Performance Evaluation of ERA5 Extreme Precipitation in the Yangtze River Delta, China

Abstract

Accurate extreme precipitation information is crucial for disaster risk management, social and economic development security, and climate change research. Taking the Yangtze River Delta (YRD), China, a high-impact area of extreme precipitation, as an example, this study evaluates the spatiotemporal performance of extreme precipitation in the latest fifth-generation reanalysis dataset from the European Centre for Medium-Range Weather Forecasts (i.e., ECMWF ERA5) for 1961–2018 based on surface observational precipitation data. The results showed that the 90th-percentile threshold of extreme precipitation extracted from ERA5 data with a daily precipitation amount >1 mm is closer to the actual observations. The ERA5 data can effectively capture the spatiotemporal patterns of the observed extreme precipitation in the YRD. The ERA5 data can successfully represent the seasonal cycle and interannual variability of daily, daytime, and nighttime extreme precipitation. However, the daytime (nighttime) extreme precipitation frequencies and amounts tend to be overestimated (underestimated) for the period 1961–2000, whereas they were significantly underestimated for the period 2000–2018. The trend estimation of seasonal and annual extreme precipitation in ERA5 needs to be improved. The ERA5 data revealed that the extreme precipitation in the YRD was dominated by large-scale precipitation, followed by convective precipitation, but their long-term trends were not clear. This study has conducted a detailed and reliable evaluation of the ERA5 extreme precipitation data. The findings serve as valuable guidance and provide accurate references to extreme climatic variables for data users and algorithm developers.