Sensitivity analysis of Look-up table for satellite-based aerosol optical depth retrieval

Abstract

Due to the high variation of aerosol in space and time, remote sensing has become an effective tool in aerosol optical depth (AOD) retrieval. Most of the satellite-based AOD retrieval methods using the radiative transfer model (RTM). RTMs help to establish a relation between AOD and the observed top-of-atmosphere (TOA) reflectance. Using look-up table-based approaches is among the commonly used methods for reducing the computational complexity in RTMs. Tuning steps for parameters in the LUTs influences computational cost and accuracy of retrieved satellite-based AOD. However, a comprehensive and quantitative assessment for the construction of LUT based on sensitivity analysis does not exist. In this paper, we present a sensitivity analysis approach called Morris one-at-a-time (MOAT) for identifying parameters that dominate model behaviour in the AOD retrieval process. We also use 8 cases to tune steps for each parameter in the look-up table. We investigate the influence of geometrical and physical parameters in the 6S RT code on radiance and reflectance. The result indicated that ground reflectance, aerosol model, and single scattering albedo are dominant parameters in LUT construction. Other geometrical and physical parameters have less impact on the simulated radiance and reflectance. Finally, we tuned proper steps for LUT construction based on the trend and behaviour of each parameter. The proposed method in this study can improve the accuracy of satellite-based AOD by constructing a LUT with a short processing time.