In recent decades, a high frequency of extreme high temperature has occurred in many regions worldwide, with serious impacts on society and the economy. As the temperature increases, the sensitivity of extreme high temperatures to changing thresholds in the northern midlatitudes exhibits a different performance response. The results of this study show that extreme high temperature in the increasing phase is more sensitive to changes in the threshold in both observations and simulations (the largest difference in the speed of temperature increase occurs at 3.5 and 25 days decade^-1), primarily in North America and central Asia. However, an obvious discrepancy appears in the time series before 1980 and the spatial scale over North America between the Berkeley Earth Surface Temperature (BEST) and Hadley Centre Global Historical Climatology Network-Daily (HadGHCND) datasets. This finding highlights the fact that the old definition of being in the increasing temperature phase in modern climate history is problematic today. At the same time, when the old base period is selected, the frequency of extreme high temperatures will become a common event (close to 98% in a year) by 2100. Using 1961–90 as the base period is not suitable for calculating extreme temperatures in the future from the perspective of adapting to climate change. The increasing temperature threshold means there will be more frequent hot days, indicating that agriculture and species will be negatively affected, more wildfires will occur, and thus risks to humanity will increase.
 

气候变化，极端高温，CMIP6，极端高温阈值