Assessing the Accuracy of Biogenic Content Estimation from Visible Derivative Spectroscopy of Sedimentary Cores from the Western Pacific

Abstract

The biogenic contents of marine sediments, such as carbonate (CaCO3) and organic carbon (TOC), provide important information about past climatic and environmental changes. For sediment cores, such as those found in the marginal seas of the western Pacific, intensive laboratory study takes considerable time and effort. The previous drilling and coring programs have developed nondestructive methods, which require less time and labor, such as those that utilize visible reflectance derivative spectra measured from the surface of sediment samples to estimate downcore biogenic content. Nevertheless, these methods have been shown to be useful only for on-site estimation of downcore samples and are not considered entirely feasible for testing samples collected from regional or larger spatial scales. The present study presents a novel protocol of spectral decomposition utilizing varimax-rotated principal component analysis (VPCA) for estimating biogenic contents of sediment samples at the basin scale. Using two sediment cores from the South China Sea (SCS) separated by 200 kilometers, we evaluated a new protocol by measuring the visible reflectance spectrum and the biogenic content. Based on six VPCA components of first derivative reflectance spectrum measurements and laboratory analyzed biogenic contents of core MD972148, a set of empirical equations for estimating CaCO3, TOC, and opal contents have been established. The equations were tested using data from core MD012396, and the new regression equations provided accurate estimations. Our study demonstrated that our new methods could achieve better estimates due to the improvement of the regression model with a reduced number of independent variables. Further, this study circumvents the limitation of applying empirical equations to sediment cores outside of the calibration range. Our present findings state that with more comprehensive and systematic reflectance spectral data, the new protocol can be used to estimate biogenic content with more regional or spatial precision in future research.