Author(s): Yongao Lu; Linglei Zhang; Min Chen; Ning Liao; Youcai Tuo; Yunxiao Jia

Linked Author(s): Min Chen

Keywords: River water temperature; Genetic algorithm; Support vector regression; River ecological management; Reservoir

Abstract: The construction of reservoirs leads to changes in the downstream river water temperature (RWT), which in turn affects biological processes in the rivers. Therefore, efficient and accurate prediction of RWT is helpful for river ecological management and protection. In this paper, support vector regression (SVR) is used to develop the machine learning prediction model of RWT downstream Fengman Reservoir (FMR), and genetic algorithm (GA) is used to optimize the model parameters to get an improved SVR model (GA-SVR). Model input indicators are screened through the influence mechanism of RWT and correlation analysis and the RWT at three stations downstream the dam is modeled. The performance of the model before and after the improvement and at different stations is compared and analyzed. The results show that GA-SVR is suitable for predicting RWT and the range of RMSE, MAE and NSE for the prediction results is 0.576~1.235

DOI: https://doi.org/10.3850/978-90-833476-1-5_iahr40wc-p1384-cd

Year: 2023