Pressure drag and momentum fluxes due to the Alps. II: Representation in large-scale atmospheric models

Abstract

A three-dimensional numerical model with interactive grid nesting is used to study the pressure drag and associated momentum fluxes due to the Alpine complex (for a specified upstream wind and static stability profile). Four basic questions are addressed which are relevant to the problems of flow description in the vicinity of mountains and the impact of orography in the context of mesoscale numerical weather prediction and general circulation models: (i) What is the dependence of basic quantities such as the pressure drag on the model's horizontal resolution? (ii) Is there a need for non-hydrostatic dynamics to account properly for the flow characteristics near steep mountains? (iii) Is there an effective role for enhanced orography such as an envelope formulation? (iv) Might a ‘gravity-wave-drag’ parametrization scheme be able to represent adequately the highest resolution fluxes and surface drag in a much coarser resolution model? A series of experiments with both hydrostatic and non-hydrostatic models at resolutions of 80, 40, 20, 10 and 5 km is analysed, and two of these experiments are repeated with an envelope orography. It is concluded that hydrostatic simulations are sufficient to describe the basic dynamics even at 5 km resolution; however, there is a strong dependence of the drag and momentum fluxes on resolution, with the drag at 80 km resolution being almost half that at the highest resolution. It is also shown that the use of an envelope orography at coarser resolutions can compensate significantly for the drag under-estimation, but is less effective in representing the vertical stress profiles or the direction of the stress vectors, which are systematically in error at the coarsest resolution. The potential role of gravity wave drag is confirmed and possible improvements considered. It is suggested that both envelope orography and gravity wave drag may make a significant contribution until resolutions are as high as 10 to 20 km.