PALEOCEANOGRAPHY SINCE THE B/M BOUNDARY

The numbers of diatom valves fluctuate significantly with depth, between ten thousand and several tens of millions of valves per gram of sediment (Fig. 2). When compared to ¹⁸O stages for the upper 24 meters below seafloor (mbsf) defined by Andreasen et al., Chap. 8 (this volume), it is clear that diatom valves decrease at the end of glacial periods. This trend is also observed at depths above 34 mbsf at Site 1020. At 34 mbsf, diatom valves are very few and then increase gradually until 10 mbsf (Fig. 3). This probably resulted from improved opal preservation. The B/M boundary was not identified at Site 1018, but the patterns of fluctuation of diatom abundance—which were very similar at Sites 1018 and 1020—were compared, and the depth of 123.5 mbsf was assumed for that event. Diatom abundance at Site 1021 is extremely low, less than ten thousand valves per gram of sediment in the upper 5 m, and very rarely identified at the depths below that. Preservation is very poor (Fig. 4).

At Site 1020 (Fig. 5), the cold-water species Neodenticula seminae dominates in the lower part (84.7-50.0 meters composite depth [mcd]), although there is an interval of low concentration of diatoms between 75.3 and 64.7 mcd. A strong peak in abundance of the freshwater group, such as Aulacoseira spp., Flagilaria spp., and Stephanodiscus spp., appears at 41.1 mcd. In the upper stages, there is a relatively high concentration and successively increasing productivity of diatoms. The top of the sample, which indicates warm conditions, corresponds to the Holocene.

After "the mid-Brunhes climatic event" around 400 ka (Jansen et al., 1986), conditions of the sea surface changed from pelagic to more coastal and productive. As shown in Figure 5, the subsequent changes in abundance of the freshwater group are in short cycles (~5 m). These may correspond to a cycle shorter than the typical glacial-interglacial cycle (such as obliquity or climatic precession) and imply that the freshwater group is transported by seasonal winds and/or enhanced precipitation because both are thought to fluctuate in precessional cycles.

At Site 1018 (Fig. 6), on the other hand, dominance of freshwater species is observed during 127-115, 95-85, and 65-52 mcd; diatoms are generally scarce at the corresponding depths of Site 1020. A short cycle cannot be observed in the fluctuating abundance of the freshwater group, but it is interesting that compared to total diatom abundance, the freshwater group increases during periods of low concentration of diatom valves (the glacial periods). If the fluctuation of diatom abundance corresponds to glacial-interglacial cycles in the entire C1n Subchron, this correlation with glacial cycles apparently resulted from falling sea level and increasing terrigenous inflow.