The nonvesicular and blocky (Lvv) texture and black to brown color of some vitric clasts suggests that they are mafic to intermediate in composition. A small percentage of these fragments (Table T1) could be attributed to magma mixing or, more likely, to incorporation of lithic fragments during eruptions. The range in the staining characteristics of the colorless glassy fragments likely reflects their range from more intermediate (Ca) to dacitic (Ca + K) to rhyolitic (K) compositions. Note that as shown in Table T3, there were only two cases where most of the fragments took only the Ca stain and many more where the majority took just the K stain. Most of the samples are mixtures of fragments with both Ca and K stain. Phenocryst assemblages in these ashes are also consistent with intermediate to felsic sources.
The ash textures observed in these samples are likely products of vesiculation of silicic magmas during Plinian to Phreatoplinian eruptions (Fisher and Schmincke, 1984; Heiken and Wohletz, 1985). Specific shapes of pyroclasts are dependent on the size, shape, and density of vesicles, which in turn reflect the magma composition, temperature, and volatile content just prior to eruption (Heiken and Wohletz, 1985). Shards can be the product of explosive disintegration of pumice during frothing or bubble coalescence (Heiken and Wohletz, 1985).
The above interpretation assumes that all the ash beds on Shatsky Rise are the distant fallout of large eruption plumes. An alternative source for submarine ash is the production of fines during abrasion of floating pumice rafts (Fisher and Schmincke, 1984). The presence of rounded pumice gravel-sized clasts in Shatsky ash-bearing units makes pumice comminution a viable source of ash. Unfortunately, we know of no studies comparing grain size or morphology of shards produced by this process as opposed to air fall processes; however, it is likely that pumice rafts would produce coarser material (Fisher and Schmincke, 1984). Based on data from Site 1208 (see below), we suspect that enrichment in vesiculated glassy fragments is also an indicator of pumice rafts.
Geographical considerations indicate several possible sources of ash delivered to Shatsky Rise and recovered during ODP Leg 198. First, airborne ash could have been transported from either the Japan or Kurile magmatic arc systems and pumice from eruptions in the Kurile arc. These were the major sources of pyroclastic debris at Site 810 that Natland (1993) considered, although both he and Cao et al. (1995), in a later study of northwestern Pacific DSDP and ODP sites, depicted the Kuroshio Current extending to Shatsky Rise from the region of the Izu-Bonin island arc. The Kuroshio Current, which passes across Shatsky Rise, is a known conduit of pumice rafts from the Mariana and Izu-Bonin magmatic arcs to the south (Lisitzin, 1972). The importance of this possible source may have been overlooked by Natland (1993) because the thick pumice deposits associated with the Sumisu Rift in the Izu-Bonin arc (Leg 126; Taylor, Fujioka, et al., 1990; Nishimura et al., 1991) had not yet been discovered (drilled) when Natland published his work. Thus, pumice and (abraded) ash derived from pumice rafts could have been transported to Shatsky Rise from the Izu-Bonin arc.
Although previous workers (e.g., Marsaglia, 1992) defined compositional modes for volcaniclastic sediments, including pumice-raft debris, along the length of the Japan and Izu-Bonin arcs, they did not use the pyroclast classification scheme outlined in this study. Therefore, we collected textural data from these samples and compared the Izu-Bonin (Leg 126) and Shatsky Rise (Leg 198) ash records. The Izu-Bonin samples, when plotted on the same ternary plots with Shatsky Rise samples (Figs. F5, F6, F7), have a higher vesiculated and lower shard proportion. The lower proportion of vesiculated fragments at Leg 198 sites could possibly be due to the physical alteration of the ash grains during transport and subsequent deposition and reworking.
The above discussion of pyroclast generation and textural attributes implies that individual eruptions may produce characteristic ash deposits. When correlating across large areas, however, one must take into consideration possible eolian fractionation, whereby wind-transported ash is sorted by size and composition. This effect has mainly been called on to explain bulk changes in ash composition as related to the concentration of denser lithic fragments and crystals nearer the vent and the more distant transport of less dense siliceous glass (see discussion in Fisher and Schmincke, 1984). What we do not know is to what degree this process affects heterogeneous compositions of glassy fragments or pyroclast shapes. For the purposes of this discussion, therefore, we have assumed that these remain constant.
As might be expected, the correlations are best between ash beds in adjacent holes but less certain between sites, particularly those from different highs on Shatsky Rise. Part of the intersite differences may be related to possible thickness variations owing to variable sinking rates of particles, pelletization in the water column, and shallow current transport, as well as reworking by bottom currents and, once the ash is deposited, the impact of bioturbative mixing on preservation potential (Fisher and Schmincke, 1984). Some of the samples analyzed from Shatsky Rise were taken from ash-filled burrows.
As mentioned above, the vitric component population at Site 1208 is generally more vesicular; furthermore, the section at Site 1208 contains a significant number of isolated pumice clasts (Cores 198-1208A-10H through 19H) (Bralower, Premoli Silva, Malone, et al., 2002). Perhaps the more vesicular character of the ashes at Site 1208 is the product of localization of pumice raft-bearing currents. Deeper currents, however, must be responsible for creating the anomalously thick pile of sediment drift cored at this site. As the source of this drift might be in part the adjacent gully (Bralower, Premoli Silva, Malone, et al., 2002), there is also the possibility of duplicating the ash record by reworking. After deposition of primary fallout across the gully and drift topography, ash might then be eroded from the gully and transported onto the pile of sediment drift. We did not see any textural evidence for this phenomenon in the sand-sized fraction.