WRF-LES over Complex Terrain Using a Velocity-Reconstruction Immersed Boundary Method

The Weather and Research Forecasting (WRF) model is increasingly being used for higher resolution atmospheric simulations over complex terrain. With increased resolution, the resolved slopes become steeper, and the native terrain-following coordinates used in WRF result in numerical errors and instability. The immersed boundary method (IBM) uses a Cartesian grid with the terrain surface represented through interpolated forcing terms. The WRF-IBM implementation of Lundquist et al. (2010, 2012) eliminates the limitations of WRF’s terrain-following coordinates and has been validated with a no-slip boundary condition for large-eddy simulation (LES) of flow over urban and idealized terrain.

This work describes the implementation of a velocity reconstruction log-law boundary condition into WRF-IBM to extend its applicability to general atmospheric complex terrain simulations. The implementation of the log-law boundary condition for WRF-IBM is validated for neutral flow over flat terrain and the complex terrain cases of Askervein Hill and Bolund Hill. This allows direct comparison between WRF and WRF-IBM for the flat terrain and moderately-sloped Askervein hill, where good agreement is achieved with observations. The Bolund Hill simulations show that WRF-IBM can handle steep topography (standard WRF fails) and compares well to observations and other previously published simulations. Overall, this log-law implementation shows acceptable performance, though the lee side representation of the flow can be potentially further improved.