Constraining ECS and TCR for 21st century for temperature forecasts and risk assessments by comparing the CMIP6 GCM simulations versus global surface temperature records

Abstract

Climatological risk assessments are currently based on simulations for the twenty-first century made using global climate models (GCMs) from the sixth Coupled Model Intercomparison Project Phases (CMIP6) by adopting several hypothetical shared socioeconomic pathways (SSPs) emission scenarios. However, there are large differences in the climatic sensitivity to atmospheric CO2 increase amongst the available climate models: for example, for the CMIP6 GCM the ECS varies between 1.8°C and 5.7°C and the TCR varies from 1.2°C to 2.8°C. There is also mounting evidence that many GCMs are operating "too hot" and are therefore unreliable for informing climate change policy for the future. Here, we assess the performance of 41 CMIP6 GCMs by ranking and comparing them using their literature-provided estimates for the equilibrium climate sensitivity (ECS) and transient climate response (TCR). We discover that the GCM sub-ensemble that performs the best in hindcasting the warming from 1980 to 2021 is that made of the GCMs with ECS ranging between 1.8 and 3.0 °C and TCR ranging between 1.2 and 1.8 °C. A total of 17 models make up this GCM sub-ensemble. The predicted warming of these models for the mid-term (2041-2060) period is 1.5–2.5°C relative to the preindustrial period (1850-1900) according to various SSP scenarios. Thus, the global aggregated impact and risk assessments assuming low to no adaptation appear, therefore, moderate, which also implies that adaptation policies may be adequate to address any unfavorable effects of future climate changes. We also discuss the additional uncertainties surrounding the warming of the Earth's surface temperature by comparing several temperature reconstructions and the warming differences observed between land and ocean to discuss the possibility that ECS and TRC could be furtherly constrained.Main references:Scafetta N (2022) Advanced testing of low, medium, and high ECS CMIP6 GCM simulations versus ERA5-T2m. Geophys Res Lett 49:e2022GL097716. https://doi.org/10.1029/2022GL097716Scafetta, N. CMIP6 GCM ensemble members versus global surface temperatures. Clim Dyn (2022). https://doi.org/10.1007/s00382-022-06493-w