Like the alteration of basaltic basement, patterns of silicate reaction within the sediments, or more specifically alteration of the abundant ash, influence the pore-water profiles at Sites 998, 999, 1000, and 1001 (Sigurdsson, Leckie, Acton, et al., 1997; see also Gieskes and Lawrence, 1981). The large volumes and frequencies of Eocene and Miocene silicic ash observed and quantified during Leg 165, both as discrete layers and fine disseminations within the host sediment, show undeniable stratigraphic correlations between their independently determined distributions and a wide range of dissolved species. For example, interstitial silica faithfully mirrors the ash distribution at Site 998 (Fig. 13), reflecting the release of silica during glass dissolution/alteration and silicate hydrolysis during alteration of the ash (Table 1). Figure 13 also documents the correspondence between ash-related silica concentration and chert distribution, which acts as a sink for the dissolved silica. Furthermore, the silicate reactions (hydrolysis; e.g., Table 1) result in alkalinity relationships that are also strongly linked to ash distribution (e.g., Fig. 14). The fine ash disseminations within the host sediment were quantified using elemental mass balances (see Sigurdsson, Leckie, Acton, et al., 1997).
A similar relationship among volcanic ash, chert, and the concentration of silica is observed at Site 1001 (Fig. 15). Although chert formation may be possible at the stratigraphic level of first appearance (~165 mbsf), pre-unconformity physical conditions (specifically higher temperatures and greater burial depths) would have favored diagenetic precipitation. Consequently, we support pre-erosional chert formation with substantial removal of overlying sediment. This is substantiated by the almost 40 Ma represented by the hiatus between the Paleogene and Neogene sections in unconformable contact at ~165 mbsf (Shipboard Scientific Party, 1997c). Of interest, the ash-related pore-water enrichment in silica that is likely responsible for chert formation is still recorded at Site 1001 (Fig. 15), which suggests that ash alteration in this core, as in the others, is ongoing. Under the conditions of slow sediment accumulation of the present study (i.e., pelagic to hemipelagic rates), diffusion is sufficiently rapid to smooth out records of nonsteady-state sinks and sources of interstitial components unless the reactions are ongoing or occurred in the recent past (Lasaga and Holland, 1976).
The silica/ash relationship is also well recorded at Site 999 (Fig. 16), where a striking correspondence is observed among ash and dissolved silica distributions and the occurrence of biogenic silica. This important relationship likely represents the enhanced preservation of biogenic opaline silica within ash-rich intervals as a consequence of silica buffering during ash alteration (e.g., smectite formation; Table 1). Furthermore, the observed enhanced preservation of carbonate microfossils proximal to ash layers attests to the buffering capacity of alkalinity production associated with ash alteration (e.g., smectite formation). These observations have important taphonomic implications. Future studies should attempt to quantify the alteration of Eocene ashes—which are now known to represent a significant fraction of the Cenozoic stratigraphic section in the Caribbean—in its role toward imprinting levels of oceanic silica and ultimately catalyzing the accumulation of the vast amount of chert that abounds in the Eocene marine record (McGowran, 1989).
Important reactions between the volcanic ash and seawater include the dissolution/alteration (hydrolysis) of volcanic glasses and crystalline phases and the back precipitation (reverse weathering) that results in zeolites and clay minerals, including smectites (Table 1). These reactions, as sources and sinks, exert a strong control on the downcore distributions of a wide range of dissolved species, including cations such as rubidium, potassium, and lithium (see Sigurdsson, Leckie, Acton, et al., 1997, for extensive details). As an example, Figure 17 shows the downcore distribution of rubidium at Site 999 relative to both discrete ash layers and dispersed ash as quantified using independent elemental mass balances. This example is included to highlight the importance of not only the individual ash "events" but also the volumetric significance of the abundant reactive ash disseminated within the host sediments, reflecting perhaps a steady background flux of ash between major eruptions and/or the effects of bioturbational mixing that may be capable of obliterating the record of smaller events.