An optimal fitting approach to improve the GISS ModelE aerosol optical property parameterization using AERONET data

Abstract

We improve the updated 2000 Goddard Institute for Space Studies, New York, ModelE aerosol optical property parameterization using an optimal fitting approach with AERONET ground measurements. The model aerosol optical properties, such as optical depth, are calculated using the aerosol mass density field from a chemical transport model, Mie scattering parameters, a prescribed dry size for each aerosol species assuming external mixing, and a hygroscopicity parameterization. A comparison between the model‐ and AERONET‐measured optical depth (AOD) and Ångström exponent (AE) indicates that the general circulation model (GCM) aerosol parameterization has a flatter AOD spectral dependence, thus a very low biased AE, which suggests that the aerosol sizes used in the model are too large. The seasonal variation of GCM AE also disagrees with that of AERONET data. On the basis of these results, we identify GCM aerosol size as the most poorly constrained parameter and develop an optimal fitting technique to adjust the GCM aerosol dry size by minimizing the total mean square error between the GCM and AERONET AOD at the six AERONET wavelengths. After adjusting the aerosol's dry size, the agreement between the GCM AE with AERONET data is improved. The fitted AOD at the six wavelengths closely matches AERONET data over most biomass burning, dust, and rural regions. The results are also greatly improved for the other aerosol types. The global distribution of the optimally fitted sizes displays regionally uniform characteristics, which allows the generation of a geographically varying size data set. Model uncertainty caused by other factors is also represented by an uncertainty parameter, which is mainly attributed to errors from aerosol mass concentration, Mie scattering parameters, relative humidity, and AERONET measurements. The relative contribution of each of these errors sources depends on the relevant aerosol type. Further comparison between the absorption optical depth and AE spectral dependence provides additional information on absorbing aerosols and GCM fine‐to‐coarse mode ratio, which will be addressed in future research.