What can we learn about orographic drag parametrisation from high‐resolution models? A case study over the Rocky Mountains

Abstract

Comprehensive high‐resolution numerical weather prediction models provide a virtual laboratory for modelling the atmospheric flow over complex mountain ranges. In this study, global and regional simulations with horizontal grid spacing ranging from 2 to 32 km, focused over the northern Rocky Mountains, are used to assess the orographic blocking and gravity wave drag parametrisations employed in the Met Office Unified Model (UM) and the European Centre for Medium‐Range Weather Forecasts Integrated Forecasting System (IFS). The total, resolved and parametrised drag components in coarse‐resolution simulations are compared with those in high‐resolution simulations, in which the orographic drag processes are better resolved. The total surface stresses and gravity wave momentum fluxes in the free atmosphere of the global 16 km UM and IFS simulations are shown to compare well with 2 km regional simulations in terms of variability and mean. While the total gravity wave momentum flux is somewhat underestimated by both global models, its vertical distribution is well captured. The “seamlessness” of the parametrisation scheme is then assessed by comparing the total orographic stress – and its components – across several horizontal resolutions of the UM. The surface stress remains relatively constant across resolutions, such that the reduction in resolved orographic stress at coarser resolutions is compensated for by an almost equivalent increase in parametrised orographic stress. However, the parametrised orographic gravity wave momentum flux in the free atmosphere remains almost constant with resolution, failing to compensate for the lack of resolved flux at coarse resolutions. This leads to an underestimation of the total gravity wave drag at coarser resolutions. Further analysis suggests that this underestimation is due to the monochromatic wave assumption made by the gravity wave drag parametrisation scheme.

COnstraining ORographic Drag Effects (COORDE): A Model Comparison of Resolved and Parametrized Orographic Drag

“…This is compensated by the IFS having slightly more parametrized gravity wave drag over this region. This is opposite to what was found in Vosper et al (2019) is much weaker than that of the resolved gravity wave drag (Figure 7(c)) in most of the models. The IFS's parametrized gravity wave drag is almost identical in LR and MR.…”

“…In parallel, a series of studies revealed that inter-model differences in drag partitioning between different processes (partly the result of repeated tuning exercises) significantly affects the large-scale circulation over the NH (Sandu et al, 2016;Pithan et al, 2016) and the circulation response to climate change (van Niekerk et al, 2017). Further work also showed that, for at least a couple of models, the change in resolved orographic drag is not accurately balanced by the change in parametrized orographic drag when horizontal resolution is varied (van Niekerk et al, 2016;Vosper et al, 2016;Sandu et al, 2019;Vosper et al, 2019). This is a key requirement for ensuring that circulation remains robust across resolutions, so that a model can be used seamlessly for a range of applications.…”

“…To compare the total (resolved plus parametrized) orographic drag from the low resolution simulations with the resolved orographic drag from high resolution simulations, we also compute the resolved surface stresses and gravity wave drag in the free atmosphere in the UM and IFS. The resolved surface stress calculation follows the method described in Smith et al (2005) and Vosper et al (2019), using:…”

Addressing the causes of large‐scale circulation error in the Met Office Unified Model

“…Without good observations of the drag resulting explicitly from orography (rather than the combined effect of orography and the land surface within the mountainous region), it is difficult to evaluate the stress from the OGWD scheme independently. There are ongoing efforts to use high-resolution simulations (e.g., Vosper et al, 2019); however, in this study the development work on the OGWD scheme focuses largely on the model's orography ancillary and subgrid orography fields. Whilst details of orography may be thought to be well known, and hence the impact of changes small, a considerable amount of subjective processing is involved, some of which has a significant impact on the magnitude of the mean and subgrid orographic fields and, as a result, on the circulation.…”

“…However, as more of the orography becomes explicitly resolved, the amount of (excessive) drag being applied by the subgrid OGWD scheme is reduced and the overall parametrized drag is lowered. Previous studies (e.g., van Niekerk et al, 2016;Vosper et al, 2019) have also shown that the resolution dependence of the scheme (i.e., the extent to which it transitions from subgrid to resolved orography as resolution increases) is suboptimal, and in the longer term the scheme needs to be revised to address this. Differences in the source datasets used by ECMWF and the Met Office also reveal some significant differences, particularly over Greenland and Antarctica, giving further evidence that the polar high pressure may be related to this.…”