The choice of the most appropriate aerosol model in a radiative transfer code

In this study, the ability of the nine aerosol models implemented in the SMARTS spectral radiative model is investigated. The use of the aerosol models is to characterize the spectral aerosol optical depth (SAOD) over Athens. The SAODs are obtained from spectral direct-beam irradiance measurements using a high resolution spectroradiometer in the spectral range 300–1000 nm. Moreover, using the least-squares fitting method the A ˚ ngstro¨m parameters a and b were obtained in the spectral interval 320–870 nm. The investigation is made in two ways: first using the aerosol optical depth at 500 nm (AOD500) and second, the A ˚ ngstro¨m turbidity coefficient b as input parameter to the SMARTS model for the quantification of the aerosol loading. The analysis shows that the use of different input parameters has a great influence on SAOD and also strongly affects the simulation of the measured SAOD. The urban and rural aerosol models seem to have a good performance on the measured SAODs while the maritime and haze ones are very inappropriate. The differences between measured-and modeled-estimated SAODs do not depend on turbidity conditions. On the other hand, when the measured a-A˚ ngstro ¨m value is similar to that derived from the aerosol model, the differences between measured and modeled SAODs are about 0. As the a values estimated through the least-squares method are very different than those derived from the aerosol models, the modeled SAODs are very far from the real ones. 2008 Elsevier Ltd. All rights reserved.