Development of the WRF-IBM model for flow over complex terrain

As model grid resolution is refined, atmospheric models can resolve increasingly steep terrain slopes. For example, at 3 km resolution, the maximum resolved terrain slope may be 4 degrees, whereas at 100 m resolution, slopes may be over 30 degrees. Terrain-following coordinates, used by models such as WRF (Weather and Research Forecasting), are unable to handle steep terrain because of the grid distortion and related numerical errors. These limitations on terrain slope have prompted the development of an alternative gridding technique known as the immersed boundary method (IBM).

IBM has been implemented into the WRF model, which eliminates conforming grids and the errors associated with terrain-following coordinates (Lundquist et al. 2010,2012). This implementation, WRF-IBM, has been validated for idealized cases and real urban cases with excellent results; however, to date WRF-IBM has been applied with idealized lateral boundary conditions and thus is not able to handle multi-scale flows with grid nesting. Furthermore WRF-IBM currently uses a no-slip boundary condition, while the use of parameterized surface fluxes is typical in atmospheric modeling.

In this work, we detail a multi-year effort to develop WRF-IBM for real, multi-scale simulations, including full atmospheric physics. Results from three aspects of this project are presented: initializing IBM domains using real meteorological and surface data, developing a nesting interface between domains using terrain-following and IBM coordinates, and modifying the IBM boundary condition to include similarity theory for surface fluxes. A companion abstract by Bao et al. gives further details. The WRF-IBM model can then be applied to problems such as atmospheric transport and dispersion, and wind energy forecasting, among others.