A bio-optical inversion scheme based on the radiative transfer equation

Ocean color inversion models are used to transform the spectral remote sensing reflectance (rrs) into inherent optical properties (IOPs, absorption, a and backscattering coefficients, bb) of the optically active oceanic constituents. Recently, a forward model based on a radiative transfer equation was developed in terms of IOPs with explicit inclusion of the volume scattering function (VSF). All terms in this forward model are expressed in measurable or derivable variables (IOPs, solar zenith angle, viewing angle and relative azimuth), shape of the VSF in backward direction and particulate backscatter ratio. The present study developed an inversion method for deriving IOPs from multispectral rrs. Three high quality datasets encompassing a wide range of optical properties were used for testing the inversion model. Test configurations involving combinations of different spectral shape models for subcomponent IOPs and particle backscatter ratio values were evaluated. Key results from the analysis are: 1) the inversion retrievals perform to comparable levels as the operational configuration of the NASA's Generalized Inherent Optical Property (GIOP) framework; 2) closure error between RT rrs and measured rrs have no significant effect on predicting the accuracy of the derived IOPs; and 3) the particle backscatter ratio is important to ZTT, but not explicitly included in other types of comparable semi-analytic algorithms. Results indicate that the IOPs derived from inversions using a fixed particle backscatter ratio set to 0.006 resulted in lower errors and the least bias compared to other settings, including dynamic settings as functions of chlorophyll. The proposed inversion method can be adapted for hyperspectral rrs inversions and applied for upcoming NASA’s Plankton, Aerosol, Cloud and ocean Ecosystem (PACE) mission.