Evaluation of the Operational Multiscale Model with Grid Adaptivity (OMEGA) for use in wind energy potential assessment in the Great Basin of Nevada

In order to further assess the wind energy potential for Nevada, the accuracy of a computational meteorological model, the Operational Multiscale Model with Grid Adaptivity (OMEGA), was evaluated by comparing data collected from a wind monitoring tower near Tonopah, NV with simulation results. Topographic forcing has a dominant role in the development and modification of mesoscale flows in regions of complex terrain, like Tonopah, especially at the level of wind turbine blade heights (~80 m). The state of Nevada is characterized by high mountains and low-lying valleys, therefore, in order to determine the wind potential for the state, meteorological models that predict the wind must be able to accurately represent and account for terrain features and simulate topographic forcing with accuracy. OMEGA incorporates an unstructured triangular horizontal grid which allows for increased flexibility and accuracy in characterizing areas of complex terrain.

The observational tower used for comparison is located at Stone Cabin, NV, and had both sonic anemometers and cup anemometers installed at heights of 40 m, 60 m, and 80 m above the surface. During a previous experiment, tower data were collected for the period February 9 through March 10, 2007 and compared to model simulations using the MM5 and WRF models. In this previous research, neither the MM5 nor the WRF showed a significant improvement in ability to forecast wind speed with increasing horizontal model resolution.

The present research evaluated the ability of OMEGA to reproduce point winds as compared to the observational data from the Stone Cabin Tower at 40 m, 60 m, and 80 m. Additionally, model sensitivity to horizontal grid resolution, initial conditions, and terrain dataset resolution were tested. OMEGA was run over five different horizontal grid resolutions with minimum horizontal edge lengths of: 18 km, 6 km, 2 km, 666 m, and 222 m. For each resolution, the model was initialized using both the Global Forecasting System (GFS) and North American Regional Reanalysis (NARR) to determine model sensitivity to the resolution of the initial conditions. Additionally, the 666 m and 222 m minimum grid resolution runs were run with both a 90 m and 1 km resolution terrain database to determine the sensitivity to terrain features. Each 30-day model run was then analyzed using statistical analysis to determine how the model generated winds compared with the observed winds. The statistical results were then compared with the results from the MM5 and WRF simulations to determine the most appropriate model for wind energy potential studies in complex terrain.