Continuing advances in computer processing power have allowed numerical models to simulate atmospheric processes at finer and finer grid spacing. However, most observation networks have not increased the corresponding spatial coverage of in situ sensors. As such, challenges have arisen when using observational data to verify the forecasts from increasingly higher-resolution models. For this study, data were collected from two observational networks, the Oklahoma Mesonet and the Wichita Mountains Micronet, to provide enhanced spatial density of in situ observations to verify hourly forecasts (out to a period of 8 hours) from a downscaled Weather Research and Forecasting Model (WRF) and the National Digital Forecast Database (NDFD) using traditional verification techniques. The downscaled WRF simulations utilized a 100 m grid spacing in an operational setting at the Weather Forecast Office of the National Weather Service in Norman, OK. The simulations were then archived for post-analysis. The NDFD, which are the official forecasts of the National Weather Service, had either 5.0 km or 2.5 km grid spacing. Air temperature at 2 m, dew point temperature at 2 m, and wind speed at 10 m were the variables analyzed over the study period, which spanned August 2010 through May 2011. In addition, unique, high impact events (e.g., fire weather) over the study period were investigated to determine the utility of the forecasts from the downscaled WRF from an operational perspective as the events evolved. Using a variety of statistical analyses, the results demonstrated that the temperature and dew point forecasts from the downscaled WRF typically yielded less overall bias than the NDFD forecasts over variable spatial and temporal domains. At the same time, the wind speed forecast bias was positive for both forecast systems.