Retrieving inherent optical properties for turbid inland waters: an improved quasi-analytical algorithm based on the linear spectral backscattering coefficient constraint

Abstract

The inherent optical property is a significant bridge between the hyperspectral remote sensing data and water color and water quality parameters. Based on the water optical radiation transfer process and existing quasi-analytical algorithm (QAA), this study provides an improved algorithm, namely a linear spectral backscattering coefficient constraint quasianalytical algorithm (LSBCC-QAA), suitable for the retrieval of inherent optical properties for turbid inland waters to address the deficiency of the QAA on the retrieval of inherent optical properties for turbid inland waters. LSBCC-QAA uses the water-leaving reflectance of the bands between 1600 and 1700 nm to estimate the water surface reflectance of the bands between 400 and 900 nm and selects 700~850 nm as the reference wavelengths to estimate the water backscattering coefficients, taking full advantage of the continuity of the backscattering coefficient spectrum. The preliminary validated results show that the particle absorption coefficient, particle backscattering coefficient and phytoplankton absorption coefficient retrieved by LSBCC-QAA are more consistent with the actual situation than those retrieved by the common QAA_v6 algorithm or QAA-Turbid algorithm. Compared with the measured particle diffuse attenuation coefficient, the error of the LSBCC-QAA retrieved particle diffuse attenuation coefficient ranges from 16.0% to 22.9%, and the average error is 18.4%.