Assessing the total mountain drag in the Met Office weather forecast model: how sensitive is it to horizontal resolution?

A desirable feature of any numerical weather prediction model is that the total drag exerted on a mountain range is independent of the horizontal resolution. At kilometre-scale resolutions almost all of the mountain drag should be explicitly resolved. At coarser resolutions less drag is explicitly resolved and so, to keep the total drag the same, the resolved drag must be "topped up" by the addition of a parametrized drag. Unfortunately, orographic drag parametrizations are typically developed and inevitably tuned for a single model resolution, and so ensuring that the total drag is independent of resolution tends not to be considered.

In this study the sensitivity to horizontal resolution of the total mountain drag in the Met Office Unified Model (UM) is assessed. The UM has therefore been multiply nested over different mountain ranges and run at horizontal resolutions ranging from 40km (the current global forecast resolution) down to 1.5km (the finest resolution now used operationally over the United Kingdom). Case studies which include significant mountain drag, e.g. associated with observed downslope windstorms, have been selected and run for each mountain range. The design of the experiments therefore represents a significant test of the UM representation of orography and, in particular, of the UM orographic drag parametrizations.

In this talk, the key results from these case studies will be described and illustrated. The sensitivity of the mountain drag to resolution will be evaluated, with consideration given both to the surface pressure drag and to the vertical profile of the momentum flux. Implications for the future development of orographic drag parametrizations will also be discussed.