Impact of land surface roughness on downslope windstorm modelling in the Arctic

Abstract

The article considers the reasons for the underestimation of the wind speed by the WRF-ARW model when simulating downslope windstorms in the Russian Arctic. Simulation results for the Tiksi windstorm, for which sensitivity tests were carried out, appeared to be weakly dependent on the initial and boundary conditions, topography resolution, and boundary layer parameterization. Wind speed underestimation was mostly related to improper land use and the highly overestimated roughness length, which are used in the model. Reduction of the roughness length in accordance with the observations leads not only to a quantitative change in the wind speed in the boundary layer, but to qualitative changes in the dynamics of the flow. Wind underestimation in simulations with the overestimated roughness was caused by the jet stream unrealistically jumping over the lee slope and wake formation in the station area, while jet stream stayed near the surface and propagated to the station area in simulations with the modified roughness length. Modification of land use and roughness length in Tiksi and other regions where downslope windstorms are observed (Novaya Zemlya, Pevek, Wrangel Island) led to a decrease in wind speed modelling error by more than 2.5 times.