Fluidal-Clast Breccia Generated by Submarine Fire Fountaining, Trooper

Journal of Volcanology and Geothermal Research 109 22001) 339±355 www.elsevier.com/locate/jvolgeores

Fluidal-clast breccia generated by submarine ®re fountaining, Trooper Creek Formation, Queensland, Australia

K. Simpson, J. McPhie*

Centre for Ore Deposit Research, University of Tasmania, G.P.O. Box 252-79, Hobart, Tasmania 7001, Australia Received 25 May 2000; accepted 18 December 2000

Abstract A distinctive monomictic breccia, composed of ¯uidal and blocky basaltic andesite clasts, occurs in a Cambro-Ordovician submarine volcanic succession in northern Queensland, Australia. Associated with this ¯uvial-clast buccia facies are coherent facies and coarse and ®ne breccia facies of the same composition. The ¯uidal-clast breccia facies is internally massive and .250 m thick, varying only in the ratio of ¯uidal clasts to blocky clasts. Fluidal clasts range in size from 2 cm to 170 cm, and have moderately to highly vesicular cores and thick 2up to 1 cm), non-vesicular, formerly glassy rims. Blocky clasts are highly vesicular to non-vesicular, ,2 cm, angular, dominantly equant or splintery in shape and identical in composition to the ¯uidal clasts. The ¯uidal clasts strongly resemble subaerial volcanic bombs and are interpreted to be the products of submarine ®re fountaining of relatively low-viscosity lava. The blocky clasts were mainly derived from disintegration of the ¯uidal clasts, by means of quench fragmentation. Coherent basaltic andesite intercalated with the ¯uidal-clast breccia represents co-genetic lavas, dykes and irregular shallow intrusions. The coarse and ®ne breccia facies is very thickly bedded, monomictic 2basaltic andesite), poorly sorted and clast supported. This facies is interpreted to have been generated by periodic gravitational collapse of unstable accumulations of the ¯uidal-clast breccia facies. Subaqueous ®re-fountain breccias are distinguished from subaerial ®re-fountain breccias by thick glassy margins on ¯uidal clasts, the lack of welding and agglutination, and the distinctive association of highly vesicular, ¯uidal clasts with non- vesicular, angular, blocky clasts. Recognition of submarine ®re-fountain breccias in volcanic successions constrains the eruption style, proximity 2tens of metres) to source and environment of deposition. q 2001 Elsevier Science B.V. All rights reserved.

Keywords: submarine ®re fountain; ¯uidal clasts; volcanic bombs; basaltic andesite; submarine volcanic succession; vesicularity; fragmentation processes

1. Introduction several other submarine volcanic successions, in Japan 2Yamagishi, 1987; Cas et al., 1996), Sweden 2Allen Volcanic breccias characterised by distinctive ¯uid- et al., 1996, 1997), Canada 2Mueller and White, ally shaped clasts occur in a Cambro-Ordovician 1992), Germany 2Schmincke and Sunkel, 1987) and submarine succession in northern Queensland, Austra- Iceland 2Kokelaar and Durant, 1983). The juvenile lia. Similar ¯uidal-clast breccias have been identi®ed on clasts are typically basaltic andesite or basaltic in the modern sea¯oor 2Smith and Batiza, 1989), and in composition; however, Archean rhyolitic examples have been recognised in Canada 2Mueller and White, * Corresponding author. 1992). As the ¯uidal clasts resemble bombs, these brec- E-mail address: [email protected] 2J. McPhie). cias are generally attributed to ®re-fountain eruptions.

0377-0273/01/$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S0377-0273201)00199-8 340 K. Simpson, J. McPhie / Journal of Volcanology and Geothermal Research 10932001) 339±355 Subaerial ®re-fountain deposits are common and In this paper, we describe the facies and facies asso- there are numerous examples in the recent volcanic ciations of the BBA and use textural characteristics to record 2Macdonald, 1972; Head and Wilson, 1987; constrain the modes of fragmentation and deposition, Head and Wilson, 1989; Wilson et al., 1995; Mattox and the facies architecture. The ¯uidal-clast breccia and Mangan, 1997). The fountain structure and erup- facies at Brittania is compared with submarine and tive products are controlled by magma viscosity, vola- subaerial ¯uidal-clast breccias, also thought to be tile content, magma volume ¯ux, accumulation rate, ®re-fountain deposits, and with other texturally simi- local temperature, and height and vigour of fountain- lar facies of different origins. ing 2Head and Wilson, 1989). In subaqueous settings, the controls on fountaining are likely to be the same, 2. Geological setting with additional important in¯uences exerted by hydrostatic pressure and magma±water interactions. The BBA is part of the Seventy Mile Range Group There is some evidence that these eruptions can occur 2SMRG; Henderson, 1980; Fig. 1) which extends at depths .1000 m although the maximum water approximately 165 km east±west and comprises four depths may be much greater 2Batiza et al., 1984; regionally mappable formations. The middle two Smith and Batiza, 1989; Clague et al., 1990; Gill et formations 2Mount Windsor and Trooper Creek al., 1990; Clague et al., 2000). Formations) are submarine volcanic successions. Subaqueous ®re fountaining has been interpreted to The other two formations 2Puddler Creek and Rollston account for a variety of different volcanic facies and Range Formations) are also submarine, but domi- clast types including hyaloclastite 2Yamagishi, 1987), nantly sedimentary. Although compositionally varied, scoria 2Gill et al., 1990), reticulite 2Siebe et al., 1995), the volcanic formations have calc-alkaline af®nities spatter and highly irregular ¯uidally shaped bombs, and are dominated by rhyolite and dacite 2Henderson, lapilli, shards 2Kokelaar and Durant, 1983; Batiza et 1986; Berry et al., 1992). Ma®c to intermediate rocks, al., 1984; Schmincke and Sunkel, 1987; Mueller and although rare in the Mount Windsor Formation, are White, 1992; Cas et al., 1996; Allen et al., 1997) and locally abundant in parts of the Trooper Creek glass spheres 2Clague et al., 1990). However, there are Formation. few systematic descriptions for the textural and facies Regional metamorphism of the SMRG ranges from characteristics of submarine ®re-fountain deposits, subgreenschist facies to amphibolite facies and is nor have any contrasts with subaerial or shallow locally overprinted by contact metamorphism asso- water equivalents been identi®ed. Correct identi®ca- ciated with Ordovician to Devonian granitoids tion of submarine ®re-fountain breccias is critical in 2Berry et al., 1992). The present distribution of the reconstructing the facies architecture of ancient volca- SMRG is the surface expression of the subvertical nic successions. They can be used to identify paleo- and south-younging limb of an east±west trending sea¯oor positions although they cannot precisely fold 2Berry et al., 1992). Despite the metamorphism constrain the water depth. Fire fountaining can gener- and deformation, the preservation of primary textures ate substantial thicknesses in short periods, signi®- is excellent and a wide variety of volcanic and sedi- cantly modifying the local sea¯oor topography and mentary facies has been identi®ed. The presence of hence facies geometry. In addition, ®re-fountain brec- pillow lava, hyaloclastite, very thick graded volcani- cias rarely extend beyond tens of metres from the clastic mass-¯ow units, peperite and massive sul®de source vent 2Kokelaar and Durant, 1983; Kokelaar, ore deposits in the volcanic formations, and marine 1986) and are thus good indicators of proximity to a fossils 2trilobites and graptolites; Dear, 1974; McClung, volcanic centre. 1976; Henderson, 1983) in the two youngest formations The ¯uidal-clast breccia and associated facies collectively indicate a submarine environment of described herein are located near Brittania Homestead deposition for the SMRG 2Berry et al., 1992). south of Charters Towers 2Fig. 1) and will be referred to as the Brittania basaltic andesite 2BBA). The BBA 2.1. Setting of the BBA was initially identi®ed in drill core and subsequent ®eld mapping allowed de®nition of three main facies. The BBA occurs near the top of the Trooper Creek K. Simpson, J. McPhie / Journal of Volcanology and Geothermal Research 10932001) 339±355 341 nswood Spur Sunrise MAGPIE

Rave 440000 WARRAWEE id intrusions 2after A Towers rs f 2A) and 2B) in relation to the Brittania HIGHWAY,REWARD, HANDCUFF Charte LIONTOWN B WATERLOO- AGINCOURT 25km

A 430000 THALANG Location of Study Area SEVENTY MILE RANGE GROUP Pentland VONIAN Creek Trooper Prospect N

Cover rocks Puddler Creek Formation - sandstone, siltstone, minor andesite and dolerite fault Lolworth-Ravenswood batholith Rollston Range Formation - volcanogenic siltstone, greywacke, minor dacite Mt Windsor Formation - dominantlyand rhyolitic andesitic volcanic volcanic facies facies with lesser dacitic, Trooper Creek Formation - rhyolitic, dacitic, and andesitic volcanic facies 420000 CAMBRIAN-ORDOVICIAN CAMBRIA ORDOVICIAN-DE TRIASSIC-TERTIARY

entRoad lopm Gregory Deve 7755000 7745000 Highway/Reward Mine

Thalanga Mine

0000 37 0000 41 Prospect Agincourt Waterloo- ne

Liontown Mi

Highway

00000 4 0000 36 10 km d

lin F Mill Waddy's 5

Bore

A 0 B 0000 39 Trafalgar 7735000 7745000 7755000 Berry et al., 1992; Stolz, 1995).Seventy The Mile Brittania basaltic Range andesite Group. occurs near the top of the Trooper Creek Formation. Inset in 2A) shows the positions o Fig. 1. Geological map of the Seventy Mile Range Group between Trafalgar Bore and Sunrise Spur 2A) and at Waddy's Mill 2B), also showing younger granito 342 K. Simpson, J. McPhie / Journal of Volcanology and Geothermal Research 10932001) 339±355 Formation 2TCF). The thickness of the TCF ranges 3. Principal lithofacies in the BBA from 500 to 4000 m 2Berry et al., 1992). The TCF is characterised by rapid facies variations and diverse The BBA comprises three facies: ¯uidal-clast brec- volcanic facies including basaltic, basaltic andesite, cia, coarse and ®ne breccia, and coherent facies. The andesitic, dacitic and rhyolitic lavas, sills, dykes and ¯uidal-clast breccia facies is volumetrically dominant volcaniclastic units plus well-bedded to massive and present throughout the strike length of the BBA. mudstone and calcareous metasedimentary lithofacies Outcrops of the coarse and ®ne breccia facies are 2Henderson, 1986; Berry et al., 1992). The TCF hosts restricted to a few areas adjacent to the ¯uidal-clast massive sul®de ore deposits at Thalanga and High- breccia facies. The coherent facies is widespread but way-Reward and numerous other massive sul®de volumetrically subordinate to the other two facies in prospects 2Berry et al., 1992). the BBA. The BBA can be traced for a minimum distance of 4.5 km along strike. The best estimates of true thick- 3.1. Fluidal-clast breccia facies ness come from two areas 2,2.5 km apart), where values of approximately 380 m and 520±630 m The ¯uidal-clast breccia facies dominates a single, were calculated. The minimum thickness in some very thick unit that was intersected in drill core over a areas is probably signi®cantly less than 380 m. true thickness of more than 250 m. The upper contact Thus, thickness varies laterally by at least 250 m. was not intersected by the drill hole. This facies is A thick succession 2.250 m) dominated by composed of basaltic andesite clasts that are either normally graded, monomictic and polymictic, felsic ¯uidally shaped and moderately to highly vesicular volcaniclastic units conformably underlies the BBA. 2up to 60% vesicles), or blocky, splintery and highly The principal components are pumice, quartz and vesicular to non-vesicular 2Fig. 2a). This facies is feldspar crystals, and lithic fragments 2felsic volcanic poorly sorted, internally massive and clast supported, clasts, volcanic sandstone and mudstone, and minor varying only in the ratio of ¯uidal clasts to blocky ma®c volcanic clasts). Grain size grades from coarse clasts. Fluidal clasts comprise 10±50% in ¯uidal breccia, with clasts up to 30 cm, to mudstone. Sedi- clast-rich domains and ,10% in ¯uidal clast-poor mentation units are 1±20 m thick and have diffuse to domains. Variations are gradational over thicknesses sharp, and in places, erosional lower contacts. These of several metres. No clasts ,0.2 cm could be distin- units are interpreted as deposits from submarine guished in this facies. Clasts are separated by coarsely volcaniclastic mass ¯ows. The components indicate crystalline carbonate that could have either ®lled derivation from a dominantly felsic, probably explo- original pore space or replaced original ®ne matrix. sive volcanic source, and the very thick graded beds As there are no relict or replacement textures in the suggest deposition below storm wave base. carbonate, it is more likely that the carbonate is in®ll- The BBA is overlain by thinly bedded mudstone ing original pore space. and thin to medium bedded sandstone of the Rollston All clasts are aphyric and most consist of relict Range Formation 2RRF). In places, ironstone 2quartz- plagioclase laths between 200 and 500 mm in varying magnetite or quartz-hematite rocks of possible hydro- abundance 2up to 60%). The laths are usually thermal exhalative origin; Henderson, 1986; Duhig et randomly oriented, but in some blocky clasts they al., 1992) occurs along the boundary between the are aligned. They are set in a microcrystalline ground- BBA and the RRF. The nature of the contact in the mass of chlorite, epidote, actinolite-tremolite, leucox- study area is unknown as it is not exposed and has not ene, and minor carbonate, sericite and quartz. been intersected in drill holes. However, regionally, Leucoxene is commonly concentrated along the the contact between the TCF and RRF is conformable margins of the ¯uidal clasts and some of the blocky 2Henderson, 1986). Turbidite and hemipelagic clasts have high concentrations of leucoxene. Plagio- mudstone facies and marine fossils, including clase is relatively unaltered, but may be partially graptolites and pelagic trilobites, constrain the replaced by albite, sericite and chlorite. Some blocky environment of deposition of the RRF to have been clasts appear to have little or no plagioclase and are relatively deep submarine 2Henderson, 1986). pervasively altered to pale green chlorite. Vesicles are K. Simpson, J. McPhie / Journal of Volcanology and Geothermal Research 10932001) 339±355 343

a b

c d

2 cm

e f

Fig. 2. Examples of breccia facies in the BBA. 2a) Fluidal-clast breccia facies. Pale, highly vesicular, irregular and ¯uidally shaped clasts, and dark grey, blocky, angular, non-vesicular to highly vesicular clasts. 2b) Large clast within ¯uidal-clast breccia facies. The clast has a highly irregular ¯uidal shape, and undeformed and slightly deformed pipe vesicles around the margin. 2c) Highly vesicular, ¯uidally shaped clast in a matrix of blocky, angular, splintery, non-vesicular to highly vesicular clasts. Vesicles in the ¯uidal clast grade from coarse and more abundant in the core to ®ne and sparse in the formerly glassy rim. The sub-planar clast margin 2RHS) implies that a portion of the clast has broken off. 2d) Fluidal-clast breccia facies. A highly vesicular, ¯uidally shaped clast occurs in a matrix of angular, blocky, non-vesicular to highly vesicular clasts. Coalesced vesicles occur in the centre of the ¯uidal clast. 2e) Sharp contact between moderately sorted ®ne breccia and moderately sorted coarse breccia. Younging is towards the bottom of the photo. AMG Zone 55, 7741432mN/431314mE. 2f) Moderately sorted coarse breccia of the coarse and ®ne breccia facies. 344 K. Simpson, J. McPhie / Journal of Volcanology and Geothermal Research 10932001) 339±355 in®lled by carbonate, chlorite and quartz, and in the breccia facies. However, there are some examples most altered samples they also contain actinolite- where the rims of ¯uidal clasts show a transition tremolite, clinopyroxene and minor epidote. Any from intact rim to jigsaw-®t rim fragments and to original glass within the clasts has been completely rotated rim fragments. altered. Immobile element ratios, especially Ti/Zr ratios, 3.1.2. Blocky clasts can be used to determine the primary composition Blocky clasts dominate the ,1-cm-size fraction of of the altered volcanic rocks in the SMRG 2Berry et the ¯uidal-clast breccia facies. They are generally al., 1992; Stolz, 1995). Whole-rock Ti/Zr ratios of the 0.2±1 cm across, with some clasts up to 2 cm. These ¯uidal-clast breccias are in the range 30±66, consis- angular, blocky or splintery clasts have curviplanar tent with basaltic andesite composition. The ¯uidal margins and can be separated into two subtypes on and blocky clasts have the same composition and the basis of vesicularity. The dominant type is poorly mineralogy, so the ¯uidal-clast breccia facies is to non-vesicular and the less abundant type is moder- considered to be monomictic. ately to highly vesicular. The poorly to non-vesicular blocky clasts have up to 5% vesicles, ranging from 0.1 3.1.1. Fluidal clasts to 0.2 mm. The moderately to highly vesicular blocky Fluidal clasts make up 1±50% of the ¯uidal-clast clasts have vesicle shapes, size and abundance similar breccia facies. They are typically elongate with long to those in the ¯uidal clasts. However, they lack ¯ui- dimensions in the range of 1±170 cm but average dal margins. Instead, the edges of the clasts are jagged approximately 7±8 cm. The largest clasts show a and cut across vesicles. crude bedding-parallel alignment of their long dimen- The blocky clasts are highly variable in terms of sions and are up to a few tens of cm thick 2Fig. 2c). vesicle size and abundance, feldspar abundance and Fluidal clasts are commonly highly irregular and alteration. Clast boundaries are dif®cult to discern, but contorted with elongate, tapered tails 2Fig. 2a,c,d). are usually marked by the presence of abundant carbo- The margins of some of the largest clasts are convo- nate. In general, the poorly to non-vesicular clasts luted and have delicate, narrow 2,2 cm across), elon- appear to be more intensely altered than both the gate protrusions. All ¯uidal clasts are moderately to moderately to highly vesicular blocky clasts and the highly vesicular 230±60% vesicles) and most show ¯uidal clasts. vesicle grading. Vesicles are larger 23±25 mm) and more abundant in the cores of clasts, and become 3.2. Coarse and ®ne breccia facies smaller 2,1±2 mm) and less abundant towards the margins 2Fig. 2c and d). Sporadically distributed, This facies is laterally continuous in patchy outcrop large 2up to 1.5 cm) vesicles that do not conform to over hundreds of metres. It is .100 m thick and has the vesicle-grading pattern are also present. Vesicles relatively sharp contacts with the ¯uidal-clast breccia can be round, irregular or elongate and are commonly facies but was not observed in contact with the coher- interconnected. In places, vesicle coalescence has ent facies. It comprises moderately to poorly sorted resulted in highly irregular vesicle shapes. Commonly coarse breccia 2clasts dominantly .1 cm) and moder- the margins 2up to 1 cm) of the ¯uidal clasts are ately sorted ®ne breccia 2clasts dominantly ,1 cm). poorly to non-vesicular. A few large clasts 225± Typically, single outcrops are exclusively either 170 cm) contain undeformed and slightly deformed coarse or ®ne breccia and beds are thicker than pipe vesicles around their margins 2Fig. 2b). outcrop dimensions, that is, thicker than 3 m. Volu- Many of the ¯uidal clasts have at least one sharp, metrically, poorly sorted coarse breccia dominates sub-planar margin that cross-cuts vesicle grading and this facies 2Fig. 2f). Only one bed contact was ¯uidal margins 2Fig. 2c). These sub-planar margins observed 2Fig. 2e); this contact is sharp and separates are angular and jagged in contrast to the unbroken, ®ne breccia below from coarse breccia above. Single smoothly curved ¯uidal margins. No ¯uidal clasts outcrops of this facies appear massive. However, ,2 cm have complete ¯uidal outlines. In general, changes from coarse to ®ne breccia occur progres- jigsaw-®t texture does not occur in the ¯uidal-clast sively across several outcrops, suggesting that the K. Simpson, J. McPhie / Journal of Volcanology and Geothermal Research 10932001) 339±355 345 very thick beds are graded. Moderately sorted coarse partially albitised. The laths are set in a microcrystalline breccia grades through poorly sorted coarse breccia to groundmass of chlorite, epidote, quartz, carbonate and moderately sorted ®ne breccia. leucoxene. Carbonate, chlorite, quartz and rare epidote, Both the coarse and ®ne breccias are clast and actinolite-tremolite in®ll the vesicles. supported and monomictic. Clasts are aphyric basaltic andesite, dominantly subangular to angular and less commonly subround. In the coarse breccia, clasts 4. Facies geometry and relationships range from ,1 to 18 cm and average 2±5 cm. In the The spatial relationships among the facies are well ®ne breccia, clasts are generally 0.1 to 1 cm but subor- constrained in the area of best exposure 2Fig. 3) and in dinate large clasts 2up to 10 cm) are sporadically drill core 2Fig. 4). The ¯uidal-clast breccia facies distributed. Clasts are non-vesicular to highly vesicu- occurs in both drill core and outcrop. It is a thick lar 2up to 60% vesicles) with spherical, elongate and unit that is cross-cut by, interbedded with, and under- highly irregular vesicles ranging in size from 0.1 mm lain by coherent facies. These two facies commonly to 1 cm. Vesicle size, shape, abundance and distribu- occur together but there are thick intervals 2up to 80 m tion vary widely from clast to clast. Clasts are either for the coherent facies and up to 115 m for the ¯uidal- non-vesicular, uniformly poorly to highly vesicular or clast breccia facies) both in drill core and in outcrop moderately to highly vesicular with vesicle-size grad- that are composed of one facies only. The coarse and ing. Most clast margins cut across vesicles and a few ®ne breccia facies was not intersected in drill core and clasts have partial ¯uidal margins. The mineralogy of no contacts with other facies are exposed. The coarse clasts in the coarse and ®ne breccia facies is the same and ®ne breccia facies is stratigraphically equivalent as that of the ¯uidal-clast breccia facies. Carbonate and laterally adjacent to the ¯uidal-clast breccia ®lls spaces between the clasts. facies, but does not occur in association with the coherent facies. 3.3. Coherent facies

The coherent facies consists of aphyric basaltic 5. Origin of BBA andesite and is dominantly massive but crude pillow-like structures are present in a few outcrops. 5.1. Modes of fragmentation Drill core intersections of the coherent facies range from 2 m to 80 m true thickness. In both drill core and The ¯uidal clasts in the BBA resemble bombs and outcrop, the coherent facies intervals are intercalated ¯uidal lapilli formed by tearing apart of relatively with the ¯uidal-clast breccia facies. Contact relation- low-viscosity lava ribbons jetted upward from vents ships vary from being sharp and planar, commonly during Hawaiian-style ®re-fountain eruptions cross-cutting stratigraphy, to sharp and highly irregu- 2Macdonald, 1972; Allen et al., 1997). The ¯uidal lar, diffuse and dif®cult to identify. At one locality shapes of the smaller clasts were partly controlled 2AMG Zone 55, 7741507mN/431391mE), there is a by surface tension and hydrodynamic forces on the chaotic mixture of coherent domains and ¯uidal-clast still-molten fragments. The outer, formerly glassy, rich domains, with lobes of coherent facies extending non- to poorly vesicular margins of the ¯uidal clasts into the ¯uidal-clast breccia facies. The coherent were probably generated by quenching on contact facies is subordinate to the ¯uidal-clast breccia facies with seawater. The grading in vesicle size in the inter- in both drill core and outcrop. iors of the ¯uidal clasts implies that the core retained The coherent facies is moderately vesicular 2up to suf®cient heat to allow further vesiculation. Jigsaw-®t, 30%) to non-vesicular. Vesicles are either evenly glassy margin fragments that surround some clasts distributed or arranged in parallel layers. They are could re¯ect expansion in response to post-deposi- spherical or elongate to highly irregular and many tional vesiculation. are interconnected, ranging in size from 0.1 to 1 cm. Clasts with partly ¯uidal and partly sub-planar Plagioclase laths between 150 and 300 mm comprise margins have undergone brittle fracture during trans- up to 90% of the rock, and are randomly oriented and port and/or deposition. Clast-to-clast collisions, clast 346 K. Simpson, J. McPhie / Journal of Volcanology and Geothermal Research 10932001) 339±355

A 7742000mN

BRDD04

0 1km Bletchington oad 432000mE 434000mE R

B 431400mE 431600mE 431800mE

? 0 100m ? Brittania Homestead ? ? ? 7741600mN

? ?

? 7741400mN ?

d a o R 7741200mN ton ing ch Blet younging

7741000mN ? ? 58

Lolworth-Ravenswood Batholith Rollston Range Formation Trooper Creek Formation volcanic breccia and granite laminated siltstone- sandstone sandstone

? Approximate contact Brittania Basaltic Andesite ? Inferred contact N Creek fluidal-clast breccia coherent andesite Road coarse and fine undifferentiated BBA breccia (no outcrop) Drill hole

Fig. 3. Geological map of the study area. 2a) Location of drill holes 2X) relative to area of detailed mapping. 2b) Detailed map showing the distribution of facies in the BBA. K. Simpson, J. McPhie / Journal of Volcanology and Geothermal Research 10932001) 339±355 347

m .5 2 83264 mm m .5 2 83264 mm

180

190

scale * 200

210

220

100 Fluidal clasts up to 20 cm

230

fluidal-clast polymictic/ monomictic 110 breccia facies volcanic breccia BBA

240 coherent facies volcanic sandstone BBA pumice breccia 120 pebbly sandstone

fine/ coarse coherent coherent rhyolite 250 dacite monomictic rhyolitic foliated/ massive 130 breccia pumice + lithic breccia

260

140

270

150

280

160

290

170

300

Fig. 4. Representative graphic log of the BBA in drill hole BRDD04. Note that the coarse and ®ne breccia facies was not intersected in drill core. The base of the BBA is at ,240 m. 348 K. Simpson, J. McPhie / Journal of Volcanology and Geothermal Research 10932001) 339±355 expansion due tovesiculation and quench fragmentation 5.2. Transport and depositional processes could have generated highly vesicular fragments partly bounded by ¯uidal margins and partly by sub-planar The monomictic nature, relatively poor sorting and surfaces. The lack ofintact ¯uidal clasts ,2cmsuggests preservation of highly irregular, delicate clasts in the that a critical size is necessary for a lava ribbon to ¯uidal-clast breccia facies suggest that these clasts survive without being completely quench fragmented have not undergone signi®cant lateral transport. into cm-size and ®ner, blocky, angular fragments. Given the close similarity of the ¯uidal clasts and The largest 225±170 cm) ¯uidal clasts probably volcanic bombs, we infer that these clasts were formed by the same mechanism envisaged for the ejected from the vent into the water column and smaller ¯uidal clasts. Bombs more than a metre long then rapidly settled onto the sea¯oor. Thus, the ¯ui- and tens of cm thick are common in proximal deposits dal-clast breccia facies is a variety of water-settled fall from subaerial ®re fountains 2Macdonald, 1972). It is deposit. Although highly vesicular, the ¯uidal clasts also possible that these large clasts are intrusions into were probably not buoyant or else, only temporarily co-genetic wet, unconsolidated ¯uidal-clast breccia so. Most would have been denser than water or rapidly 2cf. Yamagishi, 1987). The morphology of intrusive water-logged as quench fractures allowed ingress of tongues and large ¯uidal bombs are very similar in water into the interior. In contrast to subaerial bombs, two-dimensional sections. Moreover, intrusion of there is no evidence for plastic deformation of the magma into wet unconsolidated breccia could have ¯uidal clasts upon impact with the substrate. The accompanied fountaining; thus, both processes prob- force of impact with the sea¯oor would have been ably operated to produce the largest ¯uidal clasts. reduced due to the lower settling velocity of the clasts The largest ¯uidal clasts also contain pipe vesicles, in water when compared with air. In addition, the rigid which are not present in the smaller clasts. The unde- chilled margins of the ¯uidal clasts would have inhib- formed pipe vesicles imply that vesiculation occurred ited plastic deformation on impact even though the post-emplacement. Pipe vesicles form when rising gas interiors may still have been partially molten. bubbles are progressively captured by the advancing In subaerial environments, the products of ®re foun- solidi®cation front 2Philpotts and Lewis, 1987). Their tains are mostly coarse-grained and deposited very presence only in the largest clasts is mainly a conse- close to the source vent 2tens of m), building steep- quence of size, the large clasts being more effectively sided cones 2Wood, 1980). In submarine settings, thermally insulated than the small clasts and having dispersal is also likely to be restricted, with most fall- suf®cient volatile content for prolonged vesiculation. out occurring in the immediate vicinity of the vent The two categories of blocky clasts can be matched 2Kokelaar and Durant, 1983; Kokelaar, 1986; Siebe with either the moderately to highly vesicular interiors et al., 1995). Any ®ner particles that remained fully or the poorly to non-vesicular margins of the ¯uidal suspended in the water column would be deposited clasts, suggesting that most of the blocky clasts were elsewhere much farther from source. The apparently derived from post-vesiculation fragmentation of ¯ui- low content of ®ne particles in the ¯uidal-clast breccia dal clasts. Cooling contraction on contact with cold facies probably re¯ects a combination of the overall seawater was probably a major cause of disintegration dominance of coarse clasts and this sorting process. of the ¯uidal clasts. This affected the glassy margins The presence of bedding, moderate sorting, suban- of some larger ¯uidal clasts and was also responsible gular to subround clast shapes and possible grading in for the complete disintegration of any smaller the coarse and ®ne breccia facies indicate lateral 2,2 cm) ¯uidal clasts. In addition, the poorly to transport. The lack of exotic components and the non-vesicular blocky clasts could result from prefer- preservation of clasts with partial ¯uidal margins indi- ential fracturing of outer quenched rims in response to cate that the transport distance was probably short impact with other clasts or the sea¯oor. The clast and/or involved little clast interaction. Most of the assemblage can be accounted fully by these non- clasts within this facies are texturally and com- explosive fragmentation mechanisms; however, we positionally similar to clasts in the ¯uidal-clast brec- note that phreatomagmatic explosions can also cia. Others are similar to the coherent facies. The produce dense, blocky, angular and glassy fragments. coarse and ®ne breccia facies is thus inferred to be K. Simpson, J. McPhie / Journal of Volcanology and Geothermal Research 10932001) 339±355 349

Fig. 5. Schematic model for the facies relationships in the BBA. Tearing apart of low-viscosity fountaining lava produced a very thick accumulation of ¯uidal-clast breccia close to the vent. Any ®ne particles produced in the eruption remained suspended in the water column and were deposited elsewhere. Coherent facies comprising both pillow lava and shallow intrusions is associated with the thick section of ¯uidal- clast breccia. Periodic gravitational collapse of unstable piles of both ¯uidal-clast breccia and coherent facies resulted in resedimentation, forming poorly sorted, very thick beds of the coarse and ®ne breccia. the resedimented equivalent of the ¯uidal-clast brec- through the water column and many partly or comple- cia and intercalated coherent facies, generated by tely disintegrated, creating variably vesicular, blocky periodic gravity-driven collapse of unstable deposits clasts. The extensive distribution of the ¯uidal-clast of ¯uidal and blocky clasts, and coherent basaltic breccia facies, at least 4.5 km along strike, suggests andesite. The very thick beds and apparent lack of the presence of several vents. In subaerial settings, ®ne matrix are consistent with mass-¯ow resedimen- fountaining commonly occurs at vents located along tation, possibly involving a variety of coarse-clast- ®ssures 2Macdonald, 1972; Wilson and Head, 1981). dominated, modi®ed grain ¯ow 2Lowe, 1979). A similar scenario may be expected in subaqueous settings and could apply in the BBA case. 5.3. Coherent facies Despite its considerable thickness, the ¯uidal-clast breccia facies is massive and uniformly monomictic, The coherent facies is compositionally and textu- suggesting that there were no signi®cant breaks in rally identical to the basaltic andesite of the ¯uidal accumulation. However, the ¯uidal clasts in this clasts and represents co-genetic unfragmented facies show no signs of welding as is typical of magma or lava. The presence of crude pillows and thick subaerial ®re-fountain deposits 2e.g. Head and locally highly irregular contacts suggest that the Wilson, 1989). The heat loss from ¯uidal clasts during coherent facies was extrusive, or else shallowly intru- transit in the water column was evidently suf®cient to sive into unconsolidated units of the clastic facies. preclude welding. 5.4. Eruption styles The coherent facies in the BBA resulted from lava effusion and/or shallow intrusion of magma that By analogy with bombs and ¯uidal lapilli in subaer- accompanied fountaining 2Fig. 5) and may indicate ial ®re-fountain deposits, we infer that the ¯uidal-clast ¯uctuations in the magma discharge rate. In particu- breccia facies involved fountaining of relatively low- lar, relative to fountaining episodes, the lavas were viscosity magma 2Fig. 5). The ¯uidal clasts settled probably generated during periods of reduced 350 K. Simpson, J. McPhie / Journal of Volcanology and Geothermal Research 10932001) 339±355 discharge 2e.g. Grif®ths and Fink, 1992). They are isolated pillow breccia 2Carlisle, 1963) can include unlikely to be fountain-fed lavas, given the non- clasts similar in shape to bombs and ¯uidal lapilli. welded character of the associated ¯uidal-clast brec- Globular peperite, generated by shallow intrusions cia facies. into wet unconsolidated sediment, comprises irregular, lobate and ¯uidal igneous clasts 2Busby-Spera and White, 1987), which resemble bombs and ¯uidal 6. Key characteristics of subaqueous ®re-fountain lapilli. In dimensions and shape, the irregular intru- breccias sive apophyses and lobes present in subaqueous feeder dyke±hyaloclastite complexes 2Yamagishi, Fluidally shaped juvenile clasts occur in a variety of 1987) are comparable to coarse bombs 21±2 m across) volcanic facies 2Table 1): subaerial ®re-fountain that occur sporadically in the most proximal sections deposits at primary and littoral vents, submarine of ®re-fountain deposits. ®re-fountain deposits, pillow breccia, feeder dyke± Comparison of various ¯uidal-clast-bearing facies hyaloclastite complexes, and globular peperite. It is 2Table 1) suggests that there are few textural criteria clear that the mode of fragmentation Ð tearing apart for recognition of different origins. Welded ¯uidal of low-viscosity lava or magma Ð can operate in clasts and ¯uidal clasts that show accommodation to subaerial, subaqueous and shallow intrusive settings. adjacent clasts are relatively common in subaerial Subaerial ®re-fountain deposits at primary vents typi- ®re-fountain deposits but not known to occur in cally consist of highly vesicular, ragged and ¯uidally submarine facies. Dense, blocky, glassy juvenile shaped juvenile bombs and lapilli. Juvenile clasts are clasts are generally subordinate to ¯uidal clasts in generally hot when deposited and show evidence for subaerial ®re-fountain successions or else restricted plastic deformation on impact. They are commonly to single beds, whereas they are common throughout ¯attened, annealed to one another, or moulded around deposits from submarine ®re fountains. The ¯uidal the underlying topography or other clasts. Subaerial clasts 2lapilli and bombs) in subaqueous ®re-fountain bombs and ¯uidal lapilli typically show breadcrust deposits may have thicker 2up to 1 cm) glassy chilled surfaces and well-developed internal vesicle-size margins compared with chilled margins on subaerial grading, suggesting that they continued to vesiculate bombs. after deposition. Beds of welded spatter and layers of Fluidal clasts produced by subaqueous ®re foun- clastogenic lava can be intercalated with the bomb taining have vesicularities up to 60%, commonly beds, especially in proximal sections. have spindle shapes with narrow tails, and the mini- Littoral cones form where lava enters the ocean mum size of intact ¯uidal clasts is ,2 cm. In contrast, 2Moore and Ault, 1965) and also comprise thick, the ¯uidal clasts sometimes present in pillow breccias poorly bedded, juvenile deposits of bombs and ¯uidal and intrusive apophyses in feeder dyke±hyaloclastite lapilli. At the most proximal locations, juvenile clasts complexes are generally poorly vesicular, relatively are commonly welded or agglutinated and in some coarse 2.30 cm across) and do not have spindle instances, intercalated with clastogenic lavas shapes. In addition, long dimensions of large bombs 2Jurado-Chichay et al., 1996). Interaction between may be broadly bedding-parallel, whereas intrusive the lava and water is common at littoral vents, produ- apophyses and lobes may be in any orientation and cing phreatomagmatic fall and surge deposits that are interconnected in three dimensions. The igneous relatively ash-rich and that may include signi®cant clasts in globular peperite may have broadly similar proportions of non-juvenile components. Also found shapes, sizes and internal textures to ¯uidal clasts in in proximal littoral cone deposits are delicate, paper- subaqueous ®re-fountain deposits. In this case, the thin, sliver-shaped glassy clasts known as Limu-o- distinction can be made on the basis of the monomic- pele, which is formed by the bursting of lava bubbles tic composition and crude strati®cation 2if present) in due to the rapid expansion of steam 2Hon et al., 1988; ¯uidal-clast breccias of ®re-fountain origin. Mattox and Mangan, 1997). Context and facies relationships are also very The clastic facies associated with pillow lavas may important in distinguishing among the different ¯ui- contain ¯uidally shaped igneous clasts. For example, dal-clast breccia facies: ®re-fountain deposits from K. Simpson, J. McPhie / Journal of Volcanology and Geothermal Research 10932001) 339±355 351 Carlisle 21963), Dimroth et al. 21978), Schmincke and Sunkel 21987), Dolozi and Ayres 21991), Yamagishi 21991) and Walker 21992) Carlisle 21963), Kokelaar and Durant 21983), Batiza et al. 21984), Kokelaar 21986), Schmincke and Sunkel 21987), Yamagishi 21987), Cas et al. 21989), Smith and Batiza 21989), Clague et al. 21990), Dolozi and Ayres 21991), Mueller and White 21992), Cas et al. 21996) and Allen et al. 21997) References Detachment of pillows, gravitational or mechanical break-up of pillow lobes Quench fragmentation and spalling of pillow rinds, gravitational or mechanical break-up of pillows Cooling contraction granulation, phreatomagmatic eruptions, clast to clast impact, spalling of clast margins during vesiculation of interior Tearing apart of relatively low viscosity lava in a lava fountain fragmentation process ______Pillow lava, massive lava, hyaloclastite, marine sedimentary facies Pillow lava, pillow breccia, massive lava, hyaloclastite, phreatomagmatic facies, dykes, sills, marine sedimentary facies Associated facies Known or inferred 10 cm) of , co-genetic hyaloclastite Local domains with jigsaw-®t textures MonomicticMassive Massive to crudely bedded 21±5 m thick), normal and reverse graded Parallel alignment of the long dimension of clasts Poorly sorted, clast supported Matrix 2 Monomictic Massive to crudely or diffusely bedded, both normal and reverse grading Parallel alignment of the long dimension of clasts Poorly sorted, clast supported, local domains with jigsaw- ®t texture Ratio of ¯uidal clasts to blocky clasts varies Internal fabrics/textures 5m)to , 150 m . 1m± , thick Underlie, overly and lateral to pillow lava, commonly at the ¯ow fronts of pillow lava Commonly grades into pillow lava Thin sheets 2 a few hundred metre thick, within tens of m from vent May be a point source or associated with ®ssure eruptions distribution 50%) 50%) . . 5%) to highly vesicular 2 , Poorly 2 0.1 mm±2 cm diameter Ellipsoid and spherical May show vesicle grading withinfragments pillows of or pillows Ð vesiclesincrease can in either abundance inward orinward decrease Poorly vesicular, chilled margins 25 mm±3 cm) Non-vesicular to highly vesicular 2 0.1 mm±2 cm diameter Ellipsoid and spherical Non-vesicular to highly vesicular 2up to0.1±25 60%) mm diameter Spherical or irregular, coalesced Randomly distributed Typically moderately to highly vesicular,to up 60%, scoriaceous 0.1±25 mm diameter Spherical or irregular, coalesced. Vesicle grading: coarse in middle, ®netowards vesicles clast margins Elongate vesicles parallel to longof dimension clasts Poorly to non vesicular chilled margins1 2up cm) to Rare pipe vesicles in largest clasts 10 cm , 2 m long dimension . m±10 cm m± 20 cm±5 m2?) long dimension Lobate, ovoid, globular, ellipsoidal, amoeboid, spherical t Radial jointing, no breadcrust texture, no twisted forms or tails Blocky, angular, tabular Blocky angular shards or clasts, splinters, tabular shards m Clasts may be pressed togethertightly or folded, agglutinated clasts on basal surfaces 2rare) Breadcrust texture Fluidal, amoeboid, elongate, cow- dung forms, elongate tails, spindle, elongate lava ribbons, globules, ¯uidal glass shards, glass spheres, some broken clasts m ______Basalt, basaltic± andesite Basalt, basaltic± andesite, rhyolite Composition Juvenile clast shape/size/texture Clast vesicularity/distribution Thickness/ Submarine ± effusive; pillow fragment breccia Submarine ± ®re fountain Table 1 Comparison of subaerial and submarine volcanic facies thatKnown or contain ¯uidally shapedinferred clasts setting and eruption style 352 K. Simpson, J. McPhie / Journal of Volcanology and Geothermal Research 10932001) 339±355 Kokelaar 21982), Busby-Spera and White 21987) and Hunns and McPhie 21999) Yamagishi 21987) and Yamagishi 21991) References Macdonald 21972), Walker and Croasdale 21972) and Mangan and Cashman 21996) Moore and Ault 21965), Fisher 21968), Jurado- Chichay et al. 21996) and Mattox and Mangan 21997) Moore and Ault 21965) and Fisher 21968) ______Dynamic mixing of magma and wet unconsolidated sediment fragmentation process In situ quench brecciation of feeder dyke Tearing apart of low viscosity lava in a lava fountain Tearing apart of low viscosity lava in a lava fountain, bubble bursting, quenching Phreatomagmatic eruptions, some disintegration of larger clasts Hyaloclastite, pillow lava, massive lava, dykes, sills, marine sedimentary facies Dykes, sills, blocky peperite, marine sedimentary facies Associated facies Known or inferred Fountain-fed lava, scoria fall deposits, phreatomagmatic facies, dykes, lavas Phreatomagmatic facies, lava, fountain fed lavas 1cm , Monomictic Intrusive lobes may have columnar or polyhedral joints Gradation from coherent to brecciated domains Jigsaw-®t domains Ungraded, unstrati®ed, poorly sorted Intricate contact relationships between host sediment and intrusion Internal fabrics/textures Monomictic Massive to crudely bedded Proximal facies have little to no ash matrix Variably welded/ agglutinated Monomictic Variably welded/ agglutinated Proximal facies: massive to crudely bedded, little to no clasts Distal facies: poorly sorted, poorly to moderately bedded 2up to 15 cm thick beds) 1±10 m thick 10±90 m thick 10±90 m high and , , , Thin selvages at margins of lobes, or thick envelopes around coherent interiors mtotensofmthick Associated with shallow intrusions; may be highly irregular distribution Proximal, tens of m from vent May be a point source or associated with ®ssure eruptions laterally extensive over hundreds of m Proximal, tens of m from vent Con®ned to a point source 0.1 mm±1 cm diameter, commonly 0.1 mm±1 cm diameter 0.1 mm±1 cm diameter 0.1±1 mm2?) diameter , , ! Non-vesicular to moderately vesicular 220%) mm±cm diameter Randomly distributed Non-vesicular to highly vesicular 2up to 70%), pumiceous , Round to elongate and cylindrical Randomly distributed or may havesubparallel a alignment 2tube pumice) Non-vesicular to moderately vesicular 220%) Randomly distributed Variably vesicular from poorly vesicular up to 98% vesicles 2reticulite) deformed Poorly vesicular chilled margins 2typically up to several mm) Vesicle grading: coarsest in middle, becoming ®ner towards margins Variably vesicular, bimodal distribution, large 2cm) vesicles in centre Elongate, ovoid, spherical Poorly to non-vesicular chilled margins Non-vesicular to highly vesicular 2m . 1 mm±tens of cm 1/16 mm±50 cm2?) , , Apophyses, lava tongues or lobes, ®nger-like and bulbous protrusions, detached blobs, `concentric' pillows cm to tens of metres long dimension Commonly with a greater length than width Blocky, angular, splintery, bounded by curviplanar surfaces Lensoidal, bulbous, lobate, ¯uidal, ragged, globular, detached blobs cm±m Microglobular: mm scale Partly or entirely glassy Fluidal, ragged, spindle, cow-dung and fusiform bombs, ribbons, Pele's tears, Pele's hair, shards 0.5 mm± Breadcrust textures, evidence of ¯attening on impact, agglutinated clasts on basal surface Red oxidised clasts bombs, ribbons, Limu-o-Pele, Pele's hair, ¯uidal shards 1/16 mm-2 m Breadcrust textures, quench fractures, evidence of ¯attening on impact, agglutinated clasts on basal surface, chilled margins Red oxidised clasts Blocky, angular ______Basalt, andesite, dacite, rhyolite Composition Juvenile clast shape/size/texture Clast vesicularity/distributionBasalt, andesite, dacite, rhyolite Thickness/ Basalt/ andesite Basalt Fluidal, ragged, spindle, cow-dung ) continued Submarine ± effusive; feeder dyke and hyaloclastite Submarine ± shallow intrusion; globular/ ¯uidal peperite Known or inferred setting and eruption style Subaerial ± ®re fountain Table 1 2 Subaerial ± littoral cones K. Simpson, J. McPhie / Journal of Volcanology and Geothermal Research 10932001) 339±355 353 primary and littoral vents typically occur in subaerial understood. It is clear that ®re fountaining is not or very shallow water successions that may include limited to shallow water or subaerial environments. scoria lapilli fall, ash-rich phreatomagmatic surge and In the case of the BBA, the surrounding units provide fall deposits and clastogenic lavas. Subaqueous ®re- better constraints on the water depth of formation and fountain deposits, isolated pillow breccia, feeder they collectively indicate a below-wave-base, rela- dyke±hyaloclastite complexes and globular peperite tively deep water environment. occur in subaqueous host successions that may be distinguished by the presence of pillow lava, hyaloclastite and diverse subaqueous sedimentary facies. 8. Conclusions Globular peperite is also distinctive in being restricted to shallow intrusive settings. The ¯uidal-clast breccia facies was generated by fountaining low-viscosity magma in a relatively deep submarine environment. The widespread distribution 24.5 km strike length) of the ¯uidal-clast brec- 7. Signi®cance of subaqueous ®re-fountain cia facies suggests that several submarine ®re breccias fountains were erupting simultaneously. Correct identi®cation of subaqueous ®re-fountain Subaqueous ®re-fountain breccias can be distin- breccias is important in facies interpretations, as this guished from other deposits that contain ¯uidal clasts facies provides constraints on the environment of by a combination of: relatively thick 2up to 1 cm) deposition, sea¯oor position, proximity to source glassy margins on ¯uidal clasts; vesicle grading vents and possibly also water depth. In young volcanic within ¯uidal clasts 2lapilli and bombs); lack of successions, this information may be relatively easily welded, agglutinated and compacted juvenile clasts; obtained. However, such constraints are highly signif- the distinctive association of highly vesicular ¯uidal icant in ancient volcanic successions that are incom- clasts with non-vesicular, angular, blocky, formerly pletely preserved, poorly exposed, and/or affected by glassy, ®ner clasts; and the spatial association with diagenetic alteration, metamorphism and deforma- pillow lava and with equivalent resedimented facies. tion. Subaqueous ®re-fountain breccias constrain the In the Brittania area, the identi®cation of ®re-foun- paleo-environment in which they were deposited, tain breccias offers insight into the style of volcanism represent paleo-sea¯oor positions and indicate proxi- and environment of deposition. For example, the mity to eruptive vents 2tens of metres) and therefore distribution of the facies in the BBA suggests that are useful in the identi®cation of proximal settings. there were several sea¯oor vents erupting simulta- Fire-fountain breccias provide no precise limits on neously, possibly localised along a fault or ®ssure. water depth, as the effects of hydrostatic pressure on The very thick ¯uidal-clast breccia facies of the fountaining are not well understood. BBA is almost certainly proximal, deposited within tens of metres of the source vents. Identi®cation of Acknowledgements near-vent facies is important as such facies coincide with syn-volcanic heat sources and fracture networks We thank Dr S. Allen, A. Davies, Prof. R. Cas and required for circulation of hydrothermal ¯uids. Dr D. Clague for helpful comments and critical Submarine ®re-fountain breccias have been identi®ed reviews of earlier versions of this paper. Support for in several massive sul®de districts 2Allen et al., 1996; this research was provided by RGC Exploration and Cas et al., 1996; Allen et al., 1997) and they may be by the Australian Research Council's Research useful indicators of prospective areas because both Centres Program. can occur in extensional environments on the sea¯oor. The identi®cation of submarine ®re-fountain breccias currently places few 2if any) constraints on References the water depth of formation as the effects of hydrostatic pressure on submarine eruptions are not well Allen, R.L., Hundstrom, I., Ripa, M., Simeonov, A., Christofferson, 354 K. Simpson, J. McPhie / Journal of Volcanology and Geothermal Research 10932001) 339±355

H., 1996. Facies analysis of a 1.9 Ga, continental margin, back- Gill, J., Torssander, P., Lapierre, H., Taylor, R., Kaiho, K., Koyama, arc, felsic caldera province with diverse Zn±Pb±Ag±2Cu±Au) M., Kusakabe, M., Aitchison, J., Cisowski, S., Dadey, K., sul®de and Fe oxide deposits, Bergslagen Region, Sweden. Fujioka, K., Klaus, A., Lovell, M., Marsaglia, K., Pezard, P., Econ. Geol. 91, 979±1008. Taylor, B., Tazaki, K., 1990. Explosive deep water basalt in the Allen, R.L., Weihed, P., Svenson, S., 1997. Setting of Zn±Cu±Au± Sumisu backarc rift. Science 248, 1214±1217. Ag massive sul®de deposits in the evolution and facies archi- Grif®ths, R.W., Fink, J.H., 1992. Solidi®cation and morphology of tecture of a 1.9 Ga marine volcanic arc, Skellefte District, submarine lavas: a dependence on extrusion rate. J. Geophys. Sweden. Econ. Geol. 91, 1022±1053. Res. 97, 19 729±19 737. Batiza, R., Fornari, D.J., Vanko, D.A., Lonsdale, P., 1984. Craters, Head, J.W., Wilson, L., 1987. Lava fountain heights at Pu'u `O'o calderas, and hyaloclastites on young Paci®c seamounts. J. Kilauea Hawaii: indicators of amount and variations of exsolved Geophys. Res. 89, 8371±8390. magma volatiles. J. Geophys. Res. 92, 13 715±13 719. Berry, R.F., Huston, D.L., Stoltz, A.J., Hill, A.P., Beams, S.D., Head, J.W., Wilson, L., 1989. Basaltic pyroclastic eruptions; in¯u- Kuronen, U., Taube, A., 1992. Stratigraphy, structure, and ence of gas-release patterns and volume ¯uxes on fountain volcanic-hosted mineralization of the Mount Windsor structure, and the formation of cinder cones, spatter cones, root- Subprovince, north Queensland, Australia. Econ. Geol. 87, less ¯ows, lava ponds and lava ¯ows. J. Volcanol. Geotherm. 739±763. Res. 37, 261±271. Busby-Spera, C.J., White, D.L., 1987. Variation in peperite textures Henderson, R.A., 1980. Structural outline and summary geological associated with differing host-sediment properties. Bull. Volca- history for northeastern Australia. In: Henderson, R.A., nol. 49, 765±775. Stephenson, P.J. 2Eds.), The Geology and Geophysics of North- Carlisle, D., 1963. Pillow breccias and their aquagene tuffs Quadra eastern Australia. Geological Society of Australia, Queensland Island, British Columbia. J. Geol. 70, 48±71. division, pp. 1±27. Cas, R.F., Landis, C.A., Fordyce, R.E., 1989. A monogenetic Henderson, R.A., 1983. Early Ordovician faunas from the Mount Surtla-type, surtseyan volcano from the Eocene-Oligocene Windsor Subprovince, northeastern Queensland. Mem. Assoc. Waiareka-Deborah Volcanics, Otago, New Zealand, a model. Aust. Palaeontol. 1, 145±175. Bull. Volcanol. 51, 281±298. Henderson, R.A., 1986. Geology of the Mount Windsor Subpro- Cas, R., Moore, C.L., Scutter, C., Yamagishi, H., 1996. Fragmenta- vince Ð a Lower Paleozoic volcano-sedimentary terrane in tion processes and water depth signi®cance of Miocene submar- the northern Tasman orogenic zone. Aust. J. Earth Sci. 33, ine ®re fountain deposits, Ryugazaki, Hokkaido, Japan. AGU 343±364. 1996 Western Paci®c Geophysics Meeting. Abstract W 126. Hon, K., Heliker, C., Kjargaard, J.I., Limu, O., 1988. Limu O Pele: a Clague, D.A., Holcomb, R.T., Sinton, J.M., Detrick, R.S., Torresan, new kind of hydroclastic tephra from Kilauea Volcano, Hawaii. M.E., 1990. Pliocene and Pleistocene alkalic ¯ood basalts on the Geol. Soc. Am. 20, A112±A113. sea¯oor north of the Hawaiian islands. Earth Planet. Sci. Lett. Hunns, S.R., McPhie, J., 1999. Pumiceous peperite in a submarine 98, 175±191. volcanic succession at Mount Chalmers, Queensland, Australia. Clague, D.A., Davis, A.S., Bischoff, J.L., Dixon, J.E., Geyer, R., J. Volcanol. Geotherm. Res. 88, 239±254. 2000. Lava bubble-wall fragments formed by submarine hydro- Jurado-Chichay, Z., Rowland, S.K., Walker, G.P.L., 1996. The volcanic explosions on Lo'ihi seamount and Kilauea Volcano. formation of circular littoral cones from tube-fed pahoehoe: Bull. Volcanol. 61, 437±449. Mauna Loa, Hawaii. Bull. Volcanol. 57, 471±482. Dear, J.F., 1974. Lower Ordovician graptolites from the Ravens- Kokelaar, B.P., 1982. Fluidization of wet sediment during the wood area, North Queensland. In: Denmead, A.K., Tweedale, emplacement and cooling of various igneous bodies. J. Geol. G.W., Wilson, A.F. 2Eds.), The Tasman Geosyncline Ð a Soc. London 139, 21±33. Symposium. Geological Society of Australia, Queensland Divi- Kokelaar, P., 1986. Magma±water interactions in subaqueous and sion, pp. 313±317. emergent basaltic volcanism. Bull. Volcanol. 48, 275±289. Dimroth, E., Cousineau, P., Leduc, M., Sanschagrin, Y., 1978. Kokelaar, B.P., Durant, G.P., 1983. The Submarine eruption and Structure and organisation of Archean subaqueous basalt erosion of Surtla 2Surtsey), Iceland. J. Volcanol. Geotherm. Res. ¯ows, Rouyn-Noranda area Quebec, Canada. Can. J. Earth 19, 239±246. Sci. 15, 902±918. Lowe, D.R., 1979. Sediment gravity ¯ows: their classi®cation and Dolozi, M.B., Ayres, L.D., 1991. Early Proterozoic, basaltic-ande- some problems of application to natural ¯ows and deposits. Soc. site tuff-breccia: downslope, subaqueous mass ¯ow transport of Econ. Paleontol. Miner. Spec. Pub. 27, 75±82. phreatomagmatically-generated tephra. Bull. Volcanol. 53, Macdonald, G.A., 1972. Volcanoes. Prentice Hall, New Jersey, pp. 477±495. 1±510. Duhig, N.C., Stoltz, A.J., Davidson, G.J., Large, R.R., 1992. Mangan, M.T., Cashman, K.V., 1996. The structure of basaltic Cambrian microbial and silica gel textures preserved in silica scoria and reticulite and inferences for vesiculation, foam iron exhalites of the Mount Windsor volcanic belt, Australia: formation, and fragmentation in lava fountains. J. Volcanol. their petrography, chemistry, and origin. Econ. Geol. 87, 764± Geotherm. Res. 73, 1±18. 784. Mattox, T.N., Mangan, M.T., 1997. Littoral hyrdovolcanic explo- Fisher, R.V., 1968. Puu Hou littoral cones Hawaii. Geol. Rundsch. sions: a case study of lava±seawater interaction at Kilauea 57, 837±864. Volcano. J. Volcanol. Geotherm. Res. 75, 1±17. K. Simpson, J. McPhie / Journal of Volcanology and Geothermal Research 10932001) 339±355 355

McClung, G., 1976. Early Ordovician 2Late Arenigian) graptolites Stolz, A.J., 1995. Geochemistry of the Mount Windsor Volcanics: from the Cape River Beds in the Rollston Range, Charters implication for the tectonic setting of Cambro-Ordovician Towers District. Queensland Govt. Miner. J. 77, 605±608. volcanic-hosted massive sul®de mineralization in northeastern Moore, J.G., Ault, W.U., 1965. Historic littoral cones in Hawaii. Australia. Econ. Geol. 90, 1080±1097. Paci®c Sci. XIX, 3±11. Walker, G.P.L., 1992. Morphometric study of pillow-size spectrum Mueller, W., White, J.D.L., 1992. Felsic ®re-fountaining beneath among pillow lavas. Bull. Volcanol. 54, 459±474. Archaean seas: pyroclastic deposits of the 2730 Ma Hunter Walker, G.P.L., Croasdale, R., 1972. Characteristics of some basal- Miner Group, Quebec, Canada. J. Volcanol. Geotherm. Res. tic pyroclastics. Bull. Volcanol. 35, 303±317. 54, 117±134. Wilson, L., Head, J.W., 1981. Ascent and emplacement of basaltic Philpotts, A., Lewis, C., 1987. Pipe vesicles Ð an alternative model magma on the Earth and Moon. J. Geophys. Res. 86, 2971± for their origin. Geology 15, 971±974. 3001. Schmincke, H.U., Sunkel, G., 1987. Carboniferous submarine Wilson, L., Par®tt, E.A., Head, J.W., 1995. Explosive volcanic volcanism at Herbornseelbach 2Lahn-Dill area, Germany). eruptions-VIII: the role of magma recycling in controlling the Geol. Rundsch. 76, 709±734. behaviour of Hawaiian-style lava fountains. Geophys. J. Int. Siebe, C., Komorowski, J.C., Navarro, C., McHone, J., Delgado, H., 121, 215±225. Cortes, A., 1995. Submarine eruption near Socorro Island, Wood, C.A., 1980. Morphometric evolution of cinder cones. J. Mexico: geochemistry and scanning electron microscopy Volcanol. Geotherm. Res. 7, 387±413. studies of ¯oating scoria and reticulite. J. Volcanol. Geotherm. Yamagishi, H., 1987. Studies on the Neogene subaqueous lavas and Res. 68, 239±271. hyaloclastites in southwest Hokkaido. Rep. Geol. Surv. Simpson, K., McPhie, J., 1998. Characteristics of subaqueous basal- Hokkaido 59, 55±117. tic andesite ®re fountain deposits: An example from the Mount Yamagishi, H., 1991. Morphological and sedimentalogical charac- Windsor Volcanics, northern Queensland, Australia. Geol. Soc. teristics of the Neogene submarine coherent lavas and hyalo- Aust. Abstr. 49, 409. clastites in Southwest Hokkaido, Japan. Sediment. Geol. 74, 5± Smith, T.L., Batiza, R., 1989. New ®eld and laboratory evidence for 23. the origin of hyaloclastite ¯ows on seamount summits. Bull. Volcanol. 51, 96±114.