Author(s): Zihang Han; Hui Cao; Hairong Zhang; Linlin Li; Biqiong Wu; Yafang Wang; Xiaoshan Sun
Linked Author(s):
Keywords: Pangu-Weather Forecasting Model; Yangtze River Basin; Precipitation; Evaluation; SAL method
Abstract: In order to aid management personnel, weather forecasters, and relevant researchers in understanding the precipitation forecasting capabilities of the Pangu-Weather Forecasting model in the Yangtze River Basin, this study evaluates the precipitation over the Yangtze River Basin based on a typical precipitation case occurring on August 26,2023. The Pangu-Weather Model, driven by ERA5 (ECMWF Reanalysis v5) data, utilized a greedy algorithm to perform forecasting with a 24-hour step for 120-hour (5 days). Pangu’s forecast precipitation were compared against observations from CMA (China Meteorological Administration) over the Yangtze River Basin. Quantitative evaluation metrics including the bias of rainfall intensity grade, and the object-oriented forecast SAL (Structure, Amplitude, and Location) verification method, are utilized to evaluate the forecast skills. In comparison with the observations, the evaluation indicates that hindcasting with Pangu-Weather Model captures the spatial distribution features over the Yangtze River Basin reasonably. It was found that for the precipitation events on August 26 exclusively, Pangu’s simulation performance in terms of precipitation location and intensity are comparable or superior to those from the ECMWF (European Centre for Medium-Range Weather Forecasts) within the 5-day lead time. When making subjective corrections to the heavy rain or above during the flood season in the upper reaches of the Yangtze River, the precipitation location from ECMWF can be used as a reference, and the rainfall intensity can refer to the Pangu-Weather Model. This study significantly contributes to understanding the forecasting performance of precipitation in the Yangtze River Basin of the Pangu-Weather Model.
Year: 2024