Fire weather forecasts rely on numerical weather simulations where the grid size is 4 km x 4 km or larger. In areas of complex terrain, this model resolution will not capture the details of wind flows associated with complicated topography. Wind channeling in valleys, wind speed-up over mountains and ridges, and enhanced turbulence associated with rough terrain and tall forest canopies are poorly represented in current weather model applications. In this paper, we describe high resolution wind measurements over several distinct terrain configurations as a basis for evaluation of various high resolution wind modeling approaches. The field sites will include a stretch of the Salmon River which is a low vegetation steep canyon, Big Southern Butte which is a large isolated terrain obstacle with sparse vegetation, and Benton Creek in the Priest River Experimental Forest which is a mature forested mountain/valley drainage. At each site, a dense array of surface wind sensors, complemented with vertical profiling instruments, were used to map wind fields over an extended period. The high resolution models include: WindNinja, a mass consistency interpolation scheme with enhancements to use a computational fluid dynamics solver; CALMET as employed in the ClearSky a smoke dispersion forecast system; and WRF implemented at high resolution in a nest domain approach. The combination of different high resolution modeling methods with data analysis and evaluation for three very different terrain features will provide a rich and complete basis for analyzing how these high resolution methods can be used in concert with current forecast models to yield a much more detailed and accurate fire forecast.