Machine learning based aerosol and ocean color joint retrieval algorithm for multiangle polarimeters over coastal...

Area of Science:

• Earth Science
• Atmospheric Science
• Oceanography

Background:

• NASA's Plankton, Aerosol, Cloud, ocean Ecosystem (PACE) mission utilizes advanced multi-angle polarimeters (MAPs) like HARP2 and SPEXone.
• Existing retrieval algorithms face challenges with computationally expensive radiative transfer models, especially for large datasets.

Purpose of the Study:

• To expand the FastMAPOL joint retrieval algorithm for coastal waters using multi-parameter bio-optical models.
• To represent aerosols with six distinct components for improved retrieval accuracy.
• To evaluate the performance of different PACE MAP configurations for retrieving aerosol, ocean color, and surface parameters.

Main Methods:

• Developed neural network-based forward models for coastal bio-optical properties.
• Incorporated a six-component aerosol model with retrievable abundances.
• Applied the enhanced FastMAPOL algorithm to synthetic PACE mission data (HARP2, SPEXone, combined).

Main Results:

• The SPEXone-only configuration is effective for aerosol retrieval, comparable to HARP2-only.
• SPEXone offers superior ocean color retrieval due to wider spectral coverage.
• HARP2 excels in retrieving surface wind speed due to its wide field of view.
• The combined HARP2+SPEXone configuration provides the best overall retrieval performance.
• Including sun glint geometries significantly improves wind speed and absorbing aerosol retrieval.

Conclusions:

• The enhanced FastMAPOL algorithm effectively retrieves aerosol, ocean color, and surface parameters in coastal environments.
• Combined measurements from HARP2 and SPEXone offer synergistic benefits for comprehensive Earth system studies.
• The algorithm's atmospheric and BRDF correction modules enhance the utility of PACE polarimeter data for ocean biogeochemistry.