Determining the Appropriate Temporal Resolution of Short and mid-terms of Global Precipitation Forecasting Systems over Iran

Precipitation forecasting models play important role in the performance of flood and meteorological warning systems. In this research, the efficiency of five numerical weather prediction (NWP) models, which exist in the TIGGE database, are assessed to determine the best temporal resolution of forecasted datasets at distinct climate regions of Iran, during 2014-2018. Findings show that by increasing the lead time the accuracy of all forecasts decreases significantly. Moreover, most of the NWP models, especially the ECMWF and UKMO perform well, based on correlation coefficient (CC) and RMSE metrics, up to lead time of 3 days. Also, results indicate that by removing biases from the raw forecast datasets, the performance of all NWP models in different lead times increases considerably. After bias correction, the RMSE values of ECMWF, JMA, and KMA models in the lead time of 10 days reduces about 70, 65, and 73%, respectively, and, except for JMA, all NWP models perform well in most climate regions. The JMA model in humid climate zones (north and west parts of Iran) has a high level of bias and leads to unreliable forecasts.