Effect of Snow Water Equal Consideration in Runoff Prediction by Using RBF and ANFIS Models

Abstract

Abstract. Flood is one of devastating natural disasters prediction of which is significantly important. Rainfall-runoff process and flood are physical phenomena with very difficult investigation due to effectiveness of different parameters. Various methods have so far introduced to analyze these phenomena. Current study was aimed to investigate the performance of RBF and ANFIS models in simulation of rainfall-runoff process involved with Snow Water Equivalent (SWE) height in Latian watershed, Tehran province, Iran. Toward this attempt, 92 MODIS images were provided by NASA website during three water years 2003-2005, snow cover surface area in all images was extracted and finally SWE values were calculated for mentioned period. Also, precipitation height, temperature and discharge data of the study period were used for modeling. The results performance comparison of RBF and ANFIS models showed that the latter with rainfall-temperature-SWE inputs, 1-day delay, RMSE of 0.059 and R 2 of 0.656 and RBF model with rainfall-temperature-SWE inputs, 1-day delay, RMSE of 0.054 and R 2 of 0.35 had more accurate predictions than other models. It can be concluded from the results that involving SWE in the models improved their performance and increased their accuracy. Also, by comparing the results of ANFIS and RBF models, it can be concluded that ANFIS model with rainfall-temperature-SWE inputs, 1-day delay, RMSE of 0.059 and R 2 of 0.656 had better and more accurate prediction.