Comparative evaluation of IHACRES, ANN and Linear Regression to simulate runoff in Tashk- Bakhtegan basin
Subject Areas : Article frome a thesisAlireza Pilpayeh 1 , Aydin Bakhtar 2 , Akbar Rahmati 3 , Afshin Shayeghi 4
1 - Assistant Professor, Department of Civil Engineering, Parsabad Moghan Branch, Islamic Azad University, Parsabad, Iran.
2 - Graduate M.Sc. student in Water Resources Management, Department of Water Engineering, Urmia University, Urmia, Iran.
3 - Graduate M.Sc. student in the field of water resources management, Faculty of Agriculture, University of Tehran, Tehran, Iran.
4 - Graduate M.Sc. student in the field of water resources management, engineering faculty, water engineering department, Imam Khomeini International University, Qazvin, Iran.
Keywords: Neural network, flood, Artificial Intelligence, Rainfall- Runoff models,
Abstract :
Streamflow simulation is one of the most important issues in the field of hydrology and water resources. Forecasting of runoff in future periods can determine the rate of flooding and also the increasing and decreasing trend of river discharge. There are various methods to simulate or predict runoff rates, including statistical methods, rainfall-runoff models and artificial intelligence. In this study, three models consisting of neural network, IHACRES and linear regression were used to estimate runoff in Tashk-Bakhtegan basin and evaluated using statistics such as correlation coefficient, coefficient of determination and other error indices. The results showed that runoff and precipitation have a non-linear relationship and therefore, the linear regression model cannot be trusted to estimate runoff. The results also showed that, in general the neural network model due to the nonlinear relationship between precipitation and runoff compared to the other two models have good performance in estimating monthly runoff in Tashk-Bakhtegan basin.