Skill assessment of TIGGE precipitation forecasts in the four basins of Iran

Abstract

In atmospheric sciences, forecast assessment, the process of examining the quality of forecasts, is very important. For this reason, we carried out a comprehensive evaluation of the most important global numerical weather forecast (NWP) models for the Iranian basins of Karkheh, Karun, Urmia, and Sefidrood. This study aims to evaluate the output of six different NWP models from the TIGGE database in important basins of Iran, to determine the most appropriate model for each basin. The assessment period took place between 2008 and 2018 and evaluated probabilistic and non-probabilistic methods. In the Karun Basin, the global NWP models were shown to perform better than in the other three basins. Overall, the UKMO (United Kingdom Met Office) and ECMWF (European Medium-Range Weather Forecast) performed best, while the Korean Meteorological Administration (KMA) showed the poorest performance among the models in the four basins. In the 25 mm threshold, most models underestimated precipitation events. The ECMWF model showed reliable and sharp forecasts and was better able to discriminate between low, medium, and high precipitation events compared to other models. Compared to those of other models, the UKMO model demonstrated greater brier skill scores. In general, the ECMWF, UKMO, ECCC (Environment and Climate Change Canada), and National Center for Environmental Forecast (NCEP) models performed relatively satisfactory in all four basins, although they forecasted better in basins with higher mean precipitation. These models are recommended for flood warning systems, and water resource decision making in Iran. However, all forecasts will need post-processing for operational applications.