EconomicGeology

Vol 79, 1984, pp. 638-655

Gold Deposits and Hydrothermal Eruption BrecciasAssociated

with a Maar Volcano at Wau, Papua New Guinea
RICHARD H. SILLITOE,
8 WestHill Park, HighgateVillage, LondonN6 6ND, England

E. MAX BAKER,
E. M. Bakerand Associates
Pty. Ltd., 115 RossRiver Road,Mundingburra,Townsville4812, Australia

AND WILLIAM A. BROOK

GeopacificServicesPty. Ltd., P.O. Box 619, North Sydney,New SouthWales2060, Australia

Abstract

The Wau districthasproducedmuchof the 18 metrictons(0.58 millionoz) of lode gold

attributedto the Morobe goldfieldin PapuaNew Guinea.Geologicmappingandreappraisal
of the districthave shownthat Plioceneepithermalgold mineralizationwasassociatedwith
generationof a maar volcanoand associatedendogenousdomes.
Emplacement of endogenous domesanderuptionofpyroclastic flowsin the Morobegoldfield
tookplacein the intervalfrom 4 to 2.4 m.y. ago.An undateddomeat Edie Creek is spatially,
and perhapstemporally,associated with lode gold mineralization.
During the later part of this time interval,followingtilting of the ignimbritepile, a maar
wasgeneratedat Wau. The subjacentdiatremeis believedto extenddownwardfor about 1
km to a regionaldip-slipfaultzonewhichhostsdeeperlevel goldmineralizationat the nearby
Ribroastermine. Interactionof daciticmagmaand groundwater in depth on the fault zone
causeda seriesof hydrovolcanicexplosions which resultedin pyroclasticdepositsof ballistic
fall and basesurgeorigin, the latter carryingabundantaccretionarylapilli. The pyroclastics
accumulatedwithin the maar and constructedan encirclingtuff ring, only a smallpart of
which is preservedoverlyingLate Cretaceous-Paleogene phyllitesto the southand west of
the maar.
Followinginitial maarformation,hydrothermalbrecciationandgold depositionwere con-
trolled by low-angleextensionalfaultswhich developedduring incipient subsidenceof a
wedgeof poorlysupportedrockbetweenthe regionalfaultzoneandthe ring faultdelimiting
the maar. Fluid overpressures, which developedat depthsof >100 m beneathself-sealed
caprocks,triggeredhydrothermaleruptionswhichgaveriseto a ramifyingsystemof irregular
veinsandbodiesof hydrothermalbrecciasurmountedby subaeriallydepositedbrecciaaprons.
Self-sealingwas causedby early gold-bearingcalcite-manganocalcite-quartz vein mineral-
ization,nowfoundasfragments in brecciaconduitsandaprons.Laterstagesof golddeposition,
accompanied by the samegangueassemblage, occurredin gently dipping,lenticularlodes
and overlyingstockworks,manyof them hostedby hydrothermalbreccias.
During the late stagesof, and after, accumulationof pyroclasticfall and surgedeposits,
epiclasticsedimentation waswidespreadin the maar.It gaveriseto reworkedbreccias,grits,
andprominentorganic-richmudstones depositedin ephemerallakes.Hydrothermalactivity
at this time producedsubaqueous chert-pyritebedsand possiblysubaerialhot springsinter
andtravertineand,in depth,a quartz-calcite-kaolinite-smectite-(illite)
alterationassemblage.
No significantgoldis presentin anyof thesehydrothermalproducts.
Volatile-depletedmagmareachedthe surfacearoundthe perimeter of the maar during
activesedimentationand constructedtwo principalendogenousdomes,one daciticand the
other andesitic.The daciticdome, dated previouslyat 2.4 m.y., gave rise to an apron of
crumblebrecciaalongthe northwesternedge of the maar.
Accumulationof the intramaarsequencewas accompaniedand followedby widespread
subsidence and slumpingwhichresultedin steepdipsandlocallyoverturnedbeds.On com-
pletionof the filling of the maar,coherentblocksof basementphyllite slid for distances of
as much as 1 km into the maar from its steepwesternwall. Someof these slideblocksare
overlainby remnantsof the tuff ring and carry hydrothermalbrecciasand economically
importantgold mineralization,all mechanicallytransportedto their presentpositions.
At this time, an opal-pyrite-marcasitehorizonwas formingat a boiling water table, with
an overlying cristobalite-kaolinite-alunite
zone believed to be a product of acid leaching

0361-0128/84/300/638-1852.50 638
 Au DEPOSITSAND HYDROTHERMALERUPTIONBRECCIAS,PAPUANEW GUINEA 659

causedby dissolutionof boiled-offH2S in cool groundwaters.This alterationlacksgold

mineralization,exceptwhere it overprintsa gold-bearingallochthonous slideblock.
The regionaldrainagesystemeventuallybreachedthe craterrim andresultedin deposition
of channelconglomerates and fiuviolacustrine
beds,all carryingalluvialgold.Radiocarbon
datingof wood from this unit showedthat it accumulated>42,000 yearsago.
The mostrecenteventsat Wau arebelievedto be a seriesof hydrothermaleruptionswhich
constructeda compositebrecciaapronat Korangacrater on the southeastern edge of the
maar.The mostrecent eruptionis tentativelyinferredto have taken placein 1967.

Introduction lowing the location of extraordinarilyrich alluvial

MARKEDLYincreasedgold prices since 1979 have depositsat nearbyUpperEdie Creek (Fig. 3) in 1926,
focusedrenewedattentionon high-level,epithermal the district became the focus of one of the world's
preciousmetal deposits.It has recently been em- lastgreatgoldrushes(Nelson,1976). Exceptfor the
phasizedthat a significantnumberof suchdeposits war years,goldhasbeenproducedfromthe Morobe
are intimately related to a variety of little-eroded goldfieldeversince,although
sincethelate 1950s{n
volcaniclandforms(R. H. SillitoeandH. F. Bonham, reduced amounts. To date the production has
Jr., unpub.data).One of theseis a maarvolcano,the amountedto about 120 metric tons (3.8 million oz),
surfacemanifestation ofa diatreme.The writers'geo- some 15 percent of it from lode deposits(Fisher,
logic mappingand reappraisalof the Wau sectorof 1975; Lowenstein, 1982).
the Morobegoldfieldin PapuaNew Guinea(Fig. 1) Small-scaleopen pit mining at Wau is currently
have demonstratedthat epithermalgold mineraliza- conductedby New GuineaGoldfieldsLimited,since
tion accompaniedhydrothermalbrecciationas part late 1981 a subsidiaryof RenisonGoldfieldsCon-
of the late-stagedevelopmentof a typicalmaarvol- solidatedLimited of Australia, and a small amount bf
cano. alluvialgoldisrecoveredasa resultof groundsluicing
Gold was discoveredat Wau in 1922 by W. both by the companyandby privateindividuals,in-
"Sharkeye"Park, an Australianprospector,and fol- cludingtributors.Annualproductionby the company

E.14•4
• • li0n 1•
0 100 km 300
i i i

Port

FIC. 1. Location of Wau with respectto Pliocene-Pleistocene volcanicrocks(after Dow, 1977),

major faults(after Dow, 1977), and late Cenozoicplate tectonic elements(after Hamilton, 1979) of
Papua New Guinea. Volcanic outcrop A comprisesYelia and Marble Peak centers (see text).
640 SILLITOE,BAKER,AND BROOK

currentlyrunsat about0.35 metrictons(11,000 oz), heavilyonpublishedandunpublished datafor details

approximately90 percentof it fromthe UpperRidges of lodegolddepositspreviously
workedunderground
opencut(Fig.2). Althoughthisisonlysome2 percent andeitherno longeraccessible or removedby open
ofthetotalproduced eachyearasa by-product from pit mining,our work at Wau, whichincludedlogging
thePanguna porphyry copperdeposit onBougainvilleof 2,700 m of new diamonddrill core, hasresulted
Island(Fig.1), it makesa smallcontributionto Papua in redefinition
of manyofthegeologic unitsandrein-
New Guinea'spositionas the world'sninth largest terpretationof thegeologic setting,timing,andgen-
gold producer(Du Boulay,1983). Althoughnow esisof the golddeposits.Althoughpreviousworkers
overshadowed by Panguna,the Morobegoldfield clearlyrecognized the volcanicsettingof goldmin-
ranksasoneof the premiergoldfields of the western eralizationat Wau (e.g.,Fisher,1944, 1945; Rebek,
Pacificregion. 1975),andsomehavesuggested an association
with
The geologyof the Wau sector of the Morobe diatremes (e.g.,Dowet al., 1974),thecompletevol-
goldfieldis dealtwith in papersby Fisher(1944, caniccontextoforedeposition anditsanalogues else-
1945),FisherandBranch(1981),andRebek(1975), wherewerenotpreviously appreciated.
andaspects of it are alsothe subjectof severalun-
published reportsandmaps,notablythoseprepared RegionalSetting
in theearly1930sbyH. M. Kingsbury. Existing
geo- Geologicsetting
logicinformationis ably summarized in the mono-
graphby Lowenstein (1982), whichalsoprovides The southwestern partof PapuaNew Guineacom-
muchnew geochemical data. Althoughwe rely prisesa Paleozoiccrystallinebasementterrain over-

FIG.2. Obliqueaerialphotograph
looking
southovertheGoldenPeaks
(GP),GoldenRidges
(GR),
andUpperRidges(UR) sectorsof the Wau golddeposit,PapuaNew Guinea.NamieCreekandthe
millarein theforeground; theregionalfaultzoneisin theleftbackground.
Darkgrayrocks
behind
andto the left of the millarepartsof KaindiMetamorphic slideblocks.
The degraded
maarwallis
visible
between
Golden
Ridges
andUpperRidges.
Crumble
breccia
canbe seento therightof the
mill.
 Au DEPOSITS
AND HYDROTHERMAL
ERUPTIONBRECCIAS,
PAPUANEW GUINEA 641

lain by Mesozoic-Cenozoicshelf sedimentswhich, grayphyllitesbut alsoincludechlorite,chloritoidand

along its northern side, collided with a Late Creta- sericiteschists,and subsidiaryquartzitesand marbles.
ceous-Paleogeneintraoceanic island-arc system in Clastsof garnet-bearingschistwere alsoencountered
middle Tertiary times. Collision induced tectonism, in brecciasand conglomerates(see below), but the
metamorphism, ophioliteobduction,and uplift (Dow, rock was not observedin situ. The low-grademeta-
1977). morphicrocksare characterizedby foliationparallel
The Wau districtis locatedin the Owen Stanley to beddingandby numerouspodsandlensesof meta-
Ranges,approximately30 km westof the Owen Stan- morphogenic quartz.An Rb-Srisochronageof 21.0
ley faultsystem(Fig. 1), whichseparates the obducted _ 4.0 m.y. was determinedfor Kaindi samplescol-
Papuanophiolite from the Mobile Belt. The Mobile lected4 km west-northwestof Wau andis interpreted
Beltconsistsof a highlydeformedandmetamorphosed as the age of regional metamorphism(Dow et al.,
Late Cretaceousto Eocene sedimentarypile which 1974; Page, 1976), probably a late-stage facet of
is associatedlocally with basalticvolcanicrocks and middle Tertiary collisionalorogeny.
overliesoceaniccrust(Dow, 1977). Hamilton (1979) A partly subaerialvolcanoplutonicarc--the Mar-
consideredthe Mobile Belt rocksto representa m•- amuni arc (Dow, 1977)--was constructedalong the
lange terrain accretedduring northward subduction Mobile Belt in middle to late Miocene times. In the
beneaththe intraoceanic islandarcandsubsequently Papuan peninsula, the arc was probably a conse-
deformedby the continent-island arc collision.In the quenceof northwardsubductionfrom the now-buried
Wau district,the volcano-sedimentary prismis rep- Port Moresbytrench (Hamilton, 1979; Fig. 1). In the
resentedby the Kaindi Metamorphics(Figs. 3 and area of the Morobe goldfield, magmatismis repre-
4), which are dominatedby partly graphiticblue- sented by the dominantly granodioritic Morobe
batholith, which at its nearest point to Wau was ra-
diometricallydatedby the K-Ar and Rb-Sr methods
at 12 to 13 m.y. (Page, 1976). In the Wau district,
three smallbodiesof weaklyfoliatedhornblende-bio-
tite diorite (Fig. 4) are consideredas apophysesof
the Morobe batholith.
During the Plioceneto Holocene interval, major
uplift and subaerialvolcanism(Fig. 1) were wide-
spreadin PapuaNew Guinea during southwardsub-
v v v duction and included the activity, detailed below,
which gave rise to the gold mineralizationat Wau.
The OwenStanleyfault systemunderwentstrike-slip
displacementand, in commonwith the major north-
west-trendingfault mappedat Wau (Fig. 3), alsoap-
preciable normal movement.

s
Metallogenicsetting
Exceptfor the Morobe goldfield,little economic
mineralizationis known from the Papuanpeninsula
of easternPapuaNew Guinea. Minor alluvial gold
concentrations,derivedmainlyfrom middleMiocene
to Pliocenemagmaticrocks,are workedsporadically.
Severalporphyry copper prospectsof possiblePlio-
cene age, one some 75 km northwestof Wau (Fig.
0 5 10Krn 1), havebeen investigated,and a laterite nickel pros-
pect is knownin the Papuanophiolitebelt some30
km southeastof Wau at Lake Trist (Fig. 1).
Late Cenozoicvolcanotectonic
setting
Post-Pliocene
auriferous
gravels
Pliocene
I•]aria
granoo•orlte
Iow•rade
..
The Wau district is not mentioned in recent

Otibanda
Formation
Bulolo
•\ Upper
Volcanics
Edie
•ekloclas • Major faults
synthesesof Plioceneto Recent volcanismin Papua
New Guinea (e.g., Johnson,1979, 1982), probably
Dacit½-andesit½
porphyry:
mainly
endogenousd6m½s • Plioccn½ moat becauseof the absenceof either Recent activity or
any large volcanicedifice. The Wau district is part
FIG. 3. Selectedgeologicfeaturesof the Morobe goldfield, of a tectonically
complexregionsituatedbetweenthe
Papua New Guinea, modified after Dow et al. (1974). southeasternextremity of the Highlands volcanic
642 SILLITOE, BAKER,AND BROOK

v
v v
v v
V V V V V v v v v v v

VVVV tv v v v v vvvvvvvv
V
V V V V
V V V V V vv v v v
V V V
V
V V
V V Iv
v vvv v VvV
V
V V v v v v o
V V
/ v v
V V v v o
V V I v
V V V V •''•--
v v v v v
V •' 0 •
V V 0 •
}ss V V V 0• •
V V V V 0
• V V V V V V
V V V V V 0

V V V V

metres

/46 '4/'E

HOLOCENE PLIOCENE MIDDLE MIOCENE

• Landslide r-• Momb•Dior•

Endogenousdomes
• Hy•t. hermal
eruption
? PLEISTOCENE

'[-• I•edmont
fan
breccla

.•!•::•
Crumble
bmccia
Intm-maarepiclastics
& pyroclastics
Hydroth•rmoleruptionbr•.cia
Nami• Bmccio

Dac• porphyry
intru•ws
LATE CI•TACEOU5
-- PALœOGENœ

=F"•Fluviol
& fluvio-locustrine
beds
•/• Bulolo
tibonda Formation
Ignimbrit•

Ak"'"'•'•
C t/nfofil•t/Ot) • Op•/i/•
$i/,•'•
hO/'i• • f• (•]

•O. 4. Geologicmapof the Wau golddepositandvicinity,PapuaNew Guinea.

 Au DEPOSITSAND HYDROTHERMALERUPTIONBRECCIAS,PAPUANEW GUINEA 645

province, which overliesthe continentalbasement, it shouldbe stressedthat their samplingsiteswere

and the northwestern termination of the Eastern 20 km northwestof Wau. The sourceof the pyro-
Papuavolcanicprovince(Fig. 1). The EasternPapua clasticflowsthat generatedthe unit, which we pro-
volcanicprovincemaybe pairedwith southwestward poseto renamethe BuloloIgnimbrite, remainsun-
subduction of SolomonSealithosphereat the prob- proven,althougheruptionfrom,or collapseof, early
ablynow-inactive Trobriandtrench,whichdeveloped endogenous domes(seeSmithandRoobol,1982) is
in the Pliocene(Hamilton,1979; Fig. 1). The High- a distinctpossibilityandwouldfit well with the timing
lands volcanicprovince is believed by Hamilton of domeemplacementsummarizedbelow.
(1979) to be a productof southwestward subduction In the Wau district,the BuloloIgnimbriteis over-
from the eastward extension of the New Guinea lain, with little signof discordance, by anotherre-
trench alongthe north coastof PapuaNew Guinea gionallyextensiveunit, the OtibandaFormation(Fig.
(Fig. 1) which becameinactivatedby collisionwith 3), whichcomprises poorlysortedfiuviatileconglom-
the westward extension of the New Britain arc-trench erates,sandstones,and reworked tuffaceousmaterial,
system(Hamilton, 1979). Since the dormant Yelia the lastparticularlyabundantimmediatelyabovethe
andesitic volcano and the extinct Marble Peak volcanicBulolo Ignimbrite. Thin ignimbrite flows carrying
center, some95 km west-northwestof Wau (Fig. 1), fragmentsof carbonizedwoodwere alsoobservedas
are currentlyunderlainby a remnantof a southwest- interbedsin places.In the Korangacrater2 area(Fig.
dippingBenioffzone (Dent, 1976), it seemslikely 4), Kaindi Metamorphics,including garnetiferous
thatPliocenevolcanism at Wau maybe attributedto schist,dominateclastlithologiesandare accompanied
subduction at either the New Guinea or Trobriand by isolatedfragments of woodand,according to Plane
trenches. (1967), vertebrateremains.The Formationisworked
for alluvialgoldin the Korangacraterareaandelse-
Volcanic Geology
where.Plane(1967) determinedthat thispart of the
Introduction OtibandaFormationwasdepositedascoalescing fans
Our reappraisalof the Wau districtandits environs along the marginof an intermontane lake. Page and
hasresultedin recognitionfor the firsttime of a se- McDougall(1972) concludedthatthe OtibandaFor-
quenceofunweldedignimbrites,severalendogenous mation was depositednot more than 3.5 to 3.1
domes,and a maar volcano(Figs. 3 and 4). Hydro- m.y. ago.
thermalactivitywasassociated with domeemplace- Two and possiblythree varietiesof biotite-horn-
merit and maar formation. These Pliocene volcanic blendeandesiteto daciteporphyryintrusivestocks--
units and related sedimentaryrocks,of which only the Upper, Unclassified,and Lower Edie Porphy-
the maarvolcanois restrictedexclusively to the Wau ries-were distinguished in the Morobegoldfieldby
district(Fig. 4), overlie, and partly cut, the Kaindi Fisher (1945) andmost subsequent workers.All major
Metamorphicsdescribedabove. bodiesof Edie Porphyryare now recognizedas en-
dogenousdomes(Fig. 3), someof which (seebelow)
Early volcanism haveundergoneonlylimited erosion.Partsof domes
A widespreadsequenceof volcanicrocks,up to exhibit well-developedflow foliation, and a more
severalhundredmetersthick, wasmappedin parts deeply eroded one located in Upper Edie Creek
of the Morobe goldfieldand beyondby Dow et al. (Fisher'sUnclassified Porphyry)ischaracterized along
(1974) and previousworkers(Fig. 3), and was des- parts of itssteepcontactsby spectacular hydrothermal
ignatedas the BuloloAgglomerate.North of Wau, brecciaswhich carry fragmentsof both dacitepor-
the sequencedips eastwardat 40ø to 65ø. The se- phyryandKaindiMetamorphics in a rockflourmatrix.
quence in the Wau district and its environsis, how- Lithologicallysimilarhydrothermalbrecciaslocally
ever,dominated by massive,unbedded,andunwelded border restricted dikelike bodies of biotite dacite
ignimbrite, • whichcarriesuncollapsed pumicefrag- porphyry,whichintrudedthe KaindiMetamorphics
mentsand accessorylithic clasts,includingKaindi alongand immediatelywest of the major fault zone
Metamorphicsand Morobe Granodiorite, in an ash depicted in Figures 3 and 4. These truly intrusive
matrix.Interbeddedsediments werealsorecognized, daciteporphyriescommonlycarry disseminatedand
includinga distinctive in Namie veinlet pyrite, and in placesin the fault zone they
boulderconglomerate
are sericitized and silicified.
Creek,northeastof the minearea(Fig. 4). Pageand
McDougall(1972) presentedapparentK-Ar agesof We are unableto confirmthe sequenceof dacite
3.7 to 3.2 m.y. for the ignimbriteeruptionwith 3.5
m.y. selectedasa preferredmaximumage,although 2A nameintroduced
ß at least50 yearsago(N.H. Fisher,written
commun., 1983) becauseof the craterlike form of the area. Most
• Terminology usedin thisreportforvolcanic
rocksisprocess workers(Dow et al., 1974; Fisherand Branch,1981; Lowenstein,
orientedandwhereverpossiblefollowsSparksandWalker (1973) 1982) consideredit asa site of Recent volcanicactivity, whereas
and Wright et al. (1980). we favor activity of hydrothermalorigin (see below).
644 SILLITOE, BAKER, AND BROOK

porphyry emplacementproposedby Fisher (1945) The oldest recognizablerock unit directly asso-
which, on the basisof our mapping,alsoappearsto ciatedwith generationof the Wau maarisa distinctive
derive little supportfrom the detailed K-Ar dating lithic breccialong known asthe Namie Breccia.It is
studyconductedby R. W. Page(PageandMcDougall, preserved200 to 300 m west of the maar at Upper
1972; Dow et al., 1974). His biotite agesfor porphyry Ridgesand, partly aslandslidematerial, some600 m
samplesfrom the Morobe goldfieldrange from 4.2 southof the maar (Figs. 4 and 5). At both localities
to 2.4 m.y. Severalcorrespondingplagioclasesepa- it unconformablyoverliesa highly irregular surface
rates yielded older ages,apparentlya result of in- cut acrossKaindiMetamorphics.Thicknesses in excess
corporationof variousamountsof excessradiogenic of 200 m have been intersectedduring drilling at
argon (Page and McDougall, 1972). In the absence Upper Ridgesbut are suspectedto be partly a con-
of moreprecisedata,we accepthis rangeof biotite sequenceof repetitionof the succession by fiat faults
agesas broadly representativeof the time interval (Fig. 5). The Namie Brecciais composedof angular
during which dome emplacementtook place, given to rounded fragmentsof Kaindi Metamorphicsand
that the flow-bandeddaciteporphyrywhich yielded daciteporphyrysetin a gray-coloredmatrixof finely
the youngestage of 2.4 m.y. is not a late feature, as comminutedrock material(Fig. 6A). Dacite porphyry
suggested by N. H. Fisher,but is an integralpart of clastsincluderecognizablepumiceousmaterialof ei-
an endogenous dome(Fig. 4)--Fisher's (1945) Lower ther juvenile origin or derived from the Bulolo Ig-
Edie Porphyry.The age of dike intrusionalongthe nimbrite, and fiow-foliated dome material, which
majorfaultzoneisunknownbut tentativelycorrelated might suggestthe presenceof an endogenous dome
with early stagesof dome emplacement. in the Wau area prior to maar formation.Fragments
are mainly<10 cm in sizebut locallyattain>30 cm.
Maar-diatremeformation Disseminatedpyrite, at leastpartly of clasticorigin,
Maar formation at Wau was localized 600 to 1,000 is ubiquitousin the rock flour matrix.
m east of the trace of a major dip-slip fault zone, The NamieBrecciaisdominatedby a coarse,poorly
markedat surfaceby both tight gougeandopen-fault stratifiedand poorly sortedfacies(Fig. 6A) but con-
breccia.The northwest-striking faultplaneisinclined tainsinterbedsof a fine, silty,andwell-stratifiedfacies
eastwardat about 40 ø (Fig. 5) and is partly exposed (Fig. 6C). The poorly stratifiedpartscorrespondto
asfacetedspurs(Fig. 4). Fault movementis believed the explosion brecciaof Wohletz andSheridan(1983)
to have both pre- and postdatedmaar formation, and accumulatedmainly by ballistic fall of ejecta.
whichthereforeprobablytookplaceon a steep,east- The well-stratifiedpartsare thinly beddedandchar-
erly inclinedpalcoslopecomparableto that in exis- acterizedby well-developedlow-anglecrossstrati-
tence today (Fig. 3). Continuingdisplacementmay fication(Fig. 6C), aswell ashorizonscarryingabun-
be evidencedby the majorlandslidealongthe south- dantaccretionarylapilli (Fig. 6B), which exhibitthe
ern part of the fault in Figure 4. The progressive characteristic concentric internal structure. The ac-
decreasein dip of the Bulolo Ignimbrite-Otibanda cretionarylapilli unambiguously testifyto a subaerial
Formation succession northeast of the Wau area is originfor the well-stratifiedNamie Brecciaandwere
interpretedas a resultof fault-inducedtilting which probablyformedby progressiveaccretionof numer-
was largely completedprior to maar formation. ous layers of wet volcanic particles around water

A B
bend in

- - - GO•D&N
•ID GE 5
GOLDEN
.1300
130• WD10

100ø
t
7oo-I
BEDDING' •..•.".
D/,4,4,fOND
meters above seadevel
D/•/LL-/-/OLœ
•
WD
$

FIG. 5. Crosssectionof the Wau gold deposit,Papua New Guinea, along line A-B-C in Figure 4.
Legend as in Figure 4.
 rl

FIG. 6. Selectedgeologicfeaturesof the Wau golddeposit,PapuaNew Guinea.A. Varietiesof

coarse-grainedNamieBrecciafromdrill holeWD-14 (150-200 m), UpperRidges.Palefragments are
daciteporphyryandsericiteschist,anddarkfragments arephyllite,the lastdominatingin the lowest
of thethreeco•esamples. B. AccretionarylapJillin NamieBreccia,UpperRidges.C. -I•ow-angle cross
stratification
ofbasesurgeoriginin fine-grained
NamieBreccia, UpperRidges. D. Mesoscopicrecumbent
foldformedby slumpingof pyriticmudstones in the intramaarsequence. Dark-coloredbedsare rich
in syn-or diageneticpyrite. Samplefrom verticaldrill hole WD-2 (72 m), GoldenPeaks.E. Flow
foliation in the dacitic dome from the northwestern side of the maar. F. Crumble breccia from the
apronflankingthe daciticdomealongthe northwesternsideof the maar.

645
646 SILLITOE, BAKER, AND BROOK

globules.The presenceof accretionarylapilli and low- Brecciaby the dominanceof daciteporphyryclasts,

anglecrossstratification,in combinationwith features the relative paucity of Kaindi Metamorphicclasts,
such as penecontemporaneous slumpingand chan- and the much paler rock flour matrix. It is charac-
nelingof bedsandlackof sagstructures beneathlarge terizedby the sameclastsizesand degreesof clast
fragments,accordswell with an origin for the well- roundingas the Namie Brecciaand alsopossesses
stratifiedNamieBrecciaaspyroclasticsurgedeposi.ts prominentfine-grained,well-beddedhorizonswith
of the low-temperaturebase surgetype. Basesurge accretionary lapilli.Sincethesedaciteporphyry-rich
depositsare ubiquitousproductsof maar-forming brecciadeposits are absentfromthe preservedparts
phreatomagmatic or phreaticeruptions(Moore,1967; of the tuff ring andwere observedto overliea small
Fisher and Waters, 1970; Lorenz, 1973). The accre- allochthonous block of Namie Breccia within the maar
tionary lapilli may have formed in eruption columns (Fig. 5), they areattributedto laterphreatomagmatic
andthenfallenasa componentof ashor, moreprob- or phreaticexplosiveactivity, with its productsac-
ably, grew in the base surgesthemselves(Lorenz, cumulatingmainlywithin the confinesof the maar.
1974; SelfandSparks,1978). The presenceof Namie The epiclasticpart of the intramaarsequenceis
Brecciaup to at least 1.6 km from the Wau maar interbeddedwith and overliesthe pyroclasticcom-
(Fig. 4) agrees well with Wohletz and Sheridan's ponent,and rangesfrom reworkedbrecciathrough
(1979) observationthat the maximumradial distance grits and sandstones to finely laminatedmudstones,
attainedby pyroclasticsurgedepositsapproximates the lastparticularlyprominentandwidespread. The
the vent diameter. mudstonesinclude horizonsrich in plant matter as
Sincethe Namie Brecciais presentlargelybeyond well as cherty horizonswith beddedpyrite of syn-
the maar,it is interpretedasan erosionalremnantof or diageneticorigin(Fig. 6D). Travertineand sinter
a maar-encirclingtuff ring which was originally of depositedashot springapronsare alsopoorlyexposed
substantiallygreater extent. The observedremnants at two localities(Fig. 4) within the maar,althoughit
of the tuff ring were preservedby beinglet downon is not certainwhether they are part of the intramaar
the regionalfault zone, a mechanismwhich would sequenceor unconformablyoverlie it.
alsohave contributedto the steep south-westward Two of the principalendogenous domesin the Mo-
dip of the Namie Breccia(Fig. 4). Smallerbodiesof robegoldfieldabutthe Wau maar(Figs.3 and4) and
Namie Breccia at Golden Peaksand Golden Ridges were emplacedduring its development.The larger
(Figs.4 and 5) are interpretedasblocksthat slidinto one, to the northwestof the maar, is composedof
the maar at a later date (see below). We therefore unaltered biotite-hornblendedacite pori•hyry and
refinepreviousinterpretationsof the NamieBreccia exhibitswell-developedflow foliationwhich dipsin
asa volcanicbreccia(Fisher, 1944, 1945; Fisher and variousdirections,but commonlyinward, at 50 ø to
Branch, 1981) or as a diatreme breccia (Dow et al., 60 ø (Fig. 6E). The smallerdome,on the maar'ssouth-
1974), and reject Rebek's(1975) conceptof it as a ern contact,is a biotite-hornblendeandesitcporphyry
talusaccumulationpartly reworkedby water. and is characterizedby vertical flow foliationclose
The volcano-sedimentary sequence withinthe maar to its contacts,which are marked by gougedevel-
is incompletelyknownbecauseof its structuralcom- opmentin places.Two muchsmalleroutcropsof al-
plexity and the effectsof widespreadhydrothermal tered and fiow-foliatedandesitcporphyry on the
alteration.The sequenceoccupiesa surfacearea of westernsideof the maar(Fig. 4) appearto be parts
approximately1.4 X 1.4 km (Fig. 4) and definesthe of buried domeswhich were emplacedsomewhat
extent of the maar. Available evidence,both at surface earlierthanthe two principalmasses. Sillsanddikes
and from drill holes, shows that the intramaar se- of altered dacite to andesitcporphyry cutting both
quenceis delimitedby a ring fault whichdipsat 50ø the intramaar sequenceand the Namie Breccia are
to 70 ø (Figs. 4 and 5) and truncatesthe regional believed to be offshootsof the domes.The only con-
strikeanddip of the BuloloIgnimbriteandOtibanda cordantK-Ar ageson coexistingplagioclase andbiotite
Formation.The intramaarsequence dipscentripetally obtainedby Pageand McDougall(1972) in the Mo-
at anglesgreater than 30 ø, with vertical and over- robegoldfieldwere froma sampletakenwherelower
turned (Fig. 6D) bedspresentlocally. Inward dips Edie Creek is incisedinto the dacitic dome (Fig. 4).
and associated structuralcomplexitiesresultednot Their two agesof 2.4 _+0.1 m.y. effectivelydate
from regional tectonismbut from subsidenceand emplacementof the larger of the two domes,which
slumpingof unconsolidated,partly water-saturated followed the explosivestageof maar formationand
bedsduringaccumulation of materialwithin the maar. probablymuch of the intramaarsedimentation(see
The intramaar sequenceis both pyroclasticand below).
epiclasticin origin and is estimatedto exceed 200 m The southeastern side of the dacitic dome, which
in thickness.The pyroclasticpart is dominatedby a originally constituteda steep rim to the maar, is
second unit of lithic breccia of ballistic fall and base flankedby a thickandextensiveapronof daciticbrec-
surgeorigin.It is easilydistinguishedfrom the Namie cia (Fig. 4), which is believedto have accumulated
 Au DEPOSITS AND HYDROTHERMAL ERUPTION BRECCIAS, PAPUA NEW GUINEA 647

by spallingoff of the cool outer crust of the dome the remnants of one or more stream channels and a
duringits continuedgrowth--talus or crumblebrec- restrictedlake (Fig. 4). Streamchannelswere incised
cia.If suchbrecciaoriginallyflankedthe northwestern intothe intramaarsequenceandare occupiedby con-
side of the dacitic dome, it has since been removed glomerates carryinga significantproportionof Kaindi
by the deeper erosionalonglower Edie Creek, The Metamorphic(includinggarnet-bearingschist)and
breccianearthe domemarginis monomictand clast Morobe Granodioriteclasts.They are therefore in-
supported,and comprisesnonvesiculatedblocksof terpretedasproductsof an externallyderiveddrain-
daciteporphyryup to at least 2 m across(Fig. 6F). age system.In the Golden Peaks area (Fig. 4), the
Clastsizedecreases,and the amountof ashymatrix unit compriseswell-stratifiedconglomerates,sand-
increases,awayfrom the dome, in which directionit stones,andmudstones whichwere rapidly deposited
is intercalatedwith the epiclasticcomponentof the where one or more streams entered a small lake in
intramaarsequence.Pyroclasticrockserupted from the northernpart of the maar.The principalremaining
the domesflankingthe maar were not recognized, outcropof the fiuviolacustrine bedsis recumbently
and the ignimbritemappedeast of Golden Peaksis foldedandcrosscut by a northeast-trending
fault(Figs.
part of an exotic slideblock (Fig. 4; seebelow). 4 and 5). Althoughthe originof thesetwo structures
Our work at Wau hasdocumentedthe importance remainsuncertain,their positionalong the leading
of exoticslideblockswithin,andespeciallyoverlying, edgeof an exoticblockof BuloloIgnimbritesuggests
the intramaarsequence.Most of the identifiedslide that they are due to very late slidingof the block.
blocksarecomposed of Kaindiphyllite(Fig. 2), partly Prominentclastsof acid-leachedrock,includingopal,
overlainby NamieBreccia(Fig. 4). The largestblock confirmthat the fiuviolacustrinebeds postdatedal-
measures500 X 450 m at surface(Fig. 4) and was teration and mineralization. A remarkable feature of
shownin oneplaceto be 120 m thick (Fig. 5). They the fiuviolacustrine beds is the abundance of trans-
are importantasthe sourceof all gold ore minedat ported woody material, someof it partly pyritized
Golden Ridges and Golden Peaks. These Kaindi (Fig. 7B). Logsup to severalmeterslong,twigs,and
Metamorphic-Namie Breccia blocks clearly origi- leavesare all present.An earthy,but microcrystalline,
nated from Upper Ridgesand, on the basisof the dark blue mineralis sparselyscatteredthroughthe
positionof the Kaindi-Namiecontactand fault pat- unit andwasidentifiedusingX-ray diffractionanalysis
terns, may be assignedto their approximateoriginal asthe iron phosphatevivianitc.The channelfill and
positions(seebelow;Fig. 9). Sincethe maarwascon- fiuviolacustrine beds are all sluiced for their alluvial
structedon a steepmountainside, the westernwall goldcontent.Radiocarbon datingby the U. K. Atomic
of the maarcrater,stillpartlypreservedbelowUpper Energy'Authority, Harwell, of a sample of little-
Ridges(Figs. 2, 4, and 5), was probablythe most weathered wood which we collected from the flu-
unstableandthereforeprovidedthe majorityof the violacustrine bedsat GoldenPeaksyieldedan ageof
landslideblocks.However,somelandslidingalsotook •42,000 yr, the limit of the analyticprocedureem-
placefromotherpartsof the craterwalls,assupported ployed. In view of the specialcare taken becauseof
by recognitionof a block of Bulolo Ignimbrite east the suspectedantiquityof the sample,this minimum
of Golden Peaks and a block of Otibanda Formation age is preferredto that of 34,000 yr, which wasde-
north of KorangaCrater (Figs. 4 and 5). The slide termined previouslyat the AustralianNational Uni-
blocks became detached from the unstable crater versity,Canberra,for wood collectedfrom the same
wallsandsubsidedinto the maaron low-anglefaults, locality (H. A. Polach,in Fisher and Branch, 1981).
markedby substantialzonesof crushingand shearing A similar fiuviolacustrinesequenceis present in
beneaththe blocks(Fig. 7A). Upper Edie Creek (Fig. 3), where it is incisedby the
It isdifficultto equatethe intramaarsequence with present drainage.It also containsabundantwoody
the stratigraphicschemesof previous workers, al- material and scattered vivianitc, and continues to
thoughit appearsto be equivalentto Rebek's(1975) providerich alluvialgold.Wood collectedfrom this
Early Volcanics,andpart of it wasprobablyassigned unitalsoprovedtoooldto be datedby the radiocarbon
to the KorangaVolcanicsof Dow et al. (1974) and method and, like that in the fiuviolacustrine beds at
to the volcanicproductsof the Korangacrater by GoldenPeaks,hasa minimumage of •42,000 yr.
Fisher and Branch(1981). The secondand youngerfiuviatile sequence,also
Late sedimentation
worked for its alluvialgold content,occupiesan ex-
tensive area to the southeast and east of the maar
Two stagesof fiuviatilesedimentationtook place (Fig. 4), where it constitutesa broad piedmontfan.
in the Wau district after the formation of the maar It was constructed at the mountain front as a result
was essentially complete. The earlier, and topo- of renewed uplift (Fisher, 1944) on the northwest-
graphicallyhigher, is representedby a seriesof ero- trending fault zone and is incisedby the present
sional outliers up to at least 50 m thick which are drainage.
 ß

12tI I t I t I
c

FIG. 7. Additionalselectedgeologicfeaturesof the Wau gold deposit,PapuaNew Guinea.A.

Zoneof intenselow-angleshearingdevelopedin the intramaarsequence (light)beneathhydrothermal
eruptionbreccia(dark)alongtheeastern sideof theprincipalslideblock,GoldenRidges. B. Fragment
of partlypyritizedwoodin fluviolacustrine
beds,GoldenPeaks.C. Fragment of calcite-manganocalcite
vein materialin hydrothermal brecciacuttingNamieBrecciain drill holeWD-14 (102 m), Upper
Ridges. EachsmalldivisionofthescaleequalsI cm.D. Beddedhydrothermal eruptionbreccia,Golden
Ridges.
E. Crustiform
banding
in a gold-bearing
calcite-manganocalcite-quartz
vein,UpperRidges.
F. Opalinesilicahorizondeveloped
in theintramaar
sequence
exposed
in collapsed
benches
of Golden
Peaksopen cut.

648
 Au DEPOSITSAND HYDROTHERMALERUPTIONBRECCIAS,PAPUANEW GUINEA 649

Hydrothermal Brecciation, Gold Mineralization, On the basisof their lithologicand geometricchar-

and Alteration acteristics,the brecciasare assigneda hydrothermal
origin. The subaerialveneersof breccia are inter-
Hydrothermalbrecciation
preted asremnantsof compositeapronsof ejectaac-
Severalbodiesof breccia,clearlydistinctfrom the cumulatedaroundshallowventswhich were fed by
Namie Breccia,were mappedat Wau in the three the crosscutting bodies.The subaerialveneersare
centers of gold mineralizationat Upper Ridges, thereforeclassifiable ashydrothermaleruptionbrec-
GoldenRidges,and GoldenPeaks,and alsoin the cias(Lloyd,1959) or hydrothermalexplosionbreccias
Korangacraterarea(Fig.4). Thebreccias in the first (Muffler et al., 1971); the former nameis employed
three areaswere not distinguished
from Namie Brec- here in conformitywith currentusageby mostwork-
cia by previousworkers,and thoseat Korangacrater ers in New Zealand (e.g., Hedenquist, 1983). The
were either not recognizedat all or consideredas examplesdescribedby Muffler et al. (1971) in the
volcanic breccia. Yellowstonegeothermalfield, Wyoming, and com-
Althoughthe originalextentsof brecciasat Upper parableoccurrences in the geothermalsystems
of the
Ridges,GoldenRidges,andGoldenPeakshavebeen Taupo volcaniczone of North Island, New Zealand
severelyreducedby openpit mining,their overall (Lloyd, 1959; Cross,1963; Nairn and Wiradiradja,
geometriesmay still be deducedfrom existingrem- 1980; Hedenquist,1983), were generatedduringthe
nants(Fig. 4), old records,andthe recentdiamond last 15,000 yr, and the formationof somehasbeen
drilling. Overall geometricrelationspermit subdi- observed.They were generatedby shallowhydro-
visioninto massivecrosscuttingbreccia bodiesand thermal eruptionswhich causedapronsof ejecta to
weakly stratifiedaccumulations of subaerialbreccia accumulatearoundwell-definedhydrothermalerup-
(Fig. 4). Bothtypesarepresentat GoldenRidgesand tion craters; these craters were from several tens of
GoldenPeaks,but only crosscutting brecciashave metersto •1 km in diameterandup to 100 m deep.
beenrecognizedat UpperRidgesandonly subaerial In the GoldenRidgesand GoldenPeaksareas,the
brecciasat Koranga crater. hydrothermalbrecciasare restrictedto NamieBreccia
Bothbrecciatypesat UpperRidges,GoldenRidges, andKaindiMetamorphicsin the landslideblocksand,
and Golden Peaks are similar and characterizedby moreover,do not containany clastsof the subjacent
their poorlylithifiednature.They comprisea chaotic pyroclasticandepiclasticmaterialfromthe intramaar
assemblage of angularto roundedfragmentsof Namie sequence.The brecciasin thesetwo areasare there-
BrecciaandKaindiMetamorphicssetin a muddyand fore believed to have been generatedin spatialcon-
often plasticmatrix.Fragmentsof gold-bearingcal- tinuity with the Upper Ridgesbrecciasand to have
cite, manganocalcite (largelyoxidizedto manganese subsidedaspartsof the landslideblocks.The precise
oxides),andquartzcompriseup to 5 volumepercent age of brecciationis unknownbut clearly postdated
of both brecciatypes(Fig. 7C). The Korangacrater initial maar formation and predated the late land-
brecciasare lithologicallydistinctand carry promi- slidingevent.
nent fragmentsof intramaarpyroclastics,Morobe Developmentof a fumarolicvent at Korangacrater
diorite, silicified and pyritized material (including in May 1967 followed a major landslidewhich oc-
wood),andballed-upplasticmud.All brecciascontain curred during the night. Althoughactivity is widely
fragmentsup to 1 m or soin size,but lackanyessential believed to have been restrictedto landsliding(Pi-
(juvenile) material. gram et al., 1977), we tentatively suggestthat the
The crosscutting brecciasoccurasa ramifyingsys- maineventcouldhavebeena hydrothermaleruption,
tem of irregular podsand veinswhich, especiallyat with the landslidemerely being a consequenceof
Upper Ridges,preferentiallyfollow a seriesof gently oversteepening of the valleysideduringaccumulation
dippingplanes,someof whichare clearlyfaults(Figs. of water-saturatedbreccia. This reinterpretationof
4 and 5). These bodiestransectmainly the Namie the eventis basedentirelyon the presenceof poorly
Brecciabut are locally hostedby Kaindi Metamor- stratifiedmuddybrecciasin the vicinityof the vent,
phics. At Upper Ridges,breccia bodies extend to since there were no reliable witnessesand hydro-
depthsof • 100 m, andreinterpretationof old records thermal eruptionis not accompaniedby loud noise.
suggests that at greaterdepthsthey steepenand are Examinationof the areaseveraldaysafterthe inferred
transitionalto veins (Fig. 5). The subaerialbreccias eruptionrevealedthe presenceof SO2-richfumaroles
unconformablyoverlie Namie Breccia or are sepa- at temperatures of 600ø to 700øC (Pigramet al.,
rated by low-angle zones of shearingfrom either 1977), the site of which is now marked by encrus-
Kaindi Metamorphicsor the intramaarsequence.At tationsof nativesulfur.We suggest thatthesegaseous
least at Golden Ridges,the subaerialbrecciasattain emissions could have been an aftermath of the inferred
a thicknessof 60 m andare madeup of thick, nearly hydrothermaleruption,althoughPigramet al. (1977)
fiat beds defined on the basis of different clast lith- invoked spontaneouscombustion of carbonaceous
ologies(Fig. 7D). matter and pyrite asan explanation.
650 SILLITOE, BAKER, AND BROOK

The Korangacraterarea,whichisbelievedto have Edie Creek (Fig. 3). There the veins,includingmany
beenthe siteof hydrothermal
eruptionspriorto 1967, lesspersistentveinsandveinlets,mainlytransectthe
ispresentlydelimitedonthreesidesby steppedwalls, Kaindi Metamorphics(Fisher, 1939; Rebek, 1975)
producedby recent subsidence, and is breachedon in anareaadjacentto an endogenous dome,although
the fourth as a result of the 1967 landslides. It contains locallythey cut the domeitself.The veinsconstitute
a smallcold water pond and has now cooledcom- a northwest-trending belt parallelto the regionalfault
pletelyfollowingthe 1967 activity.One of the earlier zone at Wau (Fig. 3). The Upper Edie Creek veins
eruptionbreccias,in the southernpart of the Koranga carryquartz,calcite,and manganocalcite, asat Wau,
area, carriesabundantfragmentsof Morobe diorite, but containa greater abundanceof silver sulfosalts
a rockwhichdoesnot cropout aroundthe craterand (Lowenstein,1982).
is only known from a smallinlier 400 m southwest The intrusivedacite porphyries,and their imme-
and 40 m lower in elevation. It is concluded that the diatemetamorphic wallrocks,alongtheregionalfault
dioritefragmentswere carriedupwardfrombeneath zone west of Upper Ridges,are locally pervasively
the crater area by hydrothermaleruptions. sericitizedand,at the Ribroastermine(Fig. 4), have
Gold mineralization
beenworkedfor gold.There,upto 10 volumepercent
sulfidesaccompanyintensesericiticalterationlocal-
The principal hydrothermalgold mineralization ized by a crossstructure.The sulfideassemblage in-
knownin the Wau districtis closelyassociated with cludesimportantpyrite,pyrrhotite,andarsenopyrite,
the hydrothermalbrecciasat Upper Ridges,Golden and the gold is 750 fine (Lowenstein,1982); both
Ridges,and GoldenPeaks(Fig. 4). The high-grade featuresarein markedcontrastto the Wau andUpper
mineralization (10-30 g/metric ton) is present Edie Creek lodes.
mainly in gently dipping, grosslylenticularlodes, Goldmineralization at bothWau andUpperEdie
which are up to 10 m thick and 300 m longbut are Creek is believed to have been generatedduring
generallyratherdiscontinuous. Theseappearto have domeemplacement and maarformation(Fig. 8). If
beenemplacedwithin or at the baseof hydrothermal Page and McDougall's(1972) biotite agesare ac-
breccias,apparentlyof both subsurface andsubaerial cepted(seeabove),thenmineralization probablyfalls
origin.The majorveinsappearto be overlainby one within the 4- to 2.4-m.y. interval. At Wau, however,
or morelessextensiveveinsandare partlyparalleled gold introductionprecededslidingof Kaindi Meta-
by preexisting,sill-likebodiesof daciteporphyry. morphic-Namie Breccia blocks, as shown by the
In an unoxidizedstate, the veins carry calcite, abrupt truncation of gold mineralizationat their
manganocalcite, andquartz,alongwith generallymi- sheared bases.
nor pyrite, rhodochrosite,sphalerite,galena,chal- As noted above,alluvial gold in the Wau district
copyrite,tetrahedrite,andsilversulfosalts(seeRebek, is presentlocally as economicconcentrations in all
1975; Lowenstein,1982). The carbonatescommonly the clasticsedimentary unitsdistinguished
duringthe
exhibit crustiformbanding (Fig. 7E) and cockade mapping,namelythe OtibandaFormationdeposited
structure,andthe quartzis white or transparentand priorto maarformation,the fiuviolacustrinebedsand
normallyvuggy.Hypogenegoldin theveinsispresent channelconglomerates, and the piedmontfan accu-
in the native state and is probably500 to 600 fine; mulatedafter maarformation.The sourcesof gold in
theoverallAg/Auratioof theveinsis3:1 (Lowenstein, thesealluvial depositshave yet to be preciselyde-
1982). Much of the ore is oxidizedand is believed termined,but in the postmaarunits it is probably
by Lowenstein(1982) to haveundergonesupergene largely derivedfrom hydrothermalgold associated
goldenrichmentasan accompaniment to the trans- with domeemplacementandmaarformation,assup-
formationofmanganocalcite andrhodochrosite to wad portedby the relativelylow fineness (558-611) doc-
and crystallinemanganese oxides. umentedin the Korangacraterarea(Fisher,1975).
Goldmineralizationof lower gradebut commonly Gold introducedduringMorobeGranodioriteintru-
minable in bulk, with gradesof up to 5 g/metric sion(Fisher,1945) in the mid-Miocene,or evendur-
ton of gold, occupiessubstantial volumesof hydro- ing metamorphogenic quartzsegregation in the Oli-
thermallybrecciatedrock.The goldis present,with gocene-earlyMiocene, is probablythe main source
the samegangueassemblage asin the mainlodes,in of Otibandagold, which is reportedto average876
either short multidirectionalveinletsup to several fineelsewherein the Morobegoldfield(Lowenstein,
centimeters wide above the main lodes, or in vein 1982).
fragmentswithin the breccias. The majorproportionof the goldproducedfrom
At Upper Ridges, the gently dipping, breccia- the Morobe goldfieldwas dredgedfrom the Bulolo
hostedlodesaretransitionaldownwardto steep,albeit alluvials(Fig. 3) and wasderivedby erosionof the
mineralogicallysimilar, vein structures.These are Wau and Upper Edie Creek lode gold deposits
similarto the mainveinsexploitedpreviouslyat Upper (Fisher, 1945, 1975).
 AuDEPOSITS
ANDHYDROTHERMAL
ERUPTION
BRECCIAS,
PAPUA
NEWGUINEA 651

I
. - -- Dome emploccment

...... Buloio Ignimbrite

- Otibanda Formotion

--- ß Regionc•l
tilting

I I ? - Gold mincroiizotion

-- •./?
MAAR-FOI•MING
EVENTS
Hydrothcrmol
br½cciotion
I
c•tKorcmgoß

4 J 2 ! O

- • - Declteporphyryintrusion,
Ribroastergoldrn•neralizet•on
- Dletrerne/rneer/tuff-ring
forrnetmn
Hear sedirnentotlon
MAAR- FORMING - - Hydrotherrnel
brecc•et•on
+goldrn•nerohzotion
-•-Deep hydrothermel
olterobonm moor
[VœNTS
.•.. • Dome
ernp•scernent b.
- - Descent of shde blocks
- Generobon of oc•d-leachedzone
Fluvmble sed•rnentnbon
with(111uv•el
•old

FIG.8. (a)Generalized
timesequence
of volcanic
andhydrothermaleventsat Wauandvicinity.
K-AragecontrolfromPageandMcDougall (1972),(b) Amplification
of eventsassociated
withthe
formation of the Wau maar.

Hydrothermalalteration horizon is characterizedby cristobalite,kaolinite,

alunite, and much smalleramountsof iron sulfides,
In contrastto the gold mineralizationat the Ri- all confirmedby X-ray diffractionanalysisof repre-
broaster mine, which is associatedwith intense ser- sentative samples.
Cristobalite-rich rockisporousand
iciticalteration,the maingoldlodesat Wau lackany low in density.
noticeableassociatedalteration effects,althoughmi- In the centralpart of the alteredarea,restricted
nor kaoliniteandsericitewere recordedlocally.Nev- to the intramaarsequence aroundGoldenPeaksand
ertheless,two broadvarietiesof hydrothermalalter- GoldenRidges,a secondvarietyof feldspar-destruc-
ationessentially devoidof significant goldvaluesare tive alterationis foundpatchilyto depthsexceeding
recognized in associationwith the maar. 300 m. Samplesexaminedby X-ray diffractionare
A zoneof feldspar-destructive alteration,some1.5 composed of quartz,calcite,kaolinite,and smectite,
by up to 0.9 km at the surface(Figs.4 and7F), oc- and obviouslylack alunite. A pale green micaceous
cupiesinternalpartsof the maarandwasdeveloped mineralthat is presentlocallywasdeterminedto be
attheexpense oftheintramaar sequence and,locally, a mixedlayerillite. The alterationis associated with
of the main andesiticdome. It occupiesonly high veinlets of chalcedonyor calcite carrying pyrite.
groundandhasclearlybeenlargelyremovedby ero- Field evidencesupportshydrothermalalteration
sionandopenpitminingfromthelowerGoldenPeaks at severaltimesduringmaarformation(Fig. 8). The
and GoldenRidgesareas.The alterationzonedoes deep feldspar-destroying alterationappearsto be
not appearto exceed70 m in thickness andis char- partlyearly,becausedrillinghasshownthatit does
acterizedat its baseby a prominentflatly dipping not extendupwardinto either the daciticcrumble
zoneof silicificationcomposed of opal,givinga dis- breccia or the Kaindi Metamorphic-Namie Breccia
orderedcristobalite (opalCT) X-raypattern.The opal slide blocks. In contrast, the near-surfaceopalized
is locallyinvertedto chalcedony. is ac- horizon and overlyingcristobalite-kaolinite-alunite
Silicification
companied by upto 20 volumepercentofpartlycol- alteration resulted from a late-stageevent which
loform, fine-grainedpyrite and marcasitethat are clearlypostdated descent of gold-bearingslideblocks
largelyoxidizedto supergene jarositeandhematite. but predated deposition of the fiuviolacustrine
beds.
Feldspar-destructive alterationabovethe opalized The cristobalite-bearing rockshave been shownto
652 SILLITOE, BAKER, AND BROOK

contain <0.005 ppm Au, except where they were magmatic(carryingjuvenile material) origin for the
developedat the expenseof auriferoushydrothermal Wau maar is difficult to make because it is unclear
breeeia. whetherdaciteporphyryclastsare juvenile (cognate
lithics) or not (accessorylithics). In common with
Volcanic and Hydrothermal Development manydiatremes(e.g., Selfet al., 1980), the Wau dia-
In the Morobe goldfield,hydrothermalgold de- tremeisbelievedto havebeenlocalizedby the nearby
positionaccompaniedemplaeementof endogenous regionalfault zone,whichdipseastwardat about40 ø
domes,the earlier of which may have given rise to beneaththe maarandshouldthereforebe presentat
pyroelasticpumice-and ash-flowspreservedas the a depthof about1,000 m beneathit (Fig. 8). A per-
Bulolo Ignimbrite. On the basisof Page and Me- meablefault zonewouldprovideideal conditionsfor
Dougall's(1972) radiometricages,the overall max- descendingground waters to encounter ascending
imumtime spanfor gold mineralizationwas4 to 2.4 bodiesof magma.Early batchesof thismagma,which
m.y. (Fig. 8). At Wau itself, the initial generationof rose alongthe fault zone, are believed to be repre-
the maar and underlying diatreme preceded gold sentedby the dikelikebodiesofdaciteporphyrywest
mineralization,which is thoughtto have taken place of Upper Ridges(Fig. 9). Followingthe main period
prior to emplaeementof the youngestdated daeitie of hydrovolcanicactivity, residualmagmadepleted
dome 2.4 m.y. ago. Filling of the maar, including in volatilesby explosiveeruptionwasemplacedmore
eraplacement of landslideblocks,wascompletedbe- thanonceasviscousdomes,with intrusionapparently
fore depositionof the fluviolaeustrinebedsat an un- partlycontrolledby the diatremewalls.The tuff ring
known time >42,000 yr ago. andpartsof the intramaarsequence were constructed
Followingacceptedmodelsof maar-diatremefor- of ballistic fall and base surge deposits,the latter
mation (e.g., Lorenz, 1973), the Wau systemis be- resultingfrom laterallydirectedblastsof turbulent,
lieved to have been generatedby a seriesof hydro- expandedgas-soliddispersions (Moore, 1967; Sher-
volcanic(phreatieor phreatomagmatie; Sheridanand idan andWohletz, 1981). The tuff ring probablyex-
Wohletz, 1981) explosions triggeredby interaction tendedsome2 km from the perimeterof the maar,
magmaandcoolmeteoricwa- but becauseof its poor preservationpotential, it is
of hot andesitie-daeitie
ter. The choicebetween a phreatie and a phreato- now largely destroyed.

RIBROASTER-'FYI• HYDROTHERMALLY
BRECCIATED
GOLD CONDUITSWITH
DEGRADED
TUFF-RING
AUTOCHTHONOUS RECONSTRUCTED
GOLDDEPOSIT POSITIONOF
/ PRINCIPAL
SLIDE-BLOCI•

....
/ ;'/;
/ • ENDOGENOUSALLOCHTHONOUS
DOME
FLUVIO•STRINE
•OLI)DEPOSIT
HYOROTHERMAL
OVERTUItNED
BEOS
BUU:ILO
I•NiMBRITE
ERUPTION SLIDE
BLOCK
PRINCIPAL
SLIOE
BLOCK
OPALINE
SILICA

DIATREHE'.

FIG. 9. Partly reconstructedschematicsection of the Wau maar-diatreme system to illustrate

aspectsof the genetic model discussedin the text. Most of the legend as in Figure 4.
 Au DEPOSITSAND HYDROTHERMAL ERUPTION BRECCIAS,PAPUANEW GUINEA 653

During or immediately following initial explosive and dominated by meteoric water, in common with
activity, the ring fault would have formed and sub- those inferred to have been responsiblefor similar
sidenceof the enclosedrocks,includingpartsof the epithermalpreciousmetal mineralizationelsewhere
tuff ring, probably took place (e.g., Hearn, 1968; (e.g., BergerandEimon, 1982), althoughsomework-
Lorenz, 1973). Material which subsidedat this time ers suggestthat a crucial metal-bearing magmatic
may have been disaggregated by subsequentexplo- componentmayhavebeen added(e.g.,White, 1981).
sions,althoughthe NamieBrecciablock--intersected At Wau, any magmaticallyderived fluid must have
by drilling at a depth of 300 m within the maar-- beenreleasedduringfinalcrystallizationof andesitic-
survived. Once subsidencehad commenced, the cra- daciticmagmabodiesbeneaththe diatreme(Fig. 9).
ter wouldhaveundergoneprogressive
expansion
by Gold mineralizationon the upthrownsideof the re-
slumpingand sliding from its unstablewalls. gionalfault zone, at the Ribroastermine, appearsto
Hydrothermalbrecciationand gold introduction havebeen emplacedat greaterdepthby fluidschan-
were essentiallycoevalandtook placewithin the tuff neledup the fault. The fluidsare thoughtto possess
ring andunderlyingKaindiMetamorphicsof the Up- a overallparentagesimilarto thoseresponsiblefor
per Ridgesarea.In thiswedge-shaped body of rock, golddeposition at UpperRidges,but theyweremore
between the regional fault zone and the maar ring salineandin the boilingcondition(M. Jones,written
fault,gentlydippingextensional faultswere generated commun., 1983).
asthispoorlyrestrainedwedgeof rock tendedto slip Since most of the major veins have now either
downwardtoward the maar (Fig. 9). Furthermore, been removed by mining or are inaccessible,it is
the regional fault zone and associatedsubsidiary difficultto ascertainthe precisenature of their struc-
structureswere ideally locatedto act as a plumbing tural controls.However,all veins,includingearlyones
systemfor the ascentof hydrothermal fluids from fragmentedby hydrothermalbrecciation,were evi-
depth beneath the maar. dently producedby repeatedopeningand filling of
In conformity with the mechanismacceptedfor laterallyrestricted,gentlydippingfractures.As pro-
the genesisof comparablebrecciaselsewhere(Muffler posedabove,fracturesare thoughtto havebeengen-
et al., 1971; Henley and Thornley, 1979; Nairn and eratedby extensionof the rock wedge between the
Wiradiradja,1980; Hedenquist,1983), hydrothermal regionalfault zone and the ring fault owingto incip-
brecciationat Upper Ridgesis believedto have been ient downwardmovementtowardthe maar (Fig. 9).
a consequence of fluid overpressuresgeneratedbe- Explosivevolcanic activity must have continued
neathrelativelyimpermeablecaprocks.Permeability locally during early stagesof sedimentationwithin
wasreducedasa result of the self-sealingof fissures the maar, becauseof the alternation of pyroclastic
by earlydepositionof gold-bearingcalcite,mangano- and epiclasticunits,althoughhigher horizonsof the
calcite,and quartz, which are now presentas ubiq- intramaarsequencetend to be dominatedby epiclastic
uitousfragmentsin the hydrothermalbreccias.Hy- material. Maar sedimentsare mainly epiclastic,de-
draulicfracturingis thoughtlikely to havebeen trig- rived locally from within the confinesof the crater,
gered by rapid reductions of confining pressure and accumulatedpartly in shallowephemerallakes
inducedby faulting and slidingof material into the around which vegetation grew profusely. On the
maar. Decompressionin one or more fissureswould northwesternsideof the maarthey intertonguewith
havecausedwaterto flashto steam,disruptthe fissure crumble breccia accumulated on the flank of the dac-
walls and confiningrocks, and carry the resultingitic dome during its active growth. Hydrothermal
fragmentalmaterial to the surface.Continuedviolent
fluidswhich debouchedat this time gaverise to sub-
dischargeof fluidsand ejecta causedentrainmentof aqueouschert-pyrite beds and possiblyalsoto sub-
additionalmaterial from the walls of the ramifying
aerial travertine and sinter.At least part of the deep
systemof conduitsand their progressivewidening quartz-calcite-kaolinite-smectite-(illite) alteration
during the courseof the days, months,or years of probablyalsodatesfromthisperiod.The lack of gold
intermittenteruptiveactivity.In contrastto the early
in all thesehydrothermalproductssuggests their for-
hydrovolcanicstage,the loci of eruptionwere very mation from fluids different from those responsible
muchshallower(•100 m), morerestricted,andlack- for mineralizationat Upper Ridgesand Ribroaster.
ing in any direct magmaticinvolvement. The intramaar succession was deformed by con-
As suggestedfor hydrothermaleruption vents in tinued subsidenceand slumping,mainlyafter explo-
general(Henley andThornley, 1979), thoseat Wau sive activity had ceased, as documented for maars
appearto haveactedaseffectivechannelways for the elsewhere(Lorenz, 1979). Detachmentof rigid slide
focusedascentof hydrothermalfluidsduringandim- blocks from the unstable walls of the maar crater took
mediatelyfollowingbrecciation.Althoughno stable placeafterfillingof the maarwasessentially complete
isotopestudiesand only preliminaryfluid inclusion and may have causeddeformationof partsof the in-
studiesof samplesfrom Wau have been carried out tramaar sequence.Most of the recognized blocks
to date, the fluids are believed to have been dilute originatedon the oversteepenedwesternwall, from
654 SILLITOE, BAKER,AND BROOK

whichthe largesttraveledfor 600 m (Figs.4, 5, and fields Consolidated Limited. We thank R. A. Shak-
9). Movementon the regionalfault zonemayhave esby,ExplorationManager,andJ.P. McKibben,As-
triggeredthe sliding.Hydrothermalbrecciasandgold sistantExplorationManager(SouthWestPacific),for
mineralizationat Golden Ridgesand Golden Peaks permissionto presentthis summaryof the geologic
underwentmechanical transportfor distances
of up aspectsof the work. Useful discussions
were held at
to i km to their presentpositionsasintegralpartsof varioustimes with J.P. McKibben, R. M. D. Meares,
slide blocks.The gently dipping veins at Golden D. Pascoe,and J. V. Wright. Reviewsof the manu-
Ridgesand Golden Peaksbecame detached at this scriptwere generouslyprovidedby N. H. Fisher, P.
time fromtheir steeperfeederveins,whichare only Lowenstein,and R. W. Page.
present in the area of autochthonousmineralization
at Upper Ridges. August 16, 1983; January 16, 1984
In viewof itsplanar,gentlydippingform,the opal-
ized horizonis interpretedto have been a boiling
watertable, with the suprajacentcristobalite-kaolin- REFERENCES

ite-aluniteassemblage representing acidleachingby Berger, B. R., and Eimon, P. I., 1982, Comparative models of
dissolutionof boiled-offH2S in cool groundwaters epithermalsilver-golddeposits:Soc.Mining EngineersAIME
(cf.Buchanan, 1981).The zoneof acidleachingwould Preprint 82-13, 14 p.
havebeenrepresented at surfaceby steamingground, Buchanan,L. J., 1981, Precious metal depositsassociatedwith
volcanic environments in the Southwest: Arizona Geol. Soc.
andcristobalite-richrockis interpretedasa product
Digest, v. 14, p. 237-262.
of steamleaching.Shallow-levelalterationwithinthe Cross, D., 1963, Soilsandgeologyof somehydrothermaleruptions
maarappears,at leastin part,to havepostdatedarrival in the Waiotapu district: New Zealand Jour. Geology Geo-
of the majormineralizedslideblocks(Fig. 8). physics,v. 6, p. 70-87.
After alteration had ceased, the maar crater was Dent, V. F., 1976, The seismicitypatternnear Mount Yelia, Papua
New Guinea:PapuaNew GuineaGeol. SurveyRept. 76/22, 6
breachedbefore42,000 yr agoby regionaldrainage, p.
and fluviolacustrinebedswere depositedwhere an Dow, D. B., 1977, A geologicalsynthesisof PapuaNew Guinea:
externallyderivedstreamexitedontofiatsin the mid- Australia Bur. Mineral ResourcesGeology GeophysicsBull.
dle of the maar. The lithologic and possiblegeo- 201, 41 p.
chronologic betweenalluvialgold-bearing Dow, D. B., Smit, J. A. J., and Page, R. W., 1974, Wau, Papua
similarities
New Guinea: 1:250,000 geol. ser.: Australia Bur. Mineral Re-
fiuviolacustrinebeds within the maar and at Upper sourcesExpl. Notes. SB/55-14, 33 p.
EdieCreeksuggest thatthe externaldrainagereferred Du Boulay, L., 1983, Gold 1983: London, ConsolidatedGold
to abovewasan ancestralEdie Creek. Vivianitepre- Fields PLC, 55 p.
cipitationandpyritizationof woodin fiuviolacustrine Fisher, N.H., 1939, Ore geologyof the Day Dawn mine, New
bedsat both localitiesis attributedentirely to dia- -- Guinea: ECON. GEOL., v. 34, p. 173-189.
1944, Outlineof the geologyof the Morobegoldfield:Royal
genetic processes. Soc. QueenslandProc., v. 55, p. 51-58.
Epithermalgold depositionat Wau was clearly a -- 1945, The finenessof gold, with specialreference to the
shallowevent,but it is not possibleto determinethe Morobe goldfield,New Guinea. Pt. I: ECON.GEOL.,v. 40, p.
449-495.
amount of cover lost to erosion since mineralization
1975, Alluvial gold of the Morobe goldfield, in Knight,
took place. However, the preservationof subaerial C. L., ed., Economic geology of Australia and Papua New
hydrothermaleruptionbrecciaat GoldenRidgesand Guinea.I. Metals:Australasian Inst. Mining Metallurgy Mon.
GoldenPeaksstronglysuggests erosionof only a few 5, p. 1045-1047.
tensof metersof brecciaat theselocalities.Although Fisher, N.H., and Branch, C. D., 1981, Late Cainozoic volcanic
depositsof the Morobe goldfield: Papua New Guinea Geol.
erosionrates are high under the tropical climatic re- Survey Mem. 10, p. 249-255.
gime prevailingat Wau, denudationhas been mini- Fisher, R. V., and Waters, A. C., 1970, Basesurgebed formsin
mizedby bothdownfaulting
onthe regionalfaultzone maar volcanoes:Am. Jour. Sci., v. 268, p. 157-180.
and subsidence of the maar. Hamilton, W., 1979, Tectonicsof the Indonesianregion: U.S.
Geol. Survey Prof. Paper 1078, 345 p.
If our interpretationof the Korangacrater area is Hearn, B.C., Jr., 1968, Diatremes with kimberlitic affinitiesin
correct, hydrothermal brecciation was renewed at north-central Montana: Science,v. 159, p. 622-625.
Wau, thistime onthe southeastern edgeof the maar, Hedenquist, J. W., 1983, Waiotapu, New Zealand: The geo-
followingaccumulation of the piedmontfanandper- chemicalevolutionandmineralizationof an activehydrothermal
hapsevenasrecentlyas1967. The fragmentsof mas- system:Unpub.Ph.D. thesis,Univ. Auckland,242 p.
Henley, R. W., andThornIcy, P., 1979, Somegeothermalaspects
sivechalcedonic silicain the Korangabrecciaapron of polymetallicmassivesulfideformation:ECON.GEOL.,v. 74,
supportthe occurrenceof self-sealing in depth but p. 1600-1612.
contrastwith the dominanceof carbonatefragments Johnson,R. W., 1979, Geotectonicsand volcanismin PNG: A
in the older breccias. review of the late Cenozoic:BMR Jour.AustralianGeology
Geophysics,v. 4, p. 181-207.
Acknowledgments -- 1982, PapuaNew Guinea, in Thorpe, R. S., ed., Andesites:
Orogenic andesitesand related rocks: Chichester, Wiley, p.
Our work at Wau and elsewhere in the Morobe 225-244.
goldfieldwasundertakenon behalf of RenisonGold- Lloyd, E. F., 1959, The hot springsand hydrothermaleruptions
 Au DEPOSITSAND HYDROTHERMALERUPTIONBRECCIAS,PAPUANEW GUINEA 655

of Waiotapu:New ZealandJour.GeologyGeophysics,
v. 2, p. New Guinea, in 1967: BMR Jour. AustralianGeologyGeo-
141-176. physics,v. 2, p. 59-62.
Lorenz, V., 1973, On the formation of maars:Bull. Volcanol., v. Plane, M.D., 1967, Stratigraphyand vertebrate fauna of the
37, p. 183-204. Otibanda Formation, New Guinea: Australia Bur. Mineral Re-
-- 1974, Vesiculated tuffs and associated features: Sedimen- sourcesGeologyGeophysicsBull. 86, 64 p.
tology, v. 21, p. 273-291. Rebek, R. J., 1975, Edie Creek and Wau gold lodes, in Knight,
-- 1979, Phreatomagmaticorigin of the olivine melilite dia- C. L., ed., Economic geology of Australia and Papua New
tremesof theSwabian
Alb,Germany,in Boyd,F. R., andMeyer, Guinea. I. Metals: AustralasianInst. Mining Metallurgy Mon.
H. O. A., eds., Kimberlites, diatremes, and diamonds:Their 5, p. 867-872.
geology,petrology,and geochemistry:Washington,Am. Geo- Self, S., and Sparks,R. S. J., 1978, Characteristicsof widespread
phys. Union, p. 354-363. pyroclasticdepositsformedby the interactionof silicicmagma
Lowenstein,P. L., 1982, Economicgeologyof the Morobe gold- and water: Bull. Volcanol., v. 41, p. 196-212.
field, PapuaNew Guinea:PapuaNew GuineaGeol. Survey Self, S., Kienle, J., and Huot, J.P., 1980, Ukinrek maars,Alaska,
Mem. 9, 244 p. II. Depositsand formationof the 1977 craters:Jour.Volcanol.
Moore, J. G., 1967, Basesurgein recentvolcaniceruptions:Bull. Geothermal Research,v. 7, p. 39-65.
Volcanol.,v. 30, p. 337-363. Sheridan,M. F., and Wohletz, K. H., 1981, Hydrovolcanicex-
Muffler,L. J.P., White, D. E., andTruesdell,J. E., 1971, Hy- plosions:The systematics of water-pyroclastequilibration:Sci-
drothermal explosion craters in Yellowstone National Park.. ence, v. 212, p. 1387-1389.
Geol. Soc.America Bull., v. 82, p. 723-740. Smith,A. L., andRoobol,M. J., 1982, Andesiticpyroclasticflows,
Nairn, I. A., and Wiradiradja,S., 1980, Late Quaternaryhydro4 in Thorpe, R. S., ed., Andesites:Orogenicandesitesandrelated
thermalexplosionbrecciasat Kaweraugeothermalfield, New rocks:Chichester, Wiley, p. 415-433.
Zealand:Bull. Volcanol.,v. 43, p. 1-13. Sparks,R. S. J., and Walker, G. P. L., 1973, The groundsurge
Nelson,.H., 1976, Black, white and gold: Goldminingin Papua deposit:A third type of pyroclasticrock: Nature Phys.Sci., v.
New Guinea 1878-1930: Canberra, Australian Natl. Univ. 241, p. 62-64.
Press, 298 p. White, D. E., 1981, Activegeothermalsystemsandhydrothermal
Page,R. W., 1976, Geochronology of igneousand metamorphic ore deposits:ECON.GEOL.75th ANNIV. VOL., p. 392-423.
rocks in the New Guinea Highlands:Australia Bur. Mineral Wohletz,K. H., andSheridan,M. F., 1979, A modelof pyroelastic
ResourcesGeologyGeophysicsBull. 162, 117 p. surge:Geol. Soc.America Spec.Paper 180, p. 177-194.
Page,R. W., and McDougall,I., 1972, Agesof mineralizationof 1983, HydrovolcanicexplosionsII. Evolution of basaltic
goldand porphyrycopperdepositsin the New GuineaHigh- tuff rings and tuff cones:Am. Jour. Sci., v. 283, p. 385-413.
lands:ECON.GEOL.,v. 67, p. 1034-1048. Wright, J. V., Smith, A. L., and Self, S., 1980, A working ter-
Pigram, C. J., Johnson,R. W., and Taylor, G. A.M., 1977, In- minologyof pyroclasticdeposits:Jour. Volcanol. Geothermal
vestigationof hot gasemissionsfromKorangavolcano,Papua Research,v. 8, p. 315-336.