Improvements to the representation of orography in the Met Office Unified Model

Three improvements to the representation of orography for use in numerical weather- and climate-prediction models are presented. The first improvement is to replace the US Navy dataset with a new digitally generated dataset as the definition of the true earth topography. There are large differences on all scales between the two datasets and these lead to large differences in the mean and subgrid-scale fields that are derived from them. The second improvement is to filter the mean and subgrid-scale orography (SSO) fields to remove grid-scale and near-grid-scale features and thus suppress forcing on scales that the model cannot treat well. The third improvement is to implement a new, simple parametrization of the effects of SSO in which the total surface pressure drag is calculated using the analytical expression for linear hydrostatic flow over a two-dimensional ridge in the absence of friction and rotation. The surface pressure drag is partitioned into gravity-wave and blocked-flow components that depend on the Froude number of the flow impinging on the SSO. The new scheme attributes about 70% of the total drag to flow blocking. These improvements have been incorporated into a new version of the Met Office Unified Model. A series of numerical weather-prediction experiments demonstrates that the introduction of the new SSO scheme is the most significant change. In particular, significant improvements to forecast skill, attributable to the SSO scheme's flow-blocking drag component, are found at low levels in the northern hemisphere and the Tropics for an extended northern hemisphere wintertime forecast trial. Furthermore, there are no significant degradations in skill at upper levels, in the southern hemisphere or for summertime trials.