陈超辉,李崇银,谭言科,王铁. 2010. 基于交叉验证的多模式超级集合预报方法研究[J]. 气象学报, 68(4):464-476, doi:10.11676/qxxb2010.045
基于交叉验证的多模式超级集合预报方法研究
Research of the multi model super ensemble prediction based on cross validation
投稿时间:2008-07-18  修订日期:2008-10-21
DOI:10.11676/qxxb2010.045
中文关键词:  交叉验证,超级集合预报,支持向量机回归,向量EOF分解
英文关键词:Cross-validation, Super ensemble prediction, Support vector machines for regression, Vector EOF expansion
基金项目:国家科学自然基金 (40575027)
作者单位
陈超辉 解放军理工大学气象学院, 南京, 211101 
李崇银 解放军理工大学气象学院, 南京, 211101
中国科学院大气物理研究所, LASG, 北京, 100029 
谭言科 解放军理工大学气象学院, 南京, 211101 
王铁 解放军理工大学气象学院, 南京, 211101 
摘要点击次数: 3315
全文下载次数: 2956
中文摘要:
      利用AREM、MM5和WRF 3个中尺度有限区域模式,通过选取对短期天气预报影响颇大的积云参数化方案和边界层方案构成15个集合预报成员,以2003年7月汛期天气为研究对象,分别采用相关加权、多元线性回归以及支持向量机回归与“交叉验证”相结合的方法,开展有限区域模式的多模式短期超级集合预报研究。文中主要对上述3种方法的24 h降水和700 hPa流场的超级集合预报结果与多模式集合平均预报结果以及T213模式结果进行了对比分析,结果表明:(1)对于24h降水,支持向量机回归方法的超级集合预报得到的均方根误差比多模式集合平均小,各降水临界值的TS异常评分比多模式集合平均高;并且它也较相关加权法和多元线性回归的超级集合预报效果好。(2)对于700 hPa流场,对比分析各预报结果经过向量EOF分析得到的风场第1模态和第2模态表明,多模式集合平均主要使风场强度变小,多元线性回归和支持向量机回归的超级集合预报可以较好地刻画风场的强度分布,其中支持向量机回归的超级集合预报对风场强度及其区域分布的预报效果最好。(3)对于700 hPa流场,超级集合预报明显优于同期T213模式预报,从相同的预报均方根误差意义看,支持向量机回归的超级集合预报至少较T213模式预报能提前12 h。
英文摘要:
      A multi-model super-ensemble forecasting system with 15 ensemble members generated based on different cumulus parameterization and planetary boundary layer (PBL) schemes by means of changing the AREM, MM5,and WRF three limited area Mesoscale models’ configuration is employed to investigate the potential of multi model short range ensemble forecasting in the Huaihe Valley during the rainy season in China. Weighted average by correlation, multivariate linear regression and support vector machines (SVM) for regression are respectively combined with cross validation in this system in this paper to research the weather elements during the rainy season in July in 2003, which mainly contain the 24 hour accumulated precipitation and streamline field on 700 hPa. The experiment results compared with the traditional ensemble average of the 15 ensemble members and forecast by the T213 model shows that: (1) for the 24 hour accumulated precipitation, the root means square error (RMSE) and TS score of the results made by support vector machines for regression is much better than that of the multi model traditional ensemble average, and also better than that of other two methods; (2) for streamline field on 700 hPa, the first mode and the second mode created by means of vector EOF expanding all the prediction results indicate that the ensemble average decrease mainly the magnitude of wind field and multivariate linear regression and support vector machines for regression can however present a good intensity distribution map of wind vectors with the total effect made by support vector machines for regression being the best among the four means as mentioned above; (3) the results from the multi model super ensemble prediction are obviously superior to the simultaneous results created by the T213 model for the wind speed on 700 hPa so that super ensemble results by SVM can be used for dicision making in advance of the T213 model at least 12 hours in the sight of prediction root means square error.
HTML   查看全文   查看/发表评论  下载PDF阅读器
分享按钮