The record of calcium carbonate is a valuable stratigraphic and paleo-environmental tool for the study of deep-sea sedimentation. Biogenic carbonate represents both a source and a sink in the global carbon cycle, and the interpretation of carbonate records provides insights into the temporal and spatial evolution of ocean chemistry and productivity (e.g., Mayer, 1991, and references therein). Sedimentation in the Atlantic Ocean is characterized by higher carbonate content within interglacials compared to glacials. For the Leg 172 sites, carbonate content pattern matching (J. Grützner et al., unpubl. data) allows the identification of climatic cycles and correlation between holes and sites.
Traditionally, the calcium carbonate content is determined on discrete samples of cored material by pressure calcimetry or coulometric techniques (e.g., Mayer, 1991, and references therein). The methods are labor intensive and destructive. In the last decade, numerous core logging devices have been developed for measuring sediment physical properties quasicontinuously, rapidly, and nondestructively. Sedimentary components, including carbonate, were estimated with various degrees of accuracy using core logging data (e.g., Mayer, 1991; Weber, 1998). The development of empirical relationships between the diffuse spectral reflectance and carbonate content on a relatively limited, representative set of sediment samples provides a rapid, high-resolution, and noninvasive alternative to direct measurement of carbonate (Balsam and Deaton, 1996; Harris et al., 1997; Ortiz et al., 1999). The matrix-dependence of sediment spectra (Balsam and Deaton, 1996) precludes the development of a general equation for the prediction of sedimentary components. The empirical relationships, however, can have a regional applicability in areas with a similar sedimentation regime (e.g., Harris et al., 1997).
Previous carbonate content estimates were derived from sediment reflectance spectra in the visible domain or extended into near-ultraviolet and near-infrared domains on either dried samples (Balsam and Deaton, 1996) or using the Oregon State University split-core analysis track on wet core surfaces (Harris et al., 1997; Ortiz et al., 1999). During most Ocean Drilling Program cruises, however, sediment reflectance has been routinely measured over the visible spectral range with a Minolta CM-2002 handheld spectrophotometer. Balsam et al. (1997) have shown that accurate spectral data can be obtained on wet cores using the Minolta system. Since the major objective of Leg 172 was to obtain a detailed history of late Neogene paleoceanography and paleoclimate in the North Atlantic, the goal of our study was to provide high-resolution, empirical predictions of carbonate content using multiple linear regression on diffuse spectral reflectance measurements.