Abstract
Plumb's formulation of the stationary wave activity flux is used to determine how well versions of the GFDL and NCAR general circulation models simulate the sources, sinks, and horizontal propagation of atmospheric stationary waves, which play an important role in determining regional climate. The wave activity flux provides insight into the simulation of nondynamic as well as dynamic processes in these models. Model performances for current climate simulations are evaluated with respect to NMC analyses averaged over 1978–1990.
The models fare best when the stationary wave forcing is strongest, that is, in the wintertime Northern Hemisphere, where they reproduce the observed three-branch structure of upward wave activity flux. For the Northern Hemisphere summer and the Southern Hemisphere in both summer and winter, the models show less agreement with observations, although they do simulate the generally downward flux observed during Northern Hemisphere summer, which the analysis suggests is caused by convection. C02-doubling changes in the wave activity flux show little consistency between the two models. The analysis suggests that accurate modeling of stationary wave activity flux is strongly dependent on diabatic forcing, especially that occurring in storm tracks. Improving the simulation of stationary wave activity forcing requires a much better understanding of the physics governing storm tracks and latent heat release in the atmosphere, so that improvements in stationary wave simulation in these models will not occur by simply increasing model resolution.