Addressing the "Terra Incognita" - Appropriate Representation of Terrain from Mesoscale to Microscale

Mesoscale models are largely limited by the use of a terrain-following coordinate system, which breaks down in very steep complex topography. Conversely, microscale models such as large-eddy simulation (LES) codes can readily handle very complex topographies, but often do not consider atmospheric physics or real boundary conditions. These models are constrained to small geographic areas by their fine resolution, limiting their use for numerical weather prediction (NWP). The gap between the two approaches has been dubbed the "terra incognita" (Wyngaard 2004) in reference to the inability of either modeling system to represent flow features with proper parameterizations at the intermediate grid scale. An additional problem arises in the representation of topography at intermediate resolutions, which is the focus of this study. As the scales of these models converge with advances in computing, it seems logical to develop a single code that is capable of seamlessly nesting from the mesoscale down to the microscale. The Weather Research and Forecasting (WRF) model has recently been expanded to also serve as an LES model. Additionally, an immersed boundary method (IBM) has been implemented to create IBM-WRF (Lundquist et al. 2010, 2012), which allows for very steep complex terrain to be resolved within WRF. IBM-WRF is an excellent candidate for a meso-to-micro scale NWP framework. The problem that arises is that the terrain is not properly represented at intermediate resolutions. For standard WRF, as the resolved terrain begins to capture individual peaks and troughs, complex features of the terrain are realized before there is enough resolution to smoothly represent them. As the terrain is resolved further, the slopes become steep enough to discourage the use of WRF due to errors from the terrain-following coordinates. On the other hand, IBM-WRF performs well at fine resolutions, including for very steep surfaces (e.g. urban geometries). At coarser resolutions, IBM-WRF is also negatively impacted by partially resolved terrain features because of interpolation limitations. Thus, a "terra incognita", literally "unknown terrain", region arises at intermediate resolutions, in addition to issues with physical parameterizations such as turbulence. This presentation describes the development of guidelines to address this issue for intermediate terrain scales in meteorological modeling. Simulations using WRF and IBM-WRF are conducted at various resolutions to assess the performance of each. Nested simulations which transition between WRF and IBM-WRF grids are explored. This will aid in the development of an NWP framework that is appropriate for all scales.