Long-term remote tracking the dynamics of surface water turbidity using a density peaks-based classification: A case study in the Three Gorges Reservoir, China

•A water optical classification method was developed using density peaks.•The class-specific retrieval models of surface water turbidity were established.•Long-term dynamics of surface water turbidity was tracked using satellite images.•Interaction of water level and rainfall is probably the main environmental driver. Surface water turbidity (SWT), as a low-cost proxy of surface suspended sediment, is important for characterizing the hydro-ecological process and light availability in the lake or reservoir ecosystem. In this study, we proposed the combined use of HJ-1 charge-coupled device imaging and field observation to track the long-term SWT dynamics with environmental changes in Lakes Gaoyang, Hanfeng, and Changshou of the Three Gorges Reservoir, China. In situ remote sensing reflectance spectra were utilized to develop the characteristic spectral indexes for the SWT estimation in different water optical classes separated by a density peaks-based classification. Significant correlations were found between the red-, four-band, band ratio spectral indexes and SWT (determination coefficient >0.71 and root-mean-square error