Monitoring cyanobacterial blooms in China's large lakes based on MODIS from both Terra and Aqua satellites with a novel automatic approach

Cyanobacterial harmful algal blooms (CyanoHABs) pose significant environmental threats in China's large, shallow, and turbid lakes. Current satellite remote sensing methods for monitoring CyanoHABs in turbid water are limited in accuracy and frequency. The commonly used spectral index approaches often misidentify CyanoHABs by highly turbid water, and the segmentation threshold algorithms are not robust enough. Additionally, the dynamic nature of CyanoHABs requires high temporal resolution, which commonly used remote sensing data fail to provide. This study proposed a novel monitoring method utilizing a new spectral index, Anti-Turbid Algal Bloom Index (ATBI), specifically designed to minimize the impact of highly turbid water, coupled with an iterative triangle algorithm for automated ATBI threshold determination. We applied this approach to MODIS data from both Terra and Aqua satellites, achieving a temporal resolution of 0.5 days. The validation results show that the new method has a high accuracy with an F1-score of 0.85, which is higher than the commonly used algorithms with an F1-score of approximately 0.60. Utilizing this approach, we analyzed the spatiotemporal distribution of CyanoHABs in China's large lakes from 2000 to 2022, including Lakes Hulunhu, Chaohu and Taihu. Notably, Lake Hulunhu exhibited an increase in frequency and duration of CyanoHABs, while Lakes Chaohu and Taihu showed a rise in duration but a decline in frequency post 2018 and 2017, respectively. Air temperature and wind speed are critical factors influencing CyanoHABs variations in these lakes. Comparative analysis using data from Terra, Aqua, and both satellites combined demonstrated that the integrated data captures more accurate CyanoHABs information. These findings showed that our method can enhance CyanoHABs monitoring capabilities and offer insights for future research.