Verification of North American Model Output Utilizing Mesoscale Network Observations

Many agricultural and fire danger models utilize input variables from existing meteorological models output. In order to better understand possible errors associated with these input variables, output data of air temperature, relative humidity, solar radiation, wind speed, wind direction, and wind gusts from the North American Model from 6 to 48 hours for each of the four daily model runs were compared to actual observations taken at 120 locations by the Oklahoma Mesonet. Results of this study show that, in Oklahoma, model validation time and location of the grid point have a dependence upon accuracy of the model output. Additionally, limitations in the model's ability to discern smaller scale terrain and the resulting impacts are shown.