Scalar Transport and Dispersion in Complex Terrain within a High Resolution Mass-Consistent Wind Modeling Framework

In areas of complex terrain, fine-scale changes in topography or vegetation substantially alter the flow field, and in turn, the transport and dispersion of air pollutants, pheromones, or other scalars. Thus, accurate modeling of scalar transport in complex topography requires accurate prediction of the flow field at a high spatial resolution. Mesoscale weather models typically operate on horizontal grids of 4 km or larger and are not capable of handling the effects of sub-grid complex terrain, such as wind speed-up over ridges, flow channeling in valleys, flow separation around terrain obstacles, and enhanced surface roughness from vegetation. In this paper we describe a scalar transport algorithm (advection and turbulent diffusion) used with WindNinja, a high-resolution mass-consistent wind model. WindNinja operates on a terrain-following coordinate system with a hexahedral cell mesh that grows in vertical size with height above the ground. A variational calculus approach is used in WindNinja that results in fast run times on the order of one minute for a 50 km x 50 km domain and 100 m horizontal resolution. The advection-diffusion algorithm uses a first order closure scheme for turbulent diffusion, where diffusivities are parameterized based on mixing length theory and modified as a function of atmospheric stability. We initialize WindNinja simulations with output from mesoscale weather forecasts using the Weather Research and Forecasting (WRF) model to capture the large-scale atmospheric flows and stability conditions. Model performance is evaluated against field data collected under a range of conditions at different locations including a multi-day continuous tracer gas dispersion experiment in an orchard located on rolling terrain in eastern Washington and a post-wildfire PM10 monitoring campaign in SE Idaho. The combination of fast run times, low computational demands, and explicit treatment of terrain and vegetation at a high spatial resolution are expected to produce improved dispersion predictions in regions with complicated topography, especially for applications which require time-critical management decisions.