Investigating the Climate Sensitivity Differences between CESM1 and CESM2

The response of a coupled climate model to instantaneous increases in CO2 can provide information about climate feedbacks operating at a range of timescales. We derived estimates of the climate sensitivity in the Community Earth System Model (CESM) from abrupt 4xCO2 experiments using the Gregory approach (Andrews et al, 2012). These experiments show that the 2xCO2 climate sensitivity of CESM has increased from 4.1K in CESM1 to 5.3K in CESM2; the version to be used in CMIP6.

CESM2 incorporated major changes to parameterizations compared to CESM1. This includes changes in PBL turbulence, shallow convection, cloud microphysics and cloud “macrophysics”. During the development of CESM2, we conducted 4xCO2 experiments at several stages. Interestingly, early development versions of CESM2 with all parameterization changes already in place had a Gregory-method climate sensitivity close to that of CESM1. Instead, the increase in climate sensitivity appeared to accompany seemingly minor changes to the representation of aerosol lifetime, sea-ice albedos and warm rain re-evaporation.

In this presentation we attempt to disentangle the processes leading to increased Gregory-method climate sensitivity in CESM2. Because the interpretation of 4xCO2 experiments may be complicated by the fact that energetic equilibrium is not attained during the run, we also compare sensitivity estimated from 4xCO2 experiments with estimates based on alternative experimental designs, such as SST+4K or 2xCO2 Slab Ocean Model (SOM) runs. We also discuss the role of cloud-aerosol interactions in CESM and their evolution during the development of CESM2.

T Andrews, J M Gregory, M J Webb and K E Taylor. Forcing, feedbacks and climate sensitivity in CMIP5 coupled atmosphere‐ocean climate models. (2012) Geophysical Research Letters 39 L09712, https://doi.org/10.1029/2012GL051607.