Chemical and mineral compositions of sediments from ODP Site 127-797

In order to reconstruct past variations in the aeolian dust (Kosa) contribution to the Japan Sea, and to establish a direct link between terrestrial and marine climatic records, we have applied statistical procedures to distinguish and quantify detrital subcomponents within the detrital fraction of the late Quaternary hemipelagic sediments in the Japan Sea. Q-mode factor analysis with varimax and oblique rotation of the factors followed by multiple-regression analysis between mineral composition and factor loadings was conducted using six ''detrital'' elements. Four detrital subcomponents were defined, which are attributed to Kosa derived from ''typical'' loess, Kosa from ''weathered'' loess, and fine and coarse arc-derived detritus, respectively, based on comparisons with the chemical and mineral compositions of probable source materials. Using these detrital subcomponents, the variation in Kosa fraction was reconstructed for the last 200 ky. The results reveal millennial-scale as well as glacial-interglacial scale variations in Kosa contribution. Especially, millennial-scale variability of Kosa contribution suggests the presence of high frequency variation in summer monsoon precipitation in the central to east Asia during the last 200 ky.

Corrected index properties for ODP Leg 127 and 128 sites

During Legs 127 and 128, we found a systematic error in the index property measurements, in that the wet bulk density, grain density, and porosity did not satisfy well-established interrelationships. We have found that an almost constant difference exists between the weight of water lost during drying and the volume of water lost. This discrepancy is independent of volume or water content of the sample. The water losses should be equal because the density of water is close to 1.0 g/cm**3. The pycnometer wet volume measurement has been identified as the source of the systematic error. The wet volume on average is 0.2 cm**3 too low. For the rare cases when the water content is negligible, there is no offset. The source of the wet volume error results from the partial vapor pressure of water in the pycnometer cell. Newly corrected tables of index properties measured during Legs 127 and 128 are included. The corrected index properties are internally consistent. The data are in better agreement with theoretical models that relate the index properties to other physical properties, such as thermal conductivity and acoustic velocity. In future, a standard volume sampler should be used, or the wet volume should be calculated from the dry volume and the water loss by weight.