Optimal Hyperspectral Characteristic Parameters Construction and Concentration Retrieval for Inland Water Chlorophyll-a Under Different Motion States

In recent decades, the rapid expansion of phytoplankton blooms caused by lake eutrophication has led to severe ecological destruction and impeded the sustainable economic development of local regions. Chlorophyll-a (Chl-a) is commonly used as a biological indicator to detect phytoplankton blooms due to its ease of detection. To improve the accuracy of Chl-a estimation in aquatic systems, an accurate understanding of its true spectral characteristics is imperative. In this study, a comprehensive and realistic experimental scheme was designed from the perspective of real algal strains and real water states. Both in situ and laboratory-based hyperspectral data were collected and analyzed. The results show that there are huge spectral differences not only between laboratory-cultured and real algae strains, but also between static and disturbed water surface conditions. A total of ten different categories of spectral characteristics were selected in both disturbed and static states. Then, six parameters with the best models to the Chl-a concentration were identified. Finally, two linear models of the Chl-a concentration at peaks of 810 nm and 700 nm were identified as the best estimation models for the static and disturbed states, respectively. The results provide a scientific reference for the large-scale retrieval of the Chl-a concentration using satellite remote sensing data. This advancement benefits inland water monitoring and management efforts.