Postprocessed cloud radiative and microphysical output - ICON model

Asian monsoon region domain-mean storm-resolving simulation output.

With increases in computing power and the advent of a new generation of models, we are now able to simulate the global climate system at storm-resolving resolutions of 2 km. Prototypes of such simulations show promising realism in clouds and precipitation, but shortcomings in their representation of microscale processes, like the interaction of cloud droplets and ice crystals with radiation, still restrict their utility. Here, we illustrate how changes to the ice microphysics scheme can dramatically alter both the vertical profile of cloud-radiative heating and top-of-atmosphere outgoing longwave radiation in storm-resolving simulations over the Asian monsoon region: Assorted ``ice microphysical switches" can alter upper-tropospheric cloud-radiative heating by a factor of four and domain-mean outgoing longwave radiation by 30 W m^-2. Poorly-constrained parameters in the ice nucleation scheme, overactive conversion of ice to snow, and inconsistent treatment of ice crystal effective radius between microphysics and radiation all contribute to this large sensitivity in radiative outputs, whereas vertical resolution has a very limited impact. Even with deep convection explicitly simulated then, uncertainties in microscale cloud properties exert a strong control on the radiative budget that propagates to both atmospheric circulation and regional climate. These uncertainties need to be reduced to realize the full potential of storm-resolving models.