Regional Characteristics of the Northern Hemisphere Wintertime Circulation: A Comparison of the Simulation of a GFDL General Circulation Model with Observations

Abstract

The hemispheric distributions of a selected set of temporal mean, variance and covariance statistics produced by a general circulation model developed at the Geophysical Fluid Dynamics Laboratory are compared with observations. The fields presented include 1) the seasonally averaged 300 mb geopotential height and zonal wind speed, sea level pressure and 500 mb vertical velocity; 2) the root-mean-squares of 500 and 1000 mb heights, and of 850 mb temperature; 3) the correlation coefficient between the 1000 and 500 mb heights, and 4) the horizontal and vertical transports of heat by transient eddies in the lower troposphere, and the horizontal eddy transports of momentum and potential vorticity near the tropopause. The partitioning of the variance of 500 mb height according to fluctuations of different temporal and spatial scales is examined. By making use of time filters which retain fluctuations with periods between 2.5 and 6 days, the characteristics of synoptic-scale disturbances appearing in the model are studied. The regional contrasts of the observed wintertime circulation are simulated by the model. In particular, the transport properties of transient disturbances over the oceanic storm tracks and the locations of these centers of activity relative to the stationary flow field are reproduced. The agreement between model and observations substantiates some of our interpretations of the observed circulation presented in earlier works. The model simulation differs from the observed atmosphere in the following aspects: 1) the amplitude of the simulated transient fluctuations in the upper troposphere is too weak; 2) the geographic distribution of the variance associated with low-frequency planetary-scale disturbances in the model bears little resemblance to the observed pattern; 3) the observed longitudinal variations of eddy activity in the middle and high latitudes are less evident in the model results; and 4) the simulated surface lows over Iceland and the Aleutians are too deep, so that the midlatitude westerlies over most of the Western Hemisphere are too strong, and the surface circulation over the North American continent is not realistic.