Rainfall-Runoff Modeling Using Artificial Neural Networks and HEC- HMS (Case Study: Catchment of Gharasoo)

One of the methods used in various scientific fields which can simulate the complex process of rainfall – runoff is the artificial neural network model. The present study aims to review the efficiency of artificial neural network in the simulation of the rainfall – runoff process and to compare the results obtained from this investigation with that of HEC-HMS model in Gharasoo catchment area located in Ardebil province. The research data is comprised of daily rainfall and daily and momentary discharge the aforementioned river over a 30-year statistical period. According to research results, each of the methods used in this study have a considerable ability to estimate the rainfall – runoff amount of Gharasoo catchment area. Thus, with a sufficient amount of confidence, it can be argued that these methods can be of use for estimating the rainfall – runoff amount and they can be considered as good and efficient methods in this field. Among the models above, by taking into account the evaluation models, the second model of neural network (today's rainfall and yesterday's runoff) has been selected as the best model with the highest determination coefficient (0.88) and lowest error in the verification stage. In addition, the results obtained from comparing the distribution graphs of the three model indicate that the second graph of neural network has more attractive features than the other two models