Retrieval of Aerosol Single-Scattering Albedo over North Africa via Critical Reflectance

A significant source of uncertainty in the determination of global direct radiative forcing is the uncertainty in the direct forcing over land. An accurate determination of this forcing is dependent upon the availability of reliable information on aerosol absorption. Over ocean, aerosols cause a cooling at the top-of-the- atmosphere; over land, the sign of the aerosol direct effect depends on whether the aerosol is effectively brighter or darker than the underlying surface. Obtaining aerosol absorption information over land is possible using space-based sensors; however, it is complicated due to the sometimes large and spatially-varying surface signal, especially over bright desert surfaces. The aerosol critical reflectance is one parameter that can be used to determine aerosol absorption by deriving it with respect to the underlying surface properties. We estimate aerosol critical reflectance over North Africa, a source region for absorbing mineral dust and biomass burning aerosol species, using MODIS Level 1B reflectance data. Unlike single-image over-desert aerosol retrieval methods, which rely on a minimum top-of-atmosphere reflectance for information about the surface, the critical reflectance method compares two MODIS images with different aerosol loading and the same solar and viewing geometry, 16- days apart, to derive the aerosol critical reflectance. Comparisons of the critical reflectance on coincident 16- day image pairs yield basic information about the aerosol chemistry; we find a positive correlation between spectral dependence of critical reflectance and near-IR path radiance for dust, while a negative spectral dependence of critical reflectance in the visible channels indicates biomass burning aerosol. We then use this information to relate the aerosol critical reflectance to single-scattering albedo (SSA) via a DISORT-type radiative transfer code. Aerosol SSA and size estimates are compared to observations from AERONET and in situ data obtained as a part of the Dust and Biomass Experiment (DABEX) and the Langley Aerosol Research Group Experiment (LARGE), components of the African Monsoon Multidisciplinary Analyses (AMMA) campaign, during 2006. Sensitivity of the critical reflectance to assumed size distributions is also examined, and limitations of the method explored.