Experiments with a Stratospheric General Circulation Model: Part IV. Inclusion of the Hydrologic Cycle

Abstract

The hydrologic cycle has been included in a stereographic model of the Northern Hemisphere previously used as a dry model. The model had 18 vertical levels distributed between the surface and 37.5 km and was run out for annual mean conditions. The model had no topographical features. A remarkably good simulation of many of the zonal mean features of the atmosphere was obtained. The tropospheric and stratospheric jet streams were well reproduced as regards intensity although they were displaced slightly equatorward. The meridional streamfunctions in the model agreed well with observation both in intensity and structure, subject to some distortion of the Hadley cell by the equatorial wall in the model. Considerable improvements were obtained in the representation of the large scale eddy flux of relative angular momentum resulting in good agreement with observation. A detailed discussion is given concerning how the synoptic zonal and meridional wind distributions combine to produce a subtropical jet and a 3-cell mean meridional circulation pattern. The connection between the need to restrict the growth of westerly winds associated with conservation of absolute angular momentum in poleward trajectories, the production of the subtropical surface pressure highs, and the downward branch of the Hadley cell is explained, as is the location of the tropospheric jet stream. The dynamical factors which control the latitudinal extent of the Hadley and Ferrel cells are described. The production of a 3-cell mean meridional structure, and thus many of the basic characteristics of the atmosphere, are attributed to angular momentum requirements, rather than those associated with sensible or latent heat fluxes. The zonal mean temperature distribution of the moist model was in rather good agreement with observation, and not too dissimilar to that of the previous dry model. While the model produced a satisfactory tropospheric water vapor distribution its stratospheric distribution was as much as two orders of magnitude too low. This was attributed to the lack of subgrid scale vertical mixing in the model stratosphere. The stratospheric water vapor distribution was primarily maintained by a vertical flux generated by the large scale eddies, particularly at very low latitudes. The mean motions produced a net downward flux of water vapor in the vicinity of the tropical tropopause. The energy balance of the model revealed that the model lacked eddy energy, although the ratio of eddy to zonal kinetic energy was much improved compared with previous versions of this model. The energy cycle of the lower and middle stratosphere is given, and shown to be fairly similar to that of the dry model, particularly as far as the forcing terms from the troposphere were concerned. Continuous coupling was found to exist between the troposphere and the stratospheric jet in the model, emphasizing the basic homogeneity of the troposphere and the lower and middle stratosphere.