Including Ice-Cloud Longwave Scattering and Surface Spectral Emissivities in Climate Models Leads to More Impacts on Mean-State Climate than Climate Feedbacks

Climate models often ignore cloud scattering and surface emissivity in the longwave (LW) for computational efficiency. Such approximations can cause biases in radiative fluxes and affect simulated climate, especially in the Arctic because of its large sensitivity to perturbations. We implemented treatments to both physics into the Energy Exascale Earth System Model (E3SM) version 2 by DoE and assessed their impacts on the simulated mean-state global climate as well as climate feedback and sensitivity.

By turning on and off the switches in the modified E3SMv2 model, we studied the changes in mean-state climate due to cloud LW scattering and surface emissivity effects by comparing four 35-year fully-coupled simulations. Cloud LW scattering warms the entire global troposphere by ~0.4 K on average; the warming is stronger in the Arctic (~0.8 K) than in the tropics, which is a manifestation of the polar amplification phenomenon. When realistic emissivity is incorporated into the model, the surface skin temperature increases by 0.36 K instantaneously on a global average, especially in the Sahara Desert (~0.7 K) where the surface emissivity is low. Surface skin temperature, as well as surface air temperature and tropospheric temperature, further increases by 0.19 K due to the inclusion of surface spectral emissivity. The mean-state climate changes due to both effects are linearly additive. The latitudinal and seasonal pattern of surface air temperature warming resulting from both effects is very similar to the response due to CO₂ increase in the standard E3SMv2 model.

We also carried out four 35-year simulations under the abrupt 4xCO₂ scenario, with cloud LW scattering and/or surface emissivity effects on and off. Based on standard radiative kernel analysis, we found that total global-mean climate feedback does not change significantly after including either or both physics. Nevertheless, lapse rate feedback, water vapor feedback, and cloud feedbacks in the tropics have changes by up to 10%. They are primarily associated with high cloud fraction response in the upper troposphere. Our study suggests that both the cloud LW scattering effect and the surface spectral emissivity effect should be included in climate models for a faithful representation of the radiative process in the atmosphere, especially at regional scales.