Introduction

The atmospheric models described in the previous chapters simulate the evolution of an atmosphere where no exchanges occur with its surroundings. Although inertia effects dominate the evolution of the atmosphere at the synoptic scale at mid-latitudes over short time periods (24 or 36 hours), when constructing realistic forecast models allowance must be made for the physical processes not handled by the adiabatic frictionless equations.

Exchanges of momentum, heat, and water vapour between the atmosphere and its surroundings (space and Earth) are the outcome of physical processes involving far smaller space and time scales than those taken into account by the dynamic model. As these processes cannot be simulated explicitly, we try to determine where and when they occur and to calculate their average effect for each of the model’s prognostic variables (wind, temperature, water vapour). This is what is called parameterizing the effect of physical processes for scales smaller than the scales resolved by the dynamic model (or sub-grid scales). The parameterization of the effects of the sub-grid scale physical processes depends therefore on the space and time scales actually handled by the discretized model’s equations.