中文摘要：

探讨了WRF模式在风电场的风速或者功率预报中应用的可行性，主要研究和评估了WRF模式对地处东亚季风区及海陆交界的江苏如东地区夏季和冬季风速的短期预报效能。研究发现WRF模式可以比较好地预报如东站冬季的风速，24h预报的风速时间序列和观测资料的相关系数可以达到0．61，通过置信度99％的检验，48h和72h的预报与观测风速相关系数分别为0．54和0．47，也能通过置信度99％的检验；相对而言，模式对夏季风速的预报则要差一些，24h的相关系数有0．59，48h和72h的相关系数只有0．47和0．30，但仍能通过置信度99％的检验。在量值上，模式预报的风速比观测值部略偏大一些。而江苏南通市预报结果显示，模式的预报效能要比如东稍高一些，和如东类似，模式对该地冬季的预报要好于对夏季风速的预报。从更大尺度范围的分析也表明，模式对不同地区预报的准确度是不一样的，对海面以及海陆交界的海岸预报精度要高一些，在平坦的内陆地区预报也比较好，但在山区预报效能则较差。总体说来，WRF能胜任风速短期预报，值得进一步研究和应用。

英文摘要：

To discuss the possibility of applying a numerical Weather Research and Forecasting （WRF） model to wind power forecasting, the summer and winter hourly wind speed forecasting performance of the WRF model used in Rudong, Jiangsu Province, is analyzed in this paper. It is determined that the WRF model can forecast winter wind speed in Rudong 24 h in advance very well with a correlation coefficient as high as 0.61 compared to observations and is statistically signifi- cant at the 0.01 level. Similarly, the forecasts with leading times of 48 h and 72 h are also sufficiently accurate, according to the correlation coefficients. In summer, however, forecasting is less accurate, although the correlation coefficient of the 24 h leading forecast is 0.59, which is still statistically significant at the 0.01 level. The forecasting of wind speed in Nantong was also examined with observed data and was determined to be more accurate than that in Rudong in both winter and in summer. For a larger domain, the performance of wind forecasting is also analyzed. Results show that the performance of the WRF model is better for ocean than for land and is better in winter than in summer. In general, the forecasted wind speed is greater than the actual wind speed and the performance of the WRF model foresting differs significantly accord- ing to location and season. Essentially, short-term wind forecasting is valuable and is worthy of further investigation.