A Sensitivity Study of the SeaWiFS Atmospheric Correction Algorithm: Effects of Spectral Band Variations

Abstract

With the success of launch and initial data processing of Sea-viewing Wide Field-of-view Sensor (SeaWiFS), there is a great interest in ocean color community to intercompare ocean color data between different ocean color sensors. It is well known that the atmospheric correction, which removes about 90% of sensor-measured signals contributed from atmosphere in the visible, is the key procedure in ocean color imagery data processing. Therefore, it is useful to evaluate the SeaWiFS atmospheric correction algorithm applying to various ocean color sensors. The SeaWiFS atmospheric correction algorithm uses lookup tables which were generated with over ∼25,000 radiative transfer model runs for different aerosol optical and microphysical properties, solar and viewing geometries, and, in particular, at the eight SeaWiFS spectral bands. Since different ocean color sensors usually have different band spectral characterizations, it is rather difficult to apply the SeaWiFS atmospheric correction algorithm to other sensors if one needs to regenerate lookup tables at spectral bands different from SeaWiFS. In this article, we evaluate the accuracy of the SeaWiFS atmospheric correction algorithm for various ocean color sensors using the current SeaWiFS lookup tables. The focus in on the following satellite ocean color sensors: the Modular Optoelectronic Scanner (MOS), the Ocean Color and Temperature Sensor (OCTS), and the Polarization and Directionality of the Earth’s Reflectances (POLDER). These sensors have a slightly different spectral bands compared with the SeaWiFS. It was found that, with an appropriate calculation of the Rayleigh scattering contributions at each sensor’s spectral band and a simple modification in computing the diffuse transmittance of the ocean-atmosphere system, the SeaWiFS atmospheric correction applied to other sensors is as accurate as for SeaWiFS for the solar zenith angles θ 0⩽60°.