Correlation analysis of spectral reflectance in determining preliminary algorithms for water quality monitoring in Taihu Lake, China

This paper presents an empirical relationship that maybe used to potentially estimate suspended sediments, chlorophyll concentrations and yellow substances from remote sensing reflectance measurements in highly turbid water body-Taihu Lake, China. Based on the characterization of the reflectance spectra of chlorophyll and suspended sediments, the correlation between constituents concentration with spectral reflectance and reflectance ratio of arbitrary wavelength are analyzed and discussed, which provides a substantial basis for wavelength selection criterion for the inversion of water constituents concentrations in Taihu Lake. A empirical approach of relating reflectance ratio to in situ measurements through regression analysis was used. The regression results show that the ratio between the reflectance of the wavelength with higher correlation is superior to the ratio between the reflectance with lower correlation. Based on the comparison of the correlation analysis of the water constituents, the generalized band ratio of high correlation with water constituents were proposed. These results implies that the reflectance spectral features and the correlation profiles play a crucial role in providing appropriate criterion for band selection on water quality monitoring in Taihu Lake, China, which is especially constructive for remote sensing purposes and should be applied in other turbid waters