设为首页
加入收藏
Application of ECMWF Ensemble Forecast Products to Rainstorm Forecast in Guangxi
-
摘要: 基于最大相关最小冗余度算法和随机森林回归算法，该文提出一种对欧洲中期天气预报中心（ECMWF）集合预报产品进行暴雨预报的释用方法。该方法采用最大相关最小冗余度算法，对ECMWF集合预报的51个成员进行筛选，选取若干个与预报对象相关性最大、相互间冗余度最小的成员作为随机森林回归算法的输入因子。利用ECMWF集合预报降水量平均值对建模样本进行分类，使预报模型的建模样本更具有针对性。通过2012年4月—2015年12月的交叉独立样本试验预报和2016年1—9月的业务预报试验的统计结果表明：该释用方法的暴雨预报TS和ETS评分，均比采用ECMWF集合预报产品51个成员降水量预报进行插值后取平均值的释用方法分别提高了0.07和0.05以上，显示了较好的数值预报产品释用效果。
-
关键词:
- 最大相关最小冗余度算法;
- 随机森林回归算法;
- 释用
Abstract: Using the maximal correlation minimum redundancy algorithm and random forest regression algorithm, a rainstorm interpretation forecasting method with numerical prediction products is proposed based on the ensemble prediction system of European Center for Medium-Range Weather Forecasts (ECMWF). The precipitation forecast of 51 members in ECMWF ensemble prediction system are interpolated to weather stations, and then, the maximum related minimum redundancy algorithm is used to filter ensemble members. Finally, several member interpolations that have the highest correlation with the predictand and the least redundancy with each other are selected as input factors of the random forest regression algorithm. Furthemore, in order to make modeling samples of the forecast model more pertinence, the modeling samples are classified using the mean rainfall value of ECMWF ensemble prediction products of 51 members. That is, when the mean precipitation using ECMWF ensemble prediction products at a certain station is relatively large and there is a possibility of precipitation above the storm level, only historical samples containing a large amount of precipitation are selected as modeling samples of the forecasting model. Therefore, the forecasting model reduces the influence of the sunny and wet weather samples on the noise of the forecasting model, so that forecasting model focuses on the training of large precipitation samples. When the mean value of the predicted ECMWF ensemble precipitation at a certain weather station is small, all samples of the weather station (including samples of sunny days and heavy precipitation) are modeled so that the training of the forecasting model can reconcile the heavy rain samples and thus as far as possible to avoid the rainstorm of weather station omissions reported. This method is applied to 89 stations in Guangxi, and a 4-year cross-independent sample test forecast for 2012-2015 is carried out. The business test forecast is carried out in 2016. In the 4-year cross-independent sample test results, rainstorm TS and ETS scores of this method are all improved by 0.04-0.09 and 0.04-0.07, respectively, compared with the average value after interpolation using the precipitation forecast of 51 members in ECMWF ensemble prediction products. Results of the business trial in 2016 show that TS and ETS scores of the method for interpretation rainstorms TS and ETS scores are improved by 0.07 and 0.05, respectively, compared with average values of pre-interpolation methods for the precipitation forecast of 51 members in ECMWF ensemble prediction products. It shows that the proposed rainstorm precipitation method of ECMWF ensemble prediction products has advantageous effects on forecasting and practical application forecast. -
图 3 2016年8月2日20：00—3日20:00 24 h降水实况与预报
(a)实况，(b)MRMR-RFR释用方法预报，(c)AVI方法预报
Fig. 3 Observation and prediction of the case from 2000 BT 2 Aug to 2000 BT 3 Aug in 2016
(a)observation, (b)prediction of MRMR-RFR, (c)prediction of AVI
图 4 2016年8月12日20:00—13日20:00 24 h降水实况和预报
(a)实况，(b)MRMR-RFR释用方法预报，(c)AVI方法预报
Fig. 4 Observation and prediction of the case from 2000 BT 12 Aug to 2000 BT 13 Aug in 2016
(a)observation, (b)prediction of MRMR-RFR, (c)prediction of AVI
表 1 2012—2015年暴雨以上量级降水交叉独立预报TS评分
Table 1 TS of cross independent sample test forecast of rainstorm from 2012 to 2015
年份 MRMR-RFR AVI (n=10, α=15, β=10) (n=10, α=20, β=15) (n=10, α=25, β=15) TS ETS TS ETS TS ETS TS ETS 2012 0.12 0.11 0.15 0.12 0.14 0.11 0.07 0.06 2013 0.11 0.10 0.13 0.12 0.13 0.12 0.09 0.08 2014 0.11 0.11 0.17 0.14 0.13 0.12 0.08 0.07 2015 0.12 0.11 0.14 0.12 0.14 0.10 0.06 0.06 
下载: 导出CSV
表 2 不同参数n下2012—2015年暴雨交叉独立预报TS评分
Table 2 TS of cross independent sample test forecast of rainstorm under different n from 2012 to 2015
年份 MRMR-RFR (n=9, α=20, β=15) (n=10, α=20, β=15) (n=11, α=20, β=15) TS ETS TS ETS TS ETS 2012 0.15 0.13 0.15 0.12 0.14 0.12 2013 0.13 0.11 0.13 0.12 0.14 0.13 2014 0.14 0.12 0.17 0.14 0.15 0.13 2015 0.12 0.11 0.14 0.12 0.16 0.14
下载: 导出CSV
表 3 2016年1—9月单站暴雨以上量级降水业务预报TS，ETS评分
Table 3 TS and ETS of single-station forecast of rainstorm using different methods from Jan 2016 to Sep 2016
月份 MRMR-RFR (n=10，α=20, β=15) AVI 降水量不小于50 mm站次 TS ETS TS ETS 1 0.24 0.18 0.00 0.00 60 2 0.50 0.45 0.00 0.00 1 3 0.07 0.05 0.06 0.05 11 4 0.08 0.05 0.03 0.02 105 5 0.17 0.15 0.13 0.11 203 6 0.17 0.14 0.06 0.03 231 7 0.13 0.10 0.01 0.00 114 8 0.24 0.22 0.18 0.15 206 9 0.07 0.04 0.00 0.00 36 1—9 0.15 0.11 0.08 0.06 967
下载: 导出CSV
-
[1] 刘还珠, 赵声蓉, 陆志善, 等.国家气象中心气象要素的客观预报——MOS系统.应用气象学报, 2004, 15(2):181-191. http://qikan.camscma.cn/jams/ch/reader/view_abstract.aspx?file_no=20040223&flag=1 [2] 毕宝贵, 代刊, 王毅, 等.定量降水预报技术进展.应用气象学报, 2016, 27(5):534-549. doi: 10.11898/1001-7313.20160503 [3] 岳彩军, 寿亦萱, 寿绍文, 等.湿Q矢量释用技术及其在定量降水预报中的应用.应用气象学报, 2007, 18(5):666-675. doi: 10.11898/1001-7313.20070502 [4] 孔荣, 王建捷, 梁丰, 等.尺度分解技术在定量降水临近预报检验中的应用.应用气象学报, 2010, 21(5):535-544. doi: 10.11898/1001-7313.20100503 [5] 杨成荫, 王汉杰, 周林, 等.基于全场信息的数值预报产品释用方法研究.应用气象学报, 2009, 20(2):232-239. doi: 10.11898/1001-7313.20090213 [6] 刘长征, 杜良敏, 柯宗建, 等.国家气候中心多模式解释应用集成预测.应用气象学报, 2013, 24(6):677-685. doi: 10.11898/1001-7313.20130604 [7] David J S, Nusrat Y.Reliable probabilistic quantitative precipitation forecasts from a short-range ensemble forecasting system.Wea Forecasting, 2007, 22(1):3-17. doi: 10.1175/WAF968.1 [8] Nusrat Y, David J S.Reliable probabilistic quantitative precipitation forecasts from a short-range ensemble forecasting system during the 2005/06 cool season.Mont Wea Rev, 2008, 136(6):2157-2172. doi: 10.1175/2007MWR2314.1 [9] 赵声蓉, 裴海瑛.客观定量预报中降水的预处理.应用气象学报, 2007, 18(1):21-28. doi: 10.11898/1001-7313.20070104 [10] 唐晓文, 汤剑平, 张小玲.基于业务中尺度模式的配料法强降水定量预报.南京大学学报(自然科学版), 2010, 46(3):277-283. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=njdxxb201003004 [11] Jerome P C, Frederick G S.Regionalization in fine-grid GFS MOS 6-h quantitative precipitation forecasts.Mon Wea Rev, 2011, 139(1):24-38. doi: 10.1175/2010MWR2926.1 [12] Jerome P C, Frederick G S.High-resolution GFS-based MOS quantitative precipitation forecasts on a 4-km grid.Mon Wea Rev, 2011, 139(1):39-68. doi: 10.1175/2010MWR3224.1 [13] 孙靖, 程光光, 张小玲.一种改进的数值预报降水偏差订正方法及应用.应用气象学报, 2015, 26(2):173-184. doi: 10.11898/1001-7313.20150205 [14] Ding C, Peng H.Minimum redundancy feature selection from microarray gene expression data.Journal of Bioinformatics and Computational Biology, 2005, 3(2):185-205. doi: 10.1142/S0219720005001004 [15] Peng H, Long F H, Ding C.Feature selection based on mutual information:Criteria of max-dependency, max-relevance, and min-redundancy.IEEE Transactions on Pattern Analysis and Machine Intelligence, 2005, 27(8):1226-1238. doi: 10.1109/TPAMI.2005.159 [16] Hossain M A, Jia X, Pickering M.Subspace detectionusing a mutual information measure for hyperspectral image classification.IEEE on Geoscience and Remote Sensing Letters, 2014, 2(11):424-428. http://ieeexplore.ieee.org/document/6578087/ [17] Breiman L.Random forests.Machine Learning, 2001, 45(1):5-32. doi: 10.1023/A:1010933404324 [18] 方匡南.随机森林组合预测理论及其在金融中的应用.厦门:厦门大学出版社, 2012. [19] 邹亮, 黄琼, 李骜, 等.基于随机森林和富集分析的阿尔茨海默症GWA研究.中国科学(生命科学), 2012, 42(8):639-647. http://www.oalib.com/paper/4129763 [20] 李建更, 高志坤.随机森林:一种重要的肿瘤特征基因选择法.生物物理学报, 2009, 25(1):51-56. http://www.cqvip.com/QK/91590X/200901/29840838.html [21] 方匡南, 朱建平, 谢邦昌.基于随机森林方法的基金收益率方向预测与交易策略研究.经济经纬, 2010(2):61-65. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jjjw201002015 [22] Du J, Mullen S L, Sanders F.Short-range ensemble forecasting of quantitative precipitation.Mon Wea Rev, 1997, 125:2427-2459. doi: 10.1175/1520-0493(1997)125<2427:SREFOQ>2.0.CO;2 -
图(4) / 表(3)
计量
- 摘要浏览量: 3131
- HTML全文浏览量: 1245
- PDF下载量: 298
- 被引次数: 0
