Simultaneous retrieval of aerosol and marine parameters using forward and inverse coupled atmosphere-ocean radiative transfer models

The MODIS aerosol algorithm over the ocean derives spectral aerosol optical depth and aerosol size parameters from satellite measured radiances at the top of atmosphere (TOA). It is based on the addition of Apparent Optical Properties (AOPs): TOA reflectance is approximated as a linear combination of reflectance resulting from a small particle mode and a large particle mode. The weighting parameter eta is defined as the fraction of the optical depth at 550 nm due to the small mode. The AOP approach is correct only in the single scattering limit. For a physically correct TOA reflectance simulation, we create linear combinations of the Inherent Optical Properties (IOPs) of small and large particle modes, in which the weighting parameter f is defined as the fraction of the number density attributed to the small particle mode. We use these IOPs as inputs to an accurate multiple scattering radiative transfer model. We find that reflectance errors incurred with the AOP method are as high as 30% for an aerosol optical depth of 2 at 550 nm. The retrieved optical depth has a relative error of up to 8%, and the retrieved fraction eta has an absolute error of about 6%. We show that the use of accurate radiative transfer simulations and a bimodal fraction f, yields accurate values for the retrieved optical depth and the fraction f.; We extend our estimation of MODIS algorithm for 9 aerosol models (4 fine mode models and 5 coarse mode models). We explore the influence of aerosol phase function on the retrieval results, as well as how much fluctuation occured in the retrieval of optical depth and fraction eta for multiple model application. We also use the correct IOP approach to retrieve both optical depth and fraction f. We found the retrieval results are more satisfactory using IOP approach.; A simple, yet complete bio-optical model for the inherent optical properties (IOPs) of oceanic waters is developed. This bio-optical model is specifically designed for use in comprehensive, multiple scattering radiative transfer models for the coupled atmosphere-ocean system. Such models can be used to construct next-generation algorithms for simultaneous retrieval of aerosol and marine parameters. The computed remote sensing reflectance Rrs(lambda) is validated against field measurements of Rrs(lambda) compiled in the SeaBASS data base together with simultaneous chlorophyll concentrations (C) ranging from 0.03 to 100 mg·m-3. This connection between Rrs and C is used to construct a chlorophyll concentration retrieval algorithm that yields reliable results for a large range of chlorophyll concentrations.; We use a comprehensive coupled atmosphere-ocean radiative transfer code (CAO-DISORT) to develop a two-parameter bio-optical model, including a complete set of inherent optical properties (IOPs) for global oceanic waters consistent with the NOMAD data set. We apply inverse methods based on iterative cost-function minimization to retrieve the chlorophyll concentrations as well as the absorption coefficient of CDOM at 443 nm.