PNNL

Abstract

A novel method for the retrieval of aerosol optical depth (AOD) under partly cloudy conditions has been suggested. The method exploits reflectance ratios, which are not sensitive to the three-dimensional (3D) effects of clouds. As a result, the new method provides an effective way to avoid the 3D cloud effects, which otherwise would have a large (up to 140%) contaminating impact on the aerosol retrievals. The 1D version of the radiative transfer model has been used to develop look-up tables (LUTs) of reflectance ratios as functions of two parameters describing the spectral dependence of AOD (a power law). The new method implements an innovative 2D inversion for simultaneous retrieval of these two parameters and, thus, the spectral behavior of AOD. The performance of the new method has been illustrated with a model-output inverse problem. We demonstrated that a new retrieval has the potential for (i) detection of clear pixels outside of cloud shadows and (ii) accurate (~15%) estimation of AOD for the majority of them.