Improving the Surface and Cloud Longwave Radiative Couplings in the Earth System Model: A Tribute to Prof. Liou’s Works on Ice Cloud Scattering and Radiation-Topography Interaction

The intellectual legacy of Prof Liou lies not only in the field of scattering theory for non-spherical particles but also in the radiative interactions between surface and atmosphere. He has pioneered how to represent such radiative coupling between the atmosphere and complex topography (e.g., mountain terrains) in the earth system model. Motivated by these pioneering works, we have looked at two longwave physics not represented by the dominant majority of earth system models participating in the CMIP5 and CMIP6 assessments. For computational efficiency, current earth system models usually ignore cloud scattering and surface spectral emissivity in the longwave (LW). As a result, the longwave surface and atmospheric radiative couplings in the earth system models have a low fidelity compared to reality. Such approximations can cause biases in radiative fluxes and affect simulated climate, especially in the Arctic. Over the years, we implemented treatments to both longwave physics into the DoE Energy Exascale Earth System Model (E3SM) version 2. We assessed their impacts on the simulated mean-state global climate and projected future climate change in response to the increasing CO₂.

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. 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 due to the Arctic amplification phenomenon. When realistic emissivity is incorporated into the model, the surface skin temperature increases by 0.36 K instantaneously on a global average. Globally averaged surface skin temperature, surface air temperature, and tropospheric temperature further increase by ~0.19 K due to the inclusion of surface spectral emissivity. Including surface spectral emissivity leads to a more localized impact, with the most significant changes seen over the polar and desert regions. The mean-state surface climate changes due to both effects are essentially linearly additive.

We carried out simulations under the abrupt 4xCO2 scenario and found that total global-mean climate feedback does not change significantly after including either or both physics. Cloud radiative feedback is affected the most by having such two radiative processes. Our study suggests that both processes should be included for a faithful representation of the LW radiative coupling between surface and cloud.