How Well Do Nudged Hindcasts from Global Climate Models Represent Southern Ocean Clouds Observed during SOCRATES?

The Southern Ocean (SO) generates cyclonic storm systems year-round that make it a very cloudy region. It is far from natural sources of dust (the most prevalent ice freezing nucleus); it has been argued that this favors more supercooled liquid cloud layers. The SO also sees relatively little anthropogenic pollution that could enhance the concentration of cloud condensation nuclei, so it is a good environment for studying the natural cycle of aerosols and their effect on clouds. Global climate models (GCMs) tend to underestimate reflected sunlight poleward of 55^◦ S, particularly in the cold sectors of SO cyclones which are dominated by low-lying stratus and shallow cumulus. The remoteness of the SO has led to a sparsity of in-situ observations of these cold sector clouds that could be used to evaluate how well GCMs simulate supercooled liquid clouds and the aerosol-cloud interactions that determine liquid cloud droplet concentration.

Motivated by these issues, the Southern Ocean Clouds Radiation and Aerosols Transport Experiment (SOCRATES) made airborne and ship-based measurements of cold sector clouds and their interaction with the underlying ocean between Tasmania and the Antarctic coast during Jan.-Feb. 2018. The 15 SOCRATES research flights of NCAR’s G-V research aircraft sampled diverse environmental conditions, including stable and unstable boundary layers, near-surface winds up to 25 m s^-1, and persistent supercooled liquid clouds at temperatures as cold as -30^◦ C and liquid cloud droplet concentrations ranging from 20-300 cc^-1.

Here, we evaluate nudged hindcasts from the Community Atmosphere Model (CAM6) and the Geophysical Fluid Dynamics Laboratory Atmosphere Model (AM4) using time-space collocated SOCRATES airborne observations. Both global models are nudged toward global wind and temperature fields from MERRA reanalysis for the SOCRATES period. This forces a realistic evolution of storm systems and other atmospheric circulations in the GCM. Humidity, cloud fields, and aerosol fields are not nudged, so it is meaningful to compare these with the airborne observations. Those observations include vertical cloud and precipitation profiles from a cloud radar and lidar, size distributions of aerosol, liquid and frozen cloud and precipitation, CCN/IFN measurements, limited aerosol composition measurements, and meteorology. These observations will be used to evaluate the current simulation of SO boundary-layer clouds and aerosols. They will also inform future improvements of the microphysics, turbulence and cumulus parameterizations in the models.