HUANG Shitong,GENG Qinghua,WU Jing,et al.The Impact of Different Initialization Times on the Wind Speed Forecast Error of the WRF Model in the Complex Terrain Region of Southern Sichuan[J].Journal of Chengdu University of Information Technology,2026,41(02):185-191.[doi:10.16836/j.cnki.jcuit.2026.02.008]
川南复杂地形区不同起报时次对WRF模式风速预报误差的影响
- Title:
- The Impact of Different Initialization Times on the Wind Speed Forecast Error of the WRF Model in the Complex Terrain Region of Southern Sichuan
- 文章编号:
- 2096-1618(2026)02-0185-07
- Keywords:
- WRF model; initialization; wind speed forecast; forecast error; complex terrain
- 分类号:
- P456
- 文献标志码:
- A
- 摘要:
- 风电作为重要的可再生能源,在实现“双碳”目标中发挥关键作用。为探究四川南部凉山地区复杂地形条件下,不同起报时次对WRF模式风速预报误差的影响。基于2022年逐时10 m高度风速观测数据,分别评估4个起报时(00时、06时、12时、18时)的36 h预报。研究结果表明,06时起报的预报效果最佳,全年均方根误差(RMSE)为1.59 m/s,偏差(Bias)为0.87 m/s,相关系数(r)为0.38; 00时起报的RMSE为1.60 m/s,偏差(Bias)为0.90 m/s。12时和18时起报尽管相关系数较高(分别为0.39和0.38),但预报误差相对较大,12时的偏差达到0.91 m/s。在季节尺度,06时起报在全年各季表现稳定,尤其在暖季(4-10月)的RMSE低于1.0 m/s,冷季(11-3月)预报误差也较小。所以优先选择06时作为起报时间,以提高风速预报的准确性。
- Abstract:
- To explore the impact of different initialization times on wind speed forecast errors of the WRF model under the complex terrain conditions of the Liangshan region in southern Sichuan, wind speed forecasts for four initialization times(00, 06, 12, and 18 UTC)were evaluated based on meteorological data from 2022. The study found that the 06 UTC initialization time had the best forecasting performance, with an annual root mean square error(RMSE)of 1.59 m/s, bias of 0.87 m/s, and a correlation coefficient of 0.38. The 00 UTC initialization time had an RMSE of 1.60 m/s and a bias of 0.90 m/s. Although the 12 and 18 UTC initialization times had higher correlation coefficients(0.39 and 0.38, respectively), their forecast errors were relatively larger, with the 12 UTC bias reaching 0.91 m/s. From a seasonal perspective, the 06 UTC initialization showed stable performance throughout the year, especially in the warm season(April to October), where the RMSE was below 1.0 m/s, and forecast errors were also smaller in the cold season(November to March). In conclusion, it is recommended to prioritize the 06 UTC initialization time for wind speed forecasting in wind farm applications in the complex terrain areas of southern Sichuan to improve forecast accuracy.
参考文献/References:
[1] 刘轶,曾琬莹,施雅蓉,等.“双碳”背景下中国风电景观的立地布局特征及类型[J].风景园林,2023,30(11):87-95.
[2] 李强,张三.风电功率预测中的起报时间选择对预测精度的影响[J].能源研究与信息,2022,38(4):87-93.
[3] 杜梦蛟,易侃,文仁强,等.WRF模式中不同物理参数化方案组合在中国近海风能资源评估中的适用性研究[J].海洋预报,2023,40(1):65-78.
[4] 马永永,娄盼星,惠英,等.基于WRF不同边界层方案的陕西气象要素预报检验[J].陕西气象,2023(1):8-14.
[5] 张珊,王宗敏,黄刚,等.基于WRF-LES的崇礼复杂地形局地风场模拟研究[J].高原气象,2023,42(1):197-209.
[6] 罗颖,刘雨辰,米立华,等.基于WRF模拟和注意力机制的短期风速预测[J].太阳能学报,2023,44(9):302-310.
[7] 刘桂艳,连喜虎,高松,等.WRF模式对渤海和黄海海面风预报误差分析[J].防灾科技学院学报,2023,25(1):40-48.
[8] 李叶晴,师春香,沈润平,等.基于WRF-LES的门头沟地区近地面风场模拟与边界层方案敏感性研究[J].高原气象,2023,42(3):758-770.
[9] 王婧,徐枝芳,范广洲.GRAPES_RAFS系统2m温度偏差订正方法研究[J].气象,2015,41(6):753-757.
[10] 陈超君,王东海,李国平,等.冬季高海拔复杂地形下GRAPES-Meso要素预报的检验评估[J].气象,2012,38(6):657-668.
[11] 杜娟,于晓晶,辛渝,等.乌鲁木齐区域数值预报业务系统降水预报检验与评估分析[J].沙漠与绿洲气象,2016,10(6):31-40.
[12] Kalnay E.Atmospheric Modeling,Data Assimilation and Predictability[M].Cambridge:Cambridge University Press,2003:1-341.
[13] Li G,Xia X,Yang Y,et al.Impact of physical parameterizations on wind simulation with WRF V3.9.1.1 under stable conditions at planetary boundary layer gray-zone resolution:A case study over the coastal regions of North China[J].Geoscientific Model Development,2020,13(4):1211-1232.
[14] Tiesi A,Mercogliano P,Reder A,et al.Assimilation of Sentinel-1 SAR Wind Data to Improve WRF Model Simulations of Extreme Weather Events[J].Remote Sensing,2021,13(2):389.
[15] Chen K,Hong Y,Tsai W.Application of Bias Correction to Improve WRF Ensemble Wind Speed Forecast[J].Atmosphere,2022,13(9):1458.
[16] Cheng X,Zhang H,Li W.Evaluation of the WRF model in predicting near-surface wind speed over complex terrain[J].Journal of Applied Meteorology and Climatology,2018,57(9):1933-1949.
[17] Li Z,Zhu Y.Impact of initialization time on short-term wind speed prediction using WRF model in different seasons[J].Atmospheric Research,2016,170:1-12.
[18] Wang Y,Seaman N L.A comparison study of meteorological forecast error statistics in different weather regimes using the WRF model[J].Monthly Weather Review,1997,125(1):34-53.
相似文献/References:
[1]钱鑫铭,周筠珺,赵鹏国,等.一次四川盆地雷暴天气的数值模拟分析[J].成都信息工程大学学报,2019,(05):532.[doi:10.16836/j.cnki.jcuit.2019.05.015]
QIAN Xinming,ZHOU Yunjun,ZHAO Pengguo,et al.Numerical Simulation Study of Thunderstorms in Sichuan based on WRF Model[J].Journal of Chengdu University of Information Technology,2019,(02):532.[doi:10.16836/j.cnki.jcuit.2019.05.015]
备注/Memo
收稿日期:2024-08-26
基金项目:四川省科技计划资助项目(2022YFS0544、2023YFS0434)
通信作者:白磊.E-mail:caecar@hainanu.edu.cn