Atmospheric boundary-layer modeling over complex terrain

Weather prediction and climate scenario simulations nowadays rely almost exclusively on numerical modeling of the atmosphere. Several physical processes, however, occur at spatial scales that cannot be explicitly resolved by the models, even at fine resolutions, and are thus parameterized. For this reason, state-of-the-art atmospheric models include a series of empirical parameterizations that describe these processes, including the exchange of heat, moisture, and momentum between the ground and the atmosphere and within the atmospheric boundary layer. In mountainous terrain, the heterogeneity of the terrain and variations in land use related to differences in elevation strongly affect the surface energy budget, that is, radiative fluxes and turbulent fluxes of heat and moisture. The surface energy budget, however, is a major driving force for local wind circulations in mountainous terrain and for the thermal structure of the atmospheric boundary layer. The atmospheric forcing from the surface is thus particularly important, but also particularly challenging for numerical weather prediction models. First, current land cover datasets are often not up-to-date or not available at high enough resolution to correctly represent the land use in the model. As a consequence, the physical properties of some land cover types may not be representative of the real land cover at a given location. Furthermore, turbulence and land surface parameterizations are oftentimes derived from observations over flat terrain and may not necessarily be adequate for complex terrain. Studies evaluating these parameterizations in mountainous terrain, however, are scarce.

The objective of ASTER is (i) to evaluate a model’s performance in forecasting soil properties and surface and near-surface turbulent fluxes and to identify potential error sources and (ii) to evaluate the model’s sensitivity to changes or potential errors in the turbulence and land surface parameterizations or their input parameters. To this purpose, a sensitivity study based on idealized model simulations is combined with simulations of case studies over two Alpine regions, the Austrian Inn Valley near Innsbruck and the Italian Adige Valley near Bolzano. Several years of high-quality turbulence measurements are available for model comparison in these two regions, with observations at multiple sites spanning a range of different terrain orientations and land use characteristics. The overall result will thus be an identification of modeling issues in mountainous terrain related to the turbulence and land surface parameterizations that have a large impact on the forecast. Improvements in these specific areas can thus be expected to have a comparatively large positive impact on numerical model simulations and consequently weather forecasts. 

Nach oben scrollen