A synthesis of several pore-water profiles from Sites 1091, 1093, and 1094 are shown in Figure 18. The chloride profiles show clear evidence for the downward diffusion of higher salinity glacial-age seawater (McDuff, 1985). The uppermost porcellanite layer at about 68 mbsf at Site 1094 has apparently interrupted this downward diffusion and presents the intriguing suggestion that the porcellanite may have formed in the past 10 to 20 k.y. These high sedimentation-rate diatomaceous oozes were suboxic to mildly reducing. H2S was detected by smell at Sites 1091 and 1093 throughout most of these profiles, but H2S was not detected at all at Site 1094 except for a very faint whiff in one whole round near the top of the section. In addition, sulfate depletion is much lower than would be expected based on sedimentation rate, and it appears to be inversely correlated with our preliminary estimates of TOC (see site chapters for data not shown here). Phosphate profiles show little correlation with alkalinity and ammonium except at Site 1094, and dissolved manganese (Mn) is observed to varying degrees throughout the profiles. We offer the following preliminary interpretation of these observations. The highest sedimentation-rate site is Site 1093 (~25 cm/k.y.), which is located very near the contemporaneous Polar Front at ~50°S. Sedimentation rates at Site 1093 were likely least affected by glacial-interglacial migrations of the Polar Front compared to Sites 1091 and 1094 located about 3° to the north and south, respectively. Thus, the mildly reducing conditions at Site 1093 have likely persisted through glacial-interglacial cycles as evidenced by the low downcore dissolved Mn profile. Sites 1091 and 1094 have undergone much more drastic perturbations in average sedimentation rates (both ~14 cm/k.y.) and are out of phase with each other over the glacial-interglacial climate cycles. This resulted in a periodicity in the redox state of the sediments that has somehow permitted reactive Mn to persist at depth. The low sulfate reduction rates observed despite the high sedimentation rates may result from the fact that a significant, if not major, fraction of the organic carbon in these diatomaceous oozes is highly refractory opal intrinsic organic carbon that is unavailable for degradation until the opal has dissolved. Opal intrinsic organic carbon likely has a very low phosphate content, thus offering some explanation for the nature of the phosphate profiles observed in these diatomaceous oozes.
In summary, this preliminary and general interpretation of these first deep interstitial water profiles from the circumantarctic siliceous ooze belt will need to be verified and enhanced with additional shore-based analyses. Additionally, shore-based analyses of closely spaced interstitial water samples across some of the porcellanite intervals observed in the sediments at Site 1094 may offer important insight into the mechanisms involved in the transformation of diatom opal to opal-CT.