Simulation of river discharge upstream of Dez dam using metaheuristic models

Measuring river discharge has always been one of the fundamental challenges in river management. The use of accurate tools for its calculation is essential. Numerical, analytical, artificial intelligence, and empirical methods are the most common approaches for measuring daily river flow. This study develops a hybrid intelligent model based on the Support Vector Regression (SVR) approach for simulating river discharge. For this purpose, three optimization algorithms—Wavelet, Whale Optimization, and Particle Swarm Optimization—were utilized to simulate river discharge. The modeling was based on data and statistics from the hydrometric stations in the Dez basin, including Tireh Marouk, Cham Chit, Sezar, and Tang-e Panj, as a case study. Four combined input parameter scenarios from ۲۰۱۲ to ۲۰۲۲ were applied. To evaluate the performance of the models, statistical indices including correlation coefficient, root mean square error (RMSE), mean absolute error (MAE), and Nash-Sutcliffe Efficiency were used. Additionally, scatter and Taylor diagrams were employed for result analysis. Results showed that the combined scenarios in the examined models improved model performance. The evaluation metrics indicated that the SVR-Wavelet model achieved a correlation coefficient of ۰.۸۹۸-۰.۹۸۵, root mean square error (m³/s) of ۰.۰۸۸-۰.۰۰۸, mean absolute error (m³/s) of ۰.۰۴۰-۰.۰۰۴, and Nash-Sutcliffe Efficiency of ۰.۹۵۱-۰.۹۹۵ during the validation phase. Overall, the results demonstrated that the use of intelligent models based on the Support Vector Regression approach could be an effective method for ensuring the sustainability of river engineering.

کلیدواژه ها:

Hydrometric Station ، support Vector Regression ، River Discharge ، wavelet