干旱区生态环境知识资源中心

This study assessed the ability of three models of Local Linear Regression (LLR), Support Vector Machine (SVM) and Artificial Neural Network (ANN) to estimate evaporation from reservoirs as one of the critical components of hydrological cycle in arid and semi-arid regions. A case study has been carried out in the Chahnimeh water reservoirs of Zabol located in the Sistan plain of Iran. Among the models used, in terms of the evaluation criteria of root mean square error (RMSE), mean absolute error (MAE), and coefficient of determination (R-2), it is demonstrated that use of the nonlinear models of support vector machine (SVM), Broyden-Fletcher-Goldfarb-Shanno neural network (BFGSNN) and Conjugate Gradient neural network (CGNN) performed reasonably well in modeling the validation data compared to Local Linear Regression (LLR) model but both BFGSNN and CGNN failed to reach the highest possible values. In the meantime, the SVM model was able to provide more reliable estimations compared to others.