Accurate aerosol remote sensing over bright land surfaces using AATSR

This work presents a newly developed iterative algorithm to simultaneously retrieve aerosol and land surface BRDF based on accurate BRDF modeling and stringent radiative transfer, from AATSR data. Applying the algorithm to ~6 years of imagery over the Australian continent shows excellent accuracy of retrieved aerosol optical depth (AOD) and very low surface model error, with mean AOD error <; 0.03 and AOD bias <; 0.003, based on validation data from 10 sun photometer stations across the continent. The algorithm also assigns to each pixel the aerosol type from a set of candidate aerosol models. Spatial and temporal distributions of aerosol types agree well with our knowledge of major aerosol patterns across the continent, and the new method offers scope for improved characterization of the temporal and spatial distribution of aerosol over land.