The main objectives of this study are to characterize and analyze three wind datasets produced from Mesoscale wind models in the state of Arizona. These three datasets include the National Renewable Energy Laboratory (NREL) Western Wind Dataset (WRF model), the Northern Arizona University 3TIER dataset (WRF model), and a data set produced by AWS Truewind (Mesomap) covering the state of Arizona. First the accuracy of each model was tested through validation and comparison with different meteorological towers throughout Northern Arizona. Then the gridded 50 m wind speed outputs of the three models was tested and the models outputs were characterized. Additionally after reconciling spatial scales these gridded outputs were compared with one another. Yearly average changes in wind speed surfaces, and also mean difference comparison surfaces were calculated. Through this characterization patterns and trends were identified from these surfaces. Additionally a correlation analysis was conducted on these mean difference surfaces with continuous geographic data including slope and elevation, to determine whether these differences are correlated with this geographic data.

It was determined the AWS modeled dataset predicted more accurately in three out of four cases when compared to the NREL and NAU datasets. Also RMS values were lower and more favorable when and entire year of measured and modeled data was used. WRF models showed a much similar output of averaged data and average differences when compared to the AWS datasets. The AWS dataset tended to predict higher wind speeds in specific elevation bands of 1500 to 2000 meters. Additionally when the mean difference values were compared to their corresponding slope values, both WRF comparisons with the AWS model displayed positive correlation values.

By identifying some of the spatial patterns present in the model's outputs and comparison surfaces, and by determining the accuracy of each model, wind industry professionals can make informed decisions related to the use of these models as it applies to wind energy development in the region.