Numerical weather prediction models such as the Weather Research and Forecasting (WRF) model employ terrain-following coordinates to model complex topography. The signature of the topography causes distortion of the computational cells, often as high as the troposphere. This grid skewness introduces an additional truncation error, which can have devastating effects on the quality of the solution. We have implemented an immersed boundary method (IBM) in WRF, which allows complex terrain to be modeled without a coordinate transformation. IBM alleviates error associated with coordinate mapping, thereby increasing the fidelity of the solution.

IBM has been shown to effectively represent complex geometries in computational fluid dynamics codes, but additional issues arise in numerical weather prediction such as: scalar dispersion, surface energy and moisture budgets, and seamless nesting between terrain-following mesoscale simulations and those with IBM. These issues have been successfully addressed within our IBM implementation, and these features are demonstrated with three dimensional building resolving simulations using IBM within the WRF framework.

This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.