Coring at Ocean Drilling Program (ODP) Sites 1150 and 1151 has yielded the most complete recovery of Pliocene-Pleistocene marine tephras that has yet been obtained in the Japan Trench. As such, they provide the opportunity to study the effects of age on tephra geochemistry, not only as a function of temporal changes in magma composition but as influenced by postdepositional chemical alteration of glass shards. As the geochemical signature of glass shards can be moderated by medium to high degrees of postdepositional weathering, this deep record of volcanic history provides a useful opportunity to assess alteration trends and thereby the reliability of geochemical tephra correlations for the entire Quaternary and earlier. This parallels similar work by Masuda et al. (1993) in the Nankai Trough.
The geochemical instability of glass has been noted and assessed in a variety of environments (Fisher and Schmincke, 1984; Jezek and Noble, 1977). The stages and chemomechanics of hydration and subsequent devitrification are complex and poorly understood, although the progress from unweathered glass through perlite (hydrated glass-obsidian) to microcrystallite clay mineral growth and ultimate zeolitization is fundamentally a two-stage process. The first phase is the addition of water into the glass matrix. This proceeds at varying rates, dependent upon environment, glass chemistry, and temperature. Compared with basaltic glass (sideromelane), alteration of silicic glass initially involves little solid-state change (Fisher and Schmincke, 1984) as water is taken up by the glass "matrix" (Friedman and Long, 1976). At this stage, up to 3% water can be held in the glass (Friedman and Smith, 1958; Jezek and Noble, 1977; Hunt and Hill, 1993) with low ion exchange. Additional hydration subsequently takes place, initially along fracture surfaces, resulting in the "weakening and breakage of the bonds of the glass structure" (Jezek and Noble, 1977) with a tendency for clay and zeolite mineral growth during phase 2.
The initial hydration process does not result in significant ion mobility across the entire compositional range of the glass, but Na2O and K2O are particularly mobile, with a depletion of K2O and an enhancement of Na2O (Jezek and Noble, 1977). Hydration of glasses has been monitored by EPMA studies (Jezek and Noble, 1977) to demonstrate the direct (albeit nonlinear) relationship between water content and analytical (oxide sum) totals. Application of this simple relationship to glasses (shards) of different ages but in similar depositional settings can be used to investigate temporal variations in glass alteration. This has been attempted in this study of Leg 186 tephras, of which the analyzed shards span an interval of ~0-5 Ma. This age range and the relatively low geothermal gradient (Sacks, Suyehiro, Acton, et al., 2000) suggest that these tephras will lend themselves particularly well to investigation of the impact of phase 1 (i.e., hydration only) alteration.
Age-alteration relationships are investigated through depth vs. total oxide plots (Fig. F4). At present, age models lack the robustness to convert all depths to specific ages since currently details of the age-depth model for holes drilled during Leg 186 are unresolved (Sacks, Suyehiro, Acton, et al., 2000; Hayashi et al.; Li; Maruyama and Shiono, all this volume). Therefore, at this stage we feel that interpretations which utilize anything other than broadly indicative age-depth information would be premature. The total oxides data display considerable scatter, which may reflect (1) varying proportions of undetermined volatiles other than water (e.g., fluoride and chloride), (2) varying proportions of undetermined primary water, (3) varying susceptibility to phase 1 alteration as a result of varying vesicularity or shard size (e.g., Cambray et al., 1993; Dugmore et al., 1995), or (4) variability of host sediment porosity and pore water supply. However, this variability between layers is superimposed on a clear age trend, particularly apparent in the more extensive (to ~350 meters below seafloor [mbsf]) record from Site 1151. Nevertheless, although an age/depth-driven alteration trend is apparent, its magnitude over the Pliocene-Pleistocene is sufficiently small that judicious tephrostratigraphical correlation may be achieved, albeit with caution if attempts are made to link with terrestrial sites with different weathering/diagenetic regimes. Of additional interest is the absence of an age trend related to Al2O3/SiO2 ratios. Masuda et al. (1993) suggest that the Al2O3/SiO2 ratio is an index of marine burial diagenesis, due to the preferential solubility of Si relative to Al. Our data show no such change with depth (Fig. F5), implying relative geochemical resistance to alteration over Pliocene-Pleistocene timescales. From these observations we can conclude that changes have occurred over time through the addition of water to the glass "matrix," but that this addition has not yet been accompanied by geochemical breakdown through selective leaching of elements from the glass matrix. Our findings are broadly in agreement with those of Arthur et al. (1980), Cadet and Fujioka (1980), Fujioka (1986), and Fujioka et al. (1980), who show that alteration is only minimal at equivalent ages and depths at nearby DSDP Sites 56, 57, and 87B. These results are relatively encouraging for Pliocene-Pleistocene marine tephrochronology in the Japan Trench.
The maximum sample depth of tephra analyzed in this study (Site 1151) is 357 mbsf (age = ~4.9 Ma), where the overall geothermal temperature gradient is 27.7°C/km (Scientific Party, 1980). This gradient accords well with the Site 1150 data (for the upper 155 mbsf, temperature gradient = 28.9°C/km) but contrasts significantly with results from the nearby Nankai Trough (ODP Leg 131), where the geothermal gradient exceeded 111°C/km (Masuda et al., 1993). Despite this significantly greater heat flux, Masuda et al. (1993) report no significant ash alteration in the upper 500 mbsf. This is perhaps surprising, given the general consensus that geothermal gradient is the primary control on tephra alteration (see Hein and Scholl, 1978; Grechin et al., 1980; Cambray et al., 1993). This difference in reported weathering may reflect differing analytical approaches (e.g., bulk vs. grain discrete) or, more likely, may be due to the concentration of prior studies on phase 2 alteration processes contrasted by this study of Pliocene-Pleistocene tephras.