设为首页
加入收藏
Prediction of Meteorological Elements Based on Nonlinear Support Vector Machine Regression Method
-
摘要: 该文介绍了基于基本的支持向量机非线性回归方法，该方法具有解决非线性问题的能力，在数值预报解释应用技术中，对某些预报量与预报因子之间相关性不显著的要素，如风、比湿等，采用支持向量机非线性回归技术较多元回归的MOS方法更具优势；利用北京市气象局中尺度业务模式 (MM5V3) 的12：00(世界时) 起始数值预报产品和观测资料，制作北京15个奥运场馆站点6~48 h逐3 h的气象要素释用产品。对比MM5V3模式，从均方根误差的平均减小率来看，2 m温度减小12.1%，10 m风u分量减小43.3%，10 m风v分量减小53.4%，2 m比湿减小38.2%。与同期的MOS方法预报结果相比，整体预报效果SVM略优于MOS。由此可见，支持向量机非线性回归方法解决与预报因子之间非线性相关的气象要素较好，具有较高的预报优势。
-
关键词:
- 支持向量机 (SVM);
- MOS;
- 释用;
- 气象要素预报
Abstract: Nonlinear regression method based on basic support vector machine is introduced, which is able to solve nonlinear problems. The cross validation in this method makes it able to optimize the kernel function parameters.Therefore, in numerical weather prediction interpretation, nonlinear support vector machine regression technique is better than multi-variant MOS regression method when the linear relationship between the predictor and certain elements, such as wind and specific humidity, is not clear. The numerical prediction products of operational meso-scale model (MM5V3) in Beijing Meteorological Service and observations are used to make 6—48 h interpretation products with 3-hour interval of meteorological elements of 15 venues stations in Beijing. The comparison of interpretation products and MM5V3 forecast indicates that the root mean square error for 2 m temperature, 10 m wind u component, 10 m wind v component and 2 m specific humidity reduces by 12.1%, 43.3%, 53.4% and 38.2%. Compared with the prediction results of MOS, 2 m temperature, 10 m wind u compoent, v component prediction results of SVM are slightly better than those of MOS, and 2 m specific humidity prediction result of SVM is better than that of MOS.Defining the forecast with deviations no more than 2℃ as accurate for 2 m temperature, the forecast accuracy of SVM-release, MOS-release, MM5V3 model and T213 model are 66.5%, 62.2%, 58.8% and 2.5%, respectively. Forecast accuracy of 10 m wind u, v components are defined as the percentage of forecast with absolute deviations within 1 m/s, thus the forecast accuracy of SVM-release are 77.6% and 76.7%, forecast accuracy of MOS-release are 75.8% and 73.7%, forecast accuracy of MM5V3 model are 54.5% and 41.1%, and forecast accuracy of T213 model are 46.9% and 34.9%.Forecast accuracy of 2 m specific humidity is defined as the percentage of forecast with absolute deviations within 2 g·kg-1, thus the forecast accuracy of SVM-release, MOS-release and MM5V3 model are 84.9%, 67.8% and 61.7%, respectively. It shows that nonlinear support vector machine regression method is good at solving nonlinear dependence between meteorological elements and predictors, and performs better than MOS. -
图 2 SVM，MOS释用预报和MM5V3，T213模式预报北京15个奥运场馆平均2 m温度均方根误差 (a) 和预报偏差 (b) 对比
Fig. 2 The 2 m temperature root mean square error (a) and bias (b) of 15 Olympic venues in Beijing provided by SVM-release, MOS-release, MM5V3 model and T213 model
图 3 2 m温度SVM释用预报、MOS释用预报相对原模式预报均方根误差减小百分率
Fig. 3 The decreased root mean squre error ratio of 2 m temperature forecast from SVM-release, MOS-release to the counterpart of direct output of model
图 4 SVM，MOS释用预报和MM5V3，T213模式预报北京15个奥运场馆平均10 m风u分量均方根误差 (a) 和预报偏差 (b) 对比
Fig. 4 The 10 m wind u component root mean square error (a) and bias (b) of 15 Olympic venues in Beijing provided by SVM-release, MOS-release, MM5V3 model and T213 model
图 5 SVM，MOS释用预报和MM5V3，T213模式预报北京15个奥运场馆平均10 m风v分量均方根误差 (a) 和预报偏差 (b) 对比
Fig. 5 The 10 m wind v component root mean square error (a) and bias (b) of 15 Olympic venues in Beijing provided by SVM-release, MOS-release, MM5V3 model and T213 model
图 6 10 m风u分量SVM释用预报、MOS释用预报相对原模式预报均方根误差减小百分率
Fig. 6 The decreased root mean square error ratio of 10 m wind u component forecast from SVM-release, MOS-release to the counterpart of direct output of model
图 7 10 m风v分量SVM释用预报、MOS释用预报相对原模式预报均方根误差减小百分率
Fig. 7 The decreased root mean square error ratio of 10 m wind v component forecast from SVM-release, MOS-release to the counterpart of direct output of model
图 8 SVM, MOS释用预报和MM5V3模式预报2 m比湿 (a) 以及MOS和T213预报2 m相对湿度 (b) 北京15个奥运场馆平均均方根误差对比
Fig. 8 The 2 m specific humidity root mean square error provided by SVM-release, MOS-release (a), MM5V3 model, and the 2 m relative humidity root mean square error provided by MOS-release, T213 model (b) of 15 Olympic venues in Beijing
图 9 2 m比湿SVM释用预报相对MM5V3及2 m相对湿度MOS释用预报相对T213模式均方根误差的减小百分率
Fig. 9 The decreased root mean square error ratio of 2 m specific humidity forecast from SVM-release to MM5V3 model, and the decrease root mean square error ratio of 2 m relative humidity forecast from MOS-release to T213 model
表 1 利用SVM方法建模选取的预报因子
Table 1 Selected forecast factors for SVM method
预报量 三维模式预报因子 二维模式预报
因子实况因子 因子总数 预报因子 等压面/hPa 2 m温度 u和v、温度、位势高度、
相对湿度、云水975，925，850，500，
200海平面气压、
降水量2 m温度、10 m风速、10 m
风向、过去3 h总降水量132 10 m风u，v
分量u和v、位势高度 975，850，500，200 海平面气压、
降水量2 m温度、本站气压、10 m
风u分量、10 m风v分量76 相对湿度 1000，975，925，850 2 m比湿 u和v、相对湿度、
温度露点差1000，975，925，850 降水量 10 m风u分量、10 m风v分量、
2 m温度露点差、2 m相对湿度72 
下载: 导出CSV
-
[1] Klein W H, Lewis F B M, Enger I. Objective prediction of five-day mean temperatures during winter. J Meteorol, 1959, 16: 672-682. doi: 10.1175/1520-0469(1959)016<0672:OPOFDM>2.0.CO;2 [2] Glahn H R, Lowry D A. The use of Model Output Statistics (MOS) in objective weather forecasting. J Appl Meteor, 1972, 11:1203-1211. doi: 10.1175/1520-0450(1972)011<1203:TUOMOS>2.0.CO;2 [3] 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 [4] 赵声蓉.多模式温度集成预报.应用气象学报, 2006, 17(1):52-58. doi: 10.11898/1001-7313.20060109 [5] 曾晓青, 邵明轩, 王式功, 等.基于交叉验证技术的K-NN方法在降水预报中的试验.应用气象学报, 2008, 19(4):471-478. doi: 10.11898/1001-7313.20080411 [6] 陈豫英, 刘还珠, 陈楠, 等.基于聚类天气分型的KNN方法在风预报中的应用.应用气象学报, 2008, 19(5):564-572. doi: 10.11898/1001-7313.20080507 [7] 曹晓钟, 闵晶晶, 刘还珠, 等.分类与集成方法在降雨预报中的应用.气象, 2008, 34(10):3-11. doi: 10.7519/j.issn.1000-0526.2008.10.001 [8] 王迎春, 刘凤辉, 张小玲, 等.北京地区中尺度非静力数值预报产品释用技术研究.应用气象学报, 2002, 13(3):312-321. http://qikan.camscma.cn/jams/ch/reader/view_abstract.aspx?file_no=20020341&flag=1 [9] 冯汉中, 陈永义.支持向量机回归方法在实时业务预报中的应用.气象, 2005, 31(1):41-44. doi: 10.11676/qxxb2005.005 [10] 冯汉中, 杨淑群, 刘波.支持向量机 (SVM) 方法在气象预报中的个例试验.四川气象, 2005, 25(2):9-12. http://www.cnki.com.cn/Article/CJFDTOTAL-SCCX200502004.htm [11] Vapnik V N. Statistical Learning Theory. New York:John Wiley & Sons Inc, 1998. [12] Vapnik V N. The Nature Of Statistical Learning Theory. New York:Springer Verlag, 2000. [13] 陈永义, 俞小鼎, 高学浩, 等.处理非线性分类和回归问题的一种新方法 (Ⅰ)——支持向量机方法简介.应用气象学报, 2004, 15(3):345-354. http://qikan.camscma.cn/jams/ch/reader/view_abstract.aspx?file_no=20040344&flag=1 [14] 冯汉中, 陈永义.处理非线性分类和回归问题的一种新方法 (Ⅱ)——支持向量机方法在天气预报中的应用.应用气象学报, 2004, 15(3):355-365. http://qikan.camscma.cn/jams/ch/reader/view_abstract.aspx?file_no=20040345&flag=1 [15] Scholkopf B, Burges C J C, Smola A J. Advance in Kernel Methods Support Vector Learning. Cambridge:MIT Press, 1999. [16] 盛裴轩, 毛节泰, 李建国, 等.大气物理学.北京:北京大学出版社, 2003. -
计量
- 摘要浏览量: 3619
- HTML全文浏览量: 1122
- PDF下载量: 1572
- 被引次数: 0
