Evaluation of the performance of CMIP6 models in simulating extreme precipitation and its projected changes in global climate regions

Extreme precipitation events usually lead to economic, agricultural, and social losses globally. The bias of different global circulation models (GCMs) is a major challenge in the projection of extreme precipitation in different climate regions. Revealing the extreme precipitation bias of Coupled Model Intercomparison Project (CMIP) GCMs is helpful for providing a reference for predicting extreme precipitation and understanding the performance of CMIP Phase 6 (CMIP6) GCMs. Eight extreme precipitation indices were used to describe extreme precipitation based on daily precipitation data retrieved from the Global Precipitation Climatology Project (GPCP) and 19 CMIP6 GCMs. Six evaluation metrics were adopted to assess the ability of the CMIP6-determined daily precipitation to describe extreme precipitation. The results showed that half of the GCMs overestimated extreme precipitation in the Sahara, Arabian Peninsula, and Central Asia, and underestimated extreme precipitation in northern North America and northern Asia. In general, the multimodel ensemble (MME) achieved a greater performance in simulating extreme precipitation than did the individual CMIP6 GCMs in the different climate regions. The bias of extreme precipitation based on the considered CMIP6 GCMs was relatively small in tropical regions, especially in equatorial regions. In the future, extreme precipitation will increase, especially under high emission scenarios (i.e., SSP5-8.5). Global extreme precipitation will notably increase in cold and polar climate regions. Our results could improve the understanding of precipitation simulations, and they are very important for providing reliable future global extreme precipitation predictions.