Abstract—In this study, rainfall-runoff modeling was carried out in Latyan dam watershed using artificial neural networks (ANNs) and adaptive neuro-fuzzy interface system (ANFIS). For this reason, 92 MODIS instrument images have obtained from NASA website for 2003, 2004 and 2005 years. Snow cover area (SCA) was extracted from all images. Then, snow water equivalent (SWE) was computed using SCA and SWE for mentioned years. Rainfall, temperature and SWE were used as inputs for ANN and ANFIS. Root mean square error (RMSE), Nash–Sutcliffe efficiency coefficient (NS) and determination coefficient (R2) statistics are employed to evaluate the performance of the ANN and ANFIS models for forecasting runoff. Comparison of the obtained results reveals that the performance of ANN and ANFIS was very good for snowmelt runoff prediction. Based on the results of test stage, ANN with RMSE=0.04 m3 s-1, NS=0.85 and R2=0.68 and is superior to rainfall-runoff modeling than the ANFIS with RMSE=0.05 m3 s-1, NS=0.65 and R2=0.62. The combination of ANN and ANFIS by using daily SWE as input proved to be an excellent alternative to perform high quality daily snowmelt runoff prediction.

Index Terms—ANN, ANFIS, rainfall- runoff modeling, SWE, latyan watershed.

M. Vafakhah is with the Department of Watershed Management Engineering, Faculty of Natural Resources, Tarbiat Modares University, P. O. Box 46417-76489, Noor, Mazandaran Province, Iran; (e-mail: vafakhah@ modares.ac.ir).
F. Sedighi and M. R. Javadi are with Islamic Azad University, Noor Branch, (e-mail: sadighi.fatemeh@yahoo.com, javadi.desert@gmail.com).

[PDF]

Cite:M. Vafakhah, F. Sedighi, and M. R. Javadi,  "Modeling the Rainfall-Runoff Data in Snow-Affected Watershed," International Journal of Computer and Electrical Engineering vol. 6, no.1, pp. 40-43, 2014.