Abstract
The December–February climates simulated by a spectral general circulation model using different horizontal resolutions namely triangular T20, T30 and T40 truncation but with no change in model physics, are compared. Statistically significant differences are found among the climates which rival or exceed those found in the kinds of strong external forcing experiments often performed with such models. Despite the statistical significance of the differences, this model shows rather less of a tendency toward “zonalization” or westerly bias with increased resolution than has been wen in some models in the past.
The results pertaining to resolution are obtained using a version of the model which includes the gravity- wave drag parameterization which was fiat developed for and subsequently extensively used in this GCM. The effect of gravity-wave drag is to counter the tendency toward zonalization or westerly bias in the model climate. The consequences of the removal of this parameterization on the model climate is also investigated. In terms of the angular momentum budget, the paradoxical result is that there is an increase in the surface torque when the gravity-wave drag parameterization is removed. That is, the removal of a drag mechanism leads to an increase in surface stress.
Apparently the numerical solutions to the governing equations depend importantly, and in nonobvious ways, on resolution and parameterization, in particular, on the parameterization of source/sink terms in the momentum equations.
