Evaluation of SWAT and SVM Models to Simulate the Runoff of Lighvanchay River

In this study, performances of the hydrological model of Soil and Water Assessment Tool (SWAT) and support vector machine (SVM) in monthly simulating the runoff of Lighvanchai river were evaluated. After collecting the required data, the models were calibrated and verified. The SUFI-2 algorithm was used for uncertainty analysis of the SWAT model. The discharge of the Urmia lake basin was simulated using SWAT model and the results obtained for the Lighvan station were used in this study. The SVM model was applied using the rainfall and runoff data measured in the Lighvan station. In order to improve the results, the time series with different lag times were used. Three statistical criteria including coefficient of determination (R2), Nash-Sutcliffe coefficient (NS) and root mean square error (RMSE) were used to evaluate the performances of the models. The results revealed the ability of the both models in simulating the discharge of Lighvanchai river. However, the SWAT model had better performance than SVM in simulating the maximum values of the runoff of Lighvanchai river with NS and RMSE values of 0.71 and 0.41 m3s-1 respectively.