Lithology of ODP Leg 138 holes

During Leg 138, we measured reflectance spectra in the visible and near-infrared bands (455-945 nm) every few centimeters on split core surfaces from eastern tropical Pacific Ocean sediments. Here, we evaluate predictions of the content of biogenic calcite, biogenic opal, and nonbiogenic sediments from the reflectance spectra. For Sites 844 through 847, which contain a significant nonbiogenic component, reflectance spectra yielded a useful proxy for the percentages of CaCO3 over a wide range of values from nearly 0% to 100%, with root-mean-square (RMS) errors of about 9%. Direct estimates of "nonbiogenic" sediment percentages, approximated by 100 - (CaCO3 + opal), were reasonably successful (RMS error of 10%), however, were incorrect in some intervals. This suggests that mineralogy of the nonbiogenic material changes through time and that further subdivision of this component will be needed for useful estimation from reflectance. For percentages of biogenic opal, calibration equations appear to work well (RMS error of 6%) at concentrations of less than 30%, but for higher opal concentrations, reflectance equations often underestimate the true contents of opal. Improvements in multiparameter lithologic estimates from reflectance spectra may come from (1) expanding the wavelengths measured to better capture unique mineral reflectance bands, and (2) adding the ability to measure diffuse, rather than directional, reflectance to minimize the effects of surface roughness.