Abstract
The radiate transfer between the surface and atmosphere directly decides the radiant energy flux exchanges between the atmosphere and underneath surface components and, thus, plays a critical role in the surface energy process. Based on the physics of radiative transfer, once scattering is involved, the atmosphere and underneath component become an entire system and the upward and downward fluxes at the surface-atmosphere interface have to be solved simultaneously. This argument applies to any surface modules (bare land, canopy, ocean, ice, and snow) and both longwave and shortwave. In current coupling structure of E3SM or any mainstream ESMs, the coupler module passes the downward radiant flux at the surface to the surface modules, serving as the upper boundary condition for the radiate transfer solver in the modules. It also passes the upward radiant flux at the surface to the atmosphere module, serving as the lower boundary condition for the radiation solver in the atmosphere module. As a result, the aforementioned radiative coupling between atmosphere and surface is artificially cut off and it is not faithfully represented in such coupler framework.
Aforementioned issue is further compounded by another issue: the spectral radiative properties of the surface (i.e. longwave emissivity and shortwave albedo) are assumed in different ways in the surface and atmosphere modules. This inconsistency can be found in both the longwave and shortwave radiation schemes. Even the coupler ensures conservation of the broadband fluxes across the modules, the spectral decomposition of such broadband flux is not consistent across the modules. Given that fact that atmospheric absorption is highly spectrally dependent, any unrealistic spectral decomposition of upward radiant flux at surface essentially provides an incorrect lower boundary condition for the solver of atmospheric radiative transfer.
Using a couple of examples of our recent studies, namely the longwave surface spectral emissivity treatments for the high latitudes and for the Sahara and Sahel regions, I will show how such different treatments of longwave radiative coupling between surface and atmosphere can cause discernible differences in the simulated mean climate states. The physical mechanisms behind such differences will be described as well. I will then describe our work plan to improve and evaluate such representations in the E3SM. An improved representation of surface-atmosphere coupling can lead to a more faithful simulation of the climate, especially the mean climate state.