected] "


research news (jan) no.33 15/2/01 11:17 am page 1 newsletter · research news (jan) no.33 15/2/01...

Documents