The Development and assessment of a flexible inversion algorithm for Aerosol property retrieval combining passive multiangle multispectral intensity and polarization measurements

摘要

Quantifying aerosols on a global scale is extremely important due to their strong but anomalous impact on the global climate. Traditionally, the aerosols retrievals use only the intensity measurements of the scattered light. However, these measurements are less sensitive to aerosol type and also suffer contamination from ground surfaces. It is with these limitations in mind that we plan to improve the quality and scope of aerosol retrieval by making use of soon to be available polarimetric sensors such as the Aerosol Polarimetry Sensor (APS) on the GLORY satellite and combine them with other available datasets such as lidar data from the CALIPSO satellite for vertical profiling, and high-spatial-coverage intensity measurements from MODIS. To handle these extremely large sensor data sets, we will explore the capabilities of various statistical methods and even combine them to create inversion algorithms that will work best. Up to now, we worked with the simplest case, the single-scattering approximation and built a retrieval algorithm using multi-angular, multi-wavelength simulated measurements of intensity and polarization. The inversion techniques we used are the optimal estimator and the neural networks.