Earlier studies have shown that time-step sizes used by the parameterized physics in the atmosphere component of the Energy Exascale Earth System model, EAM, have large impacts on the long-term statistics of the simulated clouds and precipitation. Recent revisions in the numerical process coupling have led to significantly reduced time-step sensitivities in the subtropical low clouds, while the sensitivities related to deep convection and high clouds remain. This poster presents our progress on identifying numerical issues associated with tropical and subtropical high clouds in EAM.
We clarify how the ice-cloud related atmospheric processes are numerically coupled in EAM. Budget analyses are carried out to reveal the primary sources and sinks of ice crystal mass and number in the model. Sensitivity experiments using varied time-step sizes in different parts of EAM are conducted to help attribute the time-step sensitivities. Possible improvements in process coupling are explored using a combination of theoretical error analysis and EAM simulations.