An Improved Approach to Retrieve IOPs Based on a Quasi-Analytical Algorithm (QAA) for Turbid Eutrophic Inland Water

Absorption and backscattering coefficients, two important inherent optical properties of water, play a significant role in environmental monitoring and the study of biogeochemical cycles. In this study, the optical classification of turbid eutrophic inland Case II water was the basis for the development of an IOP-retrieval algorithm. Results show that the water can be divided into two types with different optical characteristics using the spectral slope defined as the spectrum slope of remote sensing reflectance between a 677 nm absorption valley and a 701 nm reflection peak at a threshold value of 0.32. A method of segmented simulation was proposed to model a backscattering coefficient at wavelength 400-800 nm, in which the backscattering coefficient at wavelength 400-685 nm was simulated using a power function and the backscattering coefficient at a wavelength greater than 685 nm was regarded as constant. The simulation precision of this proposed segmented simulation method was much higher than the precision using a power function. Based on the quasi-analytical algorithm (QAA) algorithm proposed by Lee et al., two algorithms using reference bands 550 and 675 nm separately for estimating a backscattering coefficient were developed. Weight coefficients of the two retrieval algorithms were also calculated based on their estimation errors. The optimal backscattering coefficient was determined through a multimodel collaborative retrieval method. Finally, the total absorption coefficient was derived from a bio-optical model introduced by Lee et al. Accuracy assessment results demonstrated that the proposed IOP-retrieval algorithm using double reference bands (QAA-DB) and based on the optical classification algorithm can be successfully applied to optically complex eutrophic inland waters with a mean absolute percentage error (MAPE) of 19.71% and root-mean-square error (RMSE) of 1.3933.