Evaluation of Extreme Climate Indices over the Three Northeastern Provinces of China Based on CMIP6 Models Outputs

Abstract

This study evaluates the performance of Global Climate Models (GCMs) in simulating extreme climate in three northeastern provinces of China (TNPC). A total of 23 GCMs from the Coupled Model Intercomparison Project Phase 6 (CMIP6) were selected and compared with observations from 1961 to 2010, using the 12 extreme climate indices defined by the Expert Team on Climate Change Detection and Indicators. The Interannual Variability Skill Score (IVS), Taylor diagrams and Taylor Skill Scores (S) were used as evaluation tools to compare the outputs of these 23 GCMs with the observations. The results show that the monthly minimum of daily minimum temperature (TNn) is overestimated in 55.7% of the regional grids, while the percentage of time when the daily minimum temperature is below the 10th percentile (TN10p) and the monthly mean difference between the daily maximum and minimum temperatures (DTR) are underestimated in more than 95% of the regional grids. The monthly maximum value of the daily maximum temperature (TXx) and the annual count when there are at least six consecutive days of the minimum temperature below the 10th percentile (CSDI) have relatively low regional spatial biases of 1.17 °C and 1.91 d, respectively. However, the regional spatial bias of annual count when the daily minimum temperature is below 0 °C (FD) is relatively high at 9 d. The GCMs can efficiently capture temporal variations in CSDI and TN10p (IVS < 0.5), as well as the spatial patterns of TNn and FD (S > 0.8). For the extreme precipitation indices, GCMs overestimate the annual total precipitation from days greater than the 95th percentile (R95p) and the annual count when precipitation is greater than or equal to 10 mm (R10 mm) in more than 90% of the regional grids. The maximum number of consecutive days when precipitation is below 1 mm (CDD) and the ratio of annual total precipitation to the number of wet days (greater than or equal to 1 mm) (SDII) are underestimated in more than 80% and 54% of the regional grids, respectively. The regional spatial bias of the monthly maximum consecutive 5-day precipitation (RX5day) is relatively small at 10.66%. GCMs are able to better capture temporal variations in the monthly maximum 1-day precipitation (RX1day) and SDII (IVS < 0.6), as well as spatial patterns in R95p and R10mm (S > 0.7). The findings of this study can provide a reference that can inform climate hazard risk management and mitigation strategies for the TNPC.