NUMERICAL MODELS | Model Physics Parameterization

Abstract

The parameterizations of physical processes that cannot be explicitly represented by numerical weather prediction models are critical model components and their improvement is one reason why numerical models are becoming more skillful. Parameterization schemes relate the effects of the chosen physical process to the model variables using an algorithmic or statistical approach. Typical parameterization schemes used in operational forecast models represent the effects of shortwave and longwave radiation, cloud cover, soil–vegetation–water–atmosphere transfer, urban areas, planetary boundary layer, convection, microphysics, and orographic drag. Brief overviews of some of these parameterizations are provided, along with a comparison of scheme results to highlight recent advances and remaining challenges.