Comparison and Evaluation of precipitation estimated by ERA-Interim, PERSIANN-CDR and CHIRPS models at the upstream of Maroon dam

Precipitation is a major component of the hydrological cycle, which has significant changes in location and time. The lack of suitable data for this parameter causes a problem in hydrological forecasts. Since satellite data provides a new solution for estimating rainfall with spatial and temporal variation, this study evaluate the accuracy of some of these data types, including high-resolution spatial data consist of ERA-Interim, CHIRPS and PERSIANN-CDR at the upstream of the Maroon Dam on daily, monthly and annual timescales. In order to evaluate gridded precipitation data and observational data from 2003 to 2014, it was considered. The results show that estimation of the annual rainfall data of the gridded precipitation models is underestimated and estimates the average annual precipitation less than the mean annual observational precipitation. In the estimation of monthly precipitation with regard to the Nash-Sutcliff coefficient at Dehno, Idenak and Margoon stations, the ERA-Interim model and at the Ghale-Raeesi station CHIRPS model indicate the best performance compared to other models. In the daily rainfall estimation, like the monthly rainfall, the best estimate at the Idenak station is the ERA-Interim model, which has a NSE of 0.63 and the best estimate of precipitation in all stations is by ERA-Interim. ERA-Interim has the best performance from the 3 gridded models in the correct detection of rainy days. The best performance of this model is in determining the correct rain days with POD = 0.53 at Idenak station.