How Much Does Ice Microphysics Matter for Simulating Cirrus Clouds?

Cirrus clouds exist across a range of scales from micrometers to kilometers, forming in minutes but persisting for days. Cloud-resolving models (CRMs), which have explicit deep convection, may more accurately represent the physical processes that can form cirrus but still have a variety of subgrid (unresolved processes) parameterizations, which may contribute to the differences in ice clouds, especially those in the upper troposphere. In previous work using global CRMs from the DYnamics of the Atmospheric general circulation Modeled on Nonhydrostatic Domains (DYAMOND) project, we find that simulated cirrus vary greatly between models. This uncertainty in simulating cirrus can have significant impacts on the top-of-atmosphere radiative energy balance. Previous studies have shown that cirrus clouds are highly impacted by the choice of microphysics scheme in models (e.g., Morrison, P3, single vs double moment), but here we quantify how changes within the P3 two-moment microphysics scheme in the Simple Cloud-Resolving E3SM Atmosphere Model (SCREAM) can impact cirrus clouds. In this study, we aim to answer the question of how much microphysical choices within a microphysics scheme matter for simulating cirrus clouds in the current and warmed climate (+4K SSTs). We run multiple simulations with changes to ice processes in SCREAM using a radiative-convective equilibrium setup. We adjust the amount of vapor deposition, sedimentation rate of ice, and ice nucleation to form an understanding of the sensitivity of cirrus to microphysical choices within P3. We find that small changes in ice microphysics can have a significant impact of local and top-of-atmosphere radiation, yet the overall texture of convection remains consistent. Future predictions of climate are also sensitive to ice microphysics and how cirrus clouds (especially in-situ or near-tropopause cirrus) change with warming. Even as models improve their resolution and need fewer parameterizations, the models still cannot resolve every processes, so we evaluate the sensitivity of cirrus clouds in a cloud-resolving model to the microphysical choices made for subgrid processes.