Determining phytoplankton community structure from ocean color at the Martha's Vineyard Coastal Observatory (MVCO)

Satellite remote sensing of ocean color allows for estimates of phytoplankton biomass on broad spatial and temporal scales. Recently, a variety of approaches have been offered for determining phytoplankton taxonomic composition or phytoplankton functional types (PFTs) from remote sensing reflectance. These bio-optical algorithms exploit spectral differences to discriminate waters dominated by different types of cells. However, the efficacy of these models remains difficult to constrain due to limited datasets for detailed validation. In this study, we examined the region around the Martha's Vineyard Coastal Observatory (MVCO), a near-shore location on the New England shelf with optically complex coastal waters. This site offers many methods for detailed validation of ocean color algorithms: an AERONET-OC above-water radiometry system provides sea-truth ocean color observations; time series of absorption and backscattering coefficients are measured; and phytoplankton composition is assessed with a combination of continuous in situ flow cytometry and intermittent discrete sampling for HPLC pigments. Our analysis showed that even models originally parameterized for the Northwest Atlantic perform poorly in capturing the variability in relationships between optical properties and water constituents at coastal sites such as MVCO. We refined models with local parameterizations of variability in absorption and backscattering coefficients, and achieved much better agreement of modeled and observed relationships between predicted spectral reflectance, chlorophyll concentration, and indices of phytoplankton composition such as diatom dominance. Applying these refined models to satellite remote sensing imagery offers the possibility of describing large-scale variations in phytoplankton community structure both at MVCO and on the surrounding shelf over space and time.