FastMAPOL: An efficient multi-angle polarimetric aerosol and ocean color inversion framework powered by a deep neural network forward model

NASA's Plankton, Aerosol, Cloud, ocean Ecosystem (PACE) mission, scheduled for launch in the timeframe of 2023, will carry a hyperspectral Ocean Color Instrument (OCI) and two Multi-Angle Polarimeters (MAP): the UMBC Hyper-Angular Rainbow Polarimeter (HARP-2) and the SRON Spectro-Polarimeter for Planetary EXploration one (SPEXone)[1]. The MAP measurements contain rich information on the microphysical properties of aerosols and hydrosols, and therefore can be used to retrieve accurate aerosol properties for complex atmosphere and ocean systems. Most polarimetric aerosol retrieval algorithms utilize vector radiative transfer models iteratively in an optimization approach, which leads to high computational costs that limit their usage in operational processing of large data volumes produced by the MAP imagers. To resolve this issue, the SPEX team represented the ocean polarimetric reflectance by a deep neural network, and coupled it with a radiative transfer model for the atmosphere[2]. The hybrid model enabled sufficient efficiency for SPEXone data processing. However, it is still challenging for HARP-2, which has a swath 15 times wider than SPEXone. In this work, we propose a deep neural network (NN) model which represent the radiative transfer simulation of a coupled atmosphere and ocean system for the HARP instrument. Through evaluation of synthetic AirHARP datasets, the NN model achieves a numerical accuracy smaller than the instrument uncertainties with a running time of 0.01s. Using the NN as forward model, we built an efficient joint aerosol and ocean color retrieval framework called FastMAPOL, evolved from the well validated Multi-Angular Polarimetric Ocean coLor (MAPOL) algorithm[3]. Retrievals are conducted on both synthetic and field measurements from AirHARP for the ACEPOL campaign in 2017. From validation with both the synthetic data and the collocated HSRL aerosol products, we demonstrate that the aerosol microphysical properties can be retrieved efficiently and within acceptable error. The FastMAPOL framework and associated experience can be used to assist practical application of polarimetric aerosol and ocean color retrieval for PACE and other satellite missions that utilize polarimeters in the retrieval of geophysical properties from Earth observations.