Evaluation of proposed SeaWiFS algorithms for the retrieval of phytoplankton pigments and suspended sediment in coastal water types

A preliminary analysis of algorithm performance has been performed for application to SeaWiFS radiometer reflectance data for the retrieval of constituents including phytoplankton pigments, suspended sediment, and dissolved organic carbon (yellow substance) in coastal, case 2, waters. Ocean waters are classified as case 1 where parameters correlate with phytoplankton pigments. In case 2 waters such parameters do not correlate requiring the development of a multiple component algorithm. The coastal waters must be characterized by local relationships which are subjected to extensive spatial and temporal variability. The spectral reflectance of the coastal water column just below the surface has been modelled using existing optical cross sections for absorption and backscatter. This model has been used to numerically simulate the reflectance variability in SeaWiFS bands over a range of coastal water types. Illustrative multiband SeaWiFS algorithms for retrieval of chlorophyll are compared to the performance obtained using maximum likelihood estimators. Substantial improvement in retrieval algorithm performance is expected when a patch of data rather than a single pixel is used to make the estimate