application of chemostratigraphy to the mungaroo formation, the ...

APPEA Journal 2010 50th ANNIVERSARY ISSUE—371

APPLICATION OF CHEMOSTRATIGRAPHY TO THE MUNGAROO FORMATION,

THE GORGON FIELD, OFFSHORE NORTHWEST AUSTRALIA

Lead authorKen

Ratcliffe

K.T. Ratcliffe1, A.M. Wright1, P. Montgomery2,A. Palfrey3, A. Vonk3, J. Vermeulen3 and M. Barrett3

1Chemostrat Inc5,850 San FelipeSuite 500 Houston, Texas, 77057United States of America2Chevron Energy Technology CompanySeafield HouseAberdeen, AB15XLUnited Kingdom3Chevron Australia Pty Ltd250 St Georges TerracePerth WA [email protected]@[email protected]@[email protected]@[email protected]

ABSTRACT

The Mungaroo Formation in the Gorgon Field is a stratigraphically complex fluvial system of Triassic age. It is also a major hydrocarbon reservoir, therefore un-derstanding its internal stratigraphic architecture is of paramount importance to exploitation of its reserves. Here, the technique of chemostratigraphy is used to construct a correlation framework for the Mungaroo Formation of the Gorgon Field.

Chemostratigraphy is a tool that employs variations in inorganic whole rock geochemistry to enable the char-acterisation and subsequent correlation of sediments. For this study, a total of 1,514 cuttings and core samples from eight wells in the Gorgon Field have been analysed. Using data derived from both claystone and sandstone lithologies, the Mungaroo Formation is divided into nine chemostratigraphic packages, 22 geochemical units and 19 sand units. Additionally, three surfaces identified as time lines (T1–T3) are geochemically defined.

Changes in values of Ga/Rb and Al2O3/(CaO+ MgO+K2O+Na2O) indicate that during deposition of the Mungaroo Formation, the paleoclimate became warmer

and wetter, resulting in increasingly intense hydrolytic weathering. Steps in the values of these ratios allow three surfaces to be identified (T1–T3), at which there is a marked and sustained change in the paleoclimate. These three surfaces represent time lines that provide a quasi-chronostratigraphic framework for the formation. Values of Cr/Al2O3, Cr/Na2O and Nb/Al2O3 are related to changes in sediment provenance and indicate that during deposition of the Mungaroo Formation the provenance became more mafic and less intermediate. It is variations in paleoclimate and provenance modelled from the geo-chemical data that allows the packages, units and sand units to be characterised and correlated.

The chemostratigraphic correlation is more detailed than is available from other stratigraphic techniques. Al-though in most instances the lithostratigraphic correlation of sand units based on wireline log correlation matches the one defined using chemostratigraphy, there are some significant differences between the two that influence reservoir models and gas production.

KEYWORDS

Chemostratigraphy, stratigraphy, whole rock geochemis-try, paleoclimate, provenance, Mungaroo Formation, Briga-dier Formation, Gorgon Field, fluvial, Northern Carnarvon Basin.

INTRODUCTION

The Gorgon Field is a super-giant gas accumulation in 200–300 m of water, 65 km to the west of Barrow Island on the North West Shelf of Western Australia (Fig. 1). The Gorgon Field is located at the southern end of the Rankin Platform in the Northern Carnarvon Basin. The Rankin Platform is an elongate, structurally high area bounding the western sides of the Barrow and Dampier sub-basins, with the Exmouth Plateau lying to the northwest. A long history of subsidence throughout the Jurassic and Cretaceous has produced the thick Jurassic and Cretaceous sediments in both the Barrow and Dampier sub-basins (McClure et al, 1988; Longley et al, 2002). Gas accumulations in the Gor-gon Field and in the other fields along the Rankin Trend are likely to have been sourced from both Jurassic and Triassic strata (AGSO/Geotech, 2000; Longley et al, 2002). The Rankin Platform consists of en-echelon horst blocks of Triassic Mungaroo Formation. There is, however, some

372—50th ANNIVERSARY ISSUE APPEA Journal 2010

K.T. Ratcliffe, A.M. Wright, P. Montgomery, A. Palfrey, A. Vonk, J. Vermeulen and M. Barrett

local preservation of Upper Triassic Brigadier Formation and Lower Jurassic Murat Siltstone below the Intra-Jurassic (Callovian) Unconformity (Fig. 2). The Gorgon Field is a large, northeast-southwest trending, fault-bounded struc-ture, about 45 km long; it ranges in width from less than 2 km in the south of the field to 8 km in the north where the subsidiary splinter blocks form bounding structural elements (Fig. 1).

Mungaroo Formation sandstones are the main reservoirs for the gas accumulations in the Gorgon Field (Sibley et al, 1999; Chevron Australia Pty Ltd, 2009). The Mungaroo Formation is a sequence of stacked reservoir quality fluvio-deltaic sandstones and claystones deposited regressively over the Locker Shale during the Middle and Late Triassic (Fig. 2; Hocking et al, 1987). The Mungaroo Formation ex-ceeds 3,000 m in thickness (Barber, 1988); approximately 2,000 m of Mungaroo Formation has been penetrated in the Gorgon Field without reaching the base. It covers a vast area, extending from onshore to several hundred kilometres offshore, and laterally over the entire North West Shelf area from southwest of the Gorgon Field, to the northwest of the North Rankin and Goodwyn fields (Fig. 1) area.

The Gorgon gas-field was discovered in 1980 with the drilling of Gorgon–1. Seven further appraisal wells were drilled in the field between 1982 and 1998. On 14 Sep-tember 2009, the final investment decision on the Gorgon Project was announced. The project’s scope includes a three-train, 15 million-metric-tonne-per-annum liquefied natural gas (LNG) facility and a domestic gas plant located on Barrow Island.

The resource is owned by a consortium of Chevron Aus-tralia (50%), Shell Development (Australia) (25%), and Mobil Australia Resources (25%); Chevron is the project operator.

Developing high resolution stratigraphic frameworks for fluvial systems such as the Mungaroo Formation is typically difficult. Biostratigraphic recovery is commonly poor, recovered forms are often facies controlled and bio-events are not numerous enough to allow correlation at the reservoir scale (Backhouse et al, 2002). The Mungaroo Formation of the Gorgon Field is typical in this respect and consequently the stratigraphic resolution of the palynological-based correlation is low. Furthermore, log signatures are often repetitive, with similar signatures being developed repeatedly through time. Historically, in the Gorgon Field, these repetitive gamma ray (GR) log signatures have resulted in pure wireline log-based lithostratigraphic sandstone correlation schemes (Sibley et al, 1999). The chemostratigraphic data presented here, however, provide a high resolution stratigraphic correlation, with a 22-fold hierarchical division of the Mungaroo Forma-tion and correlation of 19 sand units. The chemostratigra-phy therefore provides a means for enhanced inter-well reservoir correlation. Furthermore, by understanding the mineralogical and sedimentological controls on the whole rock geochemistry, it is demonstrated that the stratigraphic correlation reflects changes in paleoclimate and sediment provenance through time.

Materials

A total of 1,514 samples have been analysed from eight wells in the Gorgon Field (Table 1). The majority of these samples are from ditch cuttings (1,279), and the remainder are from core chips. Cuttings samples were analysed at 5 m spacing, and the resolution was dictated by the frequency of sample collection while drilling. Prior to preparation for induction coupled plasma (ICP) spectrometry analysis, the

Figure 1

ExmouthSub-Basin

BarrowSub-Basin

DampierSub-Basin

117°30'0"E

117°30'0"E

117°0'0"E

117°0'0"E

116°30'0"E

116°30'0"E

116°0'0"E

116°0'0"E

115°30'0"E

115°30'0"E

115°0'0"E

115°0'0"E

114°30'0"E

114°30'0"E

114°0'0"E

114°0'0"E

113°30'0"E

113°30'0"E

113°0'0"E

113°0'0"E

112°30'0"E

112°30'0"E

19°30'0"S

19°30'0"S

20°0'0"S

20°0'0"S

20°30'0"S

20°30'0"S

21°0'0"S

21°0'0"S

21°30'0"S

21°30'0"S

BarrowIsland

LowendalIslands

MontebelloIslands

DampierPeninsula

CapePreston

BurrupGas Plant

Gas p

ipelin

e

Gorgon

GoodwynNorth Rankin

Eendracht

Scarborough

Thebe

Jupiter

100

200

1000

1000

2000

2000250030003500

40004500

5000

500

1000 LegendGas accumulationOil accumulationJurassic depocentre

Major Fault(Rankin Platform only)

0 9 18 27 36 45

Kilometers140°0'0"E

140°0'0"E

120°0'0"E

120°0'0"E

10°0'0"S 10°0'0"S

30°0'0"S 30°0'0"S

Rankin Trend

RankinPlatform

Alph

a Ar

ch

Inset

N

115°

0'E

115°

0'E

20°30'S 20°30'S

299

371

227

513 514

298

585

441

370

442

586

515

443

587

297

225 226

369

300

516

228

372

444

588

512

296

584

440

368

224

155153 154 156

659658657 660

152

656

WA-37-L

WA-38-L

North Gorgon-3

North Gorgon-4

North Gorgon-2

North Gorgon-1

North Gorgon-6

Central Gorgon-1

Gorgon-1

Gorgon-3

0 5 10

Kilometers

Datum GDA94 - Projection MGA Zone 50

LegendWell

Field extent

Permit boundary

Graticular block

Figure 1. Location of the Gorgon Field and its position relative to other oil and gas accumulations, fields, major Jurassic depocentres and major Rankin Platform faults in the Northern Carnarvon Basin, North West Shelf, Australia. Inset: Gorgon Field extent showing the distribu-tion of exploration and appraisal wells, permit boundaries and graticular blocks.


Application of chemostratigraphy to the Mungaroo Formation, the Gorgon Field, offshore northwest Australia

Barrow Group

Olenekian

Tithonian

Legend Sandstone Siltstone Shale LimestoneMarl

Figure 2. Generalised stratigraphic column for the Northern Carnarvon Basin (modified after Palmer et al, 2005 and Barber, 1988).



cuttings samples were washed to remove surface contami-nation and approximately 200 chips of the predominant li-thology were picked by hand under a binocular microscope. When dealing with cuttings, poor depth control and hole condition can result in cuttings that are not representative of the formation drilled. Therefore, prior to making any chemostratigraphic interpretations, the representivity of the cuttings needs to be assessed. This is achieved here by comparing gamma values calculated from the whole rock geochemical data (ChemGR) with gamma values obtained from wireline tools. Generally, there is a close correspon-dence between the ChemGR and the downhole gamma in the wells of the Gorgon Field, which indicates that largely the cuttings chips selected for analysis are representative of the formation drilled.

The study intervals (Table 1), primarily lie in the Mun-garoo Formation, although in wells Gorgon–1, Gorgon–3, North Gorgon–6, North Gorgon–1 and North Gorgon–2, samples from the Barrow Group have been analysed. In North Gorgon–3, the Brigadier Formation has also been analysed.

Following the protocols of previously published che-mostratigraphy papers (Ratcliffe et al, 2004; Pearce et al, 2005a; Ratcliffe et al, 2006; Ellwood et al, 2008), a small number of samples have been analysed by X-ray diffraction to determine their bulk mineralogy. Such mineralogical data allow an understanding of the relationship between the whole rock geochemistry and the mineralogy to be gained.

Methodology

Chemostratigraphy, as applied here, involves the use of major and trace element geochemistry for the characterisa-tion and correlation of strata. The elemental composition of sediments is highly variable due to source composition,

facies, paleoclimate and diagenesis (Ratcliffe et al, 2007 and references cited there-in). Therefore, even apparently homogenous sequences show differences in their whole rock geochemistry, a fact that has resulted in chemostratig-raphy being extensively used in the petroleum industry to help define stratigraphic correlations between well bores (Ehrenberg and Siring, 1992; Racey et al, 1995; Preston et al, 1998; Pearce et al, 1999; Wray, 1999; Ratcliffe et al, 2004; Pearce et al, 2005a, 2005b; Ratcliffe et al, 2006). The elemental data for this paper have been acquired using inductively coupled plasma spectrometry optical emission spectrometry (ICP OES) and inductively coupled plasma spectrometry mass spectrometry (ICP MS), following a Li-metaborate fusion preparation (Jarvis and Jarvis, 1995). These analytical methods result in data for 50 elements (10 major elements, reported as oxide percent by weight [SiO2, TiO2, Al2O3, Fe2O3, MgO, MnO, CaO, Na2O, K2O and P2O5]; 25 trace elements, reported as parts per million [Ba, Be, Bi, Co, Cr, Cs, Cu, Ga, Hf, Mo, Nb, Ni, Pb, Rb, Sn, Sr, Ta, Tl, Th, U, V, W, Y, Zn, and Zr]; 14 rare earth elements [REE], reported as parts per million [La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Ho, Dy, Er, Tm, Yb, and Lu]). Precision error for the major element data is generally better than 2%, and is around 3% for the high abundance trace element data derived by ICP OES (Ba, Cr, Sc, Sr, Zn and Zr). The remaining trace elements are determined from the ICP MS and data are generally less precise, with precision error in the order of 5%. Accuracy error is ±1% for majors and ± 3–7 ppm for trace element depending on abundance. Expanded uncer-tainty values (95% confidence) that incorporate all likely errors in a statistical framework derived from 11 batches of five certified reference materials (CRMs), each prepared in duplicate, are typically 5–7% (relative) for major ele-ments and 7–12% (relative) for trace elements. The ICP OES and MS facility that produced the data presented here was granted laboratory quality system accreditation

Well Sample intervalCuttings samples Core samples Total samples

analysedper wellClaystone Sandstone Claystone Sandstone

Gorgon–1 3,545–4,400 75 50 1 14 140

Gorgon–3 3,550–4,510 85 74 20 15 194

Central Gorgon–1 3,650–4,598 122 83 2 11 218

North Gorgon–6 3,515–4,285 98 31 17 11 157

North Gorgon–1 3,430–4,495 115 85 18 22 240

North Gorgon–2 3,405–4,315 120 57 16 6 199

North Gorgon–4 3,515–4,150 87 33 36 4 160

North Gorgon–3 3,450–4,240 123 41 37 5 206

Total cuttings 1279 Total core 235

Total 1,514

Table 1. Study intervals and the number of analysed samples.



to ISO 17025:2005 which is equivalent to the ISO 9000 series but focussed on laboratory total quality systems.

Although data for a total of 50 elements are acquired, chemostratigraphic characterisation typically relies upon a relatively small number of these elements or ratios of elements (Pearce et al, 2005a; Ratcliffe et al, 2006). These key elements and element ratios, termed key indices, are used to construct a chemostratigraphic zonation scheme, which is typically hierarchical. In this paper, the following chemostratigraphic hierarchy is applied:1. Package: An interval whose claystone geochemical

composition can be differentiated from others in the study intervals.

2. Unit: An interval whose claystone geochemistry al-lows it to be differentiated from others in its parent package. These are numbered to reflect their parent package, with, for example, Unit 4.1 being the oldest subdivision of Package 4, and 4.3 the youngest subdi-vision of that package.

3. Sand Unit: These are intervals that are composed of discrete sand bodies, their nomenclature is used to reflect their position in the parent package, thus, for example Sand Unit 4a is the oldest unit in Package 4 ,and Sand Unit 4f is the youngest in that package.

CHEMOSTRATIGRAPHICCHARACTERISATION

Previous studies on clay-prone fluvial systems have shown that the claystone lithologies typically provide geo-graphically persistent geochemical markers (Cullers, 1995; Pearce et al, 1999; 2005a). Furthermore, the study intervals are largely clay-prone, with discrete sandstone intervals, resulting in the claystone dataset being notably larger than the sandstone dataset (Table 1) and since, cuttings samples were collected at 5 m intervals, the geochemical characterisation of in tervals that are less than 20–30 m thick difficult due to low sample numbers associated with such thin intervals. Therefore, data from claystone lithologies are primarily used to construct the stratigraphic correlation.

In the Gorgon Field, nine chemostratigraphic packages are defined. Packages 1–7 approximate to the Mungaroo Formation, whereas Package 8 is equivalent to the Brigadier Formation, and Package 10 represents the basal parts of the Barrow Group. A total of 22 units and 19 sand units are identified in the Mungaroo Formation of the Gorgon Field. Additionally, three regionally significant, geochem-ically-defined surfaces are defined, termed T1, T2 and T3, occurring at the top of Package 3, at the top of Package 4 and at the top of Package 6 respectively.

Geochemical characterisation of lithostratigraphic units

Since the whole rock geochemistry of a sedimentary de-posit is largely controlled by its mineralogy, it is unsurpris-ing that lithostratigraphic units typically have markedly different geochemical compositions. The Barrow Group

is readily differentiated from the sediments beneath the Intra-Jurassic Unconformity by the high MgO/Al2O3 values of its component claystones (Figs 3 and 4). The Brigadier Formation is only sampled in North Gorgon–3, where it is differentiated from the Mungaroo Formation by the high Nb/Al2O3 values of its component claystones (Figs 3 and 4).

Geochemical characterisation of the Mungaroo Formation

The Mungaroo Formation is divided into seven Packages, Package 1 being the oldest and Package 7 the youngest. Al-though each of these packages is geochemically distinctive, there is a higher order grouping of the packages, such that: Packages 1–3 lie below surface T1; Package 4 lies between surfaces T1 and T2; Packages 5 and 6 lie between surfaces T2 and T3; and Packages 7 and 8 (=Brigadier Formation) lie above surface T3 (Figs 4 and 5).

Surfaces T1, T2 and T3 are defined by sharp upward increases in values of Ga/Rb and Al2O3/(CaO+MgO+Na2O+ K2O) (Fig. 4). Although both these geochemical variables show a gradual change from low to high throughout the Mungaroo Formation and their chemical logs are serrated, there are several marked steps where both ratios sharply increase upwards and remain higher in the overlying clay-stones than the values beneath the step. It is these sharp changes that allow definition of surfaces T1, T2 and T3 (Fig. 4).

0.5 1.0 1.5 2.00.00

0.03

0.06

0.09

0.12

0.15

0.5 1.0 1.5 2.00.00

0.03

0.06

0.09

0.12

0.15

0.5 1.0 1.5 2.00.00

0.03

0.06

0.09

0.12

0.15

Gorgon-1 Gorgon-3

North Gorgon-3

Barrow Group

Barrow Group

Barrow Group

Mungaroo Fm Mungaroo Fm

Mungaroo Fm

Brigadier Fm

Brigadier Fm

(absent)

Brigadier Fm

(absent)

Barrow GroupBrigadier FormationMungaroo Formation

MgO

/Al 2O

3

Nb/Al2O3

MgO

/Al 2O

3

MgO

/Al 2O

3

Nb/Al2O3

Nb/Al2O3

Figure 3. Geochemical characterisation of the Barrow Group, Brigadier Formation and Mungaroo Formation claystones using binary plots.



Gor

gon-

1Depth

m 3600

3700

3800

3900

4000

4100

4200

4300

4400

Gam

ma

API

MgO

/Al

2O3

Ga/Rb

Al/bases

Na 2O/

Al2O

3

Na 2O/

Al2O

3

Gor

gon-

3

3500

3600

3700

3800

3900

4000

4100

4200

4300

4400

4500

Cen

tral

Gor

gon

1

3600

3700

3800

3900

4000

4100

4200

4300

4400

4500

4600

Nor

th G

orgo

n-6

3500

3600

3700

3800

3900

4000

4100

4200

4300

Nor

th G

orgo

n-1

3400

3500

3600

3700

3800

3900

4000

4100

4200

4300

4400

Nor

th G

orgo

n-2

3400

3500

3600

3700

3800

3900

4000

4100

4200

4300

Nor

th G

orgo

n-4

3500

3600

3700

3800

3900

4000

4100

Nor

th G

orgo

n-3

3400

3500

3600

3700

3800

3900

4000

4100

4200

Pack

age

10 (p

ost

UJU

)Pa

ckag

e 8

Pack

age

7Pa

ckag

e 6a

Pack

age

6Pa

ckag

e 5

Pack

age

4Pa

ckag

e 3

Pack

age

2Pa

ckag

e 1

San

dst

on

e sa

mp

leC

lays

ton

e sa

mp

le

T 3

Surf

ace

T2 S

urf

ace

T1 S

urf

ace

Off-scale

Depth

mGam

ma

API

MgO

/Al

2O3

Ga/Rb

Al/bases

Na 2O/

Al2O

3

Na 2O/

Al2O

3

Depth

mGam

ma

API

MgO

/Al

2O3

Ga/Rb

Al/bases

Na 2O/

Al2O

3

Na 2O/

Al2O

3

Depth

mGam

ma

API

MgO

/Al

2O3

Ga/Rb

Al/bases

Na 2O/

Al2O

3

Na 2O/

Al2O

3

Depth

mGam

ma

API

MgO

/Al

2O3

Ga/Rb

Al/bases

Na 2O/

Al2O

3

Na 2O/

Al2O

3

Depth

mGam

ma

API

MgO

/Al

2O3

Ga/Rb

Al/bases

Na 2O/

Al2O

3

Na 2O/

Al2O

3Depth

mGam

ma

API

MgO

/Al

2O3

Ga/Rb

Al/bases

Na 2O/

Al2O

3

Na 2O/

Al2O

3

Depth

mGam

ma

API

MgO

/Al

2O3

Ga/Rb

Al/bases

Na 2O/

Al2O

3

Na 2O/

Al2O

3

Off-scale

Off-scale

Off-scale

Off-scale

Off-scale

Off-scale

FIG

URE

4F

igu

re 4

. Che

mos

trat

igra

phic

cor

rela

tion

sum

mar

y di

spla

ying

som

e ke

y el

emen

t ra

tios

(see

Fig

. 5 f

or r

emai

ning

key

rat

ios)

. The

fir

st t

hree

che

mic

al l

ogs

are

cons

truc

ted

from

cl

ayst

one-

deri

ved

data

, the

fina

l log

is c

onst

ruct

ed fr

om s

ands

tone

dat

a. In

the

firs

t tr

ack

of c

hem

ical

logs

(G

a/R

b an

d A

l 2O3/

base

s), t

he v

alue

of A

l/bas

es h

as b

een

reve

rsed

, suc

h th

at

high

val

ues

lie t

o th

e le

ft o

f the

tra

ck. T

here

fore

red

sha

ded

part

s of

the

log

repr

esen

ts p

lace

s w

here

bot

h va

riab

les

are

high

and

blu

e sh

aded

are

as w

here

bot

h va

riab

les

are

low

. It

has

been

des

igne

d in

thi

s m

anne

r to

sho

w p

lace

s of

rel

ativ

ely

inte

nse

hydr

olyt

ic w

eath

erin

g (r

ed s

hade

d) a

nd le

ss in

tens

e w

eath

erin

g (b

lue

shad

ed).



Fig

ure

5. C

hem

ostr

atig

raph

ic c

orre

latio

n su

mm

ary

disp

layi

ng s

ome

key

elem

ent

ratio

s (s

ee F

ig. 4

for

rem

aini

ng k

ey r

atio

s). T

he f

irst

thr

ee c

hem

ical

log

s ar

e co

nstr

ucte

d fr

om

clay

ston

e-de

rive

d da

ta, t

he fi

nal l

og is

con

stru

cted

from

san

dsto

ne d

ata.

Sand

uni

ts a

re c

orre

late

d be

twee

n th

e w

ells

. Not

e th

at s

hadi

ng fo

r Pa

ckag

es 3

and

4 h

ave

been

rem

oved

for

clar

ity o

f san

d un

it co

rrel

atio

n; r

efer

to

Figu

re 4

for

Pack

ages

3 a

nd 4

cor

rela

tion.

Pack

age

10 (p

ost

UJU

)Pa

ckag

e 8

(Bri

gad

ier F

m)

Pack

age

7Pa

ckag

e 6a

Pack

age

6Pa

ckag

e 5

Pack

age

4Pa

ckag

e 3

Pack

age

2Pa

ckag

e 1

San

dst

on

e sa

mp

leC

lays

ton

e sa

mp

le

3600

3700

3800

3900

4000

4100

4200

4300

4400

3500

3600

3700

3800

3900

4000

4100

4200

4300

4400

4500

3600

3700

3800

3900

4000

4100

4200

4300

4400

4500

4600

3500

3600

3700

3800

3900

4000

4100

4200

4300

3400

3500

3600

3700

3800

3900

4000

4100

4200

4300

4400

3400

3500

3600

3700

3800

3900

4000

4100

4200

4300

3500

3600

3700

3800

3900

4000

4100

3400

3500

3600

3700

3800

4000

4100

4200

3900

Gor

gon-

1

Gor

gon-

3

Cen

tral

Gor

gon

1N

orth

Gor

gon-

6

Nor

th G

orgo

n-1

Nor

th G

orgo

n-2

Nor

th G

orgo

n-4

Nor

th G

orgo

n-3

Depth

mGam

ma

API

SandUnits

Depth

mGam

ma

API

Cr/

Al2O

3

K2O/

Al2O

3Na 2O

Rb/Cs

SandUnits

Depth

mGam

ma

API

SandUnits

Depth

mGam

ma

API

SandUnits

Depth

mGam

ma

API

SandUnits

Depth

mGam

ma

API

SandUnits

Depth

mGam

ma

API

SandUnits

Depth

mGam

ma

API

SandUnits

6b 6a

5c5d5e5f6b

5f 5e 5d 5c 5b 5a 4a4b4c

3a3ai3aii

4a4b4d4f5a5b5c5d

3a

3aii

4a4b4d5a

5a5b 4a4b4d 3a 2

2 13a4b4d

4e 4b 23a

Off-scale

Off-scale

Off-scaleNb/

Cr/

Al2O

3

K2O/

Al2O

3Na 2O

Rb/Cs

Nb/

Cr/

Al2O

3

K2O/

Al2O

3Na 2O

Rb/Cs

Nb/

Cr/

Al2O

3

K2O/

Al2O

3Na 2O

Rb/Cs

Nb/

Cr/

Al2O

3

K2O/

Al2O

3Na 2O

Rb/Cs

Nb/

Cr/

Al2O

3

K2O/

Al2O

3Na 2O

Rb/Cs

Nb/

Cr/

Al2O

3

K2O/

Al2O

3Na 2O

Rb/Cs

Nb/

Cr/

Al2O

3

K2O/

Al2O

3Na 2O

Rb/Cs

Nb/

MungarooFm

FIG

URE

5



PACKAGE 1

The claystones of this package are differentiated from the younger claystones by their high MgO/Al2O3 and Na2O/Al2O3 values. The top of Package 1 is placed at an upward decrease in values of these two variables and is generally coincident with the base of the relatively thick sand-prone interval that constitutes Package 2 (Fig. 4).

PACKAGE 2

Due to the sand-prone nature of Package 2, relatively few claystones from this package have been analysed, some-what decreasing the confidence of its characterisation; however, the claystones analysed from Package 2 generally have lower Na2O/Al2O3 and MgO/Al2O3 values than those of Package 1 and higher K2O/Al2O3 values than those of Package 3 (Figs 4 and 5). This package can be divided into three units—namely 2.1, 2.2 and 2.3—based on changes in the sandstone chemistry. Unit 2.2 has sandstones that have relatively low Na2O/Al2O values, compared to those above and below (Fig. 4).

PACKAGE 3

The claystones of Package 3 are differentiated from those immediately below by their low K2O/Al2O3 values. They can be differentiated from those above by their lower Cr/Na2O values and higher Na2O/Al2O3, Ga/Rb and Al2O3/bases values (Figs 4 and 5). The top of Package 3 is coinci-dent with the T1 Surface. There are three units associated with Package 3, namely 3.1, 3.2 and 3.3. The lower unit, Unit 3.1 encompasses Sand Unit 3a, which has a distinc-tive geochemical signature. Unit 3.2 is differentiated from Unit 3.3 by its lower Nb/Al2O3 values (Fig. 5).

PACKAGE 4

The claystones of Package 4 have values of Ga/Rb, Al2O3/bases and Cr/Na2O that are intermediate between those of the underlying and overlying packages (Figs 4 and 5). The top of Package 4 is coincident with the T2 Surface. Package 4 is further divided into five units, namely 4.1–4.5. Throughout Package 4 there is an overall increase in the values of both Ga/Rb and Al2O3/bases. In this overall trend there are a series of smaller scale widening and narrowing of the cross-over. It is these smaller scale fluctuations in the Ga/Rb and Al2O3/bases values (Fig. 4) that are used to subdivide Package 4.

PACKAGE 5

The claystones of Package 5 are differentiated from those of the underlying packages by their high values of Ga/Rb, Al2O3/bases and Cr/Na2O (Fig. 4). They are differ-entiated from the overlying claystones of the Mungaroo Formation by their low values of Cr/Na2O. Package 5 is further divided into six units, namely 5.1–5.6. Unit 5.1 is

a thin unit at the base of the package that is characterised by notably high values of Ga/Rb and Al2O3/bases. Units 5.2, 5.3 and 5.4 are defined by the low Cr/Na2O values of Unit 5.3 when compared to those above and below. The tops of Units 5.4 and 5.5 are defined by two sharp drops in values of Ga/Rb and Al2O3/bases.

PACKAGE 6 AND 6A

The claystones of Package 6 and 6a are differentiated from those below by their high values of Cr/Al2O3 and low values of Na2O/Al2O3 (Figs 4 and 5). They can be differen-tiated from those of Package 7 by their lower Ga/Rb and Cr/Al2O3 values, together with higher Na2O/Al2O3 values. Initially, in the Gorgon Field Packages 6 and 6a as defined here were combined into Package 6; however, additional well penetrations outside the Gorgon Field demonstrated that the decrease in Ga/Rb and Al2O3/bases values that characterise Package 6a on Figure 4 is a regionally ex-tensive feature. The data from these well penetrations remain confidential, but the inclusion of Package 6a typi-fies the development of chemostratigraphic frameworks as discussed further below.

PACKAGE 7

The claystones of Package 7 can be differentiated from those below by their high Ga/Rb and Cr/Al2O3 values, and low Na2O/Al2O3 values. Three units are defined in Package 7; Unit 7.1 is characterised by notably high values of Ga/Rb and Al2O3/bases and Unit 7.3 has notably higher values of Nb/Al2O3 than other units in the package. Package 7 is overlain by the Barrow Group (Package 10) in North Gorgon–4 and the Brigadier Formation (Package 8) in North Gorgon–3.

CHEMOSTRATIGRAPHIC CORRELATION

The chemostratigraphic characterisation for claystone data described above was correlated throughout the Gorgon Field (Fig. 6). The correlation clearly demonstrates that the wells of the southern part of the Gorgon Field penetrate older sequences than wells to north. It also demonstrates that progressively younger parts of the Mungaroo Forma-tion are preserved beneath the Upper Jurassic Unconfor-mity to the north; in North Gorgon–3, the Brigadier For-mation (=Package 8) is preserved. The chemostratigraphic correlation complements the limited biostratigraphic data but it is of a higher resolution and also allows a strati-graphic link between wells in the north (North Gorgon–3 and –4) and the wells to the south (Figs 4, 5 and 6). The chemostratigraphic data were acquired at the same time as a new 3D seismic survey and interpretation of that data has been iteratively combined with chemostratigraphic understandings to better constrain reservoir correlation, improve understanding of the stratigraphic controls on sand distribution and assist in clarifying reservoir com-partmentalisation.



Gor

gon-

1

Dep

th m

3600

3700

3800

3900

4000

4100

4200

4300

4400

Gam

ma

AP

I

Packages Package 10 Package 4 Package 3 Package 2 Package 1

Units

4-3

4-2

4-1

3-3

3-2

3-1

2-3

2-2

2-1

1-2

Gor

gon-

3

Dep

th m

3500

3600

3700

3800

3900

4000

4100

4200

4300

4400

4500

Gam

ma

AP

I

Packages Package 10 Package 4 Package 3 Package 2 Package 1

Units

4-3

4-2

4-1

3-3

3-2

3-1

2-3

2-2

2-1

1-2

Cen

tral

Gor

gon

1

Dep

th m

3600

3700

3800

3900

4000

4100

4200

4300

4400

4500

4600

Gam

ma

AP

I

PackagesPackage

10 Package 5 Package 4 Package 3 Package 2 Package 1

Units

5-3

5-2

5-1

4-5

4-4

4-3

4-2

4-1

3-3

3-2

3-1

2-3

2-2

2-1

1-2

Nor

th G

orgo

n-6

Dep

th m

3500

3600

3700

3800

3900

4000

4100

4200

4300

Gam

ma

AP

I

PackagesPackage

10 Package 5 Package 4 Package 3

Units

5-4

5-3

5-2

5-1

4-5

4-4

4-3

4-2

4-1

3-3

3-2

Nor

th G

orgo

n-1

Dep

th m

3400

3500

3600

3700

3800

3900

4000

4100

4200

4300

4400

4500

Gam

ma

AP

I

PackagesPackage

10 Package 5 Package 4 Package 3

Units

5-4

5-3

5-2

5-1

4-5

4-4

4-3

4-2

4-1

3-3

3-2

3-1

Nor

th G

orgo

n-2

Dep

th m

3400

3500

3600

3700

3800

3900

4000

4100

4200

4300

Gam

ma

AP

I

PackagesPackage

10 Package 5 Package 4

Units

5-6

5-5

5-4

5-3

5-2

5-1

4-5

4-4

4-3

Nor

th G

orgo

n-4

Dep

th m

3500

3600

3700

3800

3900

4000

4100

Gam

ma

AP

I

PackagesPackage

10 Package 7 Package 6a Package 6 Package 5

Units

7-2

7-1

5-6

5-5

5-4

Nor

th G

orgo

n-3

Dep

th m

3400

3500

3600

3700

3800

3900

4000

4100

4200

Gam

ma

AP

I

PackagesPackage

10 Package 8 Package 7 Package 6a Package 6 Package 5

Units

7-3

7-2

7-1

5-6

5-5

5-4

Fig

ure

6. S

umm

ary

of c

lays

tone

-bas

ed c

hem

ostr

atig

raph

ic c

orre

latio

n. C

olou

red

pane

ls a

re c

hem

ostr

atig

raph

ic p

acka

ges

(see

Fig

. 4 fo

r ke

y to

pac

kage

col

our)

and

bla

ck t

ie li

nes

are

the

top

of t

he g

eoch

emic

al u

nits

.



GEOCHEMISTRY AND MINERALOGY

The chemostratigraphic correlation presented in Figures 4, 5 and 6 is constructed using selected key element ratios. Geological interpretation of the stratigraphic scheme can be made by understanding the mineralogical controls on these variables. The most pragmatic way to determine ba-sic mineralogical control on elemental data is to directly compare mineralogical and geochemical data acquired from the same sample. Figure 7 displays mineralogical data acquired by XRD analyses (Table 2) with selected elements and element ratios. From this comparison it is apparent that:• SiO2/Al2O3 closely mimics the quartz contents of the

silty claystones;• Al2O3 closely models the total clay content of the silty

claystones;• Cs appears to be related to the proportion of illite/

smectite in the silty claystones;• Ga/Rb appears to be related to the proportion of ka-

olinite in the silty claystones;• (Fe2O3 +MgO)/Al2O3 closely mimics the siderite content

of the silty claystones; and,• Na2O is closely related to plagioclase feldspar content

of the silty claystones.

Principle components analysis (PCA)

Although direct comparison of mineralogical and geo-chemical data enables broad mineral to element associa-

tions to be determined (Fig. 7), PCA, a multivariate statisti-cal method, provides a refined understanding of element to element and therefore element to mineral relationships. The method, which is often applied to geochemical datasets (Pearce et al, 2005a; Svendsen et al, 2007; Ellwood et al, 2008; Pe-Piper et al, 2008), reduces the total number of variables in a dataset—which in this case are the element concentrations—to a smaller number of variables termed principal components (PC) (Shaw, 2003). PC1, PC2, PC3 and PC4 account for around 80% of the total variation in the entire claystone geochemical dataset. The principal component score assigned to each sample is determined from the eigenvectors (EV). The EV plots from PCA of the entire claystone dataset are depicted on Figure 8, where the closer two elements plot to one another in the appro-priate dimension, the more closely they are related to one another within the sediment. On EV1 versus EV2 (Fig. 8a) five broad element associations can be recognised: 1. Group 1: includes SiO2, the concentration of which in

the silty claystones is related to the abundance of quartz (Fig. 7), which usually occurs in the form of silt-grade grains.

0 50

SiO2/Al

2O

3

Quartz

50 - 80

0 80

Al2O

3 (%)

Total Clays

15 - 30

0 15 0 1.2

Rb (ppm) Ga/Rb

Mixed Illite / Smectite

Kaolinite / Illite Siderite Plagioclase

5 - 25 0.1 - 0.3

0 10

(Fe2O

3+MgO)/

Al2O

30 - 0.6

0 15

Na2O (%)

0 - 2

Package 5

Package 4

Package 3

Package 1

Figure 7. Comparison of selected elements and element ratios with selected mineralogical data acquired using XRD analysis. See Table 2 for XRD data. The sample order is the same as in Table 2, such that the basal sample is from 3870 m in North Gorgon-4 and the top sample is from 3680 m in Central Gorgon-1.

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

-0.2 -0.1 0 0.1 0.2

Al

Si

Ti

Fe Mn

Mg

Ca

Na K

P

Ba

Be

Cr

Ni

Sc

Sr

Zn

Zr

V

Co

Cu Ga

Rb

Y

Nb

Mo

Sn

Cs

Hf Ta

Th

U

LREEs MREEs

HREEs

EV1

EV2

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

-0.2 -0.1 0 0.1 0.2 0.3 0.4

Al

Si

Ti

Fe

Mn

Mg

Ca

Na

K

P

Ba

Be

Cr

Ni Sc

Sr

Zn

Zr

V

Co

Cu

Ga Rb

Y

Nb

Mo

Sn

Cs

Hf

Ta

Th

U LREEs

MREEs

HREEs

EV2

EV3

Group 1. Quartz

Group 2. ClayMinerals

Group 3. HeavyMinerals

Illite

Kaolinite

Non Zircon

Zircon

Group 4. Siderite

SideriteClay Minerals/ feldspars

Zircon

Quartz

Group 5. PlagioclaseFeldspar

Figure 8. Eigen vector (EV) cross plots for data derived by PCA of all claystone samples. a) EV1 vs. EV2; b) EV2 vs. EV3. LREE—light rare earth elements, MREE—middle rare earth elements, HREE—heavy rare earth elements. All major elements have been abbreviated, such that P = P2O5 etc. The red text refers to the likely mineral influence on the grouped elements.



2. Group 2: includes K2O, Rb, Sc, Cs, Be, Ga and Al2O3. Al2O3 models the total clay content (Fig. 7). Therefore, the concentrations of elements that plot in proximity to Al2O3 are also primarily controlled by clay miner-als. Generally, Ga is associated with kaolinite, hence its position in close proximity to Al2O3, whereas Rb, Cs and K2O are more common in illite/smectite, hence their separation into a subgroup on Figure 8.

3. Group 3: includes Zr, Hf, Cr, MREE’s, HREE’s, Y, LREE’s, Nb, Yb, Ta, Th and U. Zr and Hf typically reflect the abundance of the heavy mineral zircon (Armstrong et al, 2004). Therefore, other variables that plot in associa-tion with these two elements are likely associated with heavy minerals. Without corroborative heavy mineral analysis (Morton and Hallsworth, 1994), it is not pos-sibly to categorically determine which heavy minerals are controlling the elements, but of the key elements, Nb is typically associated with Ti-oxide heavy minerals such as anatase and rutile (Pearce et al, 2005a, 2005b), while Cr is more typically associated with mafic-derived heavy minerals such as Cr-spinel (Ratcliffe et al, 2007).

4. Group 4: includes CaO, Fe2O3, MnO, Ba and MgO, which are probably associated with the Fe-carbonates such as siderite. As shown on Figure 7, the relationship between these elements and siderite is not perfect and it is as-sumed therefore that multiple minerals affect these elements.

5. Group 5: includes Na2O, which is closely associated to the abundance of plagioclase feldspar (Fig. 7).When EV2 and EV3 are cross-plotted (Figure 8b) a no-

table group of elements including K2O, Rb and Cs plot in association with Na2O and SiO2. As discussed above and demonstrated on Figure 7, SiO2 concentrations reflect the amount of silt grade quarts in the silty claystones and therefore, elements plotting in proximity to SiO2 on eigen vector plots are also to some extent related to silt grade material. Therefore, while the primary control on K2O, Rb and Cs is variation in clay mineralogy, the association of these three elements with SiO2 on Figure 8b suggests that they are controlled by a silt-grade mineral, which is most likely to be K-feldspar.

Mineralogical interpretation of the key elements and element ratios

Changes in the Al2O3/(MgO+CaO+Na2O+K2O) values—or Al2O3/bases values for short—tend to match changes in the Ga/Rb values (Fig. 4). Al2O3 is a relatively stable ele-ment in settings where hydrolytic weathering is intense, whereas MgO, CaO, Na2O and K2O are readily leached in similar settings, so Retallack (1997) uses the Al2O3/bases ratio as an indicator of the degree of hydrolytic weather-ing. Although the absolute values of this variable are low, they do show a systematic increase through time, implying

Well Depth Clays Carbonates Other minerals Totals

Name m Chl Kaol Il/Mica Mx IS* Cal Dol Sid Qtz K-spar Plag Py Ba Anat Clays Carb. Other

CG–1 3,680 1 17 25 8 0 0 9 33 Tr 6 Tr Tr 1 51 9 40

CG–1 3,770 1 24 22 8 0 0 6 37 Tr 1 Tr Tr 1 55 6 39

CG–1 3,870 1 14 36 13 0 0 Tr 29 Tr 5 Tr Tr 2 64 Tr 36

CG–1 3,970 2 15 38 13 0 0 Tr 28 Tr 3 Tr Tr 1 68 Tr 32

CG–1 4,105 Tr 11 39 13 0 0 1 31 Tr 3 Tr Tr 2 63 1 36

CG–1 4,225 1 9 37 11 0 0 2 31 Tr 9 Tr Tr Tr 58 2 40

CG–1 4,450 1 6 33 10 0 Tr 3 37 Tr 10 Tr Tr Tr 50 3 47

CG–1 4,575 1 9 33 11 0 0 3 27 Tr 13 1 1 1 54 3 43

NG–1 3,675 1 13 30 9 0 0 5 31 Tr 9 Tr 1 1 53 5 42

NG–1 3,765 1 15 39 13 0 0 1 25 Tr 4 Tr 1 1 68 1 31

NG–1 3,565 1 24 25 9 0 0 3 33 Tr 2 Tr 1 Tr 59 3 36

NG–1 3,620 1 11 41 14 0 0 Tr 28 Tr 3 Tr 1 1 67 Tr 33

NG–1 3,605 1 28 28 5 0 0 3 31 Tr 1 Tr 2 1 62 3 35

NG–1 3,810 Tr 11 18 4 Tr 0 4 50 Tr Tr Tr 13 Tr 33 4 63

NG–1 3,870 1 32 24 5 0 0 Tr 33 Tr 2 Tr 1 2 62 Tr 38

Table 2. X-ray diffraction results carried out on claystone samples from the Mungaroo Formation. (GG–1—Central Gorgon–1; NG–1—North Gorgon–1; Chl—Chlorite; Kaol—Kaolinite; Ill/Mica—Illite/Mica; Mx IS—Mixed layer illite/smectite; Cal—Calcite; Dol—Dolomite; Qtz—Quartz; K-par—K feldspar; Plag—Plagioclase feldspar; Py—Pyrite; Ba—Barite; Anat—Anatase.)



that hydrolytic weathering became more intense during deposition of the Mungaroo Formation.

Ga and Rb are associated with Group 2 on Figure 8a, indicating that they are both controlled by clay mineral distribution; however, Ga is commonly enriched in kaolinite (Hieronymus et al, 2001), whereas Rb is more prevalent in illite, where it replaces K2O (Welby, 1958). Therefore, the Ga/Rb ratio reflects the kaolinite/illite ratio, which is clearly supported by Figure 7. The Ga/Rb ratio generally increases upward with a close linear relationship (R2=0.82) to Al2O3/bases, implying that as the degree of hydrolytic weathering increases, the claystones are becoming more kaolinitic. Kaolinitic claystones in a fluvial setting are typically formed in hot humid climates, while illite is more typical of drier cooler climates.

By combining Ga/Rb and Al2O3/bases it is possible to make implications about changing paleoclimate through time. Figure 4 presents the chemical logs for Ga/Rb and Al2O3/bases with the scale of latter chemical log reversed. Therefore, a negative separation between the two logs—i.e. both have relatively low values (shaded blue on Figure 4)—implies the claystones have suffered less intense (or prolonged) weathering than those claystones for which there is a positive separation (shaded red on Figure 4) of the chemical log combination.

Nb plots with a group of elements typically associated with heavy minerals on Figure 8, which suggests that its distribution in the claystones of the Mungaroo Forma-tion is related to variations in silt-sized heavy minerals. Typically, Nb occurs in association with Ti-oxide heavy minerals (Pearce et al, 2005; Ratcliffe et al, 2004, 2006), but this cannot be unequivocally demonstrated here. Nb is normalised against Al2O3 to minimise the influence of minor variations in grain size and silt versus clay contents.

The XRD data acquired from the claystones and the interpretations of the PCA results suggest K feldspar is relatively scarce in the claystones and that K2O concentra-tions are likely to be largely related to clay minerals. Illite is a potassium aluminium silicate, whereas kaolinite is an aluminium silicate, so the K2O/Al2O3 ratio reflects illite abundance relative to that of kaolinite in the claystones. As the Ga/Rb ratio reflects the inverse situation, increases in the values of this ratio tend to coincide with decreases in the K2O/Al2O3 values and vice versa.

Several factors appear to be controlling the values of the key element ratio MgO/Al2O3, including variations in the abundance of clay minerals and carbonate minerals. At some horizons over the study intervals, MgO/Al2O3 val-ues appears to be associated with K2O/Al2O3 values (e.g. Packages 3 and 4 in Central Gorgon–1), whereas there are other instances where MgO/Al2O3 values are more closely related to Na2O/Al2O3 values (e.g. Packages 1, 2 and 3 in Central Gorgon–1) (Fig. 9). These changes in association imply that MgO/Al is not controlled by a single mineral phase in the study intervals.

Na2O, as discussed above, is related to the distribution of plagioclase feldspar in the silty claystones.

Cr plots in association with the heavy mineral elements, implying therefore that it is to some extent controlled by

the presence of silt-sized heavy mineral grains. Cr is typi-cally associated with mafic-derived heavy minerals (such and Cr-spinel) (Ratcliffe et al, 2007). Therefore, the Cr/Na2O ratio is likely recording a change in mafic-derived heavy minerals and plagioclase feldspars, i.e. changing sediment provenance.

SANDSTONE DATA

There are two major changes in sandstone geochemistry in the study intervals; one is coincident with the top of Pack-age 2 and the other with the top of Package 5 (Figs 4 and 5). At the Package 2/Package 3 boundary there is marked upward decrease in Rb/Cs values and at the Package 5/Package 6 boundary there is a marked upward decrease in Na2O/Al2O3 values. Although it cannot be unequivocal-ly proven, the Rb/Cs values are most likely reflecting a change in clay mineralogy and Na2O/Al2O3 values a change in the plagioclase feldspar contents of the sandstones. These broad changes in sandstone geochemistry support the more detailed chemostratigraphic correlation afforded by the data derived from claystone lithologies, increasing the confidence in the proposed correlations.

DISCUSSION

The claystones of the three lithostratigraphic units—the Mungaroo Formation, the Brigadier Formation and the Barrow Group—each have a distinctive geochemical fin-gerprint. This is to be expected since lithostratigraphy is based upon changes in lithology and whole rock geochem-istry will change markedly between lithologies. Typically, lithostratigraphic units in hydrocarbon basins are defined using wireline logs, which commonly leads to conflicting lithostratigraphic schemes between fields in the same basin. Here it is demonstrated that chemostratigraphy has the ability to test lithostratigraphic schemes, inde-pendent of wireline log data, thereby providing a means to unify stratigraphic schemes within petroleum basin, i.e. develop stable reference stratigraphic frameworks, sensu Ratcliffe et al (2009). Marked changes in the whole rock geochemistry also occur in the Mungaroo Formation, i.e. a single lithostratigraphic unit. The changes in whole rock geochemistry are reflecting changes in weathering style that reflect long term climatic changes and provenance changes during deposition of the Mungaroo and Brigadier formations.

Definition of the highest-order correlative features—sur-faces T1, T2 and T3—is based upon sharp upward increases in the values of Ga/Rb and Al2O3/bases (Fig. 4). Both of these variables are related to changes in the intensity of weathering, which when recognised over an area the size of the Gorgon Field, are reflecting climatic changes through time. Therefore, since each of these surfaces represents a change in paleoclimate, they can be considered to be chronostratigraphic surfaces in the Gorgon Field. The ef-fects of changing paleoclimate on the Mungaroo Formation in the Gorgon Field documented here are likely to affect age-equivalent sediments beyond the Gorgon Field and



therefore have the potential to provide regional markers throughout the Triassic fluvial deltaic successions of the North West Shelf of Australia.

Definition of the packages is based upon a combination of the weathering indices and changes in Na2O/Al2O3, Cr/Na2O and Nb/Al2O3 values. The latter group of variables are all related to changes in sediment provenance that oc-curred during deposition of the Mungaroo and Brigadier formations. The Na2O/Al2O3 values are proportional to the amount of plagioclase feldspar in the silty claystones and are generally high in the older parts of the Mungaroo Formation and become lower in the younger sections. This change in plagioclase content of the silty claystones is mir-rored by changes in plagioclase content in the sandstones, as modelled using Na2O/Al2O3. Although the exact control on Cr cannot be definitively proven without heavy mineral analysis, it appears to be related to silt/fine sand-sized heavy minerals in the claystones, likely to be mafic in origin. Therefore, the Cr/Na2O ratio is probably reflect-ing the amount of mafic input compared to the amount of plagioclase feldspar and this ratio generally increases through time in the Mungaroo Formation. Again, the ex-act control on Nb cannot be determined, but it is likely related to silt and fine sand-sized heavy mineral grains (typically a Ti-oxide mineral). Through the study intervals, the Nb/Al2O3 values remain relatively low and constant, only showing a marked increase in the youngest parts of the Mungaroo Formation interval (Package 7) and the Brigadier Formation.

Changes in provenance are unlikely to be synchronous events over large distances; however, over shorter distances, such as those between the Gorgon Field wells, changes in provenance are likely to show limited diachroneity. Indeed,

none of the package boundaries cross T surfaces, which implies that there is not marked diachroneity associated with the provenance change on the scale of the Gorgon Field; however, in North Gorgon–3 the Na2O/Al2O3 values associated with Package 5 are markedly higher than is typical for Package 5 in the more southerly wells (Fig. 5). This may imply that during deposition of Package 5, the area around North Gorgon–3 was receiving sediments from a subtly different provenance to the more southerly wells. It is not possible with the present well distribution to test this hypothesis. The ability to identify changes in provenance in different geographic areas is an important feature for hydrocarbon exploration. Simplistically, it can provide information about sediment input points into ba-sins and potential reservoir fairways; however, by enabling the identification of mafic versus acidic provenances it can also provide broad information regarding regional reservoir quality.

The presence of Package 6a (Figs 4 and 5) demonstrates an important aspect of the development of chemostrati-graphic correlation frameworks as the dataset increases. Initially, the interval assigned to Package 6a was grouped together with the interval now assigned to Package 6 as one package. The decrease in Ga/Rb and Al2O3/bases seen at what is now the Package 6/Package 6a boundary was used to represent a geochemical unit boundary in a single package. At this stage of the on-going study, only two well penetrations contained Package 6: North Gorgon–3 and –4; however, analysis of additional well penetrations outside the Gorgon Field (data remain confidential) demonstrated that the decrease in weathering indices seen at what is now the Package 6/Package 6a boundary is a regional fea-ture present in multiple wells. The two units that were

Gorgon-1

Depth

3800

3900

4000

4100

4200

4300

4400

Gamma APIcl

ay

pack

ages

Pack

age

4Pa

ckag

e 3

Pack

age

2Pa

ckag

e 1

sand

uni

tsSa n d 4 e

Sa n d 4 b

Sand3a

Sand

2

Lith

stra

t

Top z 20

S Sand

Top z 10

T Sand

T su

rface

T1

Gorgon-3

Depth

3900

4000

4100

4200

4300

4400

4500

Gamma API

clay

pa

ckag

esPa

ckag

e 4

Pack

age

3Pa

ckag

e 2

Pack

age

1

sand

uni

ts

Sand3a

Sand

2Sa

nd 1

Lith

stra

t

Top z 20

S Sand

Top z 10

T SandT

surfa

ce

T1

Central Gorgon 1

Depth

4100

4200

4300

4400

4500

4600

Gamma API

clay

pa

ckag

esPa

ckag

e 4

Pack

age

3Pa

ckag

e 2

Pack

age

1

sand

uni

ts

Sa n d 4 b

Sa n d 4 a

Sand3a

Sand

2

Lith

stra

t

Top z 20

S Sand

Top z 10

T Sand

T su

rface

T1

North Gorgon-6

Depth

4000

4100

4200

4300

Gamma API

clay

pa

ckag

esPa

ckag

e 4

Pack

age

3

sand

uni

ts

Sand4b

Sand4a

Sand

3ai

i

Sand3a

Lith

stra

t

Top z 20

S Sand

Top z 10

T Sand

T su

rface

T1

North Gorgon-1

Depth

4100

4200

4300

4400

Gamma API

clay

pa

ckag

esPa

ckag

e 4

Pack

age

3

sand

uni

ts

Sand4b

Sand4a

Sand3aii

Sand

3ai

Sand

3a

Lith

stra

t

Top z 20

Q sand

R sand

S Sand

Top z 10

T Sand

T su

rface

T1

T1 Marker

Top zone 20

Top T sand

Top Q sandTop R sandTop zone 10/ Top S sand

Figure 9. Comparison of sandstone correlation derived from wireline-based lithostratigraphic correlation and from chemostratigraphy in the older parts of the Mungaroo Formation. Zone 10, Zone 20 and Q, R, S and T sands are lithostratigraphic units.



defined in Package 6 were therefore elevated to become Packages 6 and 6a.

Although the sandstone data only show two broad geochemical changes in the Mungaroo Formation (Figs 4 and 5), it is possible to place the sandstone intervals into the detailed claystone chemostratigraphic framework. By adopting this approach, together with the limited sandstone data from some of the thinner sandstones, it is possible to identify a total of 19 sand units (Fig. 5). Prior to the applica-tion of chemostratigraphy to the Gorgon Field, sandstone correlation in the Mungaroo Formation was lithostrati-graphic based upon wireline responses; sandstones occu-pying approximately the same position in the succession were correlated, resulting in a layer-cake appearance to sandstone reservoir correlations. In most of the Mungaroo Formation, the chemostratigraphic-based sandstone cor-relation shown on Figure 5 support the lithostratigraphic correlation; however, in several places, the chemostratig-raphy suggest a more complicated reservoir architecture than that suggested by the lithostratigraphic approach. By way of example, one area where the lithostratigraphy and chemostratigraphy diverge is shown in Figure 9. Sands Q, R, S and T are the lithostratigraphic units and highlight the tendency toward layer cake correlations when using wire line-based techniques. Sand 3a is present in all wells displayed, but in Gorgon–1, Gorgon–3 and Central Gorgon–1 it is equivalent to the S Sand, whereas in North Gorgon–6 and North Gorgon–1, it is equivalent to the T Sand. In North Gorgon–1, what is lithostratigraphically the S Sand (=Sand 3ai) does not correlate to other sands, i.e. it is an isolated sand body. By changing the sand correlation in this way the reservoir architecture is modified and clearly will have a marked impact upon reservoir development and future gas production.

CONCLUSIONS

Based on interpretation of whole rock geochemical data from claystones it is demonstrated that:• Each lithostratigraphic unit has a unique geochemical

signature;• The Mungaroo Formation contains three chemostrati-

graphic surfaces that approximate to chronostrati-graphic markers;

• Seven chemostratigraphic packages and 22 units are identified and correlated in the Mungaroo Formation;

• The paleoclimate during deposition of the Mungaroo Formation was becoming increasingly wet, resulting in more intense hydrolytic weathering and increased abundance of kaolinite in the claystone lithologies;

• The older parts of the Mungaroo Formation contain common plagioclase feldspar, implying an intermedi-ate igneous provenance. The intermediate provenance was replaced through time with greater influence of a mafic igneous provenance;

• The Brigadier Formation contains an unknown Nb-rich heavy mineral species, implying that the provenance of this formation is different to that of the Mungaroo Formation; and,

• In places, the chemostratigraphic sand unit correlation deviates from the lithostratigraphic sand correlation to an extent that may impact upon gas production. The variation between litho- and chemostratigraphic cor-relation are potentially vitally important for building reservoir models in the Mungaroo Formation of the Gorgon Field.

ACKNOWLEDGEMENTS

Chemostrat would like to thank David Wray and Lorna Dyer of Greenwich University for carrying out the ICP OES MS analyses and Bill Ellington of Ellington and As-sociates for carrying out the XRD analyses. We are also grateful to Chemostrat and Chevron Australia for allow-ing us time and resources to prepare this manuscript. The authors are grateful to Eric Tenthorey, Andrew Jones and Ben Van Aarssen who reviewed the original manuscript and whose comments were invaluable to the final docu-ment. The authors would like to thank the Gorgon Joint Venture partners, Mobil Australia Resources and Shell Development (Australia) for allowing permission for this paper to be published.

REFERENCES

AGSO/GEOTECH, 2000—Chararacterisation of natural gases from West Australian Basins. Australian Geological Survey Organisation and Geotechnical Services Pty Ltd, multi-client non-exclusive report, unpublished.

ARMSTRONG, A.J.S., LEE, Y.I., VERMA, S.P. AND RA-MASAMY, S., 2004—Geochemistry of sandstones from the upper Miocene Kudankulam Formation, southern India; implications for provenance, weathering, and tectonic setting. Journal of Sedimentary Research, 74, 285–97.

BARBER, P.M., 1988—The Exmouth Plateau deep water frontier: A case history. In: Purcell, P.G. and Purcell, R.R. (eds) Proceedings, The North West Shelf, Australia. Pro-ceedings of PESA Symposium, Perth, 10-12 August, 173–189.

BACKHOUSE, J., BALME, B.E., HELBY, R., MARSHALL, N.G. AND MORGAN, R., 2002—Palynological zonation and correlation of the latest Triassic, Northern Carnarvon Basin. In: Keep, M. and Moss, S.J. (eds) The Sedimentary Basins of Western Australia 3. Proceedings of the Petro-leum Exploration Society of Australia Symposium, Perth, 20–23 October, 179–201.

CHEVRON AUSTRALIA PTY LTD, 2009—Gorgon Field development plan. Confidential internal report, unpub-lished.

CULLERS, R.L., 1995—The controls on the major- and trace-element evolution of shales, siltstones and sandstones of Ordovician to Tertiary age in the West Mountains region, Colorado, USA. Chemical Geology, 123, 107–31.



EHRENBERG, S.N. AND SIRING, E., 1992—Use of bulk chemical analysis in stratigraphic correlation of sand-stones: an example from the Statfjord field, Norwegian continental shelf. Journal of Sedimentary Petrology, 62, 318–30.

ELLWOOD, B.B., TOMKIN, J.H., RATCLIFFE, K.T., WRIGHT, A.M. AND KAFAFY, A.M., 2008—Magnetic Sus-ceptibility and Geochemistry for the Cenomanian/Turonian Boundary GSSP with Correlation to Time Equivalent Core. Palaeo3, 251, 1–22.

HIERONYMUS, B., KOTSCHOUBEY, B. AND BOULÈGUE, J., 2001—Gallium behaviour in some contrasting lateritic profiles from Cameroon and Brazil. Journal of Geochemi-cal Exploration, 72, 147–63.

HOCKING, R.M., MOORS, H.T. AND VAN DE GRAAF, W.J.E., 1987—Geology of the Carnarvon Basin. Geological survey of Western Australia, Bulletin 133, 289.

JARVIS, I. AND JARVIS, K.E., 1995—Plasma spectrometry in earth sciences: techniques, applications and future trends. In: Jarvis, I. and Jarvis, K. E. (eds.) Plasma Spec-trometry in Earth Sciences. Chemical Geology, 95, 1–33.

LONGLEY, I.M., BUESSENSCHUETT, C., CLYDSDALE, L., CUBITT, C.J., DAVIS, R.C., JOHNSON, M.K., MAR-SHALL, N.G., MURRAY, A.P., SOMERVILLE, R., SPRY, T.B. AND THOMPSON, N.B., 2002—The North West Shelf of Australia—a Woodside Perspective. In: Keep, M. and Moss, S.J. (eds.) The Sedimentary Basins of Western Australia 3. Proceedings of the Petroleum Exploration Society of Australia Symposium, Perth, 20–23 October, 27–89.

MCCLURE, I.M., SMITH, D.N., WILLIAMS, A.F., CLEGG, L.J. AND FORD, C.C., 1988—Oil and gas fields in the Bar-row Sub-basin. In: Purcell P.G. and Purcell, R.R. (eds) The North West Shelf, Australia. Proceedings of the Petroleum Exploration Society of Australia Symposium, Perth, 10–12 August, 372–90.

MORTON, A.C. AND HALLSWORTH, C.R., 1994—Iden-tifying provenance-specific features of detrital heavy mineral assemblages in sandstones. Sedimentary Geology, 90, 241–56.

PALMER, N., THEOLOGOU, P., KORN, B.E. AND MUNCK-TON, T., 2005—The Wheatstone gas discovery: A case study of Tithonian and Late Triassic fluvial reservoirs. APPEA Journal, 45 (1), 333–47.

PE-PIPER, G., TRIANTAFYLLIDIS, S. AND PIPER, D.J.W., 2008—Geochemical identification of clastic sediment provenance from known sources of similar geology: the Cretaceous Scotian Basin, Canada. Journal of Sedimentary Research, 78, 595–607.

PEARCE, T.J., BESLY, B.M., WRAY, D.S. AND WRIGHT,

D.K., 1999—Chemostratigraphy: a method to improve interwell correlation in barren sequences—a case study using onshore Duckmantian/Stephanian sequences (West Midlands, U.K.). Sedimentary Geology, 124, 197–220.

PEARCE, T.J., WRAY, D.S., RATCLIFFE, K.T., WRIGHT, D.K. AND MOSCARIELLO, A., 2005a—Chemostratigraphy of the Upper Carboniferous Schooner Formation, southern North Sea. In: Carboniferous hydrocarbon geology: the southern North Sea and surrounding onshore areas. In: Collinson, J.D., Evans, D.J., Holliday, D.W. and Jones N.S. (eds) Yorkshire Geological Society, Occasional Publica-tions series, 7, 147–64.

PEARCE, T.J., MCLEAN, D., WRAY, D.S., WRIGHT, D.K., JEANS, C.J. AND MEARNS E.W., 2005b—Stratigraphy of the Upper Carboniferous Schooner Formation, southern North Sea: chemostratigraphy, mineralogy, palynology and Sm–Nd isotope analysis. In: Carboniferous hydrocarbon geology: the southern North Sea and surrounding on-shore areas. In: Carboniferous hydrocarbon geology: the southern North Sea and surrounding onshore areas. In: Collinson, J.D., Evans, D.J., Holliday, D.W. and Jones N.S. (eds) Yorkshire Geological Society, Occasional Publica-tions series, 7, 165–82.

PRESTON, J., HARTLEY, A., HOLE, M., BUCK, S., BOND, J., MANGE, M. AND STILL, J., 1998—Integrated whole-rock trace element geochemistry and heavy mineral chemistry studies; aids to the correlation of continental red-bed reservoirs in the Beryl Field, UK North Sea. Petroleum Geoscience, 4, 7–16.

RACEY, A., LOVE, M.A., BOBOLECKI, R.M. AND WALSH, J.N., 1995—The use of chemical element analysis in the study of biostratigraphically barren sequences: an example from the Triassic of the central North Sea (UKCS). In: Dun-nay, R.E. and Hailwood, E. A. (eds) Non-biostratigraphical methods of dating and correlation. Geological Society, London Special Publication 89, 69–105.

RATCLIFFE, K.T., WRIGHT, A.M., HALLSWORTH, C., MORTON, A., ZAITLIN, B.A., POTOCKI, D. AND WRAY, D.S., 2004—Alternative correlation techniques in the pe-troleum industry: an example from the (Lower Cretaceous) Basal Quartz, Southern Alberta. Bulletin of the American Association of Petroleum Geologist, 88, 1,419–32.

RATCLIFFE, K.T., HUGHES, A.D., LAWTON, D.E., WRAY, D.S., BESSA, F., PEARCE, T.J. AND MARTIN. J., 2006—A regional chemostratigraphically-defined correlation frame-work for the late Triassic TAG-I in Blocks 402 and 405a, Algeria. Petroleum Geoscience, 12, 3–12.

RATCLIFFE, K.T., MORTON, A., RITCEY, D. AND EVENCHICK, C.E., 2007—Whole rock geochemistry and heavy mineral analysis as exploration tools in the Bowser and Sustut Basins, British Colombia, Canada. Journal of Canadian Petroleum Geology, 55, 320–37.



RATCLIFFE, K.T., WRIGHT, A.M., KREMER, M., DIX, M.C.; HUGHES, S.N., REED, T. AND JACKSON, T., 2009—Work-ing toward a Stable Stratigraphic Reference Framework for Oligo-Miocene Fluvial Sequences in Cook Inlet, Alaska: Chemostratigraphic Characterization and Correlation of the Hemlock, Starichkof and Tyonek Formations. AAPG Annual Convention and Exhibition, Denver, 7–10 June, Search and Discovery Article #90090.

RETALLACK, G.J., 1997—A colour guide to paleosols. New York: John Wiley & Sons, 175.

SHAW, P.J.A., 2003—Multivariate Statistics for the Envi-ronmental Sciences. London: Hodder Arnold, 233.

SVENDSEN, J., FRIIS, H., STOLLHOFEN, H. AND HARTLEY, N., 2007—Facies discrimination in a mixed fluvio-eolian setting using elemental whole-rock geochem-istry—applications for reservoir characterisation. Journal of Sedimentary Research, 77, 23–33.

SIBLEY, D., HERKENHOFF, F., CRIDDLE, D. AND MCLER-IE, M., 1999—Reducing resource uncertainty using seismic amplitude analysis on the Southern Rankin Trend, North West Australia. APPEA Journal, 39 (1), 128–48.

SIGLEO, W.R. AND REINHARDT, J., 1988—Paleosols from some Cretaceous environments in the southeastern United States. In: Reinhradt, J. and Sigleo W.R. (eds) Paleosols and weathering through geological time: principles and applications, Geological Society of America, Special Pa-pers 216, 123–42.

WELBY, C.W., 1958—Occurrence of alkali metals in some Gulf of Mexico sediments. Journal of Sedimentary Petrol-ogy, 28, 431–52.

WRAY, D.S., 1999—Identification and long-range correla-tion of bentonites in Turonian–Coniacian (Upper Creta-ceous) chalks of northwest Europe. Geological Magazine, 136, 361–71.

Authors' biographies continued next page.

Ken Ratcliffe graduated from Imperial College of Science and Technology with an honours degree in geology in 1984. He went on to gain his PhD on sedimentological and palaeontological aspects of the Much Wenlock Limestone Forma-tion from the University of Aston in Birmingham in 1987. After lecturing

at Kingston Polytechnic, Ken moved into the oil industry in 1989, co-founding Chemostrat in 1994. He is a director of Chemostrat International Ltd, overseeing proprietary work in Australia, South America and Canada. Ken continues to publish research based around the applications of inorganic whole rock inorganic geochemical data to geosciences and to the oil industry. Member: SEPM.

THE AUTHORS

Milly Wright graduated from the University of Leicester with an honors degree in geology in 2000. She is studying for a MS at University of Houston and is hoping to extend that research and gain a doctorate at the same university. Her research is on chemostrati-graphic application to mud rocks

of the Ferron Sandstone, Utah. Milly is also a director of Chemostrat International Ltd, in charge of US-based busi-ness development and proprietary project work. Member: AAPG, CSPG and SEPM.

Paul Montgomery graduated from the University of Birming-ham with an honours degree in geological science in 1987 and a masters degree in engineering geology from Durham University in 1988. He went on to gain his PhD on the magnetostratigraphy of cretaceous chalk of southern

England from the University of Southampton in 1995. After conducting stratigraphic research at the University of East Anglia and the University of Kansas, Paul joined Chevron as stratigrapher in 2002 working in the USA, Australia and the UK. He is based in Aberdeen, working as a stratigrapher for the Chevron Energy Technology Company and is an adjunct research fellow in the Department of Geological Sciences and Geography, University of Western Australia.



Andy Palfrey graduated from The University of Leeds with an honours degree in geological sciences in 1985 and The University of London with an MSc in stratigraphy in 1994. Andy joined Chevron UK in 1990 and worked principally in appraisal and asset develop-ment roles in the UK offshore sectors. He spent six years with Chevron in

Houston where he was lead development geologist for sev-eral deepwater Angolan development projects. He is a senior development geologist with Chevron Australia, focussing on the characterisation and modelling of fluvial reservoirs of the Northern Carnarvon Basin. Member: AAPG, EAGE, PESGB and Geological Society of London.

Adam Vonk graduated from the Uni-versity of Waikato, New Zealand, with a BSc degree in 1997 and an MSc (Tech)(Hons) degree in 1999 in earth sciences. Adam’s interests include stratigraphy, sedimentology, sequence stratigraphy, basin analysis and palaeogeographic reconstructions, having worked on Neogene strata in western North Is-

land (New Zealand) sedimentary basins. Adam joined Chevron Australia in 2008 and is a geologist in the regional exploration team, working on fluvial reservoirs of the Northern Carnarvon Basin. Member: PESA and the Geological Society of New Zealand.

Jösta Vermeulen graduated from University College Dublin, Ireland, with a BSc honours degree in geology in 1994 and a MSc in petroleum geology in 1995. Jösta joined Chevron UK in 1999 and worked in exploration and asset development roles in the U.K, Irish and Norwegian offshore sectors. He is a regional geologist with Chevron

Australia, focussing on the fluvial reservoirs of the Northern Carnarvon Basin. Jösta is an honorary research fellow at the University of Western Australia. Member: PESA, PESGB and Geological Society of London.

Michael Barrett graduated from Dur-ham University in 1987 with an honours degree in geology and geophysics and gained his MSc from Imperial College RSM in 1990. Michael joined Chevron UK as a geophysicist in 1990 and worked in the UK until 1996 when he pursued an international career with Chevron Overseas Petroleum Inc. Michael has

worked in the US, Angola, Norway and Australia for Chevron and is the regional project team leader for Chevron Australia.



application of chemostratigraphy to the mungaroo formation, the ...

Documents