Can small-scale orographic gravity waves provide the missing drag in the stable boundary layer?

Currently atmospheric models for weather and climate use enhanced mixing formulations under stable conditions to increase the accuracy of atmospheric motions on the synoptic scale. This approach is required since the original (“short-tail”) mixing function lacks the necessary momentum drag to accurately represent cyclonic filling over land. This enhanced mixing function (also known as the “long-tail” function), introduces momentum drag that cannot be physically justified and deteriorates the score for near surface temperature, wind and boundary-layer height, and as such impacts on phenomena like fog and frost. Orographic gravity wave drag for stable boundary layers is hypothesized to provide the missing drag needed without the disadvantages of using an enhanced mixing function. We have include a new parametrization in the WRF model that represents the orographic drag induced by small-scale orography within the stable boundary layer and adds it to the turbulent drag induced by a short tail mixing function. To study the new parameterization we have conducted a series of model runs, with each model run simulating an 8-day period during winter . We have compared the results against those of a short-tail mixing function and a long-tail one. After a rigorous statistical comparison we have concluded that the new parametrization is able to reproduce sea level pressure, 10m wind and the cyclonic core pressure in various pressure system with higher accuracy than the other two setups. Cyclonic core pressure bias is reduced by 40% to 80% (depending on the case) compared to the short-tail setup, and sea level pressure bias is reduced by up to 1 hPa (30%) over the whole domain, resulting in even smaller biases than the long-tail scheme. This confirms our hypothesis that the new parameterization functions similarly to a long-tail mixing function, regarding the scores of sea level pressure and cyclonic core pressure for individual systems. Near surface wind bias is reduced by up to 40% compared to the long-tail and up to 20% compared to the short-tail setup, while 2m temperature bias is slightly increased (10%).