Applicability of Reanalysis Data in Analyzing the Occurrence time of Extreme Temperature Events in China

The variation in the occurrence time of extreme temperature events is a key factor in some natural disasters. Analyzing and predicting these changes can help reduce disaster losses. Reanalysis data have extensive spatial coverage and long temporal records, making them a crucial data source for climate change research. However, their suitability for analyzing extreme temperature timing variation remains unclear. In this study, 11 temperature-related time indices, such as frost start date (SD0), were used to quantify the variation in extreme temperature events occurrence time, and 4 evaluation metrics were applied to assess the applicability of reanalysis data (ERA5-Land) in simulating these variations. Results show: (1) ERA5-Land effectively captured the spatial distribution of extreme temperature events occurrence time, which exhibited the same patterns of early and late occurrence time across different regions of China as observed data. (2) ERA5-Land effectively captured the interannual variation of extreme temperature events occurrence time. The trend simulated by ERA5-Land was consistent with observations. (3) ERA5-Land has good applicability in simulating the evolution of extreme temperature events occurrence time in China. However, its applicability varies across different regions and indices. Applicability was poor for SD0, growing season startdate (SD10), and length (GSL) in the Qinghai-Tibet region, and for summer day start date (SD25) in southern China. Caution is needed when using reanalysis data in these regions. The findings help to address gaps in current climate data validations and provide a reference for analyzing changes in the occurrence time of extreme temperature events.