Using angular and spectral shape similarity constraints to improve MISR aerosol and surface retrievals over land

The Multi-angle Imaging SpectroRadiorneter (MISR) instrument on the Terra satellite has demonstrated the capability to retrieve aerosol optical depths, surface bidirectional reflectance factors, and hemispherical reflectances over a wide variety of land surface types. In particular. its multiangular imaging design has enabled the application of algorithms that minimize sensitivity of the aerosol retrievals to the brightness of the underlying surface. The novel aerosol algorithm that was developed prior to launch has had notable quantitative success. Over certain scene types, however, the approach contained obvious spatial artifacts, so a postlaunch refinement to the algorithm was implemented. It constrains the retrieved aerosol models and optical depths such that the implied angular shape of the surface hemispherical-directional reflectance factor (HDRF) is similar among all of the MISR wavelengths. This upgrade has resulted in three tangible benefits: (1) the 1 occurrence of outliers has been dramatilcally reduced, (2) correlations with AERosol RObotic NETwork (AERONET) aerosol sunphotometer data are quantitatively improved, and (3) the quality of surface products is markedly enhanced. MISR Level 2 aerosol and surface products are archived at the NASA Langley Atmospheric Sciences Data Center. Those having version numbers v0012 and higher incorporate ibis upgrade in the data processing software. (C) 2004 Elsevier Inc. All rights reserved.