Improving Wind Speed Forecasts at Wind Turbine Locations over Northern China through Assimilating Nacelle Winds with WRFDA

Abstract To improve the wind speed forecasts at turbine locations and hub height, this study develops WRFDA system to assimilate observations measured on nacelle of turbines (hereafter referred as observations) with both 3DVAR 4DVAR algorithms. Results exhibit that developed data assimilation (DA) helps in greatly improving analysis forecast speed. Among three experiments no cycling DA, 2-h 4-h last experiment generates best analysis, averaged (from T + 9 24) over all farms by 32.5% bias 6.3% RMSE. After processing into superobs, even bigger improvements are revealed when validating against either original or superobs. Taken results validated superobs an example, RMSE reduced 38.8% 12.0%, respectively. Compared best-performed (4-h superobs), following same DA strategy but using algorithm exhibits further improvements, especially for farms, changing amount enhanced farms. control experiment, reduces 54.6% (0.66 m s −1 ) 12.7% (0.34 ).