Performance assessment of GR4J rainfall-runoff model in daily flow simulation of Nirchai Watershed, Ardabil province

The necessity of precise estimation of daily discharge values and prediction of hydrological response ​​in ungagged watersheds highlights the importance of rainfall-runoff model applications. The GR4J model is an integrated rainfall-runoff model that produces reasonable results with limited input and easy accessible data. In this study, the performance of GR4J model was assessed in simulating daily flow of the Nirchai watershed, Ardabil province. Daily precipitation, evapotranspiration, and discharge data were used during the period of 2002-2012 years. After obtaining the input model parameters, the model was calibrated using trial and error process and consequently validated against different available data set. The simulation results showed that despite the over-estimation of the simulated values in validation stage, the simulation was consistent with flow variations well as predict the low flow periods in an acceptable manner. The evaluation of the results based on Nash-Sutcliffe criteria showed that the range of model accuracy varies from between 0.31-0.77 in different years. The evaluation of the results showed that the GR4J model with moderate and acceptable results can be used for daily flow simulation in similar watershed in terms of climate, physiographic and land use. The difference in elevation of the recording climatic station with the real state watershed condition may affect the accuracy of the modelling results. Despite testing the GR4J model in numerous watersheds, it seems that the area and hydrological response of the watershed are effective factors on modelling results (with daily time scale). Therefore, the fast-hydrologic response of the Nirchai watershed due to high slope amount and small area may not properly reflected in the rainfall-runoff model simulation