棉花滩水库浮游植物群落结构演替特征及风险预警研究

Research on the Succession Characteristics of Phytoplankton Community Structure and Risk Warnings in the Mianhuatan Reservoir

水华的暴发是浮游植物在适宜的水文、气象及营养盐条件下大量增殖并聚集的过程,深入探索水华暴发与相应环境因子关系可为水华风险预警提供依据。以福建省龙岩市棉花滩水库连续两年的水文、气象和水质水生态监测数据为基础,分析该水库浮游植物群落演替规律,并建立了LSTM人工神经网络模型开展水华风险预警。结果表明:棉花滩浮游植物样品中共发现6门63属,主要优势藻种为小环藻、针杆藻、小球藻、衣藻;当以日均气温、水温、风向、入库流量作为输入变量时,监测方式最为简便且输出结果最优,预测值与实测值拟合度为0.76,水华高风险时段,相对误差在0.02~0.73范围内,且模型稳定性较好。该模型有望用于棉花滩水库水华预警,为优化当地水资源管理提供手段。

The outbreak of algal bloom is a process of phytoplankton proliferation and aggregation under suitable hydrological,meteorological and nutrient conditions. Further exploration of the correlation between algal bloom and the corresponding environmental factors can provide a basis for risk warnings of algal bloom. Based on the nearly two-year monitoring data of hydrology,meteorology,water quality and water ecology of the Mianhuatan Reservoir in Longyan City of Fujian Province,the succession characteristics of phytoplankton community are analyzed in this paper. Meanwhile,the LSTM artificial neural network model is established to forecast the algal bloom. The results show that 6 phyla and 63 genera were found in the Mianhuatan Reservoir,and the main dominant species were cyclotella,synedra,chlorella and chlamydomonas. When the daily average temperature,water temperature,wind direction and inflow flow are used as input variables,the monitoring method is not only the simplest but also the best result. The fitting coefficient between the predicted value and the measured value has reached 0.76. The relative error is in the range of 0.02~0.73 in the high-risk period,which shows the stability of the model is acceptable. The model is expected to be used for risk warning of algal bloom in Mianhuatan Reservoir and provides a means for optimizing local water resources management.