Development and testing of model output statistics for wind forecasts at wind-turbine-generator sites
Objective wind speed and direction forecasts for seven projections (for time frames of +9 to +30h) were developed with the Model Output Statistics technique (MOS) for nine DOE candidate wind turbine sites in the continental United States. Forecasts were produced twice daily, at 0000 GMT and 1200 GMT, and verified against observations by statistical analyses. Though the reliability of the forecasts was degraded by the small size of the developmental data set, the technique appears sufficiently reliable to merit further investigation. The forecast verification results show a slight tendency toward underforecasting strong winds. In general, the forecasts perform better in the eastern portion of the country than in the West. The forecasts generally decrease in accuracy with time, except for sites having very regular diurnal wind patterns, and were markedly better at the 9-h forecast than at later times. Wind observations at the WTG site were found to be useful predictors for projections up to 15 hours after forecast time at some sites. Both deterministic forecasts (actual speed) and probabilistic forecasts (probability of the speed being within specified intervals) were produced. The results suggest that larger developmental samples or larger speed intervals be used to develop probabilistic wind speed forecasts.
- Research Organization:
- National Oceanic and Atmospheric Administration, Silver Spring, MD (USA). Techniques Development Lab.
- DOE Contract Number:
- AI06-79RL10046
- OSTI ID:
- 5516734
- Report Number(s):
- DOE/RL/10046-1; ON: DE82008340
- Country of Publication:
- United States
- Language:
- English
Similar Records
Investigation of boundary-layer wind predictions during nocturnal low-level jet events using the Weather Research and Forecasting model
Probabilistic Day-Ahead Forecasting Using an Analog Ensemble Approach for Wind Farm Grid Services
Related Subjects
WIND
FORECASTING
VELOCITY
CALIFORNIA
DATA COVARIANCES
KANSAS
MASSACHUSETTS
MATHEMATICAL MODELS
NEBRASKA
NEW YORK
NORTH CAROLINA
RHODE ISLAND
SOUTH DAKOTA
STATISTICS
VALIDATION
FEDERAL REGION I
FEDERAL REGION II
FEDERAL REGION IV
FEDERAL REGION IX
FEDERAL REGION VII
FEDERAL REGION VIII
MATHEMATICS
NORTH AMERICA
TESTING
USA
170604* - Wind Energy Engineering- Site Characteristics