Seasonal Simulations of the Planetary Boundary Layer and Boundary-Layer Stratocumulus Clouds with a General Circulation Model

Abstract

The UCLA general circulation model (GCM) has been used to simulate the seasonally varying planetary boundary layer (PBL), as well as boundary-layer stratus and stratocumulus clouds. The PBL depth is a prognostic variable of the GCM, incorporated through the use of a vertical coordinate system in which the PBL is identified with the lowest model layer. Stratocumulus clouds are assumed to occur whenever the upper portion of the PBL becomes saturated, provided that the cloud-top entrainment instability does not occur. As indicated by Arakawa and Schubert, cumulus clouds are assumed to originate at the PBL top, and tend to make the PBL shallow by drawing on its mass. Results are presented from a three-year simulation, starting from a 31 December initial condition obtained from an earlier run with a different version of the model. The simulated seasonally varying climates of the boundary layer and free troposphere are realistic. The observed geographical and seasonal variations of stratocumulus cloudiness are fairly well simulated. The simulation of the stratocumulus clouds associated with wintertime cold-air outbreaks is particularly realistic. Examples are given of individual events. The positions of the subtropical marine stratocumulus regimes are realistically simulated, although their observed frequency of occurrence is seriously underpredicted. The observed summertime abundance of Arctic stratus clouds is also underpredicted. In the GCM results, the layer cloud instability appears to limit the extent of the marine subtropical stratocumulus regimes. The instability also frequently occurs in association with cumulus convection over land. Cumulus convection acts as a very significant sink of PBL mass throughout the tropics, and over the midlatitude continents in summer. Three experiments have been performed to investigate the sensitivity of the GCM results to aspects of the PBL and stratocumulus parameterizations. For all three experiments, the model was started from 1 June conditions of the second year of the three-year run, and the July-mean results of each experiment were compared with the three-year composite simulated July, as well as with observations. In the first experiment, the direct interaction of the stratocumulus clouds with the boundary-layer turbulence was disabled. The results show a significant and unrealistic increase in both stratocumulus cloudiness and total cloudiness; a 22 W m?2 reduction in both the globally averaged net radiation flux at the top of the atmosphere and the total surface energy flux; and substantial changes in the relative magnitudes of the components of the surface energy flux. The primary cause of these changes is the absence of cloud-top entrainment instability in the experiment. In the second experiment, the PBL depth was fixed at a prescribed, globally uniform value. The results show a pronounced and unrealistic increase in cumulus precipitation, particularly over land; an unrealistic tendency for stratocumulus cloudiness to occur preferentially over land; and a marked shift in the surface energy balance, accompanied by stronger PBL wind speeds. In the third experiment, the diurnal cycle of solar insulation was replaced by a daily-mean ?torroidal sun.? The results show a 3% increase in the global albedo, even though the global cloudiness decreases slightly. The PBL depth increases dramatically over land and especially over desert regions. The precipitation rate sharply increases over the continents, which become sink regions for atmospheric moisture.