Item

Abstract

Mixed-phase clouds can be found between 0°C and -38°C and can consist of supercooled liquid and ice hydrometeors simultaneously. Simulating these clouds is challenging for climate models, since their partitioning into liquid and ice depends on many factors (e.g., cloud dynamics, aerosols as ice nuclei, ice multiplication processes...) and varies with cloud types, regions and seasons. To address this complex problem, we use four years of satellite and model data to analyse the cloud top phase distribution globally and regionally for different cloud types. The analysed datasets include satellite observations (Cloud_cci v2, Cloud_cci v3, CLARA-A2, CALIOP, MCD06COSP_L3 from MODIS), the global climate model CAM6-Oslo, and three convection-resolving model datasets from the DYAMOND Winter Project, i.e., ICON 5km, SCREAM 3km, and GEOS 3km.

Key findings of our study include: (1) The observational datasets clearly indicate some systematic regional dependencies of the supercooled liquid fraction (SLF) as function of cloud temperature, in particular, SLF is larger in the Southern Hemisphere than in the Northern Hemisphere, except for the continental low-level clouds, where the opposite is found, (2) In Cloud_cci v3, MCD06COSP_L3, and CAM6-Oslo, the liquid effective radius increases with decreasing SLF, while the opposite is found for the GEOS model, (3) Frequent occurrence of mixed-phase clouds with SLF ≈ 0.5 is found for Cloud_cci v3, MCD06COSP_L3, and CAM6-Oslo, (4) The global cloud-resolving models, in particular ICON and SCREAM, show most of cloudy pixels either close to SLF = 0% or to SLF = 100%, probably a sign of an efficient Wegener-Bergeron-Findeisen process.