Retrieval of aerosol optical depth over ecologically stressed east African land surfaces from top of atmosphere radiance utilizing the aerosol free vegetation index

Abstract

Traditional satellite based AOD retrieval techniques with a coarse resolution such as collection 5 hardly provides a high spatial detail suitable for accurate AOD retrieval particularly over the ecologically stressed sites of East Africa. Furthermore, AOD retrievals from Moderate Resolution Imaging Spectroradiometer (MODIS-collection 5) require a linear mixing model that utilizes the normalized difference vegetation index (NDVI) to calculate surface reflectance. Since NDVI is affected by the presence of aerosols in the atmosphere, an empirical linear relationship between short wave infrared (SWIR) and visible reflectance were estimated in order to calculate a modified aerosol free vegetation index (AFRI). Based on AFRI and the minimum reflectance technique (MRT), an improved linear mixing model was developed. Calculated Vis/SWIR ratio diverges from the one used in the hypothetical model for AOD retrieval. Likewise, the regression coefficients between retrieved AOD for the developed algorithm and AERONET were higher as compared to that of MODIS collection 5 and AERONET. These emphasize the need of using AFRI plus accurate selection of surface characteristics and aerosol type for accurate AOD retrieval over East Africa that is addressed by the developed algorithm. At low AERONET AOD values, the two algorithms i.e. MODIS and developed agree but as the AERONET AOD value increases, a discrepancy is noted between the two as a result of usingNDVI and AFRIin the MODIS and developed algorithms respectively.