On the Accuracy of SeaWiFS Ocean Color Data Products on the West Florida Shelf

Despite the importance of the West Florida Shelf (WFS) on regional ecology and local economy, systematic shelf-wide assessment of the ocean biology has not been conducted, primarily because of budgetary limitations for routine field campaigns and unknown accuracy of satellite-based data products. Here, using shipboard spectral normalized water-leaving radiance (nLw[λ]) data and chlorophyll-a concentrations (Chl-a) collected regularly during two multiyear field programs spanning \u3e 10 years, the accuracies of Sea-viewing Wide Field-of-view Sensor (SeaWiFS) standard data products were evaluated. The in situ data covered a wide dynamic range, with about one order of magnitude in nLw(490) (0.47 to 4.01 mW cm−2 μm−1 sr−1) and two orders of magnitude in Chl-a (0.07 to 10.6 mg m−3). Near-concurrent in situ and satellite nLw(λ) data showed absolute percent differences (APD) increasing from 7–9% to 10–14% when data with elevated aerosol optical thicknesses at 865 nm (τa865) were included. Most of this uncertainty, however, canceled in the maximal blue-to-green reflectance band ratios traditionally used for estimating Chl-a. SeaWiFS OC4 Chl-a showed a root mean square (RMS) uncertainty of 0.106 for log-transformed data in waters offshore of the 20-m isobath that increased to 0.255 when all data were considered. The increased likelihood for nearshore SeaWiFS Chl-a greater than ∼0.5 mg m−3 to be overestimated was shown to be caused by a variety of factors (colored dissolved organic matter [CDOM], suspended sediments, and bottom reflectance) that varied in both time and space. In the future, more sophisticated algorithms capable of taking these factors into consideration are required to improve remote determinations of Chl-a in nearshore waters of the WFS