Advanced testing of low, medium and high ECS CMIP6 GCM simulations versus ERA5-T2m

TL;DR

This study evaluates CMIP6 GCM simulations with varying ECS against ERA5-T2m data, finding high and medium ECS models overestimate warming and are unsuitable for predictions, while low ECS models are more consistent and less alarming.

Contribution

It introduces a comparative analysis of CMIP6 GCMs grouped by ECS against observational data, highlighting the limitations of high and medium ECS models for climate prediction.

Findings

High and medium ECS GCMs overestimate observed warming.

Spatial statistics reject data-model agreement over most of Earth's surface for these models.

Low ECS GCMs predict moderate warming, making them more suitable for future climate projections.

Abstract

The equilibrium climate sensitivity (ECS) of the CMIP6 global circulation models (GCMs) varies from 1.83 {\deg}C to 5.67 {\deg}C. Herein, 38 GCMs are grouped into three ECS classes (low, 1.80-3.00 {\deg}C; medium, 3.01-4.50 {\deg}C; high, 4.51-6.00 {\deg}C) and compared against the ERA5-T2m records from 1980-1990 to 2011-2021. We found that all models with ECS > 3.0 {\deg}C overestimate the observed global surface warming and that spatial t-statistics rejects the data-model agreement over 60% (using low-ECS GCMs) to 81% (using high-ECS GCMs) of the Earth's surface. Thus, the high and medium-ECS GCMs are unfit for prediction purposes. The low-ECS GCMs are not fully satisfactory yet, but they are found unalarming because by 2050 they predict a moderate warming ($\Delta T_{preindustrial\rightarrow2050}\lesssim2\:{^\circ}C$).