Diatreme Breccias at the Kelian Gold Mine

©2008 Society of Economic Geologists, Inc. Economic Geology,  v. 103, pp. 689–716

Diatreme Breccias at the Kelian Gold Mine, Kalimantan, Indonesia: Precursors to Epithermal Gold Mineralization ANDREW  G. S. DAVIES,* DAVID R. COOKE,† J. BRUCE GEMMELL, AND KIRSTIE A. SIMPSON CODES ARC Centre of Excellence in Ore Deposits, University of Tasmania, Private Bag 79, Hobart, Tasmania 7001, Australia

 Abstract Early Miocene volcanism associated with a maar-diatreme breccia complex preceded main-stage epithermal gold mineralization at the Kelian gold mine, Kalimantan, Indonesia. Prior to brecciation, andesite intrusions (19.7 ± 0.06 Ma) were emplaced into a package of felsic volcaniclastic rocks and overlying carbonaceous carbonaceous sandstones and mudstones, and a weakly mineralized geothermal system was established. Intrusion of quartz-phyric (19.8 ± 0.1 Ma) and quartz-feldspar-phyric quartz-feldspar-phyric rhyolite (19.5 ± 0.1 Ma) into the active geothermal system triggered triggered  widespread fragmentation and formation of the maa r-diatreme complex. Subsurface phreatomagmatic and phreatic explosions disrupted the preexisting hydrothermal system, producing three composite diatreme breccia bodies (the Tepu, Runcing, and Burung Breccias). The diatremes consist of polymict breccias and sandstones that contain abundant carbonaceous matrix. A distinctive facies association comprising coherent rhyolite, jigsaw-fit rhyolite breccia, and matrix-rich breccias that contain wispy  to blocky juvenile rhyolite clasts define the root zones of the diatremes. The surficial products of maar-diatreme volcanic activity at Kelian are preserved as large blocks of well-stratified breccias. They contain accretionary lapilli and were deposited by a combination of wet, pyroclastic basesurge, fallout, and cosurge fallout processes. Evidence for syneruptive resedimentation of the pyroclastic deposits is preserved in poorly stratified breccia beds. Megablocks of phreatomagmatic base-surge deposits were dropped down several hundred meters from the maar environment into the underlying diatremes.  Volcanism in the Kelian maar-diatreme maa r-diatreme complex was dominated by a combination of phreatomagmatic and a nd phreatic processes, with subordinate hydraulic, tectonic, and dry magmatic fragmentation. The carbonaceous matrix-rich diatreme breccias acted as aquicludes during subsequent hydrothermal activity, focusing fluid flow  into the wall rocks adjacent to the diatremes, where epithermal gold mineralization and hydrothermal brecciation occurred.

Introduction SOME magmatic-hydrothermal ore deposits in volcano-plutonic arcs are associated with large volcanic-hydrothermal breccia complexes (e.g., Sillitoe and Bonham, 1984; Sillitoe, 1997). Such breccia bodies have commonly been inferred to be diatremes—large-scale breccia pipes that underlie maar  volcanoes (e.g., Lorenz, 1986). In the volcanological literature, there are few descriptions of the root zones to this class of volcano, because they can only be exposed by deep-level erosion or mining activity (Clement, 1982; Cas and Wright, 1987; Lorenz and Kurszlaukis, 2007). In contrast, ore deposit geologists have provided numerous examples of what they  infer to be breccia-filled diatremes from porphyry and epithermal settings (e.g., Acupan, Philippines: Cooke and Bloom, 1990, Cooke et al., 1996; Grasberg, Indonesia: Macdonald and Arnold, 1994; Martabe, Indonesia: Sutopo et al., 2003, 2007; Wau, Papua New Guinea: Sillitoe et al., 1984; El Teniente, Chile: Cannell et al., 2005; Yanacocha, Peru: Turner, 1997; Colquijirca, Peru: Bendezú et al., 2003; Agua Rica, Argentina: Landtwing et al., 2002; Montana Tunnels, United States: Sillitoe et al., 1985; Cripple Creek, United States: Thompson et al., 1985, Thompson, 1992; Jensen, 2003; Ros R os¸ia ¸i a Montanˇ  Montana, ˇ  Romania: Wallier et al., 2006), although the supporting evidence for this genetic interpretation is not always compelling. This is because only a few descriptive papers (e.g., Wau: Sillitoe et al., 1984; Montana † Corresponding author: e-mail,

[email protected] [email protected] u.au *Present address: Teck Cominco Ltd., Vancouver, British Columbia, Canada. 0361-0128/08/3748/689-28

Tunnels: Sillitoe et al., 1985; Cripple Creek: Thompson et al., 1985) have adequately described their key textural and morphological characteristics. Kelian is a large, breccia- and vein-hosted epithermal system of Miocene age (van Leeuwen et al., 1990). Most of the mineralization at Kelian occurred subsequent to catastrophic brecciation events that produced a nested complex of carbonaceous matrix-rich breccia pipes, dikes, and beds (Davies, 2002). Open-pit mining and deep drilling have provided excellent exposures of these breccia bodies over a vertical inter val in excess of 700 m. This environment has provided an ideal opportunity for systematic mapping, description, and interpretation of the carbonaceous matrix-rich breccia complex. This manuscript documents the characteristics and facies architecture of premineralization matrix-rich breccia bodies at the Kelian gold mine and assesses their role as a precursor to the formation of a large-scale auriferous hydrothermal system. In particular, we document individual breccia lithofacies, their facies associations and spatial distributions, and interpret their origins. We comment on the implications of our results for understanding fluid flow in and around similar large-scale discordant breccia bodies in other epithermal and porphyry ore systems and how this could influence mineral exploration.

Terminology  Breccia nomenclature has been applied inconsistently in the economic geology literature. Terms that may cause some confusion are defined below. With regards to grain size and shape terms (breccia, conglomerate, sandstone, mudstone),

689

690

DAVIES ET AL.

 we follow conventional sedimentological and volcanological definitions (e.g., McPhie et al., 1993). Maars: Maar volcanoes comprise a central crater surrounded by low rims (tens of meters) of phreatomagmatic base-surge and fallout deposits and consequently have low aspect ratios (Fisher and Waters, 1970). The craters excavate country rocks as a result of phreatomagmatic eruptions (Fisher and Waters, 1970). The surface morphology and the amount of slumping of  blocks derived from the surface back into the crater can be dictated by the nature of the wall rocks (e.g., “hard-substrate” vs. “soft-substrate” maars; Auer et al., 2007). Diatreme: Diatremes are downward-tapering, subsurface  volcanic conduits, which may lie beneath maars and are formed at least in part by explosive phreatomagmatic eruptions (Lorenz, 1986; Cas and Wright, 1987; Martin et al., 2007). Dry magmatic and/or possibly phreatic (steam) explosions may also contribute to diatreme formation but in isolation do not form diatremes. Diatremes are filled by volcaniclastic deposits and collapsed wall-rock blocks (Lorenz, 1973). Their cross-sectional areas are similar to those of  maars (<100–1,500 m; Lorenz, 1986) and their vertical extents may be several kilometers, including root zones and feeder dikes (Lorenz, 1986; Lorenz and Kurszlaukis, 2007).  Juvenile clasts: Juvenile clasts are derived by fragmentation of a parental magma. Subsurface interaction of magma with ground water or hydrothermal fluid may produce distinctive ragged-shaped (“wispy”) juvenile magmatic clasts. These can occur in the root zones of a diatreme and are a key feature to recognize as they can provide evidence of a direct magmatic contribution to breccia formation and help to infer fragmentation processes. Their irregular shapes are easily modified during transport and abrasion, and so they have a low preser vation potential. Hydrothermal breccia: A hydrothermal breccia is defined here as a coarse-grained clastic aggregate generated by the interaction of hydrothermal fluid with magma and/or wall rocks, irrespective of the source of the hydrothermal fluid. Fragmentation may be explosive (e.g., phreatic or phreatomagmatic) or it may be driven by hydraulic fracturing or tectonism. Our definition is consistent with that of Sillitoe (1985). Matrix: The matrix is the fine-grained clastic components of a breccia that occurs between larger clasts. For the Kelian breccias, we have defined clastic components with a grain size of less than 2 mm as matrix. Cement: The cement is the crystalline component within a clastic rock or fracture. There are two main types of cement: minerals precipitated from an aqueous fluid (water and/or  vapor) and minerals crystallized from magma. Cement precipitated from an aqueous fluid may include ore minerals (i.e., chalcopyrite-cemented breccia) and gangue minerals (i.e., quartz-cemented breccia), whereas those from a magma produce crystalline igneous rock (i.e., rhyolite-cemented breccia). Cement precipitated from aqueous fluids is a diagnostic component of most hydrothermal breccias.

Geologic Setting and Mineralizati Mineralization on Kelian is a bulk-tonnage epithermal gold-silver deposit. Base metal sulfide-rich, Au-Ag mineralization occurs in hydrothermal breccias, veins, and disseminations (Davies et al., 0361-0128/98/000/000-00 0361-0128/9 8/000/000-00 $6.00

2008). The mine is located in East Kalimantan, Indonesia, on the island of Borneo (Fig. 1). Kelian produced more than 240 metric tons (t) of gold prior to mine closure in 2003. Kelian is hosted mostly in a structural inlier of felsic volcaniclastic rocks (the Kelian Volcanics: Davies, 2002). The inlier is surrounded and overlain by Eocene to Miocene terrestrial and shallow marine carbonaceous sedimentary rocks of  the Kutai basin succession (Moss and Chambers, 1999). U-Pb zircon dating of the Kelian Volcanics yielded an Upper Cretaceous age (67.8 ± 0.3 Ma: Davies, 2002; Davies et al., 2003). The Kelian Volcanics have been lifted along a north-trending basement fault (West Prampus fault) at its intersection  with a regional-scale, northwest-trending crustal lineament. At the surface, this lineament manifests as a series of north west-striking, strike-slip, and oblique-slip faults. The intersection of these two regional-scale structures was the locus for magma emplacement in the district during the early  Miocene. Feldspar-hornblende-phyric andesite intrusions  were emplaced in i n rhombic, extensional domains defined by  northwest- and northeast-striking faults (Davies, 2002; Fig. 2). An age of 19.7 ± 0.06 Ma was determined determined for one of the andesite intrusions by Setiabudi et al. (2001). Intrusion of  quartz-phyric rhyolite (QP) (QP) bodies occurred at at 19.8 ± 0.1 Ma, and quartz-feldspar-phyric quartz-feldspar-phyric rhyolite (QFP) at at 19.5 ± 0.1 Ma (Davies, 2002; Davies et al., 2003).

300 km

Plio-Pleistocene volcanics Oligocene - Upper Miocene flysch

N

Tertiary Ter tiary sedimentary basins Cretaceous turbidites and melange L. to U. Cretaceous granite and tonalite

8

0

4

0

L. to M. Cretaceous melange zones L. Cretaceous volcanics and ophiolites

SABAH

Ophiolite, basic / ultrabasic units Triassic - Jurassic metamorphics and volcanics; U. Triassic sediments PermoTriassic granites and metamorphics Major lineament, fault or fold South China Sea  0

4

SARAWAK

Celebes Sea 

S    a  n    g   k    u   l    i    r    a   n    g   F    a   Busang u  l    t    z  o   n  e  

Kutai Basin  0

0

0

Indo Muro

KELIAN Schwaner Mountains 

BALIKPAPAN

Masupa Ria Barito Basin 

KALIMANTAN

0

4

0

110

Muyup

0

Makassar  Straits 

0

114

  u  s   s   a   t    t  a   i  n   r   e    M   o  u  n    M

A   d    a   n    g   F    a   u   l    t  z   o   n  e   0

118

4

0

FIG. 1. Simplified geology geology of Borneo, Borneo, showing the location of of Kelian and other Miocene epithermal gold deposits (modified after Moss and Chambers, 1999).

690

DIATREME BRECCIAS AT THE KELIAN GOLD MINE, KALIMANTAN, INDONESIA

FIG. 2. Premining surface geology of the Kelian gold deposit (modified after Davies, 2002), showing the major breccia bodies. The Kelian River originally passed through the main ore zone and had to be diverted to the north, providing excellent exposures through the Runcing Breccia. Abbreviations: Quat = Quaternary, U. Cret = Upper Cretaceous. 0361-0128/98/000/000-00 $6.00

691

691

692

DAVIES ET AL.

Brecciation and mineralization occurred synchronous with, and subsequent to, rhyolite emplacement. Previous workers documented six types of breccia at Kelian (van Leeuwen et al., 1990): tuff, fault, intrusion, hydrothermal, fluidized, and muddy breccia. It is the latter group, the muddy breccias,  which are the subject of the current study. Muddy breccias  were described by van Leeuwen et al. (1990) as dark gray, massive, polymict and polyphase breccia bodies and dikes. They are aligned roughly north-south, have intrusive contact relationships, and were found to contain clasts of sediments, tuff, andesite (concentrated at contacts with andesite intrusions), and rhyolite (van Leeuwen et al., 1990). Coarse- and fine-grained varieties were observed by these previous workers, as were late-stage sandy and pebble dikes. Mineralization at Kelian has been described by van Leeuwen et al. (1990), Davies (2002), and Davies et al. (2003, 2008). Only key features are summarized below. A hydrothermal system was established prior to the formation of the muddy breccia complex, based on the presence of truncated illite-pyrite veins and associated illite alteration in breccia clasts. Main-stage gold and silver deposition commenced during the waning stages of muddy breccia formation and continued for some time after their final consolidation. Gold and sil ver are hosted by hydrothermally cemented breccia bodies, sheeted and conjugate veins, and disseminated sulfides. Davies (2002) and Davies et al. (2003, 2008) identified five stages of mineralization that progressed from pyrite- to base metal-sulfide–dominated (sphalerite + galena ± chalcopyrite) and finally to sulfosalt-dominated mineralization (proustitepyragyrite, tennantite-tetrahedrite). Deposition of gangue minerals progressed from illite-quartz to adularia and/or quartz and/or illite and finally carbonate-dominated assemblages. Stage 1 (pre-muddy breccia) mineralization produced proximal illite-pyrite-quartz–cemented veins and breccias and distal calcite-quartz ± epidote veins. Stage 2 mineralization generated pyrite-quartz-illite–cemented breccias and  veins with minor base metal sulfides in the northern part of  Kelian and adularia-quartz-pyrite in the south. A transition to abundant base metal sulfides (galena, sphalerite, and chalcopyrite) occurred from stage 2 to stage 3A. In addition to base metal sulfides, stage 3A veins and hydrothermal breccias contain ubiquitous pyrite, local sulfosalts, and abundant native gold. Stage 3B mineralization was coeval with stage 3A and was localized at depth and on the flanks of the Kelian system. It produced base metal sulfides, pyrrhotite, marcasite, and melnikovite. Boiling produced abundant bladed carbonate during stage 3C. Stage 4 sulfosalts and sulfides are intergrown with laminated and bladed rhodochrosite. Stage 5 consists of late kaolinite, covellite, digenite, and siderite, possibly  (at least in part) of supergene origin. Gold f ormed throughout stages 1 to 4 but most occurred during stages 3 and 4. Gold occurs principally as inclusions within and intergrown with pyrite, sphalerite, galena, arsenopyrite, quartz, bladed carbonate, and sulfosalts. Hydrothermal alteration is zoned about contacts, faults, breccias, and veins (Davies, 2002; Davies et al., 2003, 2008). The alteration assemblages in andesite intrusions grade from proximal quartz-illite-pyrite through illite-carbonate-pyrite, and illite-chlorite-carbonate to distal chlorite-calcite-illite. Alteration zonation in the volcaniclastic host rocks grades from 0361-0128/98/000/000-00 $6.00

proximal quartz-illite-pyrite to distal smectite-illite assemblages. Local, intense adularia- quartz-illite and/or carbonate alteration assemblages are spatially associated with adularia and carbonate cement, respectively. Alteration patterns at Kelian were controlled by lithologic units, structure, and host-rock permeability (Davies, 2002).

Carbonaceous Breccia and Sandstone Lithofacies From detailed mapping and drill core logging, we have recognized two distinctive groups of matrix-rich breccias at Kelian. Both are characterized by abundant carbonaceous components (clasts and/or matrix) derived from the surrounding sedimentary units. Collectively, these are the breccias that  were described by van Leeuwen et al. (1990) as muddy breccias and are here referred to as carbonaceous matrix-rich breccias, in order to emphasize their common, distinguishing component: carbonaceous matrix.  We have divided the carbonaceous matrix-rich breccias into two groups: unstratified (A facies) and stratified (B facies). Each group consists of several facies and subfacies (Tables 1, 2). Breccia facies A include monomict carbonaceous mudstone and sandstone-clast breccias, polymict carbonaceous matrix-rich breccias, and monomict rhyolite breccias with carbonaceous matrix. Facies have been defined by variations in (1) breccia composition (monomict vs. polymict), (2) diagnostic clast types, and (3) grain size. Subfacies are distinguished by subtle variations in the dominant clast type, other than carbonaceous mudstone and sandstone (i.e., andesite,  volcaniclastic rocks, rhyolite), or modal mineralogy of the clasts (QP vs. QFP). The breccias vary in appearance due to clast abundance, size, lithology, and degree of clast rounding but are generally black, dark gray or, where intensely illite ± quartz-pyrite altered, light gray in color. Although there are clast-supported breccias in this group, most are matrix supported and massive. The polymict facies all have similar matrix components (sand-sized fragments of quartz, polymict lithic clasts, and abundant carbonaceous mudstone and sandstone clasts) but have variable clast populations (various combinations of andesite, rhyolite, volcaniclastic rocks, earlier formed breccias, accretionary lapilli, charcoal or coal, and rare vein fragments). The B facies consist of both breccias and sandstones and have similar clast and matrix components to the A facies but differ in that they are stratified. B facies vary in the styles of  stratification and the principal clast components (rhyolite- or  wall rock-dominated). Subfacies are distinguished by grainsize variations (sandstone to breccia; Tables 1, 2). Unstratified breccias and sandstones Facies A1: These are dark gray to black, monomict, matrixrich (locally up to 60%) carbonaceous mudstone-clast breccias that contain abundant clay gouge (Fig. 3A–D; Tables 1, 2). Some A1 breccias have formed at the contact between the Runcing Rhyolite and carbonaceous mudstones (Fig. 3A), but most are spatially associated with faults (e.g., Fig. 3B). Many  A1 breccias are internally banded (e.g., aligned clasts, gouge, foliation). Sulfide vein clasts are present locally (Fig. 3C), as are late hydrothermal cements (Fig. 3D). Facies A2: These breccias occur in contact with coherent flow-banded rhyolite dikes (Fig. 4A). Facies A2 are clast-rich,

692

693

DIATREME BRECCIAS AT THE KELIAN GOLD MINE, KALIMANTAN, INDONESIA TABLE 1. Summary of Breccia Facies A and B at Kelian: Composition, Internal Organization, and Hydrothermal Features Facies and facies name

Alteration and mineralization

Subfacies

Internal organization

Clast shape and type

Matrix

A1: Monomict carbonaceous breccia

None

Monomict, jigsaw-fit, clast rotated or massive Clast to matrix supported Some imbricated clasts Local banding or foliation

Shape: angular to subrounded Type: Cms, local sulfide vein fragments

Mud- to sandsized fragments Local clay gouge

Pervasive ill-sme-pyr alt, local clay gouge

A2: Monomict,  jigsaw-fit rhyolite breccia with carbonaceous matrix

QP: QP clast bearing QFP: QFP clast bearing

Jigsaw-fit, clast to matrix supported

Shape: angular or irregular with cuspate and/or wispy margins Type: QFP, QP

Black to dark gray matrix of milled A1 facies and/or clay gouge

Weak to moderate kao ± ill alt of rhyolite clasts and ill-pyr alt of matrix

A3: Polymict, carbonaceous, matrix-supported breccia with wispy and/or blocky rhyolite clasts

QP: QP clast bearing QFP: QFP clast bearing

Polymict, massive and matrix supported

Shape: irregular to angular with wispy or cuspate margins Type: Cms, And, V, QP, QFP, charcoal, coal, vein, breccia

30–50% mud- to sand-sized fragments

Weak to intense qtz-ill-pyr alt And and V clasts locally contain stage 1A veins (i.e., some mineralization predated brecciation)

A4: Polymict, carbonaceous, matrix-supported breccia with rhyolite clasts

QP: QP clast bearing QFP: QFP clast bearing QP/QFP: QP and QFP clast bearing

Polymict, massive and matrix supported

Shape: angular to Mostly sand-sized subangular, some Up to 90% matrix subrounded in fine-grained Type: Cms ± And, V, QP, breccia QFP, coal, charcoal, AL, vein and breccia

Tepu Breccia: qtz-illpyr±crb alt , local illqtz -pyr-crb alt Qtz-ill-pyr±crb alt in Burung Breccia near 393 Breccia

A5: Polymict, carbonaceous, matrix-supported breccia

V: volcaniclastic clast dominated And: andesite clast dominated

Polymict, jigsaw-fit, clast rotated and massive

Shape: angular to subangular Type: V, And, Cms, rare C1, C7 breccia clasts

Sand-sized fragments

Weak to moderate pervasive ill-qtz-crbpyr alt Local ill-qtz-pyr±crb alt

A6: Polymict, carbonaceous, clast-supported breccia and conglomerate

QP: QP clast bearing QFP: QFP clast bearing V: volcaniclastic clast bearing

Polymict Generally clast supported Minor matrix supported

Shape: abundant subrounded to minor subangular clasts Type: QFP, Cms, And, V  

Up to 20% sand-sized fragments

A6-QFP: weak to intense ill-pyr alt A6-V: weak sme-ill alt

A7: Discordant, carbonaceous, crystal and lithic sandstone with rhyolite fragments

AL: accretionary lapilli bearing

Polymict Generally massive Locally stratified

Shape: angular to subangular Type: Cms, And, V, QFP, QP, broken qtz and fsp, local AL

Up to 100% mud- and sandsized fragments

Weak to moderate pervasive ill-pyr alt in Tepu Breccia Intense qtz-ill-pyr±crb alt in Burung Breccia

B1: Well-stratified, rhyolite-clast-rich, carbonaceous breccia and sandstone

A: medium-grained volcanic breccia B: fine-grained volcanic breccia C: volcanic sandstone

Polymict, poorly to moderately sorted Matrix to clast supported Planar, wavy, dune, and low-angle cross stratified Normal, reverse, doublegraded beds

Shape: angular to subangular Type: V, QP, Cms, minor Al

Sand-sized fragments

Weak sme-ill ± kao and trace disseminated pyr

B2: Well-stratified carbonaceous breccia and sandstone

A: medium-grained volcanic breccia B: fine-grained volcanic breccia C: volcanic sandstone

Polymict, poorly to moderately sorted Matrix to clast supported Planar, wavy, dune, and low-angle cross stratified Diffuse normal, reverse and double-graded beds

Shape: angular to subrounded Type: Cms, QP, V, abundant Al

Mud- to sandsized fragments

Weak sme-ill ± kao alt and rare disseminated pyr

B3: Weakly stratified carbonaceous breccia and sandstone

A: medium-grained Polymict, poorly to volcanic breccia moderately sorted B: fine-grained Matrix supported volcanic breccia Diffuse tabular and troughC: volcanic sandstone shaped beds, graded

Shape: angular to subrounded Type: V, Cms, QP

40% mud- to sand-sized matrix

Weak sme-ill alt

Abbreviations: And = andesite, AL = accretionary lapilli, alt = alteration, Cms = carbonaceous mudstone and/or sandstone, crb = carbonate, fsp = feldspar, ill = illite, kao = kaolinite, pyr = pyrite, QFP = quartz-feldspar-phyric rhyolite, QP = quartz-phyric rhyolite, qtz = quartz, sme = smectite, V = volcanic sandstone and/or mudstone 0361-0128/98/000/000-00 $6.00

693

694

DAVIES ET AL. TABLE 2. Summary of Breccia Facies A and B at Kelian: Contact Relationships, Morphology, and Distribution

Facies

Contacts

Morphology

Distribution

A1

Sharp or gradational into unbrecciated, laminated and thinly bedded carbonaceous mudstone and/or sandstone Crosscuts B1, B2, B3 breccias sharply

Irregular to tabular inclined to subvertical zones m to 10s of m across Forms a shell up to 200 m across around the Runcing Breccia and Runcing Rhyolite Irregular dikes (0.25–5 m thick) cut B facies on Gunung Runcing

Common in Runcing Breccia Also occurs in hanging wall to West Prampus, Burung, and North Burung faults, shallow levels of Burung Breccia, and on margins of Tepu Breccia in south wall of the East Prampus open pit

A2

Highly irregular at centimetrer scale, with transition from coherent rhyolite into  jigsaw-fit and then clast-rotated rhyoliteclast breccia over 0.5 to 10 m Complicated zones of fault gouge mingled  with rhyolite occur at margins of rhyolite dikes in fault zones

Tabular to irregular sheets along margins of fault-zone–hosted QP rhyolite dikes Tabular dikes (10 cm to 1 m width) in fault zones, fingers and irregular pods in A1, A3 and A4 breccias

Runcing and Tepu Breccias, Burung fault hanging wall, NW-trending faults, crosscutting the Burung fault Locally, in A1 and A4 breccias adjacent to the Runcing Rhyolite

A3

Convoluted, irregular and penetrating contacts with earlier breccia phases; local gradation into A2 breccias Contacts with pipes of same breccia facies are gradational over 2 to 5 m; crude increase in matrix abundance toward contact

Tepu Breccia: subvertical A3 pipes are commonly >30 m, and up to 80 m, across and crosscut A4, A3, and A2 breccia pipes Runcing Breccia: irregular pipes and pods vary from 1 to 20 m across

Common in Tepu Breccia, typically internal to the breccia body, rarely at margins Minor occurrences peripheral to Runcing QFP and QP intrusions in NW-trending faults, and NW-trending faults in the Burung fault hanging wall

A4

Sharp or diffuse, and/or erosional In the Tepu and Burung Breccias, several of the bounding surfaces are postbrecciation faults

Subvertical irregularly shaped pipes, dikes Thin breccia dikes range from centimeter to meter scale and extend tens of meters vertically Larger irregular pipes range from 10 to over 100 m across and are vertically continuous over 10s to 100s of meters

Principally in the Tepu, Burung, and Runcing Breccias Smaller bodies along NW-trending faults in East and West Prampus zones Narrow dikes (<5 m wide) in the footwall to the Burung fault

A5

Sharp to gradational over meters into A3, A4, or A1 facies, polymict hydrothermally cemented breccias, unbrecciated andesite, unbrecciated carbonaceous sedimentary rocks or unbrecciated volcaniclastic rocks

Irregular sheets and tabular, subvertical dikes 1 to 50 m across that extend along strike for up to several hundred meters

Occur at the margins of the Tepu Breccia, in particular at its southern end, and in the Burung Breccia near its margins NW-trending zones in the west wall of the northern East Prampus pit, spatially associated with NW-trending faults

A6

Sub-facies A6-QFP: Sharp, discordant and erosional contacts against all other A facies; contacts typically contorted and penetrative Subfacies A6-V: Sharp, discordant contacts  with A1 and A4 breccia, carbonaceous mudstone and volcaniclastic rocks

Subfacies A6-QFP: Isolated subvertical pipes (up to 30 m across) and narrow dikes (up to 7 m); vertical extent >75 m Subfacies A6-V: Small (<10 m) pods and dikes (<1-m width) with sharp and wavy contacts

Subfacies A6-QFP: Minor phase in Tepu and Burung Breccias, isolated pods on the north side of the Kelian River diversion Subfacies A6-V: Only observed adjacent to the Runcing Rhyolite’s SW margin, and narrow (<10 cm) dikes in the Burung fault footwall

A7

Typically sharp and less commonly gradational Sharp and generally subplanar at dike margins or sharp, contorted, and penetrative around larger pipes and at tips of narrow bodies Gradational at margins of A4 breccia bodies

Large irregular subvertical pipes up to 20 m across Narrow, <20 cm to 3 m wide, irregular dikes Local narrow (cm to m) zones at the margins of A4 breccias

Minor, but widespread, late-stage facies in the Tepu Breccia Some A7 clasts occur in A3 breccias Occurs in the Burung Breccia as a marginal phase to, and at the tips of A4 breccias, and locally as thin dikes and irregular fingers

B1

Gradational contacts between facies B1 and B2 Contacts between subfacies in individual beds are also gradational Contacts between facies are sharp where they occur as bed boundaries Contacts are locally erosional with basal scours and medium-grained breccia in scours

Well-developed centimetre- to meter-scale bedding Planar wavy planar, low-angle cross stratified and low-angle dune bed forms Diffuse normal and reverse grading; syndepositional faults, flames, and slumps Interbedded medium- and fine-grai ned breccia and volcanic sandstone

Runcing Breccia

B2

Sharp contacts at bed boundaries Erosional contacts where B1-A and B1-B breccias overlie B2 breccias Local extensive disruption by synsedimentary faults, ball and pillow and flame structures Locally gradational into B1 facies

Well-developed beds (<1 cm to 10s of cm) with normal grading and moderate sorting 10- to 20-cm beds with thin (1 cm) normally graded tops Local fine-grained breccia or volcanic sandstone tops to E1 breccia beds

Runcing Breccia

B3

Diffuse and erosional

Tepu Breccia: 30 × 10 × 75 m, crudely stratified blocks in A2 breccia pipes Runcing Breccia: erosional channels that cut  well-bedded B1 and B2 facies

Tepu and Runcing breccias

Abbreviations: QP = quartz-phyric rhyolite, QFP = quartz-feldspar-phyric rhyolite, V = volca nic sandstone and/or mudstone 0361-0128/98/000/000-00 $6.00

694

DIATREME BRECCIAS AT THE KELIAN GOLD MINE, KALIMANTAN, INDONESIA

695

FIG. 3. Breccia facies A1. A. North wall of East Prampus pit (March 1999), showing domains of A1 breccia that have cut and/or surrounded laminated carbonaceous mudstone and sandstone (CMS) in the hanging wall to the Runcing fault and 394 ore zone. The Runcing Rhyolite intrusion is surrounded by A1 breccia and carbonaceous sedimentary rocks. Light gray, bedded B facies are crosscut by dikes of A1 breccia. B. Jigsaw-fit to clast-rotated A1 breccia. Larger clasts are outlined in white, as are some internal fractures. A zone of matrix- to clast-supported A1 breccia with aligned clasts between the yellow lines is interpreted to be a fault zone (Runcing Breccia, 1090 mRL). C. Matrix-supported A1 breccia with clasts of carbonaceous mudstone, stage 2A or 3A pyrite vein fragments (PYR) and clay gouge in the matrix (Runcing Breccia, 1060 mRL). D. Clastsupported A1 breccia with minor matrix and minor carbonate (CRB) cement. There are ball and pillow structures in the bedded carbonaceous sandstone and mudstone clast (Burung Breccia, K317-188.0 m).

monomict, jigsaw-fit rhyolite breccias that have a carbonaceous mudstone matrix (Tables 1, 2; Fig. 4B-C). Rhyolite clasts (QP or QFP) are angular and blocky with curviplanar margins (Fig. 4B) or wispy with ragged margins (Fig. 4C). Some rhyolite clasts have partially wispy and partially blocky  margins. A2 breccias occur in narrow dikes (10 cm to 1 m) in preexisting faults or as irregular fingers and pods in monomict A1 or polymict A4 breccias. In several locations, a complete gradation from coherent rhyolite (Fig. 4A) to monomict  jigsaw-fit rhyolite breccia with carbonaceous mudstone matrix (Fig. 4B) to polymict carbonaceous matrix-supported breccia that contains blocky and wispy rhyolite clasts (A3 facies; Fig. 5) occurs over strike lengths of 0.5 to 10 m. Facies A3: These unsorted, matrix-supported, polymict carbonaceous breccias are characterized by texturally distinctive  wispy and blocky QP or QFP rhyolite clasts (Fig. 5A-F; Tables 1, 2). Wispy rhyolite clasts have delicate, irregular margins (Fig. 5B-D), whereas blocky rhyolite clasts have curviplanar margins (Fig. 5E-F). Some rhyolite clasts have partial wispy  0361-0128/98/000/000-00 $6.00

and partial blocky, curviplanar margins. All other clast types identified in the A3 facies are more rounded than the rhyolite clasts (Fig. 5). Apart from the common clast types (carbonaceous mudstone and sandstone, volcaniclastic sandstone, andesite, and rhyolite), unusual clast types in A3 breccias include carbonized wood (Fig. 5C), A4 breccia, A7 sandstone, and finely bedded B2 sandstone. Accretionary lapilli (2–8 mm diam) comprise up to 5 percent of some A3 breccias. Facies A4: These are the most abundant of the carbonaceous matrix-rich breccia facies at Kelian. They are poorly  sorted, carbonaceous matrix-supported polymict breccias (Fig. 6; Tables 1, 2). Discrete A3 and A4 breccia pipes can occur adjacent to each other (e.g., Fig. 6A), and the A4 breccias are similar in composition to A3 breccias (e.g., Fig. 6BC). There are, however, no wispy rhyolite clasts in the A4 breccias, and rounded rhyolite clast shapes are more common (e.g., Fig. 6D-E). Unusual, distinctive and rare clasts that have been observed in A4 breccias include accretionary  lapilli, hydrothermally cemented breccia, A3 and A1 breccia,

695

696

DAVIES ET AL.

blocks of stratified facies B breccias (Fig. 6C), and unstratified A4 breccia blocks that are up to 50 m in diameter. Andesite and volcaniclastic rock clasts that contain truncated  veins and prebreccia alteration assemblages have also been observed (Fig. 6B). There is a correlation between increased matrix abundance and clast roundness and decreased clast sizes in A4 breccias. Small dikes of A4 breccia have been observed locally (Fig. 6F). Facies A5: These occur both as discrete bodies (e.g., Fig. 7A) and with other A facies. They contain angular to subrounded clasts of carbonaceous mudstone, sandstone, volcaniclastic rocks, and andesite (Fig. 7B-C; Tables 1, 2) and are distinguished from A4 breccias by a lack of rhyolite clasts. Clasts of A1, A4, and hydrothermally cemented breccia occur locally. Facies A6: A6 breccias and conglomerates occur as irregular subvertical pipes and dikes that have crosscut and eroded A4 and A3 breccias (e.g., Fig. 8A). The A6 facies is distinguished by abundant (60–90%), subrounded to rounded QFP clasts (Fig. 8B), and ranges from clast- to locally matrix-supported, poorly to moderately sorted breccia and conglomerate (Tables 1, 2). A3-QFP dikes have cut A6 breccias locally  (Fig. 8C). Facies A7: Massive to locally stratified, poorly sorted, discordant, carbonaceous sandstone characterizes facies A7 (Fig. 9; Tables 1, 2). The most abundant components are fragments of carbonaceous mudstone and sandstone. However, fine fragments of QP and/or QFP rhyolites are ubiquitous and locally have wispy and/or curviplanar margins. Accretionary  lapilli are present locally. A7 discordant sandstone occurs as the only facies in breccia pipes that are up to 20 m in diameter, as fine-grained facies that grades into A3 or A4 facies  within larger breccia bodies, and as dikes less than 3 m wide (e.g., Fig. 9). Some A7 sandstones are laminated subparallel to the walls of the larger breccia bodies.

FIG. 4. Breccia facies A2. A. Flow-banded rhyolite dike with clots and elongate black domains of carbonaceous A1 breccia, mudstone, and fault gouge. Top of sample was located at the margin of the rhyolite intrusion,  where there is a gradation into wispy jigsaw-fit A2-QP breccia and then into A3-QP breccia (not shown; northwest East Prampus pit, 1090 mRL). B. Jigsaw-fit A2-QP breccia with QP clasts separated by a matrix of carbonaceous clay gouge and sand-sized lithic fragments. Jigsaw-fit rhyolite breccia can be traced along strike into a coherent rhyolite dike that intruded a gouge-filled fault (Runcing Breccia, 1060 mRL). C. Jigsaw-fit to clast-rotated A2-QP breccia with matrix of fine QP splinters and black carbonaceous mudstone clasts. There is a complete gradation from this breccia into coherent rhyolite (not shown). Note cuspate clast margins and similarity between the morphology of these clasts and those in the A3-QP breccias in Figure 5 (Runcing Breccia, 1060 mRL).

0361-0128/98/000/000-00 $6.00

Stratified breccias and sandstones Facies B1: These are well-stratified breccias that have angular to subangular clasts of aphanitic to finely (<1 mm) QP rhyolite, volcaniclastic mudstone and sandstone, and carbonaceous mudstone and sandstone (Fig. 10A-C; Tables 1, 2). They also contain minor accretionary lapilli (Fig. 10D). B1 breccias occur in medium to thick beds with planar, wavy planar, low-angle cross-stratified (Fig. 10A) and low-angle dune bed forms (Fig. 10B). Some beds are normally graded. Others have a core of coarser fragments and display both diffuse normal and reverse grading in a single bed (Fig. 10C). Facies B2: B2 breccias consist of angular to subangular clasts, dominated by carbonaceous mudstone and sandstone,  with lesser QP, and volcaniclastic mudstone and sandstone (Fig. 10; Tables 1, 2). Accretionary lapilli (1–8 mm diam) are locally abundant, in some cases occurring closely packed in thin (1–5 cm) beds (Fig. 10D). B2 facies are predominantly  planar bedded (Fig. 10D-E), in contrast to B1 breccias. Softsediment deformation textures, such as synsedimentary faults and flames (Fig. 10E), are common. Facies B3: Weakly stratified B3 breccias (Fig. 11; Tables 1, 2) are polymict, matrix supported, and unsorted. They occur in medium to thick beds with diffuse to sharp, planar or channel bed forms (Fig. 11). A diffuse to sharp, undulose, erosional contact separates the underlying well-bedded B1 and

696

DIATREME BRECCIAS AT THE KELIAN GOLD MINE, KALIMANTAN, INDONESIA

FIG. 5. Breccia facies A3. A. Coarse-grained marginal A3-QFP facies of Tepu breccia with large blocks of andesite (And) and wispy quartz-feldspar-phyric rhyolite clasts (QFP). Contact with coherent andesite is 1 m to right of area shown in frame (Tepu Breccia, 1000 mRL). B. Medium-grained A3-QP breccia with abundant wispy and ragged quartz-phyric rhyolite cl asts (QP) in a carbonaceous, sand-sized, lithic, and crystal matrix (Tepu Breccia, 1070 mRL). C. Fine-grained A3-QP breccia with 20% wispy and blocky rhyolite clasts and fragments of c arbonized wood (Tepu Breccia, 1070 mRL). D. Detail of ragged QP rhyolite clasts in fine- to medium-grained polymict A3-QP breccia with mudstone clasts (CMS). The two large QP clasts near the top of the photograph display small-scale jigsaw-fit texture (Tepu Breccia). E. Detail of blocky finely porphyritic rhyolite clasts in fine- to medium-grained A3-QFP breccia. There is a range of clast shapes in the rhyolite clast population—some have delicate curviplanar margins, whereas others are subrounded, as are the sedimentary clasts. The blocky clasts are interpreted to be juvenile clasts that have not been recycled, and the subrounded clasts are interpreted to be juvenile clasts that have undergone transport and abrasion possibly during multiple brecciation events. The breccia also contains clasts of  carbonaceous mudstone and sandstone (CMS) and volcanic mudstone and sandstone (V; Tepu Breccia). F. Medium- to coarse-grained A3-QFP breccia with blocky rhyolite clasts (Tepu Breccia, 1070 mRL).

0361-0128/98/000/000-00 $6.00

697

697

698

DAVIES ET AL.

FIG. 6. Breccia facies A4. A. Fault contact between A4 and A3 breccias of the Tepu Breccia (right) and volcaniclastic rocks (V) and hydrothermal breccias (left). Bottom and top of photograph are 1060 mRl and 1110 mRL, respectively (50-m vertical exposure). The massive A4 and A3 breccias in this exposure crop out at the northern margin of the Tepu Breccia. Rhyolite intrusions (R) occur in a curved, possibly faulted, contact between the A3 and A4 breccia pipes. Internal contacts between A4 and A3 facies are diffuse, irregular, and gradational in this area. B. Coarse-grained A4 breccia with large blocks of  pervasively illite-pyrite-carbonate–altered andesite that contain stage 1A pyrite-carbonate-sericite veins. The veins are truncated at clast margins and provide evidence for an early hydrothermal system (Tepu Breccia, 1060 mRL; scale bar = 10 cm). C. Medium-grained A4-QP and/or QFP breccia with rare blocks up to 5 × 3 × 4 m of stratified B1 and B2 facies (Runcing Breccia, 1090 mRL). D. Medium-grained polymict A4-QP and/or QFP breccia cut by irregular stage 3A pyrite-sphalerite  veins. The breccia contains subangular to subround clasts of carbonaceous mudstone and sandstone (CMS), volcanic mudstone and sandstone (V), andesite, and quartz-feldspar-phyric rhyolite (QFP; Tepu Breccia, drill hole K777-213.7 m). E. Fine- to medium-grained A4-QP and/or QFP breccia with subround to subangular clasts of quartz-phyric rhyolite (QP). There is a single clast of A3-QP breccia at top of frame. Light color (cf. Fig. 6D) is due to pervasive, intense quartz-illitepyrite alteration of the Burung Breccia in the vicinity of the 393 ore zone. Irregular stage 2A and/or 3A pyrite veins have crosscut the breccia (Burung Breccia, drill hole K450-291.2 m). F. A4-QFP dikes in the Runcing Rhyolite (QFP). Clasts and matrix in the dikes consist of carbonaceous mudstone and QFP fragments. QFP clasts in this example are angular but do not have wispy or curviplanar margins. The QFP clasts were derived from the local wall rocks and are not a juvenile magmatic component (Runcing Rhyolite, Kelian River diversion east end, south side, 1100 mRL). 0361-0128/98/000/000-00 $6.00

698

DIATREME BRECCIAS AT THE KELIAN GOLD MINE, KALIMANTAN, INDONESIA

699

B2 facies from overlying weakly stratified B3 breccia at one location (Fig. 11A). B3 facies also occur as large blocks (meters to tens of meters) in polymict carbonaceous matrix-rich breccia pipes (Fig. 11B). These blocks contain thick (0.75–5 m), weakly graded beds and are crosscut by A4 breccia pipes. The B3 blocks have disaggregated transitional contacts with the surrounding breccias (Fig. 11B).

Breccia Geometries, Facies Distributions, and Spatial Associations Three large, discordant bodies of carbonaceous matrix-rich breccia are exposed in the Kelian mine area: the Tepu, Burung (van Leeuwen et al., 1990), and Runcing Breccias (S. Hunt, pers. commun., 1997; Fig. 2). Other less voluminous, unnamed carbonaceous matrix-rich breccia bodies also occur at and around Kelian. Each of these breccia bodies has a complex internal arrangement of breccia facies. The composite Runcing, Tepu, and Burung Breccia bodies consist of multiple breccia pipes and dikes that each contains one or more of the A facies breccias. The bodies formed by  multiple brecciation episodes and have complex internal contact relationships. Both internal and country-rock contacts are typically gradational over meters to tens of meters, especially   where earlier breccias have been crosscut and stoped or removed by emplacement of later breccias. Some internal f acies occur as subvertical pipes that are smaller in volume than but have similar gross geometry to, the enclosing Runcing, Tepu, and Burung Breccia bodies. The Tepu and Burung Breccia geometries were constrained partly by andesite intrusions (Fig. 2). Because the Runcing Breccia was not constrained by competent andesitic bodies, it developed a more equant geometry partly controlled by and subsequently disrupted by faults. In the area of both the Tepu and Runcing Breccias, carbonaceous mudstone to fine-grained sandstone is juxtaposed against Upper Cretaceous volcaniclastic rocks (Davies et al., 2008) by a series of high-angle normal and wrench faults.  Well-developed fault breccia zones are present in the carbonaceous strata, especially in the hanging wall to t he Burung fault (Fig. 2). Fault breccias crosscut and are crosscut by monomict and polymict carbonaceous matrix-rich breccias, indicating that fault movement continued during formation of the breccias.

FIG. 7. Breccia facies A5. A. Dark gray-colored A5 breccia body in north west wall of the East Prampus pit (base of photograph at 1030 mRL, top at 1110 mRL). A prebreccia andesite intrusion (And) crops out to the left of the breccia and bedded volcaniclastic rocks (V) to the right. Left side of A5 breccia is andesite-clast dominated (A5-And subfacies), right side is volcaniclastic-clast dominated (A5-V subfacies), and center is undifferentiated A5 breccia with subequal andesite and volcaniclastic clasts. B. Coarse-grained, matrix-supported A5-A facies, with andesite (And), carbonaceous sedimentary  (Cms), and minor volcaniclastic (V) clasts. Andesite clasts are pervasively illite-pyrite-carbonate altered (East Prampus, 1060 mRL). C. Coarse-grained, clast-supported A5-V breccia. Image shows transition from carbonaceousmudstone-clast–dominated polymict breccia (left-hand side) to polymict volcanic-sandstone-clast–dominated breccia (right-hand side; Tepu Breccia, AD97127, 1090 mRL).

0361-0128/98/000/000-00 $6.00

Tepu Breccia The northeast-trending Tepu Breccia (Table 3) is located between the Crusher Andesite and Eastern Andesite intrusions (Figs. 2, 12A) and formed in a fault-bounded slice of  carbonaceous mudstone located along the East Prampus fault (Fig. 12B; Davies, 2002). The breccia body is downward tapering in cross-sectional view. Premining maps (PT Kelian Equatorial Mining, unpub. data) suggest that the Tepu Breccia flared markedly above the roof of the Crusher Andesite. An east-trending arm of the Tepu Breccia extends along the Tepu fault across the Eastern Andesite and into a mineralized hydrothermal breccia body known as the 255 Breccia (Figs. 2, 12A; Davies, 2002; Davies et al., 2008). The Tepu Breccia consists of multiple crosscutting pipes (e.g., Fig. 12C). Each pipe contains either a single breccia or, less commonly, multiple pipes and dikes of A and minor B

699

700

DAVIES ET AL.

FIG. 9. Breccia facies A7. Contact between in situ carbonaceous mudstone (CM) and a polymict A4-QP and/or QFP breccia facies of the Tepu Breccia. Both units are cut by subvertical dikes of A7 facies carbonaceous sandstone that have undergone moderate illi te-quartz-pyrite alteration (Tepu Breccia, 1070 mRL).

FIG. 8. Breccia facies A6. A. View of Tepu Breccia from 1060 RL. Sub vertical pipes of A6-QFP breccia have intruded massive A3-QFP breccia. The A6-QFP contacts are sharp, discordant, and locally contorted. B. Detail of clast-supported domain in A6-QFP breccia pipe shown in (A), with subrounded and faceted clasts of quartz-feldspar-phyric rhyolite (QFP) and minor volcanic sandstone (V) and carbonaceous sandstone (CMS) clasts set in a polymict sand-sized matrix (Tepu Breccia, 1070 mRL). C. A6-V conglomerate crosscut by a pod of A3-QFP breccia. The A3-QFP breccia has entrained rounded clasts from the A6-V conglomerate (Runcing Breccia, 1100 mRL).

breccia. Bedding in these blocks is truncated at the block margins (Fig. 11B). Formation of the Tepu Breccia postdated andesite intrusion, based on the presence of andesite clasts containing early-formed pyrite-illite veins (Fig. 6B; Table 3; Davies 2002; Davies et al., 2008). Brecciation was contemporaneous with rhyolite intrusion and early stage 2 mineralization and alteration. The juvenile magmatic component changed from QP to QFP with time (Table 3). Mineralized stage 3 carbonate  veins crosscut the Tepu Breccia (Table 3), indicating that brecciation ceased prior to stage 3 vein formation.

Runcing Breccia The Runcing Breccia is located north of the Burung-Runcing-Discovery fault system (Fig. 2). At the preserved levels, it is enclosed by carbonaceous sedimentary host rocks and unconformably overlain by Pliocene-Pleistocene mafic volcanic facies (Table 3; Fig. 13). Pipes are typically tens to less than rocks. Of the three carbonaceous matrix-rich breccia bodies at 100 m across and are vertically extensive (> several hundred Kelian, the Runcing Breccia is the largest in map view (750 × meters; Figs. 6A, 8A). Several megablocks (10 × 50 m) of  750 m; Fig. 2; Table 3). Its geometry is poorly constrained diffusely bedded B3 breccia occur in pipes of polymict A4 below approximately 200 m due to the lack of deep drilling, but 0361-0128/98/000/000-00 $6.00

700

DIATREME BRECCIAS AT THE KELIAN GOLD MINE, KALIMANTAN, INDONESIA

701

FIG. 10. Breccia facies B1 and B2. A. Low-angle cross- and planar-stratified fine- (B1-B) to medium-grained (B1-A) breccia. B. Normally graded B1 breccia interbedded with B2 breccia. Broad dune bed forms and low-angle cross stratification are present, which is c onsistent with transport in high velocity currents. C. Interbedded B1 and B2 breccias. Dark B2-C beds contain abundant accretionary lapilli. D. Detail of accretionary lapilli (AL) in thin B2-C and B1-C beds. Upper B2-C bed consists of closely packed 1- to 3-mm accretionary lapilli. E. Flames and synsedimentary faults in interbedded B1 and B2 breccia. Dark B2 beds are in general finer grained than B1 beds and commonly contain abundant a ccretionary lapilli, clasts of B2-B breccia and B2-C sandstone.

it appears to taper with increasing depth. The Runcing Breccia is surrounded by a series of arcuate faults which form a crude circular boundary around the breccia body. These normal faults are interpreted here as ring faults along which gravitational collapse of the breccia pipe has occurred (Fig. 14). 0361-0128/98/000/000-00 $6.00

The Runcing Breccia consists of a polyphase polymict breccia pipe surrounded by a shell of brecciated and in situ carbonaceous mudstone and sandstone. The breccia pipe also contains a megablock (300 × 250 × >75 m) of bedded B facies. A late-stage QFP intrusion has been emplaced mostly 

701

702

DAVIES ET AL.

A3 breccias occur as irregular pods and dikes surrounded by  A1 breccia at several locations in the Runcing Breccia. One of  the A3 breccias grades into A2 jigsaw-fit breccia and then into a coherent QFP dike extending outward from the Runcing Rhyolite. In the footwall to the Burung fault, around the edge of the Runcing Breccia, QP dikes have intruded northweststriking faults. These dikes transform along strike into A2 and then A3 breccias with wispy and dense, cuspate fragments. The dike margins have gradational contacts with carbonaceous fault gouge, defined by mingled QP and A1 breccia. Formation of the Runcing Breccia is inferred to have commenced during stage 1 mineralization. Intrusion of the rhyolite plug occurred prior to stage 3 mineralization, based on crosscutting relationships. Collapse along ring faults occurred before and probably during stage 3 mineralization but predated stage 4.

FIG. 11. Breccia facies B3. A. Weakly stratified B3-A breccia in erosional contact with bedded B1 and B2 facies. The components of the B3-A breccia are the same as those in the B1 and B2 breccias, but there is a higher proportion of sand-sized matrix due to the homogenization of generally finer grained B2 breccias with coarser grained B1 breccias (Runcing Breccia). B.  Weak stratification in B3-A breccias within the Tepu Breccia. B3-A beds are truncated by discordant A4 and A5-V breccias facies.

into a monomict A1 breccia domain. The QFP intrusion is oriented northeast and has been emplaced along the inferred projection of the prebrecciation Burung fault. A smaller QFP plug (<50 m) and several QFP dikes and pods occur in the polymict carbonaceous matrix-rich breccia at depths of 50 to 100 m below surface. 0361-0128/98/000/000-00 $6.00

Burung Breccia The Burung Breccia is the smallest (275 × 175 × >450 m) of the three carbonaceous matrix-rich breccia bodies at Kelian (Table 4). This downward-tapering pipe is located at the intersection of the northwest-striking fault corridor with the northeast-striking fault and breccia corridor that forms the core of the Kelian system (Fig. 2). The Burung Breccia consists of a polymict breccia pipe surrounded at shallow levels on the west by monomict carbonaceous breccias and in situ mudstone. The breccia tapers to ward its base into several subvertical fins that are aligned parallel to northwest-trending faults. The Burung Breccia is nested within the 393 Breccia, a hydrothermal breccia body that consists of polymict and monomict open space-filling hydrothermal breccias cemented by sulfide and carbonate minerals (paragenetic stages 2 and 3; Davies, 2002; Davies et al., 2008). The hydrothermal breccias consist of wall-rock clasts and only minor carbonaceous mudstone clasts (Fig. 15). Near surface, the 393 Breccia forms a thin (<20 m), discontinuous rind around the Burung Breccia,  which thickens and expands to the north with increasing depth. The 393 Breccia continues to taper beneath the base of the Burung Breccia and extends more than 200 m deeper. The inner rind of the 393 Breccia has diffuse contacts with the Burung Breccia and mutual clast-in-clast relationships have been observed (Fig. 15). These observations suggest that formation of the Burung Breccia and at least the initial hydrothermal breccia facies of the 393 Breccia body were linked. Similar relationships have been described from the Balatoc diatreme at Acupan, Philippines, by Sillitoe (1985). The inner rind of the 393 Breccia is cemented by stage 2A pyrite, which helps to constrain the timing of the Burung Breccia formation prior to this paragenetic stage. The Burung Breccia is crosscut by stage 3 carbonate-base metal sulfide  veins so its formation predates stage 3. Intrusion of an apparently rootless QFP rhyolite plug occurred late in the formation of the Burung Breccia. It is unclear if this plug breached the paleosurface. Several other QP and QFP plugs occur at depth in the Burung Breccia and to the south and are aligned along northeast-striking faults. These intrusions appear to have been emplaced during stage 2 mineralization. They predated stage 3 veining, based on crosscutting relationships.

702

703

DIATREME BRECCIAS AT THE KELIAN GOLD MINE, KALIMANTAN, INDONESIA TABLE 3. Characteristics and Genesis of the Tepu and Runcing Breccias Dimensions, geometry, contacts

Facies

Facies distribution

Timing relationships

Interpretation

Carbonaceous mudstone and sandstone A1 A3-QP A3-QFP A4-QP A4-QFP A4-QP/QFP A5-V A5-And A6-QFP A7 A7-AL B3

Multiple crosscutting pipes, each containing single or multiple breccia facies A1 breccias confined to northern end, at the transition into the East Prampus fault A3 and A4 breccias occur as pipes up to 100 m diameter with sharp to diffuse contacts inside the main breccia body A3 breccias contain distinctive, wispy and dense clasts of either QP or QFP rhyolite A5: concentrated along the margins of the Tepu Breccia, locally grades into in-situ brecciated wall rocks, occurs marginal to, and grades into A3 and A4 pipes A6: discrete, single-facies pipes that crosscut all facies except A7 A7: thin (0.1 to 10 cm) sandstone dikes and irregular pods that commonly contain AL

Formation of the Tepu Breccia postdated andesite intrusion and was contemporaneous with rhyolite intrusion and stage 1 to early stage 2 mineralization and alteration Juvenile magmatic component changed from QP to QFP rhyolite with time A1: oldest facies A3-QFP: crosscuts A3-QP and A4-QP, and has mutual crosscutting relationships with A4-QP/QFP B3: crosscut by A3-QP/QFP A6-QFP: crosscuts all facies except A7 A7 and A7-AL dikes and narrow pipes crosscut all other facies Local clasts of A7-AL occur in A3-QP facies All polymict facies contain clasts of illite-carbonate-pyritealtered andesite All facies crosscut by stage 3 veins

Phreatomagmatic breccia body, with subordinate phreatic and tectonic breccias A1: tectonic breccias associated with the East Prampus fault zone; precursor to the polymict phreatomagmatic breccias QP and QFP clasts: juvenile magmatic component—wispy clasts and dense clasts with cuspate margins are analogous to magmatic clasts in dispersed peperite A5: abrasion and collapse of wall rocks into the pipe, possibly also implosion of wall rocks A6-QFP conglomerate: phreatic breccias, gas streaming elutriated fines and milled clasts A7 sandstone dikes: gas streaming, may be phreatic or phreatomagmatic B3 stratified breccias: turbulent flow inside the breccia body, or collapse of surficial material into evacuated portions of the breccia pipes

Carbonaceous mudstone and sandstone A1 A3-QFP A4-QP/QFP A6-V B1 B2 B3

Polyphase polymict breccia pipe surrounded by a shell of brecciated and in-situ carbonaceous mudstone and sandstone The pipe contains a mega-block (300 × 250 × > 75 m) of bedded breccia (B facies) Polyphase A4 breccia occupies the NE two-thirds of the Runcing Breccia and contain numerous clasts (1 cm to 5 m) of bedded B1 and B2 facies breccia and sandstone, clasts of AL-bearing sandstone (either A7 or B2) A1 breccia occupies the SE third of the Runcing Breccia and has been crosscut by dikes of polymict A3-QFP, A4-QP/QFP breccia and A6-V conglomerate A3-QFP: pods and dikes around the QFP intrusion, locally surrounded by A1 breccia Mega-block of B1, B2 and B3 breccia located on the southern side of the polymict pipe A6 conglomerate dikes up to 1 m wide

Onset of Runcing Breccia formation is inferred to have been during stage 1 mineralization Intrusion of the QFP plug occurred prior to stage 3 mineralization Collapse along ring faults occurred before and probably during stage 3 mineralization, but predated stage 4 A4 breccias contain clasts of older A4 breccia and B1 and B2 breccia Blocks of B facies breccia have been down-dropped along ring faults during A4 breccia formation A6 conglomerate dikes crosscut A1 breccias A1 breccias are crosscut by A3-QFP breccias A1 and A4 breccias are baked at their contacts with the QFP intrusion B1, B2 and B3 breccias in the mega block have been tilted by intrusion of the QFP plug A1 dikes crosscut B1 and B2 breccias

Phreatomagmatic, phreatic, tectonic breccias A1: tectonic brecciation which occurred prior to phreatomagmatic brecciation, and also during collapse of the Runcing Breccia pipe on ring faults A3-QFP: phreatomagmatic breccias formed by intrusion of QFP magma into wet A1 breccia A4-QP/QFP: multiple phreatomagmatic explosions B1 and B2: phreatomagmatic eruptions—these are wet pyroclastic base surge deposits; cycling between low and high water supply produced the B1 and B2 facies respectively   B3: collapse and resedimentation of unconsolidated B1 and B2 breccias

Tepu Breccia 750 × 250 × > 500 m Downward tapering Crude elliptical shape aligned along a NE oriented long axis East-trending arm extends across the Eastern Andesite into the 255 Breccia The Tepu Breccia is aligned along the East Prampus fault Confined by Eastern and Crusher andesite intrusions

Runcing Breccia 750 × 750 × > 500 m Roughly circular and downward tapering Margins defined by ring faults Polymict A facies breccias are enclosed by a shell of A1 breccia Southern margin defined by Runcing and Discovery faults Northern margin defined by NWstriking faults QFP plug (550 × 300 m) intruded parallel to the Burung fault

Abbreviations: And: andesite; AL: accretionary lapilli; QFP: quartz – feldspar – phyric rhyolite; QP: quartz – phyric rhyolite; V: volcanic sandstone /  mudstone 0361-0128/98/000/000-00 $6.00

703

704

DAVIES ET AL.

FIG. 12. Tepu Breccia geometry and contact relationships. A. Aerial view looking to the east of the Tepu Breccia (field of   view is ~500 m horizontal; pit wall in center of view is ~ 200 m high). This view illustrates the facies distribution in the Tepu Breccia and its position between the Crusher and Eastern Andesite intrusions. Note the color difference between the two andesite intrusions. The Eastern Andesite has been intensely quartz-illite-pyrite and illite-carbonate-pyrite altered, but the Crusher Andesite is only weakly illite-chlorite-carbonate altered. It is interpreted that the relatively impermeable Tepu Breccia inhibited fluid flow to the east. Also in this view is the arm of the Tepu Breccia that extends along the Tepu fault into the 255 Breccia. Blue lines highlight the margins of several discrete breccia pipes in the Tepu Breccia. White lines are nonstructural contacts and red lines are faults and fault contacts. Yellow dashed line (X-X’) indicates location of graphic section illustrated in Figure 13). B. Intrusive contact between carbonaceous mudstone and the Crusher Andesite (left) and breccia contact between mudstone and A4-QP and/or QFP breccia (right). The mudstone has a 1-m-wide baked zone adjacent to the intrusion. The mudstone is inferred to be the protolith to the Tepu Breccia and is only preserved in this location (1070 mRL). C. A5-V breccia developed on contact between A4-QP and/or QFP breccia and the Upper Cretaceous Kelian Volcanics (KFV). The contact between the two breccia facies is in part gradational but also shows pods of A4 breccia in the A5 breccia (1070 mRL). 0361-0128/98/000/000-00 $6.00

704

705

DIATREME BRECCIAS AT THE KELIAN GOLD MINE, KALIMANTAN, INDONESIA Log with grainsize (mm)

X

m

0.5 2

8 32 64

Log with grainsize (mm) m

0.5 2

8 32 64

Log with grainsize (mm) m

0.5 2

8 32 64

Approximately 5% wispy QP clasts More andesite and volcanic sandstone and mudstone clasts, subangular to subrounded Some andesite and volcanic clasts contain illite-pyrite veins that are truncated at clast margins

Crusher Andesite 10

110

20

120

A3-QFP

30

130

A3-QP 10-20%, <1 cm, wispy to angular QP clasts; 50-60%, subround to round carbonaceous sedimentary clasts; 10% volcanic sandstone and andesite clasts Carbonaceous mud matrix Several clasts of coal or carbonised wood observed at 150 m Many blocks and clasts of A7-AL facies, some clasts are finely and diffusely bedded Diffuse contact marks increase in wispy QP clast size and abundance 30% fine QP wispy clasts Absence of A7-AL clasts

210

A3-QP Polymict clast-supported breccia with strongly elongated to wispy QP clasts Wispy QP clasts are deformed around larger clasts of andesite, volcanic sandstone and

220

230

A7-AL Carbonaceous sandstone with accretionary lapilli

40

140

240

A7-AL Carbonaceous sandstone with abundant accretionary lapilli

1 m wide dike contains clasts of A3-QP, A3-QFP, and A6-QFP blocks in an A7-AL matrix

A6-QFP 50

60

A3-QFP 40-50% matrix of lithic sandstone 20% QFP rhyolite clasts QFP rhyolite clasts are subrounded to wispy, with smaller clasts more wispy-shaped Carbonaceous sedimentary clasts are sub-rounded to rounded Up to 20% QFP wispy rhyolite clasts as <1 cm fragments Bimodal clast distribution (<1.5 cm and >3 cm)

150

70

80

Decreasing QFP clasts increasing carbonaceous mudstone and sandstone Increasing abundance of matrix

250

160

A6-QFP Clast-supported, QFP rhyolite 7090% / carbonaceous mudstone 10-30% / 10% sandstone matrix Subrounded sedimentary clasts QFP clasts are subangular to subrounded Average clast size is 2 cm Cross cuts A7 and A6 facies

X’

A7-AL - dike

A7-AL - dike A3-QP 170

180

10-20%, 0.5-2 cm wispy QP clasts 10-20% volcanic sandstone and mudstone clasts 60% matrix and carbonaceous sedimentary clasts Increased abundance of carbonaceous sedimentary matrix and clasts More blocks of A7 facies More clasts of volcaniclastic rocks More clasts of andesite Less common wispy QP clasts (average 5%)

A7-AL dike 90

100

A3-QP finely quartz-phyric rhyolite clasts: (10-20% wispy-shaped QP clasts typically <1 cm; 5-10% blocky QP clasts, 1-10 cm) 50-60% round to subround carbonaceous sedimentary clasts, 0.2-4 cm diameter 10% Volcanic siltstone and sandstone clasts, <5 cm, subround to subangular 30-40% matrix

190

Local concentrations of wispy QP clasts (20-30%)

200

FIG. 13. Graphic section from mapping traverse X-X' through core of the Tepu Breccia at 1070 mRL. Refer to Figure 12 for section location.

0361-0128/98/000/000-00 $6.00

705

706

DAVIES ET AL.

FIG. 14. Schematic illustration of Runcing Breccia formation (plan view). A. Intrusion of QP rhyolite dikes along north west-striking faults and the Burung fault triggered phreatomagmatic explosions at the fault intersection. Polymict facies A breccias were emplaced into the carbonaceous mudstone wall rocks. B. After the main brecciation event(s) collapse of the breccia pipe and inferred overlying maar breccias occurred along ring faults. The megablock of B facies may have collapsed several hundred meters into the breccia pipe, based on depth estimates from fluid inclusion studies (Davies 2002, Davies et al., 2008). The ring faults nucleated on preexisting northeast- and northwest-striking faults. This resulted in downdropping of a segment of the Burung fault along the Runcing and Discovery faults. This was a period of major subhorizontal extension at Kelian (Davies, 2002) so faults may have been under extension during collapse. C. The Runcing QFP intrusion was emplaced along the projection of the Burung fault. Minor phreatic and phreatomagmatic explosions took place during this period. Doming of the bedded breccias occurred during QFP emplacement.

0361-0128/98/000/000-00 $6.00

706

707

DIATREME BRECCIAS AT THE KELIAN GOLD MINE, KALIMANTAN, INDONESIA TABLE 4. Characteristics and Genesis of the Burung Breccia Dimensions, geometry, contacts

Facies

Facies distribution

Timing relationships

Interpretation

Carbonaceous sandstone and mudstone A1 A4-QP A4-QP/QFP A5-V A5-And A6-QFP

The polymict breccia pipe is surrounded at shallow levels on the west by monomict carbonaceous breccias and in situ mudstone The Burung Breccia is dominated by an A4-QP/QFP breccia pipe A5 breccias occur parallel to the western Burung Breccia margin, in the footwall to the North Burung fault In situ and clast-rotated A1 breccias occur on the west side of the Burung Breccia between the A4-QP/QFP polymict breccia pipe and the North Burung fault A1 breccias grade laterally into unbrecciated carbonaceous mudstone A6-QFP conglomerate containing faceted clasts of pyritized QFP occurs in a narrow pipe (<10 m) at deeper levels

Formation of the Burung Breccia commenced during stage 1 and continued into early stage 2 A1: partly crosscut by A4 or grade laterally into A4 breccias Rare A3-QP breccia clasts occur in the A3-QP-QFP breccia A6-QFP crosscuts all other facies QFP and QP rhyolite plugs and dikes have intruded the Burung Breccia Minor clasts of hydrothermal breccias, which have been cemented by stage 2 cements, occur in the Burung Breccia, indicating partial overlap with hydrothermal breccia formation Contacts with the enclosing hydrothermal breccias of the 393 Breccia are diffuse and have been overprinted by stage 2A quartz-pyrite Polymict facies of the Burung Breccia are intensely quartzillite-pyrite altered and moderately carbonate altered Stage 2A, 3A, B, C, and 4 veins crosscut the Burung Breccia

A1 breccias formed by tectonic brecciation along the North Burung fault and by implosion around the margins of the phreatomagmatic breccia pipe A4 QP and A4 QP-QFP breccias are phreatomagmatic, formed by intrusion of QP and QFP magmas into wet, A1 facies carbonaceous breccia and wet faults in mudstone QP and QFP clasts are the juvenile magmatic component in A4 breccias, but cuspate and wispy clast margins have been milled A5 breccias formed by tectonic brecciation on the North Burung fault, progressive brecciation of wall rocks adjacent to A4 facies breccias, including implosion and collapse, and during phreatic brecciation associated with either the Burung or 393 Breccia complexes A6-QFP phreatic breccias formed during the late stages of QFP intrusion into the Burung Breccia Phreatic brecciation and implosion of enclosing wall rocks produced the hydrothermal breccias that were cemented during stage 2A—these are the earliest facies of the 393 Breccia

Burung Breccia 275 × 175 × >450 m N- to NE-trending long axis, with a crude rectangular shape in plan view Tapers downward Becomes elongate  with depth along NE axis  Western margin defined by North Burung fault Southeast margin confined by the Eastern Andesite below ~1000 mRL Above 1000 mRL, the Burung Breccia flares to the north The Burung Breccia is enclosed by a conical shell of hydrothermally cemented breccias that comprise the 393 Breccia A late-stage, 90 × 40 m QFP plug has intruded into the Burung Breccia

Abbreviations: And = andesite, QFP = quartz-feldspar-phyric rhyolite, QP = quartz-phyric rhyolite, V = volcanic sandstone and/or mudstone

Facies Interpretations Several origins for the carbonaceous matrix-rich breccias at Kelian have been proposed by previous workers. They  have been interpreted as lahars, volcanic breccias, and intrusive breccias (Ferguson, 1986), and as hydrothermal breccias produced by late-stage intrusion of magma into the hydrothermal system (Lawless, 1988). Van Leeuwen et al. (1990) proposed that the muddy breccia was emplaced during the main hydrothermal event by fluidization and diapiric emplacement of the mud component of the mudstone and/or sandstone sequence. They considered fluidization to have been the result of either tectonic forces, increased heat from the hydrothermal system, or depressurization resulting from hydrothermal eruptions. R.H. Sillitoe (1993, unpub. report for PT Kelian Equatorial Mining) reinterpreted the muddy  breccias and associated felsic intrusions to be components of  a maar-diatreme complex. Large blocks of bedded carbonaceous matrix-rich breccias with low-angle cross stratification and accretionary lapilli in the Gunung Runcing area were interpreted to be pyroclastic surge deposits and to indicate 0361-0128/98/000/000-00 $6.00

a prior connection to the paleosurface. He also suggested that clasts of aphanitic felsic rock in the Runcing diatreme  were fragments of the igneous intrusion that generated the diatreme. The carbonaceous matrix-rich breccia bodies at Kelian share many features with breccia pipes that have been interpreted as diatremes (e.g., Montana Tunnels: Sillitoe et al., 1985; Cripple Creek: Thompson et al., 1985; Thompson, 1992; Wau: Sillitoe et al., 1984; Balatoc: Cooke and Bloom, 1990). Diatremes are inferred to form by phreatomagmatic plus magmatic and/or phreatic brecciation processes (Lorenz, 1973, 1986; Lorenz and Kurszlaukis, 2007); identification of  the juvenile magmatic component in the phreatomagmatic deposits is essential for this interpretation. The Kelian carbonaceous matrix-rich breccia bodies have large vertical extents, taper downward, contain abundant matrix clasts derived from the adjacent wall rocks, distinctive juvenile magmatic clast components (QP, QFP), and large blocks of surface-derived stratified breccia that contain low-angle dune forms and accretionary lapilli. In isolation, many of the breccia facies at Kelian are not diagnostic

707

708

DAVIES ET AL.

FIG. 15. Diamond drill hole K422 graphic log illustrating relationships between the Burung Breccia and enclosing hydrothermal breccias of the 393 Breccia. The 393 Breccia consists of multiple hydrothermal breccia phases. Its earliest phase, which is cemented by stage 2A quartz and pyrite, formed during the late stages of phreatomagmatic brecciation in the Burung Breccia. 0361-0128/98/000/000-00 $6.00

708

DIATREME BRECCIAS AT THE KELIAN GOLD MINE, KALIMANTAN, INDONESIA

of any single brecciation process. However, in combination  with facies distribution, facies associations, breccia body  geometry, and wall-rock relationships, the origins of the A and B facies can be determined. Table 5 summarizes key  textures and spatial associations used to interpret the origins of individual carbonaceous matrix-rich breccia facies at Kelian, based on the classification scheme of Sillitoe (1985).

709

 A1 breccias A1 breccias are interpreted to have formed by a combination of tectonic and phreatic and/or hydraulic processes (Table 5). The tectonic interpretation is based on their discordant nature, spatial association with faults, local occurrence of  clast imbrication, and presence of shear fabrics and gouge, and tectonic fragmentation was most likely triggered by seismic

TABLE 5. Summary of Breccia Facies Interpretations Facies

Interpretation(s)

Key features

A1

Tectonic breccia

Discordant; spatial association with faults; planar A1 breccia zones surrounded by in-situ carbonaceous mudstone; shear fabrics and fault gouge in the A1 breccias Dikes of A1 breccia have crosscut other breccias; broad areas of fragmentation not aligned along fault trends or tabular in shape; spatial association with margins of QFP intrusion in Runcing Breccia; occur at the margins of A4 breccia pipes against in-situ brecciated wall rocks—hydraulic implosion suspected; occur at contacts between hydrothermal breccias and mudstone units

Phreatic and/or hydraulic breccia

A2

Phreatomagmatic breccia and/or peperite

Discordant; jigsaw-fit textures and sedimentary matrix; facies association of QP or QFP rhyolite dikes, A2-QP, or -QFP breccia and A3-QP or -QFP breccia

A3

Phreatomagmatic breccia and/or peperite

Discordant; distinctive QP and QFP clast morphology (wispy or dense and cuspate) is inconsistent with rounding of other clasts in A3 breccia; abundance of QP and QFP clasts not consistent with QP and QFP abundance in wall rocks; consistent phenocryst abundances in felsic clasts and clast textures across individual A3 breccia pipes suggests common and contemporaneous origin for the distinct clasts; interpretation of wispy and c uspate clasts as juvenile magmatic component; high degree of fragmentation and abundance of wall-rock fragments consistent with phreatomagmatic breccias; facies associations

A4

Phreatomagmatic breccia ± phreatic reworking

Presence of andesite and exotic rhyolite clasts similar to the intrusions at Kelian; alteration assemblages and veins in the andesite clasts eliminated these as juvenile components; facies association with A3 breccias and rhyolite dike-A2-A3 facies association lead to interpretation of QP and QFP clasts as juvenile component; high degree of fragmentation and abundance of  wall-rock fragments consistent with phreatomagmatic breccias

A5

Phreatic breccias, indirectly phreatomagmatic, and/or hydraulic (implosion) breccias

Discordant; location at margins of phreatomagmatic breccia pipes and parallel to wall-rock contacts Gradational contacts with both the phreatomagmatic breccias and wall rocks—brecciation may have been by collapse or implosion Facies association with A3 and A4 breccia—phreatomagmatic, phreatic, and hydraulic implosion Facies association with hydrothermal breccias adjacent to the Burung Breccia—phreatic and hydraulic implosion As above, but with addition of fault gouge or sheared fabrics Location along projection of faults at margins of breccia bodies

Tectonic breccia A6

Phreatic or intrusion-related phreatic

Abundance of one clast type—QFP or QP rhyolite; spatial association with late-stage QP and QFP intrusions; faceted clast shapes and lack of matrix; discordant (pipe and dike) geometries

A7

Phreatomagmatic or phreatic breccia and sandstone

High degree of fragmentation; accretionary lapilli indicate wet gas-rich transport; plane and ripple laminations parallel to dike walls; same composition as A4 breccias, but finer grain size; discordant geometry  Not possible to distinguish definitively between phreatic and phreatomagmatic processes

B1

Phreatomagmatic base surge deposit (low water/magma ratio)

Low-angle dune bed forms, diffuse grading Cuspate clast margins, relict perlite, and interpretation of QP juvenile clast component Abundance of juvenile clasts and relatively coarse grain size consistent with low water/magma ratio Facies association with B2 facies, A4 breccia pipes, and QFP intrusions

B2

Phreatomagmatic base surge and cosurge fallout deposits (high  water/magma ratio)

Low-angle dune bed forms—consistent with base surge Graded accretionary lapilli beds with constant bed thickness—consistent with cosurge fallout Abundance of wall-rock clasts, accretionary lapilli, soft-sediment deformation structures, and relatively fine grain size—consistent with low water/magma ratio Cuspate clast margins, relict perlite, and i nterpretation of QP juvenile clast component Facies association with B1 facies, A4 breccia pipes, and QFP intrusions

B3

Phreatomagmatic

Diffuse bed forms, compositional and textural similarities to A4 breccias Internal position of blocks within Tepu Breccia; facies associations with deep A4 and A3 breccias Massive to diffuse bedding; channel bed forms that crosscut B1 and B2 breccias; lack of high-velocity bed forms Facies association with B1 and B2 facies; equivalent components to B1 and B2 breccias but more rounding

Resedimented base surge deposits (phreatomagmatic)

Abbreviations: QFP = quartz-feldspar-phyric rhyolite, QP = quartz-phyric rhyolite 0361-0128/98/000/000-00 $6.00

709

710

DAVIES ET AL.

ruptures. In the case of the Runcing Breccia, A1 breccias are spatially associated with the margin of a QFP intrusion, where phreatic brecciation is inferred to have occurred due to boiling of pore waters (Davies, 2002). Adjacent to the QFP intrusion, anastamosing A1 dikes have crosscut the bedded B facies megablock and can be traced back to the brecciated mudstone surrounding the intrusion. The irregular and penetrative contacts of the A1 breccia dikes suggest that the components of  the A1 breccias were forcefully injected into the polymict carbonaceous matrix-rich breccia. The observed textures are attributed to fluidization of sediment due to the intrusion of felsic magma and boiling of pore water, in a manner akin to that described by Hanson and Wilson (1993).

explosive and nonexplosive fragmentation. Although gradations between A3 breccia, A2 breccia, and coherent rhyolite have been observed locally, typically the A3 breccias have no identifiable magmatic roots. It is interpreted that rhyolite disaggregation into wispy and blocky clasts generally resulted in fragmentation of the upper parts of the intruding magma batch into the unconsolidated, wet, subsurface breccia. Explosive magma-wet sediment interaction was sufficiently energetic to disperse the clasts away from the coherent rhyolite feeder into the overlying breccia. The A2 jigsaw-fit breccias (e.g., Fig. 4B) are interpreted to be the preserved roots of the A3 breccias and their rare occurrence most likely relates to destruction of the magmatic root zone by explosive fragmentation at the current level of exposure. The predominance of subrounded clasts in the A3 breccias (Fig. 5D-E) suggests that they resulted from greater clast transport and abrasion than the juvenile rhyolite clasts (cf. Roache et al., 2000), perhaps due to clast recycling during multiple brecciation events (e.g., Houghton and Smith, 1993) with introduction of new juvenile clasts in each event. Preservation of the wispy and blocky juvenile magmatic clasts in A3 breccias (Fig. 5B, D, F) is interpreted to indicate only minor clast transport for these clasts. In the A4 breccias, compositionally and texturally equivalent juvenile magmatic components occur; however, they lack the wispy or cuspate morphology of those in the A3 facies. A4 breccias are interpreted to be the product of progressive transport and abrasion of the A3 breccia components, either in single brecciation events or during repeated brecciation cycles (either phreatic or phreatomagmatic; Houghton and Smith, 1993). Lithification is not likely to have occurred between these cycles, as clasts of A3 or A4 breccia are rarely observed in A3 or A4 breccia. Instead, the unconsolidated breccia deposits are inferred to have been disaggregated and reworked during subsequent events. Some A2 and A3 breccias are interpreted here to be varieties of peperite (e.g., Busby-Spera and White, 1987; Hanson and Wilson, 1993), produced by intrusion of rhyolite magma into (1) wet, unconsolidated fault breccia and gouge, (2) wet A1 breccia, (3) wet carbonaceous mudstone, or (4) wet A4 breccia. Blocky and angular clast domains in A2 breccia (Fig. 4B) are similar to textures described by Hanson and Wilson (1993) as closely packed peperite and by Busby-Spera and  White (1987) and Hanson and Hargrove (1999) as blocky  peperite. The A3 breccias are matrix supported and lack the  jigsaw-fit textures of the A2 breccias, indicating that disaggregation of the rhyolite magma was more advanced in the A3 breccias. Some A3 breccias have similar internal organization to the dispersed peperite of Hanson and Wilson (1993). A2 and A3 breccias containing wispy rhyolite clasts (e.g., Figs. 4C, 5B, D) are interpreted to be varieties of fluidal peperite (cf. Busby-Spera and White, 1987; Hanson and Hargrove, 1999). Roache et al. (2000) described similar wispy and blocky rhyolite clasts set in a matrix of polymict breccia and inferred that the wispy clast morphology was produced by  squashing of plastic rhyolite clasts by the relatively coarse grains of the host breccia.

 A2, A3, and A4 breccias Breccia facies A2, A3, and A4 are interpreted to have formed primarily by subsurface explosive phreatomagmatic brecciation (Table 5). This process of direct magma-water interaction incorporates a juvenile magmatic component into the resulting deposits (Sheridan and Wohletz, 1981). Rhyolite clasts (QP or QFP) in the carbonaceous matrix-rich breccias at Kelian (e.g., Figs. 4B-C; 5B, D) were first interpreted to be  juvenile magmatic clasts by R.H. Sillitoe (1993, unpub. report for PT Kelian Equatorial Mining). Our observations support his findings. The QP and QFP rhyolite clasts are texturally and compositionally (based on modal mineralogy) equivalent to the QP and QFP rhyolite intrusions. The spatial association and, in some locations, gradational transitions from in situ rhyolite to A2 and then A3 breccias indicate that there is a direct link between magma intrusion and generation of QP and QFP rhyolite clasts. This type of evidence is seldom preserved in breccia pipes interpreted to have phreatomagmatic origins (e.g., Montana Tunnels, Montana: Sillitoe et al., 1985; stage 3 fragmental porphyry, Black Cloud mine, Colorado: Hazlitt and Thompson, 1990). More commonly, an igneous-clast component is inferred to be juvenile (i.e., fragments of the parental magma, such as dacite porphyry clasts at Wau: Sillitoe et al., 1984). There is a problem with the occurrence of subsurface phreatomagmatic explosions (e.g., Wohletz, 1986): how does magma mix with water efficiently enough to drive explosive “fuel-coolant” interactions? At Kelian, the A2 breccias occur  where rhyolite dikes have intruded (1) preexisting fault zones consisting of monomict and polymict fault breccia and fault gouge, (2) A1 breccia, and (3) A4 breccias. In all of these sites, the host rock is interpreted to have been unconsolidated or poorly consolidated, because sharp intrusive contacts are rare, whereas irregular interfingering and contorted contacts are common. We propose that rhyolitic magmas were emplaced into poorly consolidated or unconsolidated, earlyformed breccias that had undergone preintrusion disaggregation either by tectonic, phreatic, or earlier phreatomagmatic brecciation events, which facilitated efficient mixing of the fuel and coolant (Sheridan and Wohletz, 1981) to result in subsurface phreatomagmatic fragmentation. Interaction between the intruding rhyolite and wet, unconsolidated, or poorly consolidated sediment could have re-  A5 breccias sulted in purely explosive phreatomagmatic fragmentation,  Where A5 breccias occur as marginal phases to A3 and A4 nonexplosive quench fragmentation, or a combination of both rhyolite-bearing breccias, they are interpreted to have formed 0361-0128/98/000/000-00 $6.00

710

DIATREME BRECCIAS AT THE KELIAN GOLD MINE, KALIMANTAN, INDONESIA

711

during the phreatomagmatic explosions that generated the A3 B facies and A4 breccias by a combination of (1) wall-rock collapse B1 and B2 facies contain the same juvenile magmatic cominto open space at the margins of A3 and A4 breccia pipes, (2) ponent as the A3-QP subfacies (QP rhyolite). Planar, dune, fragmentation of wall rocks by mechanical abrasion, and/or and minor low-angle cross beds are consistent with deposition (3) hydraulic implosion or phreatic brecciation in response to from turbulent, gas-rich flows at the transition from low- to pressure gradients between the wall rocks and evacuated high-flow regimes (Walker, 1984; Valentine and Fisher, 2000; pipes (Table 5). In the Burung Breccia, the A5 breccias are Table 5). Accretionary lapilli (Fig. 10D) and soft-sediment or not spatially associated with rhyolite intrusions and are, there- plastic deformation structures (Fig. 10E) are consistent with fore, interpreted to have formed by explosive phreatic frag-  wet surge deposits. B1 and B2 are interpreted to have been mentation and tectonic fragmentation on late-stage normal deposited at the surface by a combination of wet, pyroclastic faults. The occurrence of clasts of hydrothermally cemented base-surge fallout and cosurge fallout, generated by explosive breccias in the A5 breccias, and vice versa, indicates that for- phreatomagmatic eruptions (Fisher and Waters, 1970; mation of some A5 breccias occurred during the post-  Walker, 1984). Cosurge fallout is an important process for dephreatomagmatic phase of breccia development. positing fine (sand- to silt and/or clay-sized) grains and abundant accretionary lapilli in uniformly thick beds up to a few   A6 conglomerates and breccias centimeters thick (Walker, 1984). Accretionary lapilli-rich B1 The A6 facies is distinct from all other A facies as it is dom- and B2 beds probably formed as fallout of fine ash after pasinated by subrounded to rounded clasts and has only minor sage of the turbulent wet surge. Deposition is inferred to have matrix. The high degree of clast rounding is inconsistent with occurred in a wet, terrestrial environment based on the presin situ fragmentation and indicates significant clast abrasion. ence of interbedded carbonaceous mudstone (with wood The paucity of matrix suggests that the sand- to mud-sized fragments) in the sandstones and breccias. B facies blocks component was removed (winnowed?) during breccia forma- then slumped into the diatremes, including the megablock in tion. Localization of A6 breccias above a QFP intrusion in the the Runcing Breccia. Burung Breccia and the dominance of rhyolite clasts imply  B3 breccias are interpreted to have formed by syneruptive that the rhyolite intrusions triggered localized phreatic explo- resedimentation of the B1 and B2 facies or collapse of the insions, focused in and above the intrusions (Table 5). Gas ferred maar deposits into the maar crater (Table 5). Blocks streaming resulted in milling of the clasts and removal of the containing B3 beds in the Tepu Breccia are interpreted to fine matrix component. These breccias may be equivalent to have been deposited in the breccia pipe by slumping of mathe intrusion-related phreatic breccias described by Sillitoe terial back down the evacuated conduit. (1985), which have been described elsewhere as pebble dikes (e.g., Gustafson and Hunt, 1975). Genetic Model for Diatreme Formation at Kelian  A7 sandstones The facies of the Tepu, Burung, and Runcing Breccias and Based on their contact relationships and geometry (Table the surrounding sedimentary rocks and cogenetic intrusive 2), the A7 facies discordant sandstone and accretionary lapilli- rocks record a complex sequence of tectonic brecciation and bearing sandstone formed in the subsurface. Since the com- phreatic and phreatomagmatic brecciation and eruption. ponents and textures of the A7 facies are identical to those of  Collectively, these breccias are interpreted to be the prodthe A3 and A4 breccias, their origin is inferred to be equiva- ucts of maar-diatreme volcanism and related tectonism and lent (Table 5). A7 sandstone is considered to have formed hydrothermal activity. The following section outlines a seduring high-water/magma ratio explosions and/or as a result quence of events that could have produced these breccias of highly efficient fuel-coolant mixing. This resulted in the and highlights the key evidence for each phase of volcanism anomalously intense fragmentation of wall rock and magma. and brecciation.  Water supply and magma-sediment mixing may have become Ground preparation: Eocene to Oligocene carbonaceous greater with time due to the increased abundance of the sedimentary strata were juxtaposed against Upper Cretaceous sand- to mud-sized component in the pipes after successive  volcaniclastic rocks by normal and strike-slip faults prior to or explosion events. The occurrence of abundant accretionary  during the Miocene (Fig. 16A). lapilli is consistent with deposition from wet, gas-rich particle  Andesite intrusions: Intrusion of late Miocene andesitic systems (Walker, 1984). Accretionary lapilli are sometimes stocks was controlled by the same faults that were active durused to indicate formation in a subaerial environment; how- ing ground preparation (Fig. 16B). Early phreatic breccias ever, they have been identified in subvolcanic breccias at probably formed at the andesite margins at this time, and Mount Leyshon (Wormald, 1991), in gas segregation pipes in some A1 breccias formed in the carbonaceous sedimentary  the Oruanui Ignimbrite in New Zealand (Self, 1983), and in rocks by phreatic, tectonic, or hydraulic processes. Evidence dikes cutting breccias at the Rain mine (Williams et al., 2000), for a prediatreme hydrothermal system is preserved in later Cripple Creek (T. Thompson, pers. commun., 2003), and formed carbonaceous matrix-rich breccias as altered andesitic Lihir (Carman, 1994, 2003). It is possible that the accre- and volcaniclastic rock clasts that contain truncated stage 1 tionary lapilli at Kelian formed in a surficial setting and were  veins (e.g. Fig. 6B). then reworked into the subsurface environment, but we preRhyolite intrusions and commencement of volcanism: Finefer the interpretation of a subsurface environment of forma- grained QP intrusions were emplaced along faults into the tion, based on the lack of broken accretionary lapilli in the A7 fault-bounded carbonaceous sedimentary blocks and domains dikes and pipes. of A1 breccia at 19.8 ± 0.1 Ma (Davies, 2002; Fig. 16C). The 0361-0128/98/000/000-00 $6.00

711

712

DAVIES ET AL.

main phreatomagmatic phase was triggered by intrusion of  QP into the already-active hydrothermal system. It is not known if the initial explosions were phreatic, in advance of  the ascending magma, or phreatomagmatic. The subsurface products of phreatomagmatic explosions at Kelian were the discordant, monomict to polymict facies A breccias and sandstones. The A4 breccias probably underwent greater degrees

of transport and abrasion or recycling (e.g., Houghton and Smith, 1993) than the A3 breccias because the latter have preserved the delicate-textured juvenile magmatic clasts. At the surface, eruptions contemporaneous with the subsurface explosions produced base- and cosurge fallout deposits. B1 and B2 facies in the Runcing Breccia are interpreted to be  wet base- and cosurge fallout deposits on the basis of their

FIG. 16. Schematic model for the formation of the Kelian diatremes, showing their relationships to enclosing units, the structural framework, and present-day erosion level. A. Ground preparation: downfaulting juxtaposed Eocene-Oligocene carbonaceous sedimentary rocks against Upper Cretaceous volcaniclastic rocks (Kelian Volcanics). Preserved remnants of  fault-bounded blocks have been sheared and tectonically brecciated. B. Intrusion of Miocene andesitic stocks. Andesitic intrusions were localized by same fault system as in (A). The early hydrothermal system at Kelian is interpreted to have commenced during or after emplacement of these intrusions and before formation of the carbonaceous matrix-rich breccias. Ingress of ground water deep into the carbonaceous sedimentary rocks may have occurred due to andesite-related brecciation, along the bounding faults for the sedimentary units. Arrow and bubbles indicate possible locations of phreatic explosions at this time. C. Quartz-phyric rhyolite dikes intruded along the same fault system as previous events, encountering ground water in the fault zone and brecciated sedimentary rocks. Phreatomagmatic explosions produced massive, unstratified polymict carbonaceous matrix-rich facies A breccias in the subsurface (thick arrows) and wet pyroclastic base-surges and fallout eruptions (facies B1 and B2) at the surface (thin arrows). Synvolcanic resedimentation of volcaniclastic deposits into maar (dashed arrows) also occurred (facies B3). D. Continued phreatomagmatic explosions and eruptions excavated the conduit, widening the diatreme, increasing its depth, and producing multiple crosscutting breccias. Excavation of the diatreme accompanied by collapse of the maar and upper conduit along ring faults and preexisting structures. Blocks of wall rocks, earlier diatreme phases and bedded pyroclastic deposits slumped several hundred meters into the conduit, based on fluid inclusion estimates of the depth of hydrothermal activity (van Leeuwen et al., 1990; Davies, 2002; Davies et al., 2003). These megablocks of facies B breccias have been preserved locally (e.g., Runcing bedded facies) but most were probably disaggregated and incorporated into later diatreme breccias. E) Late-stage intrusion of QFP plugs into the Burung and Runcing Breccias accompanied by minor phreatomagmatic and widespread phreatic eruptions. Maar sediments were fluidized and pumped through older breccia phases along contemporaneous faults. Collapse of the maar and diatreme along ring faults and other structures continued. F. Miocene mineralization (hydrothermal brecciation and veining) occurred in the wall rocks around the Kelian diatreme breccias. The ore deposit was then uplifted and exhumed. Mafic lava and volcaniclastic deposits overlie a Plio-Pleistocene paleosurface. 0361-0128/98/000/000-00 $6.00

712

DIATREME BRECCIAS AT THE KELIAN GOLD MINE, KALIMANTAN, INDONESIA

713

FIG. 16. (Cont.)

bed forms and occurrence of cuspate QP fragments and the presence of accretionary lapilli. Diatreme excavation: Ground-water infiltration to greater depths would have been facilitated by excavation of the conduit and relatively permeable breccia deposit infilling (Fig. 16D). Phreatomagmatic eruptions thus initiated at progressively deeper levels under hydrostatic conditions. Gravitational collapse of the walls is an integral process in the progressive widening of maars and diatremes (Self et al., 1980; Sillitoe et al., 1984; Lorenz, 1986). At Kelian, collapse occurred by spalling of wall rocks into the evacuated pipe and by subsidence of the crater walls or crater-fill deposits via normal movements on ring faults (cf. Sillitoe et al., 1984; Lorenz, 1986) during and after eruptions. Syn- and posteruptive collapse of the eruptive breccia facies and collapse and resedimentation of volcaniclastic deposits into the maar probably  filled the diatreme between eruptive events and contributed to recycling of clasts in subsequent eruptions. Blocks of B facies slumped back into the pipes at this time, including a megablock of bedded breccia in the Runcing Breccia. Subsidence and phreatic explosions: Continued subsidence occurred during intrusion of late-stage QFP domes into the Runcing and Burung Breccias at 19.5 ± 0.1 Ma (Davies, 2002; Fig. 16E). Ragged QFP clasts surrounded by brecciated black, carbonaceous mudstone have been observed within 100 m of the QFP contacts and indicate that minor phreatomagmatic eruptions may have occurred during intrusion. Phreatic eruptions during this phase of magmatism resulted in formation of the A6 conglomerate pipes and fluidization of brecciated carbonaceous sediments. These fluidized monomict breccias (A1 facies) crosscut all polymict facies in the Runcing Breccia, including the bedded polymict facies. Postdiatreme evolution: Main-stage gold mineralization was associated with widespread hydrothermal brecciation and 0361-0128/98/000/000-00 $6.00

 veining. Mineralization at Kelian mostly postdated diatreme formation because the auriferous pyrite-base metal ± carbonate veins crosscut diatreme breccias, and no clasts of base metal ± carbonate mineralization have been identified in the diatreme breccias (van Leeuwen et al., 1990; Davies, 2002). Uplift and erosion produced a Pliocene-Pleistocene surface recorded by a paleosol over the Runcing Breccia. This erosional surface is overlain by mafic lavas and associated resedimented volcaniclastic and mafic pyroclastic rocks. Erosion by  the Kelian River and its tributaries exposed the upper levels of the diatreme complex in the Runcing area and deeper levels in the Tepu and Burung areas (Fig. 16F). Implications for models of diatreme volcanism In the traditional models for diatreme formation (e.g., Lorenz, 1973, 1986, Lorenz and Kurszlaukis, 2007), phreatomagmatic and dry magmatic eruptions are triggered when magma ascends to shallow depths (constrained by the available heat energy and hydrostatic head) and explosively interacts with shallow ground waters (e.g., Fig. 17A). Interaction between magma and superheated liquid water is the most productive mechanism for generating phreatomagmatic explosions due to the large volume changes that occur on vaporization (White and Houghton, 2000). As the explosions continue, a conduit is excavated progressively and the water table is depressed around the evacuated cone. Drying out of  the aquifer can cause cycling between phreatomagmatic and dry magmatic explosions and can result in surge deposits intercalated with ballistic fallout layers (Cas and Wright, 1987). At Kelian, felsic magmas are inferred to have intruded an active hydrothermal system, rather than a cold ground-water aquifer. The Kelian geothermal system was probably at or near its boiling point, or may even have been overpressured, during initial rhyolite intrusion. We infer that the addition of  magmatic heat into the geothermal system triggered hybrid

713

714

DAVIES ET AL.

A

‘wet’ pyroclastic eruptions

0m

Water table depressed Increasing eruption initiation depth with time

> 2500 m

No direct link to mineralization

B Magma intrusion triggers hybrid phreatomagmatic and phreatic explosions

250°C 300°C

Abundant hot fluids in active hydrothermal system, at or near critical point

C Large scale hydrothermal explosions and brecciation

Catastrophic disruption of chemical and physical conditions in hydrothermal system

Phreatomagmatic explosions through active system trigger syn- and post-diatreme hybrid phreatic explosions

    g       n      i     x      i     m      d      i      u       l      F

250°C 300°C Mineralization in wall rocks Breccia pipe inhibits fluid flow - hydrothermal system enhanced in wall rocks

D

 

Hydrothermal  system  advance 

    m      0       0       5       2

Structurally controlled mineralization at margins of diatreme

Catastrophic volatile loss/pressure reduction Hydrothermal  system collapse 

(i) Magma intrusion into early hydrothermal system at lowwater rock ratios, possibility of fluid contributions from hydrous magma and magmatichydrothermal brecciation

0361-0128/98/000/000-00 $6.00

(ii) Late magma intrusion into active high-T (300-350°C) hydrothermal system causes sub-surface phreatomagmatic explosions, generates huge volumes of steam and results in maar - diatreme volcanism

phreatomagmatic and phreatic explosions at depth (Fig. 17BD). Magma intrusion catastrophically disrupted the hydrology, P-T conditions, and thermal and chemical gradients of  the geothermal system. We believe that the initial sites of  phreatomagmatic explosions were deep in the hydrothermal system, rather than shallow as suggested for diatremes related to cold ground-water aquifers (e.g., Lorenz, 1973). This interpretation is based on the observations of root zone feeder dikes, truncated veins within altered clasts in the breccias (e.g., Fig. 6B), and paleodepth estimates from our fluid inclusion studies (Davies et al., 2008). Drying out of the hydrothermal system may have been difficult to achieve at these depths (estimated to be several hundred meters on fluid inclusion evidence: Davies et al., 2008), resulting in a predominance of wet eruptive products derived from phreatomagmatic and phreatic, rather than magmatic explosions. We infer that the onset of widespread phreatic brecciation at Kelian was contemporaneous with the waning phases of  phreatomagmatic brecciation, as shown by observed paragenetic relationships between veins, facies A, and early hydrothermal breccias (e.g., Tables 3, 4). This situation is comparable to the effects documented when a basaltic dike intruded the active Rotomahana-Waimangu geothermal system in New Zealand during the early 1900s. An initial phreatomagmatic eruption was followed by several years of  phreatic explosive activity and then a return to boiling hot spring conditions (Nairn, 1979; Simmons et al., 1993). Kelian may contain the subsurface products of this type of transition from phreatomagmatic to phreatic conditions. Implications for fluid flow and mineralization Although the diatreme breccias formed prior to the main stages of hydrothermal mineralization and alteration at Kelian, they are only weakly mineralized and altered. Instead, strongly mineralized hydrothermal breccias and veins formed peripheral to the diatreme breccias in the surrounding wall rocks (Davies, 2002; Davies et al., 2008). It appears that the carbonaceous matrix-rich breccias at Kelian acted as aquicludes during main-stage hydrothermal activity, causing fluids to be focused into the wall rocks outside of the main diatreme breccia pipes. The relative impermeability of the diatreme breccias with respect to the adjacent Cretaceous volcaniclastic rocks and Miocene andesite intrusions appears to have been caused by the widespread distribution of muddy  carbonaceous matrix through the breccia bodies. This situation contrasts markedly with other diatreme-related hydrothermal systems, such as the Cripple Creek deposit in Colorado, where mineralization and alteration was restricted almost exclusively to the diatreme breccia pipes, rather than the adjacent Precambrian granites that form the wall rocks to that deposit (Thompson et al., 1985; Thompson, 1992; Jensen, 2003). There are also cases where mineralization and alteration occur both in diatreme breccias and adjacent wall

FIG. 17. Evolution of phreatomagmatic breccia pipes and hydrothermal systems as a result of magma intrusion into: (A) cold meteoric water, (B) and (C) shallow (<1–2 km) geothermal waters, and (D) deep (>2 km) magmatichydrothermal fluids.

714

DIATREME BRECCIAS AT THE KELIAN GOLD MINE, KALIMANTAN, INDONESIA

rocks (e.g., Acupan: Cooke et al., 1996; Yanacocha: Turner, 1997; Colquijirca: Bendezú et al., 2003; Martabe: Sutopo et al., 2003, 2007). It is clear that hydrothermal activity in and around diatremes is strongly controlled by the local facies architecture, fracture network, and composition of the individual breccia bodies, and exploration for mineral resources near these large volcanic structures needs to take into account such variability.

715

Stuart Masters are thanked for the help that they provided at various stages of the project. We thank Jim Mortensen for dating the Kelian Volcanics. We thank Rodney Allen and Ray Cas for their detailed reviews, which have helped to improve this manuscript significantly, and Pete Hollings for useful advice on formatting issues. The senior author was the recipient of an Overseas Postgraduate Research Scholarship (OPRS) from the University of Tasmania, which is gratefully acknowledged. This research was also supported by CODES, the Australian Research Council’s Centre of  Excellence in Ore Deposits.

Conclusions As inferred by previous workers (Lawless, 1988; R.H. Sillitoe, 1993, unpub. report for PT Kelian Equatorial Mining; Sillitoe, 1994; van Leeuwen, 1994), the carbonaceous matrix- March 15, 2007; June 11, 2008 rich breccias at Kelian are in part phreatomagmatic and REFERENCES formed in a maar-diatreme complex. However, the origins of  A. Martin, U., and Németh, K, 2007, The Fekete-Hegy (Balaton Highthe breccias are more varied than originally thought. The for- Auer, land Hungary) “soft-substrate” and “hard-substrate” maar volcanoes in an mation of the Burung, Runcing, and Tepu Breccias involved aligned volcanic complex—implications for vent geometry, subsurface phreatomagmatic, phreatic, hydraulic, magmatic, and tecstratigraphy and the palaeoenvironmental setting: Journal of Volcanology  and Geothermal Research, v. 159, p. 1–3. tonic brecciation. These processes overlapped, and it is likely  Bendezú, R., Fontboté, L., and Cosca, M., 2003, Relative age of Cordilleran that many of the breccias preserved at Kelian are the product base metal lode and replacement deposits, and high sulfidation Au–(Ag) of hybrid phenomena. epithermal mineralization in the Colquijirca mining district, central Peru: Phreatomagmatic brecciation was caused by the intrusion Mineralium Deposita, v. 38, p. 683–694. of rhyolitic magma into an active hydrothermal system. The Busby-Spera, C.J., and White, J.D.L., 1987, Variation in peperite textures associated with differing host-sediment properties: Bulletin of Volcanology, v. presence of juvenile magmatic clasts (QP or QFP rhyolite) 49, p. 765–777.  with delicate wispy and cuspate margins, in situ QP and QFP Cannell, J., Cooke, D.R., Walshe, J.L., and Stein, H., 2005, Geology, minerbreccias, association with QP and QFP dikes, and base-surge alization, alteration, and structural evolution of the El Teniente porphyry  deposits are the key evidence for a phreatomagmatic origin. Cu-Mo deposit: ECONOMIC GEOLOGY, v. 100, p. 979–1003. The in situ, jigsaw-fit A2 breccias and rhyolite dikes are in- Carman, G.D., 1994, Genesis of the Ladolam gold deposit, Lihir Island, Papua New Guinea: Unpublished Ph.D. thesis, Victoria, Australia, Monash terpreted to be the root zones to the phreatomagmatic brecUniversity, 263 p. cias and to be “frozen” examples of the magma-sediment mix- ——2003, Geology, mineralization and hydrothermal evolution of the ing event. The transition from in situ rhyolite dikes through Ladolam gold deposit, Lihir Island, Papua New Guinea: Society of EcoA2 breccias and into the juvenile clast-bearing A3 breccias nomic Geologists Special Publication 10, p. 247–284. has not been documented previously in the subsurface parts Cas, R.A.F., and Wright, J.V., 1987, Volcanic successions: Modern and ancient. A geological approach to processes, products and successions: Lonof such breccia complexes, perhaps due to the generally poor don, Allen and Unwin, 528 p. preservation potential for these facies. Clement, C.R., 1982, A comparative geological study of some major kimberStructure played a key role in the localization of the dialite pipes in the Northern Cape and Orange Free State: Unpublished Ph.D. treme breccias. Volcanic and sedimentary rock units were juxthesis, Cape Town, South Africa, University of Cape Town, 432 p. taposed by faults and brecciated tectonically. Felsic magmas Cooke, D.R., and Bloom, M.S., 1990, Epithermal and subjacent porphyry  mineralization, Acupan, Baguio district, Philippines: A fluid-inclusion  were emplaced along northeast- and northwest-striking and paragenetic study: Journal of Geochemical Exploration, v. 35, p. faults, and ring-fault development was controlled by preexist297–340. ing faults and structural weaknesses in the Runcing area. Par- Cooke, D.R., McPhail, D.C., and Bloom, M.S., 1996, Epithermal gold mineralization, Acupan, Baguio district, Philippines: Geology, mineralization, tial collapse of the Runcing Breccia dropped megablocks of  alteration, and the thermochemical environment of ore deposition: E COphreatomagmatic base-surge deposits several hundred meNOMIC GEOLOGY, v. 91, p. 243–272. ters down from the surface into the diatreme. Davies, A.G.S., 2002, Geology and genesis of the Kelian gold deposit, East Diatreme formation continued into the early stages of  Kalimantan, Indonesia: Unpublished Ph.D. thesis, Tasmania, Australia, phreatic brecciation and main-stage mineralization at Kelian University of Tasmania, 404 p. Davies, A.G.S., Cooke, D.R., and Gemmell, J.B., 2003, The Kelian breccia (described in detail in Davies, 2002, and Davies et al., 2008). complex: A giant epithermal gold-silver deposit in Kalimantan, Indonesia, The hydrothermal system was catastrophically and irre in Eliopoulos, D., ed., Mineral exploration and sustainable development:  versibly altered as a result of diatreme formation, and this Proceedings of the Seventh Biennial SGA Meeting: Rotterdam, Millpress, event established the setting for the subsequent formation of   v. 1, p. 465–468. a large epithermal gold deposit hosted primarily by hy- Davies, A.G.S., Cooke, D.R., Gemmell, J.B., van Leeuwen, T., Cesare, P., and Hartshorn, G., 2008, Hydrothermal breccias and veins at the Kelian drothermal breccias and veins.  Acknowledgments  We thank PT Kelian Equatorial Mining (KEM) and PT Rio Tinto Indonesia for their research and travel funds, pro vision of site access, and logistical support in Indonesia. Theo van Leeuwen, Steve Hunt, Kerry O’Sullivan, Roger Norris, Pat Cesare, Greg Hartshorn, John Miller, John Sanderson, Seno Aki, Yudhi Nurchahayna, Ewa Rappe, and 0361-0128/98/000/000-00 $6.00

gold mine, Kalimantan, Indonesia: Genesis of a large epithermal gold deposit: ECONOMIC GEOLOGY, v. 103, p. 717–757. Ferguson, K.J., 1986, The Kelian gold prospect, Kalimantan Indonesia [abs.]: International Association of Volcanology and Chemistry of the Earth’s Interior, International Volcanological Congress, Auckland-Hamilton-Rotorua, New Zealand, Abstracts, p. 293. Fisher, R.V., and Waters, A.C., 1970, Base-surge bed forms in maar volcanoes: American Journal of Science, v. 268, p. 157–180. Gustafson, L.B., and Hunt, J.P., 1975, The porphyry copper deposit at El Sal vador, Chile: ECONOMIC GEOLOGY, v. 70, p. 857–912.

715

716

DAVIES ET AL.

Hanson, R.E., and Hargrove, U.S., 1999, Processes of magma/wet sediment interaction in a large-scale Jurassic andesitic peperite complex, northern Sierra Nevada, California: Bulletin of Volcanology, v. 60, p. 610–626. Hanson, R.E., and Wilson, T.J., 1993, Large-scale rhyolite peperites (Jurassic, southern Chile): Journal of Volcanology and Geothermal Research, v. 54, p. 247–264. Hazlitt, J.S., and Thompson, T.B., 1990, Breccia bodies in the Leadville district, with emphasis on occurrences in the Black Cloud mine, Lake County, Colorado: ECONOMIC GEOLOGY MONOGRAPH 7, p. 180–192. Houghton, B.F., and Smith, R.T., 1993, Recycling of magmatic clasts during explosive eruptions: Estimating the true juvenile content of phreatomagmatic volcanic deposits: Bulletin of Volcanology, v. 55, p. 414–420. Jensen, E.P., 2003, Magmatic and hydrothermal evolution of the Cripple Creek gold deposit, Colorado: Unpublished Ph.D. thesis, Arizona, University of Arizona, 499 p. Landtwing, M.R., Dillenbeck, E.D., Leake, M.H., and Heinrich, C.H., 2002, Evolution of the breccia-hosted porphyry Cu-Mo-Au deposit at Agua Rica, Argentina: Progressive unroofing of a magmatic-hydrothermal system: ECONOMIC GEOLOGY, v. 97, p. 1273–1292. Lawless, J.V., 1988, Punctuated equilibrium and paleohydrology: New  Zealand Geothermal Workshop, 10 th, Auckland, New Zealand, 1988, Proceedings, p. 165–170. Lorenz, V., 1973, On the formation of maars: Bulletin of Volcanology, v. 37, p. 183–204. ——1986, On the growth of maars and diatremes and its relevance to the formation of tuff rings: Bulletin of Volcanology, v. 48, p. 265–274. Lorenz, V., and Kurszlaukis, K., 2007, Root zone processes in the phreatomagmatic pipe emplacement model and consequences for the evolution of maar-diatreme volcanoes: Journal of Volcanology and Geothermal Research, v. 159, p. 4–32. Macdonald, G.D., and Arnold, L.C., 1994, Geological and geochemical zoning of the Grasberg igneous complex, Irian Jaya, Indonesia: Journal of Geochemical Exploration, v. 50, p. 143–178. Martin, U., Németh, K., Lorenz, V., and White, J.D.L., 2007, Introduction: Maar-diatreme volcanism: Journal of Volcanology and Geothermal Research, v. 159, p. 1–3. McPhie, J., Doyle, M., and Allen, R., 1993, Volcanic textures: University of  Tasmania, Centre for Ore Deposit and Exploration Studies, 198 p. Moss, S.J., and Chambers, J.L.C., 1999, Tertiary facies architecture of the Kutai basin, Kalimantan, Indonesia: Journal of Asian Earth Sciences, v. 17, p. 157–181. Nairn, I.A., 1979, Rotomahana-Waimangu eruption, 1886: Base-surge and basalt magma: New Zealand Journal of Geology and Geophysics, v. 22, p. 363–378. Roache, M.W., Allen, S.R., a nd McPhie, J., 2000, Surface and subsurface facies architecture of a small hydroexplosive, rhyolitic centre in the Mesoproterozoic Gawler Range Volcanics, South Australia: Journal of Volcanology and Geothermal Research, v. 101, p. 237–259. Self, S., 1983, Large-scale phreatomagmatic silicic volcanism: A case study  from New Zealand: Journal of Volcanology and Geothermal Research, v. 17, p. 433–469. Self, S., Kienle, J., and Huot, J.P., 1980, Ukinrek maars, Alaska, II. Deposits and formation of the 1977 craters: Journal of Volcanology and Geothermal Research, v. 7, p. 39–65. Setiabudi, B.T., Campbell, I.H., and Palin, M.J., 2001, Age of calc-alkaline magmatism associated with the Kelian gold deposit, Indonesia: An example of zircon U-Pb dating by excimer laser ablation ICP-MS [ext. abs.]: 2001 A Hydrothermal Odyssey, New Developments in Metalliferous Hydrothermal Systems Research, Townsville, Queensland, Australia, 17–19th May, EGRU Contribution 59, Extended Conference Abstracts, p. 185–186. Sheridan, M.F., and Wohletz, K.H., 1981, Hydrovolcanic explosions: The systematics of water-pyroclast equilibration: Science, v. 212, p. 1387–1389.

0361-0128/98/000/000-00 $6.00

Sillitoe, R.H., 1985, Ore-related breccias in volcanoplutonic arcs: E CONOMIC GEOLOGY, v. 80, p. 1467–1514. ——1994, Indonesian mineral deposits—introductory comments, comparisons and speculations: Journal of Geochemical Exploration, v. 50, p. 1–11. ——1997, Characteristics and controls of the largest porphyry copper-gold and epithermal gold deposits in the circum-Pacific region: Australian Journal of Earth Sciences, v. 44, p. 373–388. Sillitoe, R.H., and Bonham, H.F.J., 1984, Volcanic landforms and ore deposits: ECONOMIC GEOLOGY, v. 79, p. 1286–1298. Sillitoe, R.H., Baker, E.M., and Brook, W.A., 1984, Gold deposits and hydrothermal eruption breccias associated with a maar volcano a t Wau, Papua New Guinea: ECONOMIC GEOLOGY, v. 79, p. 638–655. Sillitoe, R.H., Grauberger, G.L., and Elliott, J.E., 1985, A diatreme-hosted gold deposit at Montana Tunnels, Montana: E CONOMIC GEOLOGY, v. 80, p. 1707–1721. Simmons, S.F., Keywood, M., Scott, B.J., and Keam, R.F., 1993, Irreversible change of the Rotomahana-Waimangu hydrothermal system (New  Zealand) as a consequence of a volcanic eruption: Geology, v. 21, p. 643–646. Sutopo, B., Jones, M.L., and Levet, B.K., 2003, The Martabe gold discovery: A high sulphidation epithermal gold-silver deposit, North Sumatra, Indonesia: NewGen Gold 2003 Conference, Perth, Western Australi, Proceedings, p. 147–158. Sutopo, B., Gemmell, J.B., Levet, B.K., Jones, M.L., and Harlan, J.B., 2007, The Martabe high sulfidation Au-Ag deposits, Sumatra, Indonesia [abs.]: Arizona Geological Society, Ores and Orogenesis Abstract Volume, p. 92–93. Thompson, T.B., 1992, Mineral deposits of the Cripple Creek district, Colorado: Mining Engineering, v. 44, p. 135–138. Thompson, T.B., Trippel, A.D., and Dwelley, P.C., 1985, Mineralized veins and breccias of the Cripple Creek district, Colorado: E CONOMIC GEOLOGY,  v. 80, p. 1669–1688. Turner, S.J., 1997, The Yanacocha epithermal gold deposits, northern Peru: High-sulfidation mineralization in a flow dome setting: Unpublished Ph.D. thesis, Colorado, Colorado School of Mines, 341 p.  Valentine, G.A., and Fisher, R.V., 2000, Pyroclastic surges and blasts: Encyclopedia of Volcanoes, p. 571–580.  van Leeuwen, T.M., 1994, 25 years of mineral exploration and discovery in Indonesia: Journal of Geochemical Exploration, v. 50, p. 13–90.  van Leeuwen, T.M., Leach, T.M., Hawke, A.A., and Hawke, M.M., 1990, The Kelian disseminated gold deposit, East Kalimantan, Indonesia: Journal of  Geochemical Exploration, v. 35, p. 1–61.  Walker, G.P.L., 1984, Characteristics of dune-bedded pyroclastic surge bedsets: Journal of Volcanology and Geothermal Research, v. 20, p. 281–296.  Wallier, S., Rey, R., Kouzmanov, K., Pettke, T., Heinrich, C.H., Leary, S., O’Connor, G., Tˇ amas¸, C.G., Vennemann, T., and Ullrich, T., 2006, Magmatic fluids in the breccia-hosted epithermal Au-Ag deposit of Ros¸ia Montanˇ a, Romania: ECONOMIC GEOLOGY, v. 101, p. 923–954.  White, J.D.L., and Houghton, B., 2000, Surtseyan and related phreatomagmatic eruptions: Encyclopedia of Volcanoes, p. 495–511.  Williams, C.L., Thompson, T.B., Powell, J.L., and Dunbar, W.W., 2000, Goldbearing breccias of the Rain mine, Carlin trend, Nevada: E CONOMIC GEOLOGY, v. 95, p. 391–404.  Wohletz, K.H., 1986, Explosive magma-water interactions: Thermodynamics, explosion mechanisms, a nd field studies: Bulletin of Volcanology, v. 48, p. 245–264.  Wormald, P.J., 1991, The Mt. Leyshon gold mine (NE Queensland), an intrusive breccia and igneous complex: World Gold Conference 1991, Proceedings, Australasian Institute of Mining and Metallurgy, Cairns, p. 223–232.

716