Research Abstracts 1999-2000:

Report No. 2000-01

Barbour, P.L. and S.N. Walker, "Site Wind Forecasts using a Regional Weather Prediction Model," OSU Wind Research Cooperative, WRC Report No. 2000-01, August 17, 2000, 31 pp., $22.50.

A regional, mesoscale, atmospheric model is used to produce wind forecasts at four well-exposed wind sites along the coast of central Oregon. Forecasts for a 36 hour period were produced daily for an extended period enabling forecast accuracy to be evaluated using a relatively large number of case studies. Error statistics were evaluated to determine the general accuracy and characteristics of the site wind forecasts produced in this manner. A number of important characteristics were identified. Forecasts of wind speed contained some large errors and showed only marginal skill for many sites. The diurnal variation in the winds was predicted, however, the forecast errors varied significantly depending on the site and the projection time of the forecast. Forecasts of the v (south to north) wind component were highly correlated with observations while the u (west to east) component forecasts were not. In addition, large errors were apparent in the initial conditions at most of the sites, indicating that improvements in the model application should be made. Two simple linear adjustment methods were also tested. These adjustment methods are an attempt to account for any systematic bias that might exist in the model forecasts. The adjustments were computed based on the relationship between the predicted wind speed and that observed at a site over a limited calibration period consisting of either ten, twenty or thirty days. The results suggest that a considerable improvement can be achieved for sites that show some form of a bias. Of the four sites used in this study, the RMSE values were reduced by as much as 20% through the use of a simple adjustment.