Friction Melting, Catastrophic Dilation and Breccia Formation Along Caldera Superfaults

By Kokelaar, Peter

At Glencoe caldera volcano, friction melts and magmas transformed explosively to froth or spray where they encountered rapid decompression in dilatant sections of superfaults. The friction melts were blasted upwards, plastering free surfaces, and were rapidly followed by fragmented magma and then liquid magma that formed intrusions. Irregular contacts of fault intrusions record explosive disruption of hydrothermal systems cut by the dilational faults, with lithic breccias removed from the rapidly formed cavities before the arrival of melt spray. Layers of lithic breccia are common in ignimbrites around calderas and usually show incorporation of hot and hydrothermally altered rocks. Such layers may specifically reflect initial superfault dilation during caldera collapse. This paper proposes two previously unrecognized processes resulting from rapid, large-scale faulting: decompression- fragmentation of friction melt and explosive excavation of a decompressed hydrothermal system. Caldera collapse during large- magnitude volcanic eruptions can involve fault displacements of many hundreds of metres in hours or days and, where the rocks are robust and remain in contact, friction melting can occur; typical rates of displacement on such ‘superfaults’, in the range – 0.1-10 cm s^sup – 1^, are sufficient to cause this, and they could be faster (Spray 1995, 1997). Caldera faults commonly dip outwards, especially at shallow levels (Branney 1995), or are non-planar, such that subsidence towards the magma chamber causes the fault surfaces to separate locally and form voids into which fluids can flow. Such voids are readily exploited by caldera-fault intrusions. Lithic breccias are normally erupted along faults at caldera volcanoes, for example at Campi Flegrei (Italy), Crater Lake (USA) and Santorini (Greece), where they evidently derive from considerable depths and include hydrothermally altered fragments (e.g. Druitt 1985; Druitt & Bacon 1986; Rosi et al. 1996). Such lithic breccias within intracaldera and outflow ignimbrites are often interpreted as marking caldera collapse, but how they formed and were transported through an eruptive conduit has not been addressed closely. This paper relates such lithic breccia layers to explosive conduit processes that accompany caldera collapse.

Geological setting. Rugged terrain at Glencoe in Scotland exposes inner parts of a Siluro-Devonian caldera volcano set on a foundation of metasedimentary rocks (Fig. 1). Intra-caldera volcanic rocks are

Clough et al. (1909) showed that the fault intrusions are discontinuous and in northern outcrops they detected an outer ‘Early Fault’ and an inner ‘Main Fault’ (Fig. 1), both locally intruded and with ‘flinty crush-rock’ along contacts. The flinty rock was subsequently interpreted as pseudotachylyte, formed by frictional melting (Bailey 1960), although this was disputed and alternative origins involving particles streaming in a flow of magmatic gas were proposed (Reynolds 1956; Roberts 1966). Excellent exposures of the faults, flinty crush-rock and fault intrusions occur at the peak named Stob Mhic Mhartuin (Fig. 1).

Peripheral caldera-fault and intrusion system. Outer elements of the caldera-fault system define an incomplete polygon, 14 km x 8 km, with additional major strands to the SW and north (Fig. 1). What was formerly described in terms of a ‘ring’ is not a continuous annular structure, but comprises a linkage of faults. Different faults were active and intruded by magma at different times and some were reactivated. Along northern and eastern sections the faults dip outwards, and elsewhere they are subvertical or dip steeply inwards. The minimum cumulative subsidence within the caldera was 1.4 km and late subsidence along the northern faults considered here was

The fault intrusions,

The contrast between the irregular outer contacts and the planar inner contacts of the fault intrusions has not been understood. The planar inner contacts, in places lined with fault rocks for hundreds of metres, are fault surfaces, but the outer contacts reflect substantial removal of country rock. Given that the volcanotectonic subsidence was many tens to several hundreds of metres, the inner fault surfaces were formed at significantly shallower levels than the juxtaposed outer irregular contacts.

Flinty crush-rock. Flinty crush-rock, an extremely fine-grained, black to brown rock with undulose and swirling laminations, occurs widely as a centimetre-scale veneer along the contacts of fault intrusions, and in veins that extend irregularly away from these contacts, especially outwards, away from the volcano centre. A thin band of reddish porphyritic rhyolite commonly lies between the flinty crush-rock and the main body of the fault intrusion.

At Stob Mhic Mhartuin (Figs 1 and 2), a lenticular body of porphyritic monzodiorite,

The rhyolite adjacent to the flinty crush-rock is mostly only a few centimetres thick, although locally

At the highly uneven outer contact of the fault intrusion, flinty crush-rock with rhyolite also occurs, but here the succession of lithologies that exists widely along the inner contact, including microbreccias and ultracataclasite, is only locally present (Fig. 2). Mostly the veneer of flinty crush-rock and rhyolite, or solely rhyolite, coats irregular blocky contacts and penetrates for several metres along fractures oriented at high angles to the general fault orientation (Fig. 2). Irregular veins,

Emplacement model. It is proposed that the flinty crush-rock (pure end-member) originated by friction-induced melting during early movement on steep sections of faults, and that both the friction melt and the rhyolite were transported upwards as fragmented liquids during fault dilation and consequent catastrophic decompression (Fig. 4). In friction melting, the least robust minerals such as amphiboles and micas melt preferentially, because they are most susceptible to grain-size reduction (Maddock 1992; Spray 1992). Hence friction melts are commonly more ferromagnesian and contain more water than their protoliths (e.g. Allen 1979). Those at Glencoe would have had sufficient dissolved water (>/=0.5- 1 wt%) to become fragmented by catastrophic volatile exsolution, and they would initially have had viscosity akin to that of basalt at 1200 [degrees]C, even when their protoliths were of intermediate to acid composition (Spray 1993). A melt with dissolved volatiles will transform into foam or spray by explosive volatile exsolution when decompressed (e.g. Martel et al. 2001). Decompression will occur where a melt enters rapidly dilating fractures or releasing bends on active faults, or during subsidence on outward-dipping caldera faults. Foam or spray of disrupted melt will be entrained by expanding gas and impregnate any accessible cavity, where, at high temperatures, it will coat surfaces and coalesce into a continuous liquid, perhaps initially vesicular. Decoupled gas will continue upwards whereas trapped gas will escape via some combination of shear-related permeability and diffusion (e.g. Stasiuk et al. 1996). The rhyolite represents fragmented magma that invaded the fault cavity immediately after the first friction melt was deposited and before arrival of the magma to form the fault intrusion. The presence of feldspar phenocrysts in the flinty crush-rock and rounded quartz grains in the rhyolite (Fig. 3) is the result of the mixing of coexisting fragmented melts; the fragmentation of the feldspar phenocrysts is like that resulting from explosive magma fragmentation and particulate transport (e.g. Best & Christiansen 1997; see Allen & McPhie 2003). That the deposited friction melt and rhyolite had closely similar rheologies is indicated by the patterns of intimate mingling of the two (Fig. 3), like two paints stirred together. The proposed emplacement model can account for both the evident mobility of the material that formed the flinty crush-rock and rhyolite, and the lithological zonation (Fig. 2).

The friction melt represented by the flinty crush-rock around the Main Fault intrusion probably formed at a level significantly deeper than the present outcrop. The flinty crush-rock along the outer, irregular contacts cannot have formed there and the outward dip of the footwall fault-plane would have led to dilation with termination of frictional heating soon after initiation of slip. The thorough mixing of fragmental friction melt and rhyolite implies that the components ascended some considerable distance together before deposition upon the walls of the developing fault cavity; they may have been blasted up the fault for several hundreds of metres.

The irregular outer intrusion contacts are interpreted as having developed through fault dilation in the presence of hot, pressurized water. At Stob Mhic Mhartuin, hydrothermal Iy altered breccias predate downthrow on the Main Fault and it is postulated that part of a hydrothermal system here boiled, disintegrating the (outer) hanging-wall rocks, following rapid dilation on the Main Fault (Fig. 4b). Rapid dilation of a fault that transects a permeable hydrothermal system will cause explosive transformation of superheated water to steam and vigorous expansion of vapour, with locally steep pressure gradients in fluid pathways around the low- pressure cavity. Such explosivity and fluid flow would blast the hot- water-bearing country rock towards the dilating fault. The lining of the irregular outer contact with flinty crush-rock and rhyolite (Fig. 2) shows that excavation occurred before the arrival of the fragmented melts, which is strong evidence that it occurred at an explosive rate in early stages of dilation (Fig. 4). In contrast, the relatively cool and dry rock mass that subsided from much shallower levels was less fragmented and a planar fault surface was thus presented to the inner side of the intrusion.

Surface evidence for this explosive activity is not seen at Glencoe, because the volcanic succession preserved there mainly predates the peripheral fault and intrusion system (Kokelaar & Moore 2006). Nevertheless, in caldera-related ignimbrites, layers of lithic breccia with hydrothermally altered clasts probably originate from, and thus mark the time of, initial dilation of caldera superfaults.

Discussion. Reynolds (1956) and Roberts (1966) doubted that a friction melt could be sufficiently mobile to be injected /=0.1-10 cm s^sup -1^) is sufficient to cause frictional melting (Spray 1995, 1997), pseudotachylyte is to be expected where robust rocks have been involved, as well as intrusive welded tuffs. Although spherulitic devitrification textures occur (rarely), traces of original vitroclastic or globular particles in the flinty crush-rock and rhyolite are not evident. High temperatures already at the site, or resulting from the intrusion, probably caused complete particle coalescence and diffusive gas loss.

Conclusions. Field relationships around caldera-superfault intrusions at Glencoe show evidence of almost simultaneous faulting- melting and dilation-explosion accompanying caldera collapse. Decompressed friction melts closely followed by magmas were blasted as spray along dilatant parts of faults, plastering the walls of the cavities, where there was coalescence and localized viscous flow. Before arrival of the melts, hot-waterbearing wall rocks exploded into the dilatant faults, adding space for subsequent passive intrusions of liquid magma and supplying debris for lithic breccias. Such aggressive excavation along dilatant faults is likely to be a major mechanism for supply of lithic breccias into eruption conduits during caldera-forming eruptions. Consequently, lithic breccia layers in ignimbrites can indicate the time of initial dilation of caldera superfaults. It is conceivable that rock-bursting, as a result of sudden loading in the hanging wall, may contribute to the excavation (see Bjornerud & Magloughlin 2004), although no clear evidence for this has been recognized and it is uncertain whether this could operate at scales and rates sufficient to account for the rock removed.

The mobility of friction melts formed by faulting can be far greater than previously conceived, although migration over tens to perhaps hundreds of metres may only occur in superfault settings via melt fragmentation. Flinty crush-rock is known from contacts of outward-dipping fault intrusions at other deeply dissected caldera volcanoes; for example, in the Huaura Ring Complex of the Peruvian Coastal Batholith (Bussell et al. 1976) and at Mount Aetna Cauldron, Colorado (Johnson et al. 1989), where it is intimately mingled with igneous constituents and is difficult to interpret without invoking substantial mobility, as at Glencoe.

The author is grateful to M. Branney, M. Howells, J. Spray and V. Troll for comments that improved early versions of this paper, and to K. Lancaster for drafting the figures.

References

ALLEN, A.R. 1979. Mechanism of frictional fusion in fault zones. Journal of

Structural Geology, 1, 231-243.

ALLEN, S.R. & MCPHIE, J. 2003. Phenocryst fragments in rhyolitic lavas and lava domes. Journal of Volcanology and Geothermal Research, 126, 263-283.

BAILEY, E.B. 1960. The Geology of Ben Nevis and Glen Coe and the Surrounding Country, 2nd. Memoir of the Geological Survey, Sheet 53, Scotland.

BEST, M.G. & CHRISTIANSEN, E.H. 1997. Origin of broken phenocrysts in ash-flow tuffs. Geological Society of America Bulletin, 109, 63-73.

BJORNERUD, A. & MAGLOUGHLIN, J.F. 2004. Pressure-related feedback processes in the generation of pseudotachylytes. Journal of Structural Geology, 26, 2317-2323.

BRANNEY, MJ. 1995. Downsag and extension at calderas-new perspectives on collapse geometries from ice-melt, mining, and volcanic subsidence. Bulletin of Volcanology, 57, 303-318.

BUSSELL, M. A., PITCHER, W.S. & WILSON, P.A. 1976. Ring complexes of Peruvian Coastal Batholith – long-standing sub-volcanic regime. Canadian Journal of Earth Sciences, 13, 1020-1030. CAMACHO, A., VERNON, R.H. & FITZ GERALD, J.D. 1995. Large volumes of anhydrous pseudotachylyte in the Woodroffe Thrust, Eastern Musgrave Ranges, Australia. Journal of Structural Geology, 17, 371-383.

CLOUGH, C.T.H., MAUFE, H.B. & Bailey, E.B. 1909. The cauldron subsidence of Glen Coe and the associated igneous phenomena. Quarterly Journal of the Geological Society of London, 65, 611-678.

CUREWITZ, D. & KARSON, J.A. 1999. Ultracataclasis, sintering, and frictional melting in pseudotachylytes from East Greenland. Journal of Structural Geology,2l, 1693-1713.

DRUITT, T.H. 1985. Vent evolution and lag breccia formation during the Cape Riva eruption of Santorini, Greece. Journal of Geology, 93, 439-454.

DRUITT, T.H. & BACON, CR. 1986. Lithic breccia and ignimbrite erupted during the collapse of Crater Lake Caldera, Oregon. Journal of Volcanology and Geothermal Research, 29, 1-32.

JOHNSON, CM., SHANNON, J.R. & FRIDRICH, CJ. 1989. Excursion 15?: roots of ignimbrite calderas: batholithic plutonism, volcanism and mineralization in the Southern Rocky Mountains, Colorado and New Mexico. In: Chapin, CE. & Zidek, J. (eds) Field excursions to volcanic terrones in the western United States, Volume 1: Southern Rocky Mountain Region. New Mexico Bureau of Mines and Mineral Resources Memoir, 46, 275-302.

KOKELAAR, B.P. & MOORE, I.D. 2006. Classical Areas of British Geology: Glencoe Caldera Volcano, Scotland. British Geological Survey, Keyworth.

LIN. A. 1999. Roundness of clasts in pseudotachylytes and cataclastic rocks as an indicator of frictional melting. Journal of Structural Geology, 21, 473-478.

MADDOCK, R.H. 1992. Effects of lithology, cataclasis and melting on the composition of fault-generated pseudotachylytes in Lewisian gneiss, Scotland. Tectonophysics, 204, 261-278.

MARONE, C & SCHOLZ, CH. 1989. Particle-size distribution and microstructures within simulated fault gouge. Journal of Structural Geology, 11, 799-814.

MARTEL, C, DINGWELL, D.B., SPIELER, O., PICHAVANT, M. & WILKE, M. 2001. Experimental fragmentation of crystal- and vesicle-bearing silicic melts. Bulletin of Volcanology, 63, 398-405.

MOORE, I. & KOKELAAR, P. 1997. Tectonic influences in piecemeal caldera collapse at Glencoe Volcano, Scotland. Journal of the Geological Society, London, 154, 765-768.

MOORE, I. & KOKELAAR, P. 1998. Tectonically controlled piecemeal caldera collapse: a case study of Glencoe volcano, Scotland. Geological Society of America Bulletin, 110, 1448-1466.

REYNOLDS, D.L. 1956. Calderas and ring-complexes. Verhandelingen van hel Koninklijk Nederlands Geologisch Mijnbouwkundig Genootschap, 16, 355-379.

ROBERTS. J.L. 1966. Emplacement of main Glencoe fault-intrusion at Stob Mhic Mhartuin. Geological Magazine, 103, 299-316.

ROSI, M., VEZZOLI, L., ALEOTTI, P. & DECENSI, M. 1996. Interaction between caldera collapse and eruptive dynamics during the Campanian Ignimbrite eruption, Phlegraean Fields, Italy. Bulletin of Volcanology, 57, 541-554.

SIBSON. R.H. 1975. Generation of pseudotachylyte by ancient seismic faulting. Geophysical Journal of the Royal Astronomical Society, 43, 775-794.

SPRAY, J.G. 1992. A physical basis for the frictional melting of some rock-forming minerals. Tectonophysics, 204, 205-221.

SPRAY, J.G. 1993. Viscosity determinations of some frictionally generated silicate melts-implications for fault zone rheology at high-strain rates. Journal of Geophysical Research: Solid Earth, 98, 8053-8068.

SPRAY, J.G. 1995. Pseudotachylyte controversy-fact or friction. Geology, 23, 1119-1122.

SPRAY, J.G. 1997. Superfaults. Geology, 25, 579-582.

STASIUK, M.V., BARCLAY, J., CARROLL, M.R., JAUPART, C, RATTE, J.C, SPARKS, R.S.J. & TAIT. S.R. 1996. Degassing during magma ascent in the Mule Creek vent (USA). Bulletin of Volcanology, 58, 117-130.

Received 26 April 2006; revised typescript accepted 13 February 2007.

Scientific editing by Ken McCaffrey

PETER KOKELAAR

Earth and Ocean Sciences Department, University of Liverpool, Liverpool L69 3BX, UK

(e-mail: p.kokelaar@liv.ac.uk)

Copyright Geological Society Publishing House Jul 2007

(c) 2007 Journal of the Geological Society. Provided by ProQuest Information and Learning. All rights Reserved.