A Global Low-Order Spectral General Circulation Model. Part II: Diagnosis of the Seasonal Energetics

Abstract

The purpose of this paper is to document the seasonal variation of the atmospheric energy cycle simulated by a low-resolution general circulation model in order to assess the model's strengths and weaknesses and better define its usefulness for future sensitivity studies. The numerical model is a global, spectral, primitive equation model with five equally-spaced sigma levels in the vertical and triangular truncation at wavenumber 10 in the horizontal. Included in the model are: orography; time-varying (but prescribed) sea surface temperature, snow cover, and solar declination angle; parameterizations of radiation, convection, condensation, diffusion, and surface transports; and a surface heat budget. The model simulates the observed vertically-integrated energetics of the troposphere reasonably well and the values are comparable to other general circulation models. At northern extratropical latitudes, the very long waves, wavenumbers 1 through 3, which are associated with thermal and orographic zonal asymmetries, dominate the model energy cycle. The annual cycle prevails, with maximum values approximately one month after the winter solstice. These model aspects are similar to those documented observationally. Analysis of the time variation of the model energetics details the presence of asymmetric and steplike variations superimposed on the annual cycles at northern extratropical latitudes. The. model statistics also reveal important latitudinal variations. In particular, more sinusoidal seasonal cycles and a predominance of energy at wavenumbers 4 through 7 are found at southern extratropical latitudes, and the thermally direct east-west circulations, associated with land-sea and precipitation zonal asymmetries, are important in the tropical energy cycle. The major discrepancies in the model simulation of the atmospheric energy cycle are the overestimation of zonal and eddy available potential energy and the conversion between them, an exaggerated semiannual component in the seasonal cycle of eddy available potential energy, a deficiency of eddy kinetic energy at the jet core level, and an underestimation of the transients. The model limitations leading to these discrepancies are discussed, as well as improvements made in subsequent versions of the model.