| description abstract | The performance of a recent version of the general circulation model used at the Goddard Laboratory for Atmospheric Science is evaluated with particular emphasis on its behavior at high latitudes of the Northern and Southern Hemispheres. A January?February climatology for the model was constructed by averaging eight 30-day means, each of which spanned the period from 15 January to 14 February. A mean July climatology was similarly defined on the basis of seven 30-day averages, each spanning the period 1?31 July. Model-generated sea level pressure, 500 mb geopotential, and surface air temperature are compared with observed long-term climatologies. Sensible heat, evaporative, and radiative fluxes at the surface, and radiative fluxes at the top of the atmosphere also are compared with observed data. In the Northern Hemisphere the major features that are satisfactorily simulated include the position and intensity of the Aleutian and Icelandic lows in winter; the central Arctic pressure distribution during winter and summer; and the summertime North Atlantic and North Pacific high pressure regimes. Sensible and evaporative heat fluxes and radiation budget parameters are not unreasonable, but a rigorous comparison is difficult because of data deficiencies. The most notable shortcomings of the model include its weak wintertime Asiatic high, and missing meridionality of the 500 mb flow over the North Pacific. The GCM is less successful in simulating the observed climatology of the Southern Hemisphere. The 500 mb circumpolar flow is adequate, but the model does not successfully reproduce the stationary low pressure centers at the surface around the Antarctic continent. Simulated energy flux components do not disagree in any substantial way with the sparse observations that exist. The performance of other major GCM's at high latitudes is summarized, and some implications for climatic experiments of GCM performance at high latitudes are discussed. | |
