| Abstract |
The lack of well-preserved carbonate in much of the Arctic marine environment dictates the need for alternative methods of paleoceanographic reconstruction. The broad variety of physical properties measurements makes them well suited for use in a variety of environments, but they provide unique opportunities when employed in the Arctic. Because Arctic sediment is introduced and reworked by a variety of mechanisms, the signature from multiple processes becomes intermixed with the sediment. Many of these processes operate in other ocean basins, while some function only in Polar Regions. A strategy to address this mixing problem is to employ spectrally-resolved physical properties measurements, or to use multiple methods in conjunction to generate multivariate data sets, which can differentiate concurrent processes. Data of this type is well suited to multivariate analysis techniques such as sample-based or variable-based, varimax-rotated, principle component analysis (VPCA). These are methods that decompose the data matrix to infer process from orthogonal functions. The method is applied to cores from the Chukchi sea to document that visible derivative spectroscopy provides a powerful means of reconstructing sediment provenance. In the Chukchi Sea, diffuse spectral reflectance provides a proxy to monitor variations in Holocene flow through the Bering Strait.
|
