Tuesday, 9 January 2018: 2:00 PM
Room 19AB (ACC) (Austin, Texas)
This study investigated the performances of new developed high-resolution WRF-RTFDDA-based deterministic numerical weather prediction (NWP) system that began real-time operational forecast for wind energy prediction at CEPRI in May 2017. The model system is with grid intervals of 3km, covering the power grid operation of the State Grid Corporation of China (SGCC). The system runs one cycle every 6 hours, with 72h forecasts. The model forecast of surface (10m AGL) winds and near-surface later winds are verified with the ~1600 standard surface weather stations over China and tall met-towers available at wind farms and weather centers. The verification is performed on the operational forecast archive from 2017 summer and fall. The statistics of the system performance is calculated on a station-by-station basis as well as for the domain and regional average of traditional verification metrics including bias, root mean square error (RMSE), mean absolute error (MAE), and the correlation between observation and model outputs. The result shows significant geographical, diurnal, seasonal differences in all metrics. In particular, large differences in wind forecast error are found for a short distance near wind farms. Furthermore, regime-based weather pattern classifications are performed for a northwestern China region where large wind farm clusters are developed. Wind forecast errors characteristics for different weather regimes are studied to support the model system refinement and wind forecast error correction through statistical post-processing.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner