Radiative Transfer and Inverse Model Development for Polarimetric Aerosol Remote Sensing
Aerosols and their interactions with clouds are the major uncertainty sources of climate forcing. A polarimeter that measures both radiance and polarization of scattered light has been proved to be highly valuable to improve the remote sensing ability. In this talk, I will introduce the progress of our group on radiative transfer and inversion model development, with applications to aerosol retrievals using the observations of JPL’s airborne polarimeter - Multiangle SpectroPolarimetric Imager (AirMSPI) and the observations of French satellite-borne polarimeter - POLDER/PARASOL. I will a) introduce the developement of a Markov chain model for computing polarized radiative transfer from ultraviolet to infrared. This model accounts for non-spherical particles with random and preferred orientations and the cross-polarization in their extinction, and is combined with neural-network for improving computational efficiency; b) discuss the benefits of using real multi-angle polarimetric measurements to constrain aerosol properties; c) introduce an correlation-based inversion method for improving aerosol retrieval accuracy and efficiency; and d) demonstrate the observation of cloudbow shift from polarimetric imaging of an stratocumulus cloud, and the resultant estimate of pixel-scale cloud-top droplet size distribution.