Sensitivity of orographic drag to model resolution

The current, non-hydrostatic version of the Met Office Unified Model (UM) became operational in August 2002, including a new sub-grid scale orography (SSO) parametrization scheme which accounts for the drag due to low-level flow blocking. The UM is used for both mesoscale and global NWP forecasts and also climate simulations and so is routinely run at horizontal resolutions ranging between 12km and 300km. Therefore the SSO parametrization scheme plays a very important role in the correct prediction of winds over a wide range of scales. The 8 November 1999 northerly foehn case which occured during MAP involved flow-splitting around the Alps as a whole and around individual peaks, with mountain waves forced aloft at horizontal scales as small as 6km, making this the perfect testbed for the new SSO parametrisation scheme. Mesoscale simulations of this case have been carried out at horizontal resolutions between 60km and 4km in order to assess the total drag predictions. As the resolution increases the resolved drag should converge towards its real value while the parametrized drag should reduce to zero.

The magnitude of the resolved drag is found to increases monotonically with resolution over the studied range. The lack of convergence suggests that individual peaks and valleys, which are not yet fully resolved, are as important as longer orographic features in producing drag. It therefore appears necessary to extend the study to even higher resolutions to achieve a "truth" simulation. Momentum flux profiles indicate that the largest proportion of the drag is felt below the level of the highest mountain peaks, demonstrating the importance of representing drag due to low-level flow blocking for the correct prediction of large scale winds.