Remote sensing estimation of Chlorophyll a and suspended sediment concentration in turbid water based on spectral separation

Abstract

Chlorophyll a and suspended sediment are important indicators of water quality, and remote sensing estimation of them is difficult due to the optical complexity of turbid water. The spectrum above water surface is influenced by phytoplankton, suspended sediment and colored dissolved organic material in water, thus spectral separation is important before estimating one specific component. Based on the field experiment of pond water and Taihu lake, China, this study calculated the Gaussian parameters of Chlorophyll a (Chla) and suspended sediment (SS) through spectral decomposition, and then these parameters were used to separate the mixed spectrum of water samples from pond water and Taihu lake. After spectral separation, the Chla estimation model based on the peak height at 650nm has high accuracy (R2=0.78, RMSE=4.80mg/m3), better than the band-ratio model; the SS estimation model based on the peak height at 811nm (R2=0.82, RMSE=6.80mg/L) performs better than the single-band model. Results in this study indicate that spectral separation based on Gaussian parameters is a good method for Chla and SS estimation in turbid lake water.