short-wave infrared reflectance investigation of sites of

ASTROBIOLOGYVolume 4, Number 3, 2004© Mary Ann Liebert, Inc.

Research Paper

Short-Wave Infrared Reflectance Investigation of Sites ofPaleobiological Interest: Applications for Mars

Exploration

ADRIAN BROWN,1 MALCOLM WALTER,1 and THOMAS CUDAHY1,2

ABSTRACT

Rover missions to the rocky bodies of the Solar System and especially to Mars require light-weight, portable instruments that use minimal power, require no sample preparation, andprovide suitably diagnostic mineralogical information to an Earth-based exploration team.Short-wave infrared (SWIR) spectroscopic instruments such as the Portable Infrared MineralAnalyser (PIMA, Integrated Spectronics Pty Ltd., Baulkham Hills, NSW, Australia) fulfill allthese requirements. We describe an investigation of a possible Mars analogue site using aPIMA instrument. A survey was carried out on the Strelley Pool Chert, an outcrop of stro-matolitic, silicified Archean carbonate and clastic succession in the Pilbara Craton, interpretedas being modified by hydrothermal processes. The results of this study demonstrate the ca-pability of SWIR techniques to add significantly to the geological interpretation of such hy-drothermally altered outcrops. Minerals identified include dolomite, white micas such as il-lite-muscovite, and chlorite. In addition, the detection of pyrophyllite in a bleached andaltered unit directly beneath the succession suggests acidic, sulfur-rich hydrothermal activitymay have interacted with the silicified sediments of the Strelley Pool Chert. Key Words: Pil-bara—Short-wave infrared spectroscopy—Mars—Stromatolites—Archean. Astrobiology 4,359–376.

359

INTRODUCTION

SHORT-WAVE INFRARED (SWIR) reflectance spec-troscopy has gained recognition in the explo-

ration and mining community due to its speed,simplicity, and ability to characterize alterationzones around ore bodies. Recent SWIR studies(Huston et al., 1999; Thompson et al., 1999; Yanget al., 2000, 2001; Herrmann et al., 2001; Bierwirthet al., 2002; Thomas and Walter, 2002) have em-

phasized the instrument’s ability to detect alter-ation minerals such as white micas and chlo-rites. Each of these studies was conducted usingthe Australian-built Portable Infrared MineralAnalyser (PIMA) instrument.

SWIR spectroscopy instruments have beensuggested as an instrument for use on futurelanded missions to Mars (see, e.g., Blaney, 2002).To test the SWIR technique in a geologic settingcomparable to the ancient flood basalts of Mars,

1Australian Centre for Astrobiology, Macquarie University, North Ryde, New South Wales, Australia.2CSIRO Division of Exploration and Mining, ARRC Centre, Technology Park, Western Australia, Australia.

fieldwork was undertaken in the arid, 3.5-Ga Pil-bara Craton of Western Australia. Over 250 SWIRspectra were obtained by a hand-held PIMA IIspectrometer at an outcrop of a heavily silicified,stromatolitic Archean carbonate-chert successionto simulate the investigation of such an outcropby a rover on Mars.

To simulate the rover responding to commandsfrom an Earth-based exploration team, spectrawere taken only on accessible weathered sur-faces of the outcrop, and care was taken to targetmineral assemblages that could be distinguishedvisually. This simulated a situation whereby a re-mote command team would possess only pan-oramic color images from the rover to select lo-cations for collection of spectra.

GEOLOGICAL SETTING

The North Pole Dome in the East Pilbara Gran-ite Greenstone Terrane (Van Kranendonk, 2000)is a structural dome of bedded, dominantly maficgreenstone sequences (the Warrawoona Group)that dip gently away from a central monzogran-ite. The monzogranite has been interpreted as asyn-volcanic laccolith, a product of diapiric up-rise and consanguineous magmatism (Van Kra-nendonk, 2000), though this hypothesis has re-cently been challenged because of a gravimetricsurvey that suggests the granite may be a shal-low feature underlain by basaltic material (Ble-wett et al., 2004). Minor occurrences of felsic vol-canic rocks interbedded with the greenstones arecapped by cherts that indicate hiatuses in vol-canism (Van Kranendonk, 2000).

The dating of a stratigraphic column of the War-rawoona Group is presented in Table 1. Stromato-

lite and putative microfossil occurrences have beendocumented within the Warrawoona Group atthree distinct stratigraphic levels—within theDresser Formation, Apex Chert, and Strelley PoolChert (Dunlop et al., 1978; Walter et al., 1980;Awramik et al., 1983; Lowe, 1983; Schopf, 1993;Ueno, 1998; Hofmann et al., 1999; Van Kranendonk,2000; Ueno et al., 2001a,b; Van Kranendonk et al.,2003). The North Pole Dome is interpreted as anearly setting for life (Groves et al., 1981; Buick, 1990).

The rocks of the Warrawoona Group are dom-inated by thoelitic and komatiitic volcanic suc-cessions, which have been suggested as an ana-logue for the flood basalts of Mars (Baird et al.,1981; Reyes and Christensen, 1994; Mustard andSunshine, 1995; Christensen et al., 2000). Al-though the Earth and Mars have experiencedvastly different weathering environments in thepast 3.5 billion years, the ancient age of theArchean rocks of the North Pole Dome makesthem a compelling analogue for similarly agedparts of the southern highlands on Mars. Thepreservation state of the Warrawoona group isgenerally excellent, in contrast with other EarlyArchaean terranes. Metamorphism has not ex-ceeded greenschist facies throughout the NorthPole Dome, and is most commonly at prehnite-pumpellyite facies (Van Kranendonk, 2000). Thepresence of putative microfossils and stromato-lites at the North Pole Dome makes it an ideal testbed for astrobiological techniques designed to ex-amine ancient fossilized life.

A number of contributions have been made re-garding the depositional setting of the Warra-woona Group. Some researchers have suggesteda shallow marine environment based on the pres-ence of pillow basalts, sedimentary analysis, andRare Earth Element (REE) systematics (Buick and

BROWN ET AL.360

TABLE 1. STRATIGRAPHIC COLUMN OF WARRAWOONA GROUP UNITS

PRESENT AT THE NORTH POLE DOME (VAN KRANENDONK, 2000)

Age (Ga) Unit Fossil assemblages

Euro Basalt3.458 Strelley Pool Chert Well-preserved conical stromatolites (Hofmann et al., 1999)

Panorama FormationApex Basalt Microfossils within Apex Chert (Shopf, 1993)

3.470 Duffer FormationDresser Formation Domical stromatolites and microfossils (Walter et al., 1980;

Awramik et al., 1983)Mt. Ada Basalt

3.515 Coonterunah Group

Dates are derived from U-Pb isotopes from zircons in the units and are accurate to approximately 3 million yearsfrom Thorpe (1992).

Dunlop, 1990; Van Kranendonk et al., 2003),whereas others have suggested an Archean mid-ocean ridge setting due to geochemical analysisof basaltic successions and apparent similaritiesbetween hydrothermal alteration patterns at theNorth Pole Dome and in modern mid-ocean ridgedepositss (Kitajima et al., 2001). The most recentinterpretations invoke a caldera-like environmenton an oceanic plateau based upon interpretationof volcanic conduits in the north of the North PoleDome (Van Kranendonk and Pirajno, 2004).

Barley (1984) suggested that silicification andcarbonate-chlorite alteration of North Pole Domegreenstones were the result of low temperature,low pressure (“epithermal”) hydrothermal alter-ation. Silicification associated with epithermalsystems in early Archean greenstone terrains isalmost ubiquitous (De Wit et al., 1982; Gibson etal., 1983). Others have concluded that silicifica-tion in the Dresser Formation and Strelley PoolChert occurred very early after the deposition ofa shallow subaqueous to subaerial evaporite se-quence (Lowe, 1983; Buick and Dunlop, 1990).The process of early silicification is critically im-portant for preservation of delicate biogenicstructures such as microbial mats.

The Trendall Locality, within the Strelley PoolChert in the southwest part of the North PoleDome, contains a silicified carbonate-clastic suc-cession with well-preserved stromatolites (Hof-mann et al., 1999). The origin of these stromatolitesis controversial because of their ancient age andapparent lack of preserved microfossils. The three-dimensional conical morphology of the laminaeand REE analysis suggesting deposition inArchean seawater conditions support a biotic ori-gin during normal marine stromatolite precipita-tion (see, e.g., Fig. 3g) (Hofmann et al., 1999; VanKranendonk et al., 2003). However, Lindsay et al.(2003) have proposed the structures formed abiot-ically by direct carbonate precipitation from sub-seafloor syn-depositional hydrothermal activity.They propose that evidence from cavity filling tex-tures suggest CO2-rich hydrothermal fluids wereresponsible for deposition of dolomite, with no bi-otic mediation, and later silicification took place asthe hydrothermal activity waned.

In any case, it is clear that the Trendall localityhas been affected by hydrothermal activity, eitherby syn-depositional activity, post-depositional ac-tivity, or both processes. We adopt the view thathydrothermal alteration mediated the depositionof at least some parts of the Strelley Pool Chert.

METHODOLOGY

Reflectance spectroscopy utilizing the SWIR re-gion of the electromagnetic spectrum (1.3–2.5�m) often exploits absorption bands due to bondsbetween hydroxyl (OH�) ions and nearby cationsin the crystal lattice. Minerals that contain hy-droxyl ions are commonly associated with hy-drothermal alteration zones that formed whenhot (�50°C) water entraining solutes passedthrough rock pores or fissures.

Figure 1 displays the SWIR spectra of several al-teration minerals typical of hydrothermal systems.The absorption bands around 2.2 �m are causedby a combination of the v2 fundamental stretchingvibration mode of the OH� hydroxyl ion with theAl-OH bending mode in the crystal lattice of eachmineral (Hunt, 1979). The central wavelength ofthe absorption band varies slightly because of thetype of cation (for example, Fe2� or Mg substitut-ing for Al) ionically bonded to the hydroxyl ion.This spectral characteristic allows the determina-

SWIR SPECTROSCOPY: APPLICATIONS TO MARS 361

FIG. 1. SWIR spectra of common alteration minerals.Courtesy of ISPL (http://www.intspec.com).

tion of relative proportions of Mg or Fe2� to Al inwhite micas like muscovite-phengite. Another sig-nificant absorption band in the SWIR region is thecarbonate (CO3

2-) vibration mode at 2.32 �m, theposition of which also shifts wavelength with vary-ing amounts of Ca and Fe2� relative to Mg in mag-nesite, dolomite, and calcite (Gaffey, 1986).

Table 2 gives a partial list of alteration miner-als discernible using SWIR spectra, as well as de-tails regarding their common mode of occurrence(Thompson et al., 1999).

Mapping alteration systems surrounding orebodies has long been a pursuit of economic geol-ogists (Meyer and Hemley, 1967). By recognizingdistinctive mineralogies that typically form zoneswithin hydrothermal alteration systems, suitable“vectors to ore” can be determined (Galley, 1993).For example, the presence of white mica orsericite is often the result of the breakdown of pla-gioclase when an igneous rock is hydrothermallyaltered. By mapping the occurrence of whitemica, the direction of flow in hydrothermal veinsor alteration zones can be delineated. AirborneSWIR studies are a particularly effective tool forregional mapping of white mica or sericite veins(Cudahy et al., 2000; Brown, 2003), and correla-tion on the ground is possible using hand-held orrover-mounted instruments such as the PIMA.

The PIMA instrument measures reflected lightfrom an internal light source at SWIR wave-lengths between 1.3 and 2.5 �m. The instrumentmust be in direct contact with the sample duringanalysis, since it uses an internal light source for

illumination. This configuration makes it an idealinstrument for analyzing outcrop surfaces (as inthis study), rock chips, or crushed powders. Theinstrument integrates the reflected light from asmall region of approximately 10 mm in diame-ter in front of the detector.

The spectral resolution of the PIMA is 7–10 nm,and the spectral sampling interval is 2 nm. The unithas a signal-to-noise ratio of between 3,500 and4,500 to 1 (see http://www.intspec.com). Measure-ments typically take 1–2 min to acquire, depend-ing on the selected integration time. Following eachspectrum collection cycle, the instrument automat-ically carries out a reflection calibration against aknown internal standard contained within thePIMA. Wavelength calibration is carried out peri-odically by comparing the spectrum of the internalstandard with a known baseline spectrum pro-vided by the manufacturer. The instrument oper-ates from a 12-V NiCd battery. Internal tempera-ture and battery status are measured and reportedto an attached WinCE palm computer or laptop PC.Spectra from the PIMA instrument can be down-loaded for further analysis using programs such asMicrosoft Excel or The Spectral Analyst (TSA).

PIMA instruments have been manufactured byIntegrated Spectronics Pty Ltd. (Baulkham Hills,NSW, Australia) since 1991. Approximately 250units are in service throughout the world. Twomodels were produced: the PIMA II and the up-graded PIMA SP model. Further PIMA modelsare under development (T. Cocks, personal com-munication).

BROWN ET AL.362

TABLE 2. PARTIAL LIST OF MINERALS DETECTABLE USING SWIR INSTRUMENTS

Mineral Standard formula Example mode of occurrence

Chlorite (Mg,Al,Fe)12[(Si,Al)8O20](OH)16 Diagenesis, metamorphism, andhydrothermal alteration

Illite-muscovite K2Al4(Si6Al2)O20(OH)4 Diagenesis, metamorphism, andhydrothermal alteration

Pyrophyllite Al2Si4O10(OH)2 Diagenesis and hydrothermal alteration

Saponite (Ca,Na)0.67(Mg,Fe)6(Si,Al)8O20(OH)4 � 8H2O Weathering, sedimentary, andhydrothermal alteration

Calcite CaCO3 Ca-rich hydrothermal alterationand marine sedimentary

Dolomite CaMg(CO3)2 Mg alteration and diagenesisGypsum CaSO4 � 2H2O EvaporiteEpidote Ca2Fe3�Al2Si3O12(OH) Ca-rich hydrothermal alterationSerpentine Mg3Si2O5(OH)4 Alteration of ultramafic rocksTalc Mg3(Si4O10)(OH)2 Alteration of ultramafic and carbonate rocksJarosite KFe3(SO4)2(OH)6 Weathering of Fe sulfide-bearing rocks

The standard formula is from Deer et al. (1992). Example mode of occurrence is modified from that of Thompsonand Thompson (1996).

In this study, the TSA version 4 program wasused to automatically identify the minerals in theacquired SWIR spectra based on a comparison ofthe acquired spectrum to a library of pure end-member spectra. TSA uses a proprietary algo-rithm developed at CSIRO to identify the pre-dominant one or two SWIR-active mineralswithin the sample. If the spectrum lacks sufficientinfrared-active features to identify the phase(s),the rock is declared “aspectral.” This automatedidentification procedure simulates a scenariowhere a rover independently determines the min-erals present in a rock at a landing site and re-ports the results of its survey to a remote scienceteam. To confirm the accuracy of the automatedidentification program, the same spectra weremanually inspected using The Spectral Geologistsoftware, and visual assessments were conductedto interpret the mineral assemblage by relatingthe shape and wavelength of bands to knownspectra, for example, those provided in an ac-companying mineral library.

TSA uses a database of 500 samples of 42 SWIRactive library (or “endmember”) minerals (Pon-tual, 1997), which are used to determine the clos-est match for unknown sample spectra. Similardatabases have been developed for reflectancespectroscopy techniques by the U.S. GeologicalSurvey (Clark et al., 1990).

Reflectance spectra are often characterized bytheir albedo, which is the overall reflectance re-sponse of a sample (i.e., dark samples have a lowalbedo). SWIR reflectance spectra display a spec-tral continuum that is shaped like an invertedhull. The continuum or “hull shape” may be rel-atively flat or significantly curved, dependingupon the albedo of the specimen. Many softwareproducts, including TSA, provide a method forremoval of the spectral continuum to improve theidentification of spectral absorption bands. Thisnormalization method, termed “convex hull re-moval” (Clark et al., 1987), was applied to all ofthe spectra acquired in this study.

RESULTS

This survey covered a 50- � 30-m section ofoutcrop at the Trendall Locality, an outcrop of the Strelley Pool Chert. To support the collectionof spectra, 1,500 digital photographs were takenof the outcrop; these are available online athttp://aca.mq.edu.au/abrown.htm. Figure 2 dis-plays a geological map interpreted from thesephotos of the Trendall Locality.

A visual assessment was made of rock types atthe Trendall locality before spectra were taken.This simulated the classification of rocks in the


TABLE 3. VISUALLY DETERMINED CHARACTERISTICS OF MINERALS AT TRENDALL OUTCROP

Category name Description

Basalt Overlying basaltic sequence, green-gray chlorite richMudstone Partially silicified layer with predominantly white clasts in

microcrystalline gray chert, occasionally pure gray chert withoutvisible clasts

Pebble conglomerate Conglomeratic with clasts of black chert, some up to 2 cm across, withsome fine white clasts in coarse-grained gray groundmass

Boulder conglomerate Black microgranular quartz clasts with surficial red to purple iron stainingBlack chert Black microgranular quartz, smooth-textured, displaying conchoidal

fracture, possibly kerogenousBlack and white chert As for black chert, but with discontinuous bedding conformable white

chert layers. Occurs in wide laminae and fine laminae formsQuartz Quartz vein, usually cutting through black chert layering. Textures

commonly radiate in towards the center of veins and displaychalcedonic texture typical of vug-filling quartz

Radiating crystal splays Silicified, subvertical radiating crystal splaysPlanar carbonate Brown dolomite, roughly textured with fine grain size, planar layering

evident with pervasive conical stromatolitesSiliceous planar laminate As for carbonate layer, but dolomite partially replaced by silica. Color of

unit varies from light brown to yellow, presumably varying with degree of silicification. Stromatolitic conical planar layering pervasivethroughout the unit

Pyrophyllite schist Pervasively altered unit, slight green in color and golden brownweathered schist. Quartz grains (relicit amygdales) present. Occurs inbulbous (surrounded by white groundmass) and foliated forms

BROWN ET AL.364

FIG

. 2.

Geo

logi

cal

map

of

the

Tre

nd

all

Loc

alit

y m

odif

ied

aft

er V

an K

ran

end

onk

an

d H

ick

man

(20

00)

by

usi

ng

dig

ital

im

ages

of

the

outc

rop

.


vicinity of a robotic rover by remote scientists us-ing panoramic images. Similar panoramic imageswere used intensively in NASA’s Pathfinder (Mc-Sween et al., 1999) and MER (Squyres et al., 2003)missions. On the basis of this assessment, rockswere categorized according to color and texture,as indicated in Table 3. Examples of units presentat the locality are shown in Fig. 3.

Following the rock type and unit categoriza-tion based on visual evidence, the area was sur-veyed with the PIMA spectrometer in the man-ner a rover might examine an outcrop. This wasdone by acquiring several spectra of each unit. Atleast seven spectra were acquired from each rockunit, sufficient to determine its spectral charac-teristics and identify any significant changes be-tween spectra taken from the same unit.

The spectral features of each unit are discussedbelow. Apart from the overlying basalt unit andthe underlying pyrophyllite schist unit, all unitsare part of the Strelley Pool Chert.

Basalt

The overlying Euro Basalt is characterized bya light brown weathering rind and a green chlo-rite-rich mineral assemblage. It is fine grainedand largely eroded at this location. Many sam-ples are present as partially buried bouldersrather than as competent outcrop.

The TSA analysis of the PIMA spectra of thebasalt unit consistently identified Mg-chloritewithin the unit (Table 4). An example spectrumis shown in Fig. 4. The chlorite was identified byits diagnostic absorption band centered at 2.25�m, lack of a feature at 1.55 �m (typical of epi-dote), and the presence of associated chlorite fea-tures at 2.0 and 2.33 �m (McLeod et al., 1987).

Mudstone

The mudstone unit is characterized by a grayfine-grained cherty groundmass, containing mil-limeter-sized white clasts that commonly displaya sugary texture indicative of silicification. It liesnear the top of the Trendall locality, among stratathat have been classified as part of a clastic se-quence based on previous research at this local-ity (Van Kranendonk and Hickman, 2000).

The spectra from the mudstone unit showedvariable signatures, though most displayed asymmetric absorption band diagnostic of illite-muscovite at 2.2 �m. The spectra also demon-strated variation in the amount of bound water,

characterized by differences in the shape of theabsorption band at 1.91 �m. Illite generally has adeeper absorption band at 1.91 �m than mus-covite. A typical spectrum of the mudstone unitis shown in Fig. 4. Visual interpretation of themudstone spectra suggests a white mica is defi-nitely present within the unit, a conclusion basedon the regular appearance of the Al-OH absorp-tion band at 2.2 �m.

Pebble conglomerate

The pebble conglomerate unit is characterizedby a gray-green coarse-grained groundmass con-taining large (up to 2 cm) ovoid clasts. The clastsare often white and sometimes derived from un-derlying units, primarily black chert and mud-stone. This unit is also part of the clastic succes-sion at the Strelley Pool Chert (Van Kranendonkand Hickman, 2000). The presence of rip-up clastsof underlying units argues for a high-energy sub-aqueous depositional environment.

Since the conglomerate is constituted by clastsfrom the mudstone unit (among others), it is notsurprising that the spectra of the pebble conglom-erate unit were similar to those of the mudstoneunit. It appeared, however, that the pebble con-glomerate contained less water, an interpretationbased on the presence of a weaker 1.9 �m absorp-tion band, and the fact that this band was charac-terized by a flatter hull shape. The unit was vari-ably chloritized—strong chlorite absorption bandsat 2.25 and 2.33 �m were present in some samples,but absent in the majority of samples (Table 4).

Boulder conglomerate

This unit consists of large (up to 5–10 cm)clasts, commonly of black chert, but also includ-ing black and white layered chert clasts and pla-nar layered carbonate. The clasts often displayred to purple staining due to the presence of ironoxides.

This unit displays overall low reflectance (typ-ically �20%). Most spectra contained a water ab-sorption band at 1.9 �m that included bound wa-ter (centered at 1.915 �m) and unbound water(centered at 1.93 �m). Unbound water is typicalof free, adsorbed, and trapped water (as expectedfor chert with �1% H2O), whereas bound watercan be associated with minerals like illite andsmectites (Aines and Rossman, 1994). Additionalmineral absorptions of the spectra of this unit in-

FIG. 3. Representative images of the units de-scribed in this paper as observed at the outcrop.Geological hammer is shown for scale. a: Widelaminae black and white chert (i) and fine lami-nae black and white chert (ii). b: Basalt. c: Pla-nar carbonate with stromatolites. d: Mudstone.e: Boulder conglomerate. f: Bulbous pyrophyl-lite schist in white groundmass (i) and foliatedpyrophyllite shist (ii). g: Siliceous planar lami-nate displaying conical stromatolites.

ai aii

b c

e

fii

g

d

fi

366 BROWN ET AL.

TA

BL

E4.

MA

PO

FSP

EC

TR

AL

VE

RSU

SV

ISU

AL

CA

TE

GO

RIE

SA

TT

RE

ND

AL

LL

OC

AL

ITY

Uni

tSI

LIC

AC

AR

BC

AR

BC

AR

BC

AR

BC

AR

BSU

LFSU

LFW

MIC

AW

MIC

AW

MIC

AC

HLO

RC

HLO

RC

HLO

RM

ON

TK

AO

LK

AO

LK

AO

LT

otal

Bas

alt

38

011

Mud

ston

e22

282

557

Pebb

le1

712

21

13

27co

nglo

mer

ate

Bou

lder

11

31

39

cong

lom

erat

eB

lack

che

rt1

31

813

Bla

ck a

nd1

11

14

8w

hite

che

rtQ

uart

z3

17

11R

adia

ting

21

12

11

23

13cr

ysta

lsp

lays

Plan

ar2

181

32

911

652

carb

onat

eSi

liceo

us4

52

11

14

41

1639

plan

arla

min

ate

Pyro

phyl

lite

11

54

20

13sc

hist

Opa

lD

olm

Mag

nSi

dA

nkC

alc

Na-

Alu

K-A

luPy

roM

usc

Ill

Fe-C

hlIn

t-C

hlM

g-C

hlM

ont

Hal

loy

Kao

lN

acA

spec

tral

253

All

asse

ssm

ents

wer

e m

ade

auto

mat

ical

ly u

sing

TSA

com

pute

r co

de.

Min

eral

gro

ups

are

abbr

evia

ted

in u

pper

case

and

app

ear

at t

he t

op o

f th

e ta

ble.

See

tex

t fo

r m

an-

ual

spec

tra

iden

tifi

cati

on a

nd i

nter

pret

atio

n. S

ILIC

A �

silic

a gr

oup,

CA

RB

�ca

rbon

ate

grou

p, S

UL

F �

sulf

ate

grou

p, W

MIC

A �

whi

te m

ica

grou

p, C

HL

OR

�ch

lori

tegr

oup,

MO

NT

�m

ontm

orill

onit

e gr

oup,

KA

OL

�ka

olin

gro

up, D

olm

�d

olom

ite,

Mag

n �

mag

nesi

te, S

id �

sid

erit

e, A

nk �

anke

rite

, Cal

c �

calc

ite,

K-A

lu �

pota

s-si

um a

luni

te,

Na-

Alu

�so

diu

m a

luni

te,

Pyro

�py

roph

yllit

e, M

usc

�m

usco

vite

, Il

l �

illit

e, F

e-C

hl �

iron

-ric

h ch

lori

te,

Int-

Chl

�in

term

edia

te c

hlor

ite,

Mg-

Chl

�cl

inoc

hlor

e, H

allo

y �

hallo

ysit

e, K

aol �

kaol

init

e, N

ac �

nacr

ite.

cluded symmetric bands at 2.2 �m due to thepresence of illite-muscovite and coupled 2.25 and2.33 �m absorptions indicative of chlorite.

Since TSA could not identify any characteris-tics in spectra from this unit that contained blackchert clasts, they were designated by the programas aspectral (Table 4). That is not surprising giventhe fact that chert is inactive in the SWIR. TSAroutine did identify opaline silica in one samplebecause of the presence of a broad Si-OH ab-sorption band at 2.215–2.25 �m.

The presence of lithified planar laminate car-bonate, black chert clasts, and quartz clasts withinthis conglomerate (see Fig. 3e) confirmed that itis older than the planar carbonate unit and thatsome of the black chert had been lithified priorto the deposition of this unit. The lack of planarsiliceous clasts suggests that the silicification ofthe planar carbonate occurred after the conglom-erate was formed.

Black chert

The black chert unit consists of massive mi-crogranular quartz. The black color is due to mi-nor amounts of kerogen within the chert (Katoand Nakamura, 2003). At the Trendall Locality,massive black chert occurs typically as smooth,fine-grained quartz layers 5–10 cm thick whenbedding conformable (often interleaved withother units, such as the mudstone unit), or asmassive crosscutting veins 30–50 cm thick whenoriented approximately normal to the strati-graphic layering. For the purposes of this study,the black chert laminations associated with whitequartz layers were analyzed separately andgrouped as the “black and white chert” unit (dis-cussed below).

The PIMA spectra of the black chert unit wereextremely dark (generally �10% reflectance) andshowed much weaker (if present) water and min-eral absorption bands compared with the otherunits studied. TSA assigned a majority of thespectra as aspectral (Table 4). The spectra dis-played low signal-to-noise ratio because of theirextremely low albedo. In low-albedo situations,the imposition of deviations due to instrument-related noise upon a small signal makes identifi-cation of absorption bands extremely difficult,even after convex hull removal.

Black and white chert

The black and white chert unit is present in twoforms (see Fig. 3a). The fine layered form is re-stricted to a relatively small area of the outcropin the center of Fig. 2, and consists of microgran-ular black chert interstratified at a millimeterscale with microgranular white chert in parallellaminae that display flat bottoms and curvedtops. The morphology of this unit suggests directperiodic sedimentary deposition in a quiet envi-ronment. The coarse layered form is laminar, andthe width of laminae can range up to 5–10 cm.The wide laminae are subparallel to other unitsat the locality, and often appear to radiate fromthe vertical black dykes described previously. Forthis reason, the wide laminae black and whitechert is likely a result of hydrothermal injection.Both forms of black and white layered chert dis-play similar smooth microgranular texture. Inmost cases, the black chert surrounds the whitechert. Both forms are generally found below theclastic upper units and above the carbonate andsiliceous planar laminar units at the Trendall Lo-cality.

BROWN ET AL.368

FIG. 4. Convex hull-removed example spectra of unitsfound at the Trendall Locality.

The PIMA spectra of both forms of the blackand white cherts could not be separated and arereported together here. The black and white chertshowed higher albedo than the spectra of theblack chert unit, though the water and mineralabsorption bands were generally of similar rela-tive intensity. TSA classified most of the spectraas aspectral (Table 4), even though absorptions at2.2 and 2.3 �m were apparent. These small fea-tures are most likely due to small amounts ofwhite mica and carbonate.

Quartz

Where quartz occurs in small (less than 4 cmwide) veins and displays irregular surface tex-tures (unlike the smooth chert units), it was iden-tified as a separate unit for the purposes of thisstudy. The veins, which were not mapped, are notresolvable in Fig. 2 because of their small size.However, they all occurred within the black andwhite chert unit and in most cases represent vug-filling quartz that grew into cavities.

The PIMA spectra of these quartz veinsshowed a much stronger development of the un-bound water absorption band centered at 1.93�m, as shown by a representative spectrum inFig. 4. The broad and more defined character ofthis band contrasts with the weak and generallynarrow bands due to bound water (e.g., typical ofthe black cherts discussed previously). Somequartz spectra were very dark with albedos�10%, similar to those of the black cherts. A smallnumber of the samples were identified by TSA asopaline silica (Table 4), based on the presence ofan absorption band at 2.25 �m. Some spectrashowed symmetric Al-OH absorption bands, pos-sibly related to illite-muscovite. However, mostof the samples were classified by the TSA pro-gram as aspectral. The presence of opal is mostlikely due to late-stage weathering, evidenced bythe sinuous, crosscutting, and vug-filling natureof the quartz veins.

Radiating crystal splays

Radiating crystal splays were identified by earlier researchers (Van Kranendonk and Hick-man, 2000) and were likened by them to beds ofaragonite deposited in modern day travertine (see,e.g., Jones et al., 1997). Recent trace element stud-ies suggest the unit represents a dolomitized re-placement of radiating crystal fans, interpreted assecondary crystal growth below the sediment–wa-ter interface (Van Kranendonk et al., 2003).

Although visibly silicified, the spectra of the unitidentified dolomite within the unit because of thepresence of a wide absorption band at 2.31 �m.

Planar layered carbonate

Planar laminated carbonate beds with pervasiveconical stromatolites (Hofmann et al., 1999) are pre-sent in the southeast part of the outcrop (bottomleft of Fig. 2). At point A on Fig. 2, the unit is lat-erally crosscut by a subvertical black chert, whichdivides well-preserved planar laminated carbon-ate from the siliceous planar laminated unit.

The spectra of the planar layered dolomite dis-played a bimodal nature (Fig. 5). Where stroma-tolite laminae had been preserved, the presenceof Al-OH bonding indicative of kaolin was re-vealed by a deep asymmetric absorption at 2.2�m. This was accompanied by a relatively weakcarbonate absorption band at 2.31 �m. Wherethere was no preserved stromatolite laminae, thespectra lacked an absorption band at 2.2 �m anddisplayed a relatively strong carbonate absorp-tion at 2.31 �m.

As shown in the graph in Fig. 5b, the antipa-thetic nature of the strength of the absorptionbands at 2.2 and 2.31 �m in the planar carbonatelayer demonstrates that the presence of a kaolinmineral on the surface of the outcrop is obscur-ing the carbonate beneath it. The kaolin is mostlikely a weathering product or desert varnishcaught in the closely spaced laminae followingaeolian movement. This hypothesis was strength-ened when the fresh underside of a hand samplewas examined with the PIMA and found to bedolomite with no trace of kaolin, even on the pre-served stromatolitic laminae, thus showing thekaolin was not an original part of the stromato-lite laminae. This process of desert varnish trap-ping by stromatolitic laminae may be relevant onthe windswept plains of Mars.

The nature of the weathering or varnish and itsexclusive distribution on the exposed laminaethat protrude beyond the pure carbonate is notentirely clear, but may provide one method fordetecting preserved stromatolitic laminae in out-crops of martian rocks.

Beneath the planar layered carbonate lies a re-gion devoid of stromatolite laminae. Irregularsinuous veins of quartz penetrate this region in asubvertical orientation, and terminate at a thinlayer of black chert just below the lowest stro-matolite laminae. The presence of unlaminatedcarbonate beneath the stromatolite layer suggests


carbonate deposition commenced prior to perva-sive stromatolite growth. The origin of the quartzveins is not entirely clear, though they could berelated to hydrothermal heating.

The radiating crystals region discussed earliermay be analogous to the unlaminated carbonateregion due to their similar orientation (beneaththe planar layered carbonate and siliceous units)and lack of stromatolitic features.

Siliceous planar laminate

This unit was visually assessed as the silicifiedportion of the planar carbonate layer based on thepresence of stomatolitic laminae of similar verti-

cal and horizontal dimensions. The dominantcolor of the unit is dark gray to white. A smoothmicrocrystalline siliceous texture exists betweenthe stromatolitic layers.

Horizontally, the siliceous planar laminate isseparated from the planar carbonate regions by avertical black chert dyke (point A on Fig. 2). Thesiliceous planar laminate also occurs above thecarbonate planar unit, and is separated by a hor-izontal black chert dyke. The presence of theblack dykes surrounding the carbonate layer sug-gests that the silicification process was impededby the black chert that must have been in posi-tion before the silicification event occurred.Though the vertical black dykes at the Trendall

BROWN ET AL.370

FIG. 5. a: Example spectra taken from preserved stro-matolite laminae (above) and carbonate regions (below).b: Graph displaying antipathetic relationship of abun-dance of carbonate and kaolin within the planar carbon-ate region. See text for explanation.

a

b

Locality may have been responsible for trans-portation of some hydrothermal fluids, such asthose containing kerogen that contributed to thehorizontal wide layered hydrothermal injectionblack and white cherts discussed earlier, it is un-likely that they were fluid conduits responsiblefor silicification of the planar layered carbonates,which was probably a later event.

The partially silicified nature of this carbonateunit was confirmed by the PIMA spectra, manyof which showed minor absorption bands in-dicative of dolomite at 2.31 �m. Several spectracontained an Al-OH related symmetric absorp-tion band at 2.2 �m, indicative of the presence ofwhite mica. About half the spectra of this unitwere designated by the TSA program as aspec-tral (Table 4).

Areas of intense red, brown, and yellow dis-coloration occur sporadically within the generallywhite siliceous unit. Several spectra from this re-gion of the outcrop showed the distinctive bandsof goethite at approximately 1.65 and 1.75 �m.The sporadic nature of the discoloration suggestsrelatively recent weathering may be responsiblefor these features.

The PIMA spectra of the iron-oxide stainedstromatolitic laminae showed weak evidence forkaolin with absorption bands centered at 2.165and 2.209 �m. Generally the spectra of the stro-matolitic laminae were found to lack diagnosticabsorption bands, but all displayed moderatelyhigh albedo.

Pyrophyllite schist unit

Stratigraphically beneath and to the south eastof the area depicted in Fig. 2 lies a bleached brownto golden colored, heavily foliated schist that ex-tends vertically for approximately 50 m beneaththe Strelley Pool Chert. There is no outcrop at thecontact between the chert and schist units. How-ever, as the pyrophyllite schist unit converges onthe base of the Strelley Pool Chert the alterationstyle changes from platy, foliated brown togolden outcrop to smaller bulbous outcrop sur-rounded by a white fine-grained groundmass(see Fig. 3f). The sample locations are shown inthe inset of Fig. 2.

The spectra taken of the bulbous and foliatedforms of the pyrophyllite schist unit clearlyshowed the presence of pyrophyllite, indicatedby a diagnostic symmetric and sharp absorptionband centered at 2.165 �m, accompanied by a

smaller shoulder absorption band at 2.195 �m. Arepresentative example of this spectrum is shownin Fig. 4. Pyrophyllite is a typical alteration min-eral resulting from acidic hydrothermal fluidsrich in sulfur (high-sulfidation) (White andHedenquist, 1990). No clear linkage has been es-tablished between the high-sulfidation alterationbeneath the Strelley Pool Chert and the paragen-esis of units within it. However, since no trace ofpyrophyllite is found in the Euro basalt unit over-lying the Strelley Pool Chert, it is likely that itformed prior to the pyrophyllite schist unit andacted as a barrier to the later high-sulfidation hy-drothermal fluids. An alternative explanation isthat the Strelley Pool Chert units were laid downon a partially eroded phyrophyllite schist unit,though no evidence for an unconformity has beenreported between the units.

Variation in octahedral Al of white mica

Tschermak substitution in phyllosilicate min-erals, where two Al atoms substitute for a Si atomand either a Fe or Mg atom in the crystal lattice[Al2Si-1(Fe,Mg)-1], can be recognized in SWIRspectra by analysis of the Al-OH absorption bandposition at around 2.2 �m (Duke, 1994). An in-crease in the Al content in a crystal structure con-taining Al-OH bonds is reflected by a decrease inthe wavelength of the 2.2 �m absorption band.The wavelength variations typically range from2.217 to 2.197 �m.

All units displaying the Al-OH absorption bandaround 2.2 �m were analyzed for variations in thewavelength of the band minima. No significantTshermak variations were found within the unitsat the Trendall Locality, though small variationswere found between units. Table 5 summarizes thedifferences in average position of the Al-OH ab-sorption bands within Al-OH bearing units.

The lack of Tschermak variation within unitsat the Trendall locality does not rule out varia-tions in the same units some distance from thisoutcrop. These results suggest that at the scale ofthis outcrop, approximately 50 m across, therewere no significant temperature variations thatwould have affected the abundance of Al withinhydrothermal white mica.

Geochemical analysis

To support the PIMA mineral identification inthe samples studied, x-ray diffraction (XRD) spec-tra were taken of selected samples from the Tren-


dall locality. Because of the historical importanceand current state of preservation of the site, sam-pling is discouraged by the Geological Survey ofWestern Australia, and so only limited samplingwas undertaken. Rietveld analysis (Rietveld,1969) was carried out by Sietronics Pty Ltd. (Can-berra, ACT, Australia) using a Bruker AXS D4 x-ray diffractometer, and all other samples wererun at CSIRO Exploration and Mining using aPhillips PW 1050 diffractometer using CuK � ra-diation. Each run was conducted with the x-raydiffractometer set to the following levels: 40 kV,40 mA, step size of 0.03°, and count time per stepof 1.8 s.

Two samples of the planar carbonate unitwere analyzed. The results (Fig. 6) showed thepresence of quartz and dolomite. No other min-erals were detected. Rietveld analysis was carriedout on one of the samples using the Siroquantprogram (Taylor, 1991), which indicated that thesample contained 45 wt% quartz and 55 wt%dolomite. The XRD analysis supported the SWIRdetection of dolomite in the planar carbonateunit. One sample of the pyrophyllite schist unitwas analyzed, and pyrophyllite and quartz wereidentified. Quartz was identified as the sole min-

eral component in samples of the black and whitechert, mudstone, and silicified chert.

DISCUSSION

Previous studies have demonstrated the abil-ity of SWIR techniques to map alteration miner-als in hydrothermally altered terrain (see, e.g.,Thompson et al., 1999). This study has shown theutility of using a SWIR instrument to identifymineral phases indicative of hydrothermal alter-ation in a hydrothermally altered, silicified, stro-matolitic carbonate and clastic succession.

Highly silicified environments are typical ofArchean greenstone terrains that have undergonemoderate hydrothermal alteration (Gibson et al.,1983; Duchac and Hanor, 1987; Van Kranendonkand Pirajno, 2004). Archean greenstone terrainsoften contain cherts that preserve textures andfabrics indicative of past biological activity suchas stromatolites (De Wit et al., 1982). Some work-ers have indicated that chert deposits may beideal places to search for fossilized life on Mars(Walter and Des Marais, 1993).

Similarities between hydrothermal events in thePilbara and Mars

It has been postulated that impact craters (New-som et al., 2001) and sites of gully formation (Brak-enridge et al., 1985; Gulick, 1998) may be sites onMars where hydrothermal waters may have pen-etrated the martian regolith. Sites of hydrothermalactivity on Mars may be characterized by silicifi-cation and cherty sediments such as those foundat the Trendall Locality. Such deposits would beideal sites for the preservation of microbial matfabrics (Walter and Des Marais, 1993).

Although SWIR instruments are limited intheir ability to examine low-albedo black cherts,we have shown that the SWIR technique canidentify minerals such as carbonates and phyl-

BROWN ET AL.372

TABLE 5. VARIATIONS IN WAVELENGTH OF BAND MINIMA FOR AL-OH ABSORPTION

BAND FOR AL-OH-BEARING UNITS AT THE TRENDALL LOCALITY

Unit Al-OH band minima (�m) Number of samples with Al-OH feature

Mudstone 2.207 52Pebble conglomerate 2.205 23Planar carbonate 2.207 20Siliceous planar laminate 2.209 10Pyrophyllite 2.197 11

FIG. 6. XRD spectra of the stromatolitic carbonate unit,showing peaks for dolomite.

losilicates and be used to characterize the degreeand nature of alteration of such minerals, a criti-cal precursor to the localization and discovery oflife or preserved fossils. In the course of a typicalrover mission on Mars, and particularly if a roverwere to encounter a site of astrobiological inter-est on Mars, the minerals present at the site wouldneed to be quickly surveyed to localize areaswhere further, more time-consuming analysescould be carried out. A SWIR instrument such asthe PIMA could serve this role for a rover oper-ating in such a reconnaissance mode.

In preparation for the 2003 Mars ExplorationRover missions, NASA science and engineeringteams used an instrument called IPS, which is sim-ilar to PIMA, on its evaluation rovers (Haldemannet al., 2002). This instrument differs from PIMA inthat the instrument does not have to be in contactwith the rock in order to obtain a spectrum. Thelight of the sun is used by IPS as an illuminationsource. The instrument was found to add consid-erably to the geologists’ ability to characterize min-eralogically distinct geological sequences in a rover-remote science team scenario (Jolliff et al., 2002).

This study shows that automatic rock classifi-cation using an SWIR reflectance spectroscopy in-strument such as PIMA (Pedersen et al., 1999;Moody et al., 2001; Gulick et al., 2003) wouldprove useful in future extraterrestrial robotic mis-sions. Mineral identification algorithms such asthose used by TSA may prove useful in such asystem, though this study has highlighted the dif-ficulty of automatic recognition of low-albedo re-flectance spectra (such as those of black cherts),due to low achievable signal-to-noise ratio.

The ability of the PIMA to identify mineralmodes makes it ideal to support elemental abun-dance instruments such as the Alpha Proton X-ray Spectrometer (Golombek, 1998) or Laser In-duced Breakdown Spectrometer (Wiens et al.,2002). The minerals modes provided by a PIMAspectrum could be quantified using a standardnormalization technique when combined with anelemental abundance measurement. Unusual hy-droxyl-bearing mineral assemblages, such asjarosite and saponite, can be detected unambigu-ously by the PIMA.

CONCLUSIONS

This study has demonstrated the ability ofSWIR reflectance spectroscopy to identify miner-

als in an Archean hydrothermally altered envi-ronment. Minerals identified included carbonatessuch as dolomite, white micas such as illite-mus-covite, opaline silica, clays, and chlorite. Theseidentifications supplemented visual assessmentsand aided in the geological interpretation of theoutcrop under study.

The SWIR spectroscopic method is ideal formounting on a small rover for remote field analy-sis of collected samples. It is lightweight and haslow power requirements, being able to be oper-ated with a 12-V battery. With an instrument suchas the PIMA, calibration and spectrum collectioncan be completed in around 2 min. The avail-ability of a large body of research and spectral li-braries coupled with advanced algorithms inpackages such as TSA make the technique a low-risk supplementary and reconnaissance instru-ment suitable for mineral detection of samplesthat require virtually no sample preparation. Wehave demonstrated its utility even in highly sili-cified environments.

The results of this study show that SWIR spec-troscopy contributed to the classification and in-terpretation of the geology of the Strelley PoolChert. New observations made using the PIMAinstrument contributed the following informa-tion:

1. The presence and extent of carbonate and par-tially replaced dolomitic carbonate was con-firmed.

2. Preserved stromatolitic laminae at the TrendallLocality are covered by a kaolin-rich desertvarnish, most probably trapped by the stro-matolitic laminae, in contrast to the purelydolomitic nature of the carbonate ground-mass.

3. Pyrophyllite was detected in an extensively al-tered and foliated unit (pyrophyllite schist) be-neath the silicified outcrop, which suggeststhat acidic, high-sulfidation hydrothermal al-teration has taken place directly beneath theStrelley Pool Chert.

ACKNOWLEDGMENTS

This research would not have been possiblewithout the generous assistance of the GeologicalSurvey of Western Australia. Assistance in thefield from Naomi Mathers was greatly appreci-ated. Assistance with XRD analysis from Michael


Verrall, CSIRO, was appreciated. Reviews fromMartin Van Kranendonk, Kai Yang, and SherryCady were helpful and enlightening and pro-duced a much-improved manuscript. Abigail All-wood is thanked for her comments on the manu-script. CSIRO Division of Exploration and Miningis thanked for the loan of its PIMA II field spec-trometer. Jonathon Huntington and Peter Masonat CSIRO are thanked for their assistance withprocessing and interpretation of SWIR spectra.

ABBREVIATIONS

PIMA, Portable Infrared Mineral Analyser;REE, Rare Earth Element; SWIR, short-wave in-frared; TSA, The Spectral Analyst; XRD, x-ray dif-fraction.

REFERENCES

Aines, R.D. and Rossman, G.R. (1984) Water in mineral?A peak in the infrared. J. Geophys. Res. 89, 4059–4071.

Awramik, S.M., Schopf, J.W., and Walter, M.R. (1983) Fil-amentous fossil bacteria from the Archaean of WesternAustralia. Precambrian Res. 20, 357–374.

Baird, A.K. and Clark, B.C. (1981) On the original igneoussource of martian fines. Icarus 45, 113–123.

Barley, M.E. (1984) Volcanism and hydrothermal alter-ation, Warrawoona Group, East Pilbara. In Archean andProterozoic Basins of the Pilbara, Western Australia, editedby J.R. Muhling, D.I. Groves, and T.S. Blake, GeologyDepartment (Key Centre) and University Extension,The University of Western Australia, Perth, pp. 23–26.

Bierwirth, P., Huston, D., and Blewett, R. (2002) Hyper-spectral mapping of mineral assemblages associatedwith gold mineralization in the Central Pilbara, West-ern Australia. Econ. Geol. 97, 819–826.

Blaney, D.L. (2002) Visible to short wavelength infraredspectroscopy on rovers: Why we need it on Mars andwhat we need to do on Earth. In LPI Contribution No.1148: Workshop on Mars Infrared Spectroscopy: From The-ory and the Laboratory to Field Observations, edited by L.Kirkland, J. Mustard, J. McAfee, B. Hapke, and M. Ram-sey, Lunar and Planetary Institute, Houston, p. 2010.

Blewett, R.S., Shevchenko, S., and Bell, B. (2004) The NorthPole Dome: A non-diapiric dome in the Archaean Pil-bara Craton, Western Australia. Precambrian Res. 133,105–120.

Brakenridge, G.R., Newsom, H.E., and Baker, V.R. (1985)Ancient hot springs on Mars: Origins and paleoenvi-ronmental significance of small Martian valleys. Geol-ogy 13, 859–862.

Brown, A.J. (2003) Hyperspectral mapping of an ancienthydrothermal system [abstract # 12979]. Astrobiology 2,635.

Buick, R. (1990) Microfossil recognition in Archean rocks:An appraisal of spheroids and filaments from a 3500m.y. old chert-barite unit at North Pole, Western Aus-tralia. Palaios 5, 441–459.

Buick, R. and Dunlop, J.S.R. (1990) Evaporitic sedimentsof early Archean age from the Warrawoona Group,North Pole, Western Australia. Sedimentology 37, 247–277.

Christensen, P.R., Bandfield, J.L., Smith, M.D., Hamilton,V.E., and Clark, R.N. (2000) Identification of a basalticcomponent on the Martian surface from Thermal Emis-sion Spectrometer data. J. Geophys. Res. Planets 105(E4),9609–9621.

Clark, R.N., King, T.V.V., and Gorelick, N. (1987) Auto-matic continuum analysis of reflectance spectra. In JPLPublication 87-30: Proceedings of the Third Airborne Imag-ing Spectrometer Data Analysis Workshop, Jet PropulsionLaboratory, Pasadena, CA, pp. 138–142.

Clark, R.N., King, T.V.V., Klejwa, M., and Swayze, G.A.(1990) High spectral resolution reflectance spectro-scopy of minerals. J. Geophys. Res. 95(B), 12653–12680.

Cudahy, T.J., Okada, K., and Brauhart, C. (2000) Target-ing VMS-style Zn mineralization at Panorama, Aus-tralia, using airborne hyperspectral VNIR-SWIRHyMap data. In International Conference on Applied Ge-ologic Remote Sensing, 14th Las Vegas, Nevada, Proceed-ings, Environmental Research Institute of Michigan, LasVegas, pp. 395–402.

De Wit, M.J., Hart, R., Martin, A., and Abbott, P. (1982)Archean abiogenic and probable biogenic structures as-sociated with mineralized hydrothermal vent systemsand regional metasomatism, with implications forgreenstone belt studies. Econ. Geol. 77, 1783–1802.

Deer, W.A., Howie, R.A., and Zussman, J. (1992) An In-troduction to the Rock Forming Minerals, Longman, Har-low, UK.

Duchac, K.C. and Hanor, J.S. (1987) Origin and timing ofthe metasomatic silicification of an early Archean ko-matiite sequence, Barberton Mountain Land, SouthAfrica. Precambrian Res. 37, 125–146.

Duke, E.F. (1994) Near infrared spectra of muscovite,Tschermak substitution, and metamorphic reactionprogress: Implications for remote sensing. Geology 22,621–624.

Dunlop, J.S.R., Muir, M.D., Milne, V.A., and Groves, D.I.(1978) A new microfossil assemblage from the Ar-chaean of Western Australia. Nature 274, 676–678.

Gaffey, S.J. (1986) Spectral reflectance of carbonate min-erals in the visible and near infrared (0.35–2.55 mi-crons): Calcite aragonite and dolomite. Am. Mineral. 71,151–162.

Galley, A.G. (1993) Characteristics of semi-conformablealteration zones associated with volcanogenic massivesulfide districts. J. Geochem. Explor. 48, 175–200.

Gibson, H.L., Watkinson, D.H., and Comba, C.D.A. (1983)Silicification: Hydrothermal alteration in an Archeangeothermal system within the Amulet Rhyolite Forma-tion, Noranda, Quebec. Econ. Geol. 78, 954–971.

Golombek, M.P. (1998) The Mars Pathfinder mission. Sci.Am. 279(1), 40–49.

BROWN ET AL.374

Groves, D.I., Dunlop, J.S.R., and Buick, R. (1981) An earlyhabitat of life. Sci. Am. 245(4), 56–65.

Gulick, V.C. (1998) Magmatic intrusions and a hy-drothermal origin for fluvial valleys on Mars. J. Geo-phys. Res. Planets 103(E8), 19365–19387.

Gulick, V.C., Morris, R.L., Gazis, P., Bishop, J.L., Alena,R., Hart, S.D., and Horton, A. (2003) Automated rockidentification for future Mars exploration missions [ab-stract 2103]. In 34th Lunar and Planetary Science Confer-ence Abstracts, LPI Contribution No. 1156, Lunar andPlanetary Institute, Houston.

Haldemann, A.F.C., Baumgartner, E.T., Bearman, G.H.,Blaney, D.L., Brown, D.I., Dolgin, B.P., Dorsky, L.I.,Huntsberger, T.L., Ksendzov, A., Mahoney, J.C., Mc-Kelvey, M.J., Pavri, B.E., Post, G.A., Tubbs, E.F., Arvid-son, R.E., Snider, N.O., Squyres, S.W., Gorevan, S., Klin-gelhofer, G., Bernhardt, B., and Gellert, R. (2002) FIDOscience payload simulating the Athena Payload. J. Geo-phys. Res. Planets 107(E11), article number 8006.

Herrmann, W., Blake, M., Doyle, M., Huston, D., Kamprad,J., Merry, N., and Pontual, S. (2001) Short wavelength in-frared (SWIR) spectral analysis of hydrothermal alter-ation zones associated with base metal sulfide depositsat Rosebery and Western Tharsis, Tasmania, and High-way-Reward, Queensland. Econ. Geol. 96, 939–955.

Hofmann, H.J., Grey, K., Hickman, A.H., and Thorpe, R.I.(1999) Origin of 3.45 Ga coniform stromatolites in War-rawoona Group, Western Australia. GSA Bull. 111,1256–1262.

Hunt, G.R. (1979) Near infrared (1.3–2.4 �m) spectra ofalteration minerals—potential for use in remote sens-ing. Geophysics 44, 1974–1986.

Huston, D., Kamprad, J., and Brauhart, C. (1999) Defini-tion of high-temperature alteration zones with PIMA:An example from the Panorama VHMS district, centralPilbara Craton. AGSO Res. Newslett. 30, 10–12.

Jolliff, B., Knoll, A., Morris, R.V., Moersch, J., McSween,H., Gilmore, M., Arvidson, R., Greeley, R., Herkenhoff,K., and Squyres, S. (2002) Remotely sensed geologyfrom lander-based to orbital perspectives: Results ofFIDO rover May 2000 field tests. J. Geophys. Res. Plan-ets 107(E11), article number 8008.

Jones, B., Renaut, R.W., and Rosen, M.R. (1997) Verticalzonation of biota in microstromatolites associated withhot springs, North Island, New Zealand. Palaios 12,220–236.

Kato, Y. and Nakamura, K. (2003) Origin and global tec-tonic significance of Early Archean cherts from the Mar-ble Bar greenstone belt, Pilbara Craton, Western Aus-tralia. Precambrian Res. 125, 191–243.

Kitajima, K., Maruyama, S., Utsunomiya, S., and Liou, J.G.(2001) Seafloor hydrothermal alteration at an Archaeanmid-ocean ridge. J. Metamorph. Geol. 19, 581–597.

Lindsay, J.F., Brasier, M.D., McLoughlin, N., Green, O.R.,Fogel, M., McNamara, K.M., Steele, A., and Mertzman,S.A. (2003) Abiotic Earth—establishing a baseline forearliest life, data from the Archean of Western Australia[abstract 1137]. In 34th Lunar and Planetary Science Con-ference Abstracts, LPI Contribution No. 1156, Lunar andPlanetary Institute, Houston.

Lowe, D.R. (1983) Restricted shallow-water sedimenta-tion of early Archean stromatolitic and evaporitic strataof the Strelley Pool Chert, Pilbara Block, Western Aus-tralia. Precambrian Res. 19, 239–283.

McLeod, R.L., Gabell, A.R., Green, A.A., and Gardavski,V. (1987) Chlorite infrared spectral data as proximityindicators of volcanogenic massive sulfide mineraliza-tion. In Pacific Rim Congress 87, Gold Coast, 1987, Pro-ceedings, Australasian Institute of Mining and Metal-lurgy, Parkville, Australia, pp. 321–324.

McSween, H.Y., Murchie, S.L., Crisp, J.A., Bridges, N.T.,Anderson, R.C., Bell, J.F., Britt, D.T., Bruckner, J.,Dreibus, G., Economou, T., Ghosh, A., Golombek, M.P.,Greenwood, J.P., Johnson, J.R., Moore, H.J., Morris, R.V.,Parker, T.J., Rieder, R., Singer, R., and Wanke, H. (1999)Chemical, multispectral, and textural constraints on thecomposition and origin of rocks at the Mars Pathfinderlanding site. J. Geophys. Res. Planets 104(E4), 8679–8715.

Meyer, C. and Hemley, J.J. (1967) Wallrock alteration. InGeochemistry of Hydrothermal Ore Deposits, edited byH.L. Barnes, Holt, Rinehart and Wilson, New York, pp.166–235.

Moody, J., Silva, R., and Vanderwaart, J. (2001) Data fil-tering for automatic classification of rocks from re-flectance spectra. In Seventh ACM SIGKIDD Interna-tional Conference on Knowledge Discovery and DataMining, ACM, San Francisco, pp. 347–352.

Mustard, J.F. and Sunshine, J.M. (1995) Seeing throughthe dust—martian crustal heterogeneity and links to theSNC meteorites. Science 267, 1623–1626.

Newsom, H.E., Hagerty, J.J., and Thorsos, I.E. (2001) Lo-cation and sampling of aqueous and hydrothermal de-posits in martian impact craters. Astrobiology 1, 71–88.

Pedersen, L., Apostolopoulos, D., Whittaker, W., Cassidy,W., Lee, P., and Roush, T. (1999) Robotic rock classifi-cation using visible light reflectance spectroscopy: Pre-liminary results from the Robotic Antarctic MeteoriteSearch program [abstract 1340]. In 30th Lunar and Plan-etary Science Conference Abstracts, LPI Contribution No.964, Lunar and Planetary Institute, Houston.

Pontual, S. (1997) G-Mex Vol. 1: Special Interpretation FieldManual, Ausspec International, Pty Ltd., Kew, Aus-tralia.

Reed, M.H. (1983) Seawater-basalt reaction and the originof greenstones and related ore deposits. Econ. Geol. 78,466–485.

Reyes, D.P. and Christensen, P.R. (1994) Evidence for ko-matiite-type lavas on Mars from Phobos Ism data andother observations. Geophys. Res. Lett. 21, 887–890.

Rietveld, H.M. (1969) A profile refinement method for nu-clear and magnetic structures. J. Appl. Crystallogr. 2,65–71.

Schopf, J.W. (1993) Microfossils of the Early ArchaeanApex Chert: New evidence of the antiquity of life. Sci-ence 260, 640–646.

Squyres, S.W., Arvidson, R.E., Baumgartner, E.I., Bell, J.F.,Christensen, P.R., Gorevan, S., Herkenhoff, K.E., Klin-gelhofer, G., Madsen, M.B., Morris, R.V., Rieder, R., andRomero, R.A. (2003) Athena Mars rover science inves-tigation. J. Geophys. Res. 108, article number 8062.


Taylor, J.C. (1991) Computer programs for standardlessquantitative analysis of minerals using the full powderdiffraction profile. Powder Diffract. 6, 2–9.

Thomas, M. and Walter, M.R. (2002) Application of hy-perspectral infrared analysis of hydrothermal alterationon Earth and Mars. Astrobiology 2, 335–351.

Thompson, A.J.B. and Thompson, J.F.H. (1996) Atlas of Al-teration: A Field and Petrographic Guide to HydrothermalAlteration Minerals, Mineral Deposits Division, Geolog-ical Association of Canada, St. John’s, Newfoundland.

Thompson, A.J.B., Hauff, P.L., and Robitaille, A.J. (1999)Alteration mapping in exploration: Application of short-wave infrared (SWIR) spectroscopy. SEG Newslett. 39,16–27.

Thorpe, R.I., Hickman, A.H., Davis, D.W., Mortensen,J.K., and Trendall, A.F. (1992) U-Pb zircon geochronol-ogy of Archaean felsic units in the Marble Bar region,Pilbara Craton, Western Australia. Precambrian Res. 56,169–189.

Ueno, Y. (1998) Earth’s oldest bacteria (3.5 Ga) fromW.Australia and the carbon isotope signature [abstract].GSA Abstracts Programs 30, A-98.

Ueno, Y., Isozaki, Y., Yurimoto, H., and Maruyama, S.(2001a) Carbon isotopic signatures of individual mi-crofossils(?) from Western Australia. Int. Geol. Rev. 43,196–212.

Ueno, Y., Maruyama, S., Isozaki, Y., and Yurimoto, H.(2001b) Early Archaean (ca. 3.5 Ga) microfossils and 13Cdepleted carbonaceous matter in the North Pole area,Western Australia: Field occurence and geochemistry.In Geochemistry and the Origin of Life, edited by S.Nakashima, S. Maruyama, A. Brack, and B.F. Windley,Universal Academy Press, Tokyo, pp. 203–236.

Van Kranendonk, M.J. (2000) Geology of the North Shaw1:100 000 Sheet, Geological Survey of Western Australia,Department of Minerals and Energy, Perth.

Van Kranendonk, M. and Hickman, A.H. (2000) ArchaeanGeology of the North Shaw Region, East Pilbara GraniteGreenstone Terrain, Western Australia—A Field Guide, Ge-ological Survey of Western Australia, Department ofMinerals and Energy, Perth.

Van Kranendonk, M.J. and Pirajno, F. (2004) Geochem-istry of metabasalts and hydrothermal alteration zonesassociated with ca. 3.45 Ga chert �/- barite deposits:Implications for the geological setting of the Warra-

woona Group, Pilbara Craton, Australia. Geochem. Ex-plor. Environ. Anal. 4, 253–278.

Van Kranendonk, M., Webb, G.E., and Kamber, B.S. (2003)Geological and trace element evidence for a marine sed-imentary environment of deposition and biogenicity of3.45 Ga stromatolitic carbonates in the Pilbara Craton,and support for a reducing Archaean ocean. Geobiology1, 91–108.

Walter, M.R. and Des Marais, D.J. (1993) Preservation ofbiological information in thermal-spring deposits—de-veloping a strategy for the search for fossil life on Mars.Icarus 101, 129–143.

Walter, M.R., Buick, R., and Dunlop, J.S.R. (1980) Stro-matolites 3400–3500 Myr old from the North Pole area,Western Australia. Nature 284, 443–445.

White, N.C. and Hedenquist, J.W. (1990) Epithermal en-vironments and styles of mineralization: variations andtheir causes, and guideline for exploration. J. Geochem.Explor. 36, 445–474.

Wiens, R.C., Arvidson, R.E., Cremers, D.A., Ferris, M.J.,Blacic, J.D., Seelos, F.P., and Deal, K.S. (2002) Combinedremote mineralogical and elemental identification fromrovers: Field and laboratory tests using reflectance andlaser-induced breakdown spectroscopy. J. Geophys. Res.Planets 107(E11), article number 8004.

Yang, K., Huntington, J.F., Brown, P.R.L., and Ma, C.(2000) An infrared spectral reflectance study of hy-drothermal alteration minerals from the Te Mihi sectorof the Wairakei geothermal system, New Zealand. Geothermics 29, 377–392.

Yang, K., Browne, P.R.L., Huntington, J.F., and Walshe,J.L. (2001) Characterising the hydrothermal alterationof the Broadlands-Ohaaki geothermal system, NewZealand, using short-wave infrared spectroscopy. J.Volcanol. Geotherm. Res. 106, 53–65.

Address reprint requests to:Adrian Brown

Australian Centre for AstrobiologyMacquarie University

North Ryde, NSW 2109, Australia

E-mail: [email protected]

BROWN ET AL.376

short-wave infrared reflectance investigation of sites of

Documents