Rinehart and Ross (1957) mapped the Bishop Tuff within the Casa Diablo Mountain Quadrangle. Putnam (1960) discussed the evolution of the Owens River and Rock Creek Gorges. Sheridan (1965) described the mineralogy and development of the ash-flow zones. Dalrymple et al. (1965) established a 0.71 Ma potassium-argon age of the Bishop Tuff. Bateman (1965) mapped Bishop Tuff of the southern Volcanic Tableland and discussed its structural and erosional modifications. Sheridan (1970) reported "fossil" fumaroles and relationships to ash-flow zones. Crowder and Sheridan (1972) mapped ash-flow sub-units of the Bishop Tuff in the White Mountain Peak Quadrangle. Ragan and Sheridan (1972) analyzed the compaction by welding of the tuff. Bailey et al. (1976) presented the relationship of the Bishop Tuff to regional volcanism and Long Valley caldera. Hildreth (1979) showed that the Bishop Tuff successively tapped deeper, less evolved levels of a zoned magma chamber, and identified several flow "lobes". Horn and Sheridan (1983) calculated pyroclastic flow emplacement temperatures modelled from density profiles. Hildreth and Mahood (1986) analyzed the lithic inclusions to infer an eruption sequence and possible source vents of the Bishop Tuff.
The Bishop Tuff is located in a geologically young and historically active area of volcanism and tectonism. This area in eastern California, adjacent to Nevada (Fig. 1), lies at the west edge of the Basin and Range Province. Extensional tectonics created topography that not only controlled the paths of pyroclastic flows that deposited the Bishop Tuff, but also faulted and warped the ignimbrite after it had cooled.
Pre-Tertiary rocks of the region include folded late Pre-Cambrian or early Cambrian to Triassic sedimentary and Mesozoic volcanic rocks of the White Mountains. Metamorphosed equivalents of the earlier sedimentary sequence are preserved as roof pendants and isolated bodies in the Sierra Nevada. In Mesozoic time the area was intruded by felsic quartz-bearing granitic plutons, ranging from granodiorite to alaskite (Bateman 1965), of the Sierra Nevada batholith, the root of a magmatic arc. The batholith was uplifted in Tertiary time and stripped of its cover, producing a range with relief of a few thousand feet. Final uplift was achieved before the onset of Pliocene glaciation (Christensen 1966).
By approximately 3 Ma the White Mountains and Sierra Nevada stood high over the intervening Owens Valley. Long Valley formed an embayment into the north-south trend of the Sierra Nevada, an "encroachment of Basin and Range extension into thick and stable Sierra Nevada crust" (Hill et al. 1985). A magma body below Long Valley would be the source of later volcanism. Late-Tertiary volcanism was largely basaltic, but also included andesitic and rhyodacitic compositions. Though not direct products from the Long Valley magma chamber, these eruptions represent early differentiation of a mafic parent (Bailey 1987). The most extensive exposures of basalt, a lava plain covering some 770 km2, are north of Adobe Valley (Gilbert et al. 1968). In the Owens River Gorge basaltic lavas and cinder layers blanket a granitic basement ridge. The basalts had flowed upon an eroded granitic terrain of moderate relief.
The rhyolite complex of Glass Mountain represents the first volcanic products from the Long Valley magma chamber (Fig. 1). Two periods of eruption, 2.13 to 1.20 Ma and 1.10 to 0.79 Ma (Metz and Mahood 1985), produced more than 15 km3 of domes, lava flows, breccias, pumice, and ash. High-silica rhyolite compositions, 76.8 to 77.9% SiO2 (Metz and Mahood 1985), include black crystal-poor obsidian, pink and grey lava, and white pumice. The eruptions were not sufficient to trigger caldera collapse (Metz and Mahood 1985).
Pleistocene to Recent age glacial deposits are found in the vicinity of the Sierra Nevada. Pleistocene Sherwin Till forms a thick deposit near the head of Rock Creek Gorge, where it overlies the granitic and metamorphic basement. It also occurs below the Bishop Tuff, as observed in the roadcut of US 395 known as the "Big Pumice Cut" (Sharp 1968) and in a tunnel excavated for the Los Angeles Department of Water and Power aqueduct (Putnam 1960). Weathering of granitic boulders in the till indicates a gap of tens of thousands, but less than 100,000 years, between the glaciation and the Bishop Tuff eruption (Sharp 1968). Sherwin deposits also are found between the basalts and the Bishop Tuff in the Owens River Gorge. The sandy, sorted matrix indicates that these are outwash deposits rather than till (Sharp 1968).
The Bishop Tuff was deposited more than 700,000 years ago, starting with a Plinian eruption followed by a series of pyroclastic flows from multiple source vents within Long Valley. The caldera, a 17 by 32 km elliptical structure, formed as the roof foundered above and collapsed into a partially depleted magma chamber.
Following the collapse of the caldera, Long Valley was filled by a Pleistocene lake (Mayo 1934). The waters reached a maximum elevation indicated by gravel capped terraces east of the caldera, and was emptied by 50,000 to 100,000 years B.P. (Bailey 1987) . The evacuation of Long Valley Lake caused or contributed to the incision of Owens River Gorge (245 m deep) and Rock Creek Gorge (185 m deep) into the plateau of Bishop Tuff to the southeast. The mechanism by which lake drained has not been resolved; it has been attributed to both headward growth of the Owens River (Mayo 1934; Putnam 1960) and lake overflow (Rinehart and Ross 1957; Bailey et al. 1976). Following caldera collapse volcanism continued in Long Valley with smaller volume rhyolite eruptions. Resurgent doming of the western part of the caldera occurred 40,000 to 100,000 years after the caldera formed (Bailey et al. 1976). Later activity included rhyolitic and rhyodacitic magmas (including the 0.18 Ma Mammoth Mountain complex), and basalts of the Devil's Postpile and June Lake area (Bailey et al. 1976). Holocene eruptions of the Inyo and Mono chain of craters and domes occur in an arc that intersects the northwest margin of Long Valley caldera, possibly the initiation of a new caldera cycle.
Pyroclastic flows originate from explosive fragmentation of volatile-rich magma. These gravity driven flows owe their mobility to rapidly expanding gases, and are capable of traveling hundreds of kilometers. The height at which the flows originate is probably the dominant factor controlling flow distance (Malin and Sheridan 1982). Once emplaced, cooling and compaction of the hot deposits can alter an initial homogeneous mass of glass, ash, and crystals into a layered unit of rock indistinguishable from one deposited as a lava flow. Such textural changes might be considered metamorphic (Sheridan 1965). Smith (1960a, 1960b) and Ross and Smith (1961), based on years of field observations, outlined the basis for modern concepts of the processes which produce zonation in welded tuffs. Riehle (1973) mathematically modelled the cooling and compaction of rhyolitic ash-flows.
Tuff is an excellent insulator, retaining heat in thick deposits that compact, or weld, under its own weight or overburden. The high temperatures also promote secondary crystallization within the sheet-like deposits. The ultimate product is a layered formation with varying degrees of compaction and crystallization. The Bishop Tuff (Fig. 3) displays all of the ash-flow zones described by Smith (1960b).
Welding of ash-flow tuffs is the cohesion of the hot glassy fragments, ranging from simple sticking together to plastic deformation that effectively reduces loose ash and pumice to a rock of nearly zero porosity (Smith 1960a). While a number of variables influence welding, the observed textural changes can be interpreted in terms of two: emplacement temperature and thickness of the unit (Smith 1960a). A thin sheet, despite a small load, may weld if it is emplaced at a high temperature. A thick sheet, however, may weld at lower temperatures due to insulation and greater load stress. Smith's (1960b) diagrams of zonation in sheets of various thickness and emplacement temperatures have served as excellent models for observed ignimbrites.
Compaction by uniaxial load stress is the mechanism of welding (Ragan and Sheridan 1972). The "Y"-shaped glass shards are compressed, aligned in the plane perpendicular to the load stress, and may bend around more rigid objects such as crystals or lithic inclusions. Carr (1981) used scanning electron microscopy to observe incipient welding between shards as lines of imperfections that develop near dust and gas at the juncture. Increasing degrees of welding reduces porosity, producing a denser rock. The pumice lapilli flatten and darken in color, their vesicles are squeeze closed, and ultimately reduce to obsidian-like lenses or "fiamme".
Three welding zones are described; zones of dense welding, partial welding, and no welding (Smith 1960b). The idealized cooling of a single ash-flow contains a central densely welded zone that grades to partly welded and non-welded zones toward the top and bottom surfaces and at distal margins. A "cooling unit" (Smith 1960b), represents one or more pyroclastic flows that cool as a package, with the zonation pattern just described. Smith (1960b) demonstrated how the zones pinch out with distance from the source, as functions of both decreasing thickness and temperature. Typically, zone boundaries are gradational, and not all zones are always present.
The non-welded zone most closely represents the original texture of the deposit. Incipient welding may be present but only recognizable microscopically. The density ranges from that of pumice (0.8 g/cm3) to weakly indurated tuff (1.4 g/cm3), greater than 40% porosity assuming bulk density of 2.45 g/cm3 (Ragan and Sheridan 1972). An entire ash-flow may be non-welded, or the zone may only be found at the top and bottom of the sheet and its thin margins. Generally unconsolidated and subject to rapid erosion, the non-welded zone is usually only found in young ash-flow tuffs. The densely welded zone is defined by porosities of less than 10% (densities from 2.2 to 2.45 g/cm3) and represents compaction of unconsolidated material by a factor of greater than two (Ragan and Sheridan 1972). Its position, the hottest part of the sheet, is displaced below the center due to the greater rate of heat loss at its upper surface (exposed to the atmosphere) relative to the heat loss at its lower surface. The partly welded zone is transitional from non-welded tuff to densely welded tuff with a range of densities from 1.4 to 2.2 g/cm3.
Two processes produce crystalline zones in welded tuffs, devitrification and vapor-phase crystallization. Both yield similar minerals: silica polymorphs and alkali feldspar. The distinction is that devitrification forms crystals within the glass fragments while vapor-phase crystallization occurs in the pore space between fragments and in pumice vesicles. An idealized ash-flow sheet has a crystalline core inside of a vitric envelope (Smith 1960b), demonstrated by the Bishop Tuff (Sheridan 1965) and the Yucca Mountain Tuff, Nevada (Lipman and Christiansen 1964).
Like any glass, vitric tuff is metastable and over time it crystallizes. Lofgren (1971) produced spherulites and axiolites of quartz, alkali feldspar, plus minor pyroxene, muscovite, and zeolites through experimental devitrification of natural rhyolite glass. Initial crystallization of anhydrous minerals increases the amount of H2O diffusing out of the glass. An H2O rich "halo" lowers the viscosity of the surrounding glass, promoting a spherical area of crystallization (Friedman and Long 1984). Complete crystallization to this "spherulitic stage" (Lofgren 1971) may drive off H2O that escapes as fumarolic gas (Friedman and Long 1984).
In rhyolitic welded tuffs, the typical products of devitrifcation are fine intergrowths of cristobalite and alkali feldspar (Smith 1960b), often only identifiable from x-ray analyses. In the glass shards axiolitic crystals grow inward, often leaving a central black line or discontinuity (Ross and Smith 1961). The Bishop Tuff includes intergrown fibers of sanidine and cristobalite, plus coarser quartz and feldspar (Gilbert 1938; Sheridan 1965). The conditions that promote welding, high temperatures and thick deposits, are also conducive to devitrification. Therefore, the densely welded zone is often devitrified.
The zone of vapor-phase crystallization is indicated by the growth of tridymite and alkali feldspar within open spaces (Smith 1960a). Compaction in the zones of welding "squeezes" out volatiles, which migrate upward and induce crystallization upon reaching the more porous partly- or non-welded zones. Volatiles include primary magmatic gases from the pyroclastic flows, vaporized ground water, and gases diffused from the glass following devitrification to anhydrous minerals. Bulk chemical composition of the tuff is not grossly altered by crystallization (Sheridan 1970; Lipman and Christiansen 1964), implying only small scale element transport. Since vapor phase crystallization requires room to grow, it does not occur in the densely welded zone. Most vapor-phase crystallization takes place in the upper non-welded or partly welded zone (Fig. 3). This relationship, and the fact crystallization prior to welding would produce rigid objects that inhibit compaction, suggests that the crystallization occurs after welding (Sheridan 1970).
Vapor-phase crystallized tuff is a brilliant white, brittle rock. Fenner (1948) used the name "sillar" for Pleistocene tuffs in the Arequipa region of Peru which were indurated solely by crystallization. However, Smith (1960a) suggests that the Peruvian tuff was partly welded before vapor-phase crystallization. Smith and Bailey (1966) report tridymite, alkali feldspar, and iron oxides from the vapor-phase zone of the Bandelier Tuff. Carr (1981) observed alkali feldspar, cristobalite, and mica as vapor-phase minerals in the upper non-welded zone of the Matahina ignimbrite (New Zealand). Mafic phenocrysts of biotite, hornblende, and orthopyroxene may be replaced by opaque oxides (Smith 1960b). Gilbert (1938) described rims of clear feldspar on sanidine crystals, only in the upper portions of the Bishop Tuff, suggesting "some attack by rising gases has affected the feldspars after the emplacement of the deposit."
Smith (1960b) noted a third crystallization zone, since "any ash-flow that contains hot gas or is emplaced at temperatures high enough to crystallize on cooling should give off gas at its surface." The zone of fumarolic alteration results from a form of vapor-phase crystallization, with very intense local alteration near the vents. These fumaroles were outlets for the vapor-phase zone, where gases migrate through early-formed fractures in the upper part of the sheet. The vapor-phase includes H2O, HF, HCl, H2SO4, and CO2 (Sheridan 1970). A fumarole in the Valley of Ten Thousand Smokes, Alaska, was 99.7% steam, exiting originally at an estimated temperature of 800-900 ¡C (Lovering 1957). Like the vapor-phase zone, crystallization in fumarolic zone produces tridymite and alkali feldspar, plus cristobalite, a number of iron phases and several other minerals (Table 1). While the fumarolic zone should occur in most welded ash-flows, Smith and Ross (1961) noted a lack of actual examples. Since then, however, a number of ash-flows with fumarolic zones have been described. In the Bishop Tuff, past fumarolic activity is marked by small round hills or "fumarolic mounds" (Putnam 1960; Bateman 1965; Sheridan 1970). The crystallization near the vents locally indurated the surrounding non-welded tuff. These areas persist as positive features once erosion stripped the surrounding loose un-altered tuff. Griggs (1922) described similar features, "pimple hills", in the Valley of Ten Thousand Smokes Tuff. Although Griggs (1922) attributed the hills to sand accumulations near the condensing steam of fumaroles, Sheridan (1970) and Keith (1984) suggest that the "pimple hills" were indurated by crystallization and thus are analogs for the Bishop Tuff mounds. Fumarolic "pipes" and alteration have also been reported in the Bandelier Tuff (Smith and Bailey 1966), the pinnacles at Crater Lake (Williams 1942), and the Rio Caliente ignimbrite (Mahood 1980; Wright 1981).
The Bishop Tuff includes 300 km3 of air-fall deposit and 500 km3 of pyroclastic flows, of which 350 km3 is buried by alluvium within Long Valley caldera (Bailey et al. 1976). The present compacted thickness is 150-200 m. This rhyolitic tuff, 75 to 77% SiO2 (Sheridan 1965; Hildreth 1979), has been dated at 0.74 Ma by K-Ar (most recently by Izett et al. 1988) and fission track methods (Izett and Naeser 1976).
The Bishop Tuff represents withdrawal from the top of a single magma body below Long Valley. Hildreth's (1979) Fe-Ti oxide magmatic temperatures and trace element data show the eruptions tapped a compositionally zoned magma chamber. The earliest Bishop Tuff sampled magma at a temperature of 720 ¡C while the later eruptions originated from deeper levels where the temperatures were 790 ¡C (Hildreth 1979). The Long Valley magma was also zoned with respect to H2O, which increased from 4 to 6% toward the top of the magma chamber (Anderson et al. 1989).
Crystals of the Bishop Tuff include quartz, sanidine, oligoclase, and biotite, both in pumice and in the matrix. Lesser amounts of zircon, apatite, and magnetite also occur. Pyroxenes are present in trace amounts for flows erupted above 737 ¡C and allanite is present for the ones erupted at less than 763 ¡C (Hildreth 1979). Lithic inclusions of the pre-Tertiary basement range in abundance from 1 to 3% (Hildreth and Mahood 1986), representing wall rocks of the conduit or fragments plucked by pyroclastic flows.