Arctic synoptic regimes: Comparing domain-wide Arctic cloud observations with CAM4 and CAM5 during similar dynamics

Arctic cloud properties, variability, and sensitivity to surface conditions are strongly dependent on the synoptic regimes in which they exist, and proper evaluation of modeled cloud properties against observations requires that the dynamic and thermodynamic environment be carefully accounted for. In this study, a k-means clustering algorithm is used to sort Arctic clouds observed by CALIPSO into four distinct synoptic regimes: three regimes largely associated with mid-tropospheric subsidence or very weak uplift and separated by differences in lower tropospheric stability; and one regime associated with mid-tropospheric uplift. Boundary layer clouds are present within the three subsidence regimes with increases in stability relating to lower altitudes of the maximum cloud percentage. In going from sea ice to open ocean, the cloud amount increases at the lowest atmospheric level for all subsidence regimes except the least stable regime, and the height of maximum low clouds rises for all subsidence regimes except the most stable regime. Within each synoptic regime, clouds diagnosed from the CALIPSO simulator in CAM4 and CAM5 forecast runs are then compared with observations. Compared with CAM4, CAM5 produces boundary layer clouds that are more similar to the observations, and more closely captures the observed sensitivity of cloud properties to the underlying surface type. However, CAM5 overestimates cloud percentages within the uplift regime. The improved boundary layer turbulence and cloud microphysical parameterizations in CAM5 yield a model that exhibits an improved sensitivity of Arctic low-level clouds to changes in lower tropospheric stability and surface type.