A bio-optical numerical approach for remote retrieval of total suspended matter from turbid waters

Abstract

A novel semi-analytical algorithm was developed to retrieve total suspended matter (TSM) in turbid waters. A turbidity index developed from remote sensing reflectance (R sub(rs)) of three bands of visible spectrum was employed to estimate TSM. Adopting a partial differentiation approach, wavelength Lambda sub(1) at which maximum absorption of TSM was identified between 650 nm and 750 nm. The second (Lambda sub(2)) and third (Lambda sub(3)) wavelengths were identified through radiative transfer modelling and parameterization of particulate backscattering to minimize the effect of other optically active substances and account for backscattering effects, respectively. Thus, the wavelengths identified were 679 nm (Lambda sub(1)), 695 nm (Lambda sub(2)), and 704 nm (Lambda sub(3)). Regressing the index with 50 randomly chosen in situ data points (Zuary River estuary, Goa) resulted in a best-fit polynomial form of algorithm. This algorithm was validated with a different set of in situ data points (n = 116), and resulted in a correlation coefficient, r = 0.88. In addition, a comparative analysis of the developed algorithm with forty-one empirical and semi-analytical models of TSM indicated their non-suitability in varying optical conditions. The study further pointed out the significance of 695 nm as an inseparable band of any future optical sensor to retrieve TSM from remotely sensed data.