PNNL

Abstract

Mixed-phase clouds modulate the water and energy cycles of high-latitude regions, yet their liquid-ice phase partitioning has long been poorly represented in large-scale models as well as in the emerging global convection-permitting models at kilometer-scale resolutions. This study identifies key deficiencies in cloud parameterizations that continue to challenge convection-permitting models and underscores a parallel need for process-oriented observations. Here, we assess the Simple Cloud-Resolving E3SM Atmosphere Model (SCREAM) in simulating Arctic mixed-phase clouds using large-eddy simulations, satellite data, and ground-based observations from the DOE ARM Cold-Air Outbreaks in the Marine Boundary Layer Experiment. SCREAM simulates nearly completely frozen clouds, at odds with observations that clearly show both liquid and ice. Moreover, although SCREAM reproduces boundary layer cloud deepening seen by satellites, it generates a persistent ice cloud aloft that is not observed. Such biases are largely due to the unreasonably strong Wegener–Bergeron–Findeisen (WBF) process that converts liquid to ice excessively. In addition, the temperature-deterministic deposition ice nucleation scheme produces over-abundant ice at cold temperatures, which leads to subsequent WBF killing liquid and the high-level ice clouds, thus biasing the simulated cloud forcing and top-of-atmosphere radiative fluxes. The proposed physically-based improvement of the subgrid cloud overlap treatment substantially increases supercooled liquid content and notably improves cloud-top phase partitioning, aligning better with observations. Such improvement also converges with increasing horizontal resolution within a certain range of temperatures. This study enhances the understanding of the physical processes governing mixed-phase cloud simulations, and provides guidance on the improvements for the next-generation km-scale modeling.