UC Santa Cruz

Earth's climate is intimately associated with biogeochemical processes of the sea. Biological Oceanography explores mechanisms controlling carbon uptake by phytoplankton, carbon transfer through biogeochemical processes, and energy flow through ecosystems. Satellite Oceanography affords a synoptic view of the sea surface and reveals underlying physical, chemical, and biological processes. Since the advent of ocean color satellites in 1978, ocean color algorithms evolved from quantifying phytoplankton biomass to addressing more complex bio-optical and oceanographic problems: characterizing inherent optical properties of the water column, estimating primary productivity, and detecting water masses. Locating a water mass, tracking its changes, and discriminating its constituents using bio-optical algorithms are the three objectives of this dissertation. The first objective identifies the location of the Columbia River Plume (CRP) by using light absorption by chromophoric dissolved organic matter (a _CDOM) as an optical proxy for salinity. It relates in situ measurements of (a _CDOM to salinity using linear regression analysis, then computes "synthetic" salinity using MODIS-Aqua satellite imagery. The algorithm is robust at predicting salinity of the CRP on the Oregon and Washington shelf. The second objective identifies sub-mesoscale features within the CRP and tracks their changes in space and time. It employs k-means clustering and discriminant function analysis to identify water types from bio-optical and environmental input variables using in situ and MODIS-Aqua satellite observations. The algorithm is robust at identifying features in satellite and mooring data, consistent with measured and modeled water masses in previous work. The third objective involves development of an optical model (PHYDOTax) that discriminates phytoplankton taxa contained within an algal bloom. A hyperspectral ocean color signature-library for known phytoplankton (dinoflagellates, diatoms, haptophytes, cryptophytes, chlorophytes, cyanophytes, and phycocyanin-containing eukaryotes) was developed and then PHDYOTax decomposed ocean color spectra for culture mixtures and field samples into constituent taxa. PHYDOTax is robust at discriminating phytoplankton taxa and is one of the first algorithms to distinguish dinoflagellates from diatoms in ocean color data. These algorithms are new tools for the oceanographic community to constrain the location of carbon uptake and transfer through space and time in the CRP, and to partition energy flow through different phytoplankton-taxon dominated ecosystems.