Bio-optical model for Chesapeake Bay and the Middle Atlantic Bight

Retrievals of bio-optical properties from satellite measurements in Case 2 waters depend on algorithms that account for multiple constituents affecting spectral quality of the upwelling light flux. Semi-analytical (SA) models are suitable for this purpose, but must be parameterized with in situ data, particularly in estuarine and coastal waters. We examined spatial and temporal variability of biooptical properties in Chesapeake Bay and the adjacent Middle Atlantic Bight (CB/MAB) to parameterize and validate the Garver/ Siegel/Maritorena (GSM01) model. Several years (1996–2002) of data on inherent and apparent optical properties confirmed high scattering and strong absorption by dissolved and particulate components that did not co-vary. These data, consisting of the chlorophyll (chl a)-normalized phytoplankton absorption coefficient, aph * (l), and the spectral slope of absorption due to dissolved and detrital materials, Scdm, were used to optimize model parameters of GSM01 and produce a version of the model tuned for CB/ MAB we have designated GSM01-CB. Performance of GSM01-CB was measured against the globally optimized version of GSM01 and the empirical algorithm, OC4v.4, using both in situ and satellite-derived radiances from the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) as inputs to the model. GSM01-CB outperformed OC4v.4 in retrieving chl a in CB, and both SA models outperformed OC4v.4 in MAB. GSM01 and GSM01-CB also returned reasonable estimates of other bio-optical products, including the absorption coefficient for dissolved and detrital materials, acdm(443), and the particulate backscatter coefficient, bbp(443). We present alternatives to the parameterization of GSM01-CB to account for interannual variability of aph * (l) in CB using empirical relationships with key variables that regulate phytoplankton dynamics in the estuary (i.e., freshwater flow and nutrient loading), and to include regional gradients of Scdm in MAB.