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

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.