Simulations performed with the default versions of the Energy Exascale Earth System Model (E3SMv1) and the Community Earth System Model (CESM2) produce relative humidity values greater than 110% with respect to liquid water and greater than 200% with respect to ice in instantaneous model output. These extreme values of supersaturation are found primarily in the extratropical and polar regions. Supersaturation has an impact on latent heat release, aerosols, clouds, and radiation, so properly representing it in our models is important.

To understand the impact of process coupling and the choice of model timestep on these extreme values of supersaturation we perform extensive sensitivity tests using single column simulations in both E3SMv1 and CESM2. We find that decreasing process timesteps decreases the magnitude and frequency of extreme values of relative humidity, especially when the microphysical timestep is decreased. Guided by these simulations, we run simulations with the global models to better understand how reducing the microphysical timestep impacts values of extreme supersaturation.

Although values of extreme supersaturation can be reduced by altering model timesteps, strong supersaturation with respect to liquid water still remains. Although it is physical for supersaturation with respect to liquid water to exist, a saturation adjustment which removes these supersaturations after the microphysics has been used in previous versions of CESM. We run 10-year simulations with both the CESM2 and E3SMv1 with this saturation adjustment turned on and off to see the impact of completely removing supersaturation with respect to liquid water after microphysics. Applying this saturation adjustment decreases liquid water path towards the poles, which is associated with an increase in shortwave cloud forcing.