Author(s): Xiaxia Li; Minghao Ji; Chao Wang; Yao Cheng; Hongtao Li; Yong Wu; Lixin He; Xiaohui Lei; Bin Chen; Beibei Chai

Linked Author(s): Xiaohui Lei

Keywords: Yangtze River basin; Algae; Structural equation model; Neural network; Algae predict

Abstract: In recent years, lake eutrophication has become a global focus, but the Yangtze River basin, as the largest river in China and the third largest basin in the world, its eutrophication problem still exists. As an important indicator of nutrient status of water, it is still challenging to predict algae and study the driving factors of their growth. Mantel and structural equation model (SEM) were used to study phytoplankton in the Tuojiang River Basin, and it was found that the environmental factors affecting phytoplankton community growth in the actual sampling data were basically consistent with the national monitoring data at the same time scale in neighboring areas, and the growth of phytoplankton was mainly limited by phosphorus, and the ratio of nitrogen to phosphorus was the main factor affecting phytoplankton growth. The use of national monitoring data has certain research significance and reference value. Therefore, this paper took the Yangtze River Basin as the research object, and the time scale was from May 31,2021 to May 23,2023. Structural equation model (SEM) wass used to study the water environment indicators (water temperature (WT), PH, dissolved oxygen (DO), electrical conductivity and turbidity) of lake water bodies at 9 representative national monitoring stations in the basin. The path relationship between water quality indexes (potassium permanganate index (COD), total nitrogen (TN), total phosphorus (TP), ammonia nitrogen (HN3-N), nitrogen to phosphorus ratio (N/P) ) and phytoplankton index chlorophyll a (Chla). According to the analysis results, the control measures of each monitoring point were put forward. According to structural equation model results, the training set and test set were divided by data driven models (LSTM, CNN-LSTM, CNN-BiLSTM) that had a great impact on Chla, and the Chla content was accurately predicted. The results showed that the main factors affecting algae growth in the Yangtze River Basin were nitrogen/phosphorus ratio (N/P), total phosphorus (TP), dissolved oxygen (DO) and water temperature (WT). Through algorithm selection, CNN-BiLSTM model was finally determined to be the most suitable for algae prediction in the Yangtze River Basin. This study aims to provide new ideas for algae prediction methods and provide theoretical basis for water quality improvement, pollution control and algal control technology in the Yangtze River basin through the parallel analysis of measured data, monitoring data and model prediction.

DOI: https://doi.org/10.3850/iahr-hic2483430201-144

Year: 2024