[1]黄小燕,赵华生,黄颖,等.遗传—神经网络集合预报方法在广西热带气旋降水预报中的应用[J].自然灾害学报,2017,(06):184-196.[doi:10.13577/j.jnd.2017.0621]
 HUANG Xiaoyan,ZHAO Huasheng,HUANG Ying,et al.An genetic neural network ensemble prediction model for tropical cyclones rainfall in Guangxi[J].,2017,(06):184-196.[doi:10.13577/j.jnd.2017.0621]
点击复制

遗传—神经网络集合预报方法在广西热带气旋降水预报中的应用
分享到:

《自然灾害学报》[ISSN:/CN:23-1324/X]

卷:
期数:
2017年06期
页码:
184-196
栏目:
出版日期:
2017-12-08

文章信息/Info

Title:
An genetic neural network ensemble prediction model for tropical cyclones rainfall in Guangxi
作者:
黄小燕1 赵华生1 黄颖1 林开平2 何立1
1. 广西区气象减灾研究所, 广西 南宁 530022;
2. 广西区气象台, 广西 南宁 530022
Author(s):
HUANG Xiaoyan1 ZHAO Huasheng1 HUANG Ying1 LIN Kaiping2 HE Li1
1. Guangxi Research Institute of Meteorological Disasters Mitigation, Nanning 530022, China;
2. Guangxi Meteorological Observatory, Nanning 530022, China
关键词:
热带气旋降水预报数值预报产品多维尺度遗传-神经网络集合预报
Keywords:
tropical cyclone (TC)precipitation forecastnumerical forecast productmultidimensional scale (MDS)genetic neural network ensemble prediction (GNNEP)
分类号:
P457;X43
DOI:
10.13577/j.jnd.2017.0621
摘要:
利用多维尺度分析可以从一系列数据集的相似性信息中发掘其中的潜在结构信息,并通过样本点间的相似度构建相似矩阵,再将相似矩阵映射到低维欧氏距离空间获取新的特征的能力。论文以1980—2015年共36年的广西热带气旋逐日降水量为基础,综合考虑热带气旋降水的数值预报产品物理量预报因子,采用多维尺度分析的预报因子信息数据挖掘技术,以进化计算的遗传算法,生成期望输出相同的多个神经网络个体,建立了一种新的非线性人工智能集合预报模型,进行遗传—神经网络的广西热带气旋降水集合预报模型研究。遗传—神经网络的集成个体的输入因子是通过选出相关程度较高的数值预报产品的物理量场格点因子,同时网络的输出是通过多维尺度降维方法对初选预报因子群进行合理的降维处理来实现的。通过对2011—2015年影响广西的22个热带气旋共94个独立样本的试验预报的统计结果表明,对于暴雨以上量级的预报,广西89站的预报平均TS评分达到了0.3;同区域同样本的对比分析表明,新预报模型5年的TS评分比欧洲细网格的TS评分提高了15%以上;进一步对大量级的降雨落区的预报对比分析表明,新方案的预报效果比欧洲中心的预报更理想。
Abstract:
We developed a novel nonlinear artificial intelligence ensemble prediction (NAIEP) model based on multiple neural networks with identical expected output created by using the genetic algorithm (GA) of evolutionary computation. We extracted the potential structure information from the similarity information of series of data sets by multidimensional scaling (MDS) analysis, and the similarity matrix is constructed by the similarity between the sample points. Then the similarity matrix is mapped to the low dimensional Euclidean distance space to obtain the new feature. We set up the Genetic Neural Network Ensemble Prediction (GNNEP) model based on data of the daily precipitation of tropical cyclones (TC) that occurred in Guangxi in the period 1980-2015 for 36 years. The predictors were selected by MDS in the Numerical Weather Prediction products to predict the TC precipitation for each station. The input factors of GA neural network are selected by selecting the physical quantities of the products with higher degree of correlation. At the same time, the output of the network is realized by the dimensionality reduction of the primary forecasting factor group by MDS. The results of the experimental forecast of 94 independent samples of 22 tropical cyclones in Guangxi from 2011 to 2015 show that the average TS score of the 89 stations in Guangxi is 0.3 for the forecast of the magnitude above the heavy rain. The comparative analysis shows that the new forecast model for 5 years TS score than the European fine grid TS score increased by 15% or more under the condition of the same region and samples, and further large magnitude of rainfall area forecast comparison analysis shows that the forecasting effect of the new scheme is more ideal than by the European Center.

参考文献/References:

[1] 孙永刚,李彰俊,孟雪峰,等. 天气动力学组合因子在MOS降水预报中的应用[J]. 气象,1998,24(2):27-30. SUN Yonggang,LI Zhangjun,MENG Xuefeng,et al. The application of composite factors of synoptic dynamics to the MOS precipitation forecasting method[J]. Meteorological Monthly,1998,24(2):27-30. (in Chinese)
[2] 陈力强,韩秀君,张立祥. 基于MM5模式的站点降水预报释用方法研究[J]. 气象科技,2003,31(5):268-272. CHEN Liqiang,HAN Xiujun,ZHANG Lixiang. Study of station precipitation forecasting method based on MM5 model output[J]. Meteorological Science and Technology,2003,31(5):268-272. (in Chinese)
[3] 曾晓青,邵明轩,刘还珠,等. 基于交叉验证技术的KNN方法在降水预报中的试验[J]. 应用气象学报,2008,19(4):471-478. ZENG Xiaoqing, SHAO Mingxuan, LIU Huanzhu, et al. Forecasting precipitation experiment with KNN based on crossing verification technology[J]. Journal of Applied Meteorological Science,2008,19(4):471-478. (in Chinese)
[4] 高守亭,冉令坤,李娜,等. 集合动力因子暴雨预报方法研究[J]. 暴雨灾害,2013,32(4):289-302. GAO Shouting,RAN Lingkun,LI Na,et al. The "Ensemble Dynamic Factors" approach to predict rainstorm[J]. Torrential Rain and Disasters,2013,32(4):289-302. (in Chinese)
[5] 龚佃利,王以琳,谢考宪. 山东飞机增雨降水区分级预报方法研究[J]. 应用气象学报,2001,12(增刊):139-145. GONG Dianli, WANG Yilin, XIE Kaoxian. A study of discrimination method for artifical precipitation areas in Shangdong Province[J]. Quarterly Journal of Applied Meteorology, 2001, 12(Supplement):139-145. (in Chinese)
[6] ZHAO Huasheng,JIN Long,HUANG Ying,et al. An objective prediction model for typhoon rainstorm using particle swarm optimization:neural network ensemble[J]. Natural Hazards,2014,72(2):427-437.
[7] 周家斌,张海福. 一种汛期降水分布的客观集成预报方法[J]. 应用气象学报,2000,11(增刊):93-97. ZHOU Jiabin,ZHANG Haifu. An objective method of ensemble forecast about precipitation distribution in rainy season[J]. Quarterly Journal of Applied Meteorology,2000,11(Supplement):93-97. (in Chinese)
[8] 陈力强,周小珊,杨森. 短期集合预报中定量降水预报集合方法初探[J]. 南京气象学院学报,2005,28(4):543-548. CHEN Liqiang,ZHOU Xiaoshan,YANG Sen. A quantitative precipitation forecasts method for short range ensemble forecasting[J]. Journal of Nanjing Institute of Meteorology,2005,28(4):543-548. (in Chinese)
[9] 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[J]. 气象,2008,34(10):3-11. CAO Xiaozhong,MIN Jingjing,LIU Huanzhu,et al. Application of classification and integration to rainfall forecast[J]. Meteorological Monthly,2008,34(10):3-11. (in Chinese)
[10] 黄刚,张立凤,王秋良,等. 基于集合预报产品的降尺度降水预报试验[J]. 气象科学,2012,32(5):508-514. HUANG Gang,ZHANG Lifeng,WANG Qiuliang,et al. Precipitation forecasting experiments based on the products of ensemble forecast by using downscaling method[J]. Journal of the Meteorological Sciences,2012,32(5):508-514. (in Chinese)
[11] 李俊,杜钧,陈超君. "频率匹配法"在集合降水预报中的应用研究[J]. 气象,2015,41(6):674-684. LI Jun, DU Jun, CHEN Chaojun. Applications of "frequency-matching" method to ensemble precipitation forecasts[J]. Meteorological Monthly,2015,41(6):674-684. (in Chinese)
[12] 王雨. 2004年主汛期各数值预报模式定量降水预报评估[J]. 应用气象学报,2006,17(3):316-324. WANG Yu. Evaluation of quantitative precipitation forecasting of multiple NWP models in summer of 2004[J]. Journal of Applied Meteorological Science,2006,17(3):316-324. (in Chinese)
[13] 杨森,陈力强,周晓珊. 数值降水预报定量集成方法的业务应用[J]. 气象与环境学报,2012,28(5):14-18. YANG Sen,CHEN Liqiang,Zhou Xiaoshan. Application of an ensemble method of quantitative precipitation forecasting[J]. Journal of Meteorology and Environment,2012,28(5):14-18. (in Chinese)
[14] Glahn,H.R.and D.A.Lowry. The use of model output statistics (MOS) in objective weather forecasting[J]. J.APPL.Mteor,1972,11(8):1203-1211.
[15] Ashok Kumar,Pvarinder Maini,S.V. Sinhg. An operational model of forecasting probability of precipitation and Yes/No forecast[J]. Weather and Forecasting,1999,14(1):38-48.
[16] 周兵,赵翠光,赵声蓉. 多模式集合预报技术及其分析与检验[J]. 应用气象学报,2006,17(增刊):104-109. ZHOU Bing,ZHAO Cuiguang,Zhao Shengrong. Multi-model ensemble forecast technology with analysis and verification of the results[J]. Journal of Applied Meteorological Science,2006,17(Supplement):104-109. (in Chinese)
[7] 梁钟清,金龙,巩远发. 低纬度地区局地暴雨的神经网络预报方法研究[J]. 热带气象学报,2009,25(4):158-464. LIANG Zhongqing,JIN Long,GONG Yuanfa. The research on methods of forecasting local rainstorms in low latitude area using the neural network[J]. Journal of Tropical Meteorology,2009,25(4):158-464. (in Chinese)
[18] 吴木贵,江彩英,张信华,等. 交叉熵神经网络及其在闽北大雨以上降水预报中的应用[J]. 南京信息工程大学学报:自然科学版,2012,4(3):220-225. WU Mugui,JIANG Caiying,ZHANG Xinhua,et al. Application of BP neural network using cross-entropy to 96 hours forecast of heavy precipitation event in northern Fujian province[J]. Journal of Nanjing University of Information Science and Technology:Natural Science Edition,2012,4(3):220-225. (in Chinese)
[19] Davision M L. Multidimensional Scaling[M]. New York:Wiley,1983.
[20] Quinlan j R. Induction of decision trees[J]. Machine Learning,1986,1:81-106.
[21] 魏莱,王守觉,徐菲菲. 基于带权多维尺度变换的奇异值挖掘[J]. 计算机科学,2008,35(1):190-192. WEI Lai,WANG ShouJue,XU FeiFei. Outliers mining via weighted multidimensionality scaling[J]. Computer Science,2008,35(1):190-192. (in Chinese)
[22] Haupt S E,Young G S,Allen C T. Validation of a receptor-dispersion model coupled with a genetic algorithm using synthetic data[J]. Journal of Applied Meteorology and Climatology,2006,45(3):476-490.
[23] 张韧,洪梅,王辉赞,等. 基于遗传算法优化的ENSO指数的动力预报模型反演[J]. 地球物理学报,2008,51(5):1346-1353. ZHANG Ren,HONG Mei,WANG Huizan,et al. Retrieval of the non-linear dynamic forecast model of El Nino/La Nina index based on the genetic algorithm optimization[J]. Chinese J. Geophys,2008,51(5):1346-1353. (in Chinese)
[24] Kalnay E,Kanamitus M,Kigler R,et al. The NCEP/NCAR 40-year reanalysis project[J]. Bull Amer Meteor Soc,1996,77:437-471.
[25] Tucker-Drob E M,Salthouse T A. Methods and measures:confirmatory factor analysis and multidimensional scaling for construct validation of cognitive abilities[J]. International Journal of Behavioral Development,2009,33(3):277-285.
[26] Marshalek B,Lohman D F,Snow R E. The complexity continuum in the radex and hierarchical models of intelligence[J]. Intelligence,1983,7(2):107-127.
[27] Snow R E,Kyllonen P C,Marshalek B. The topography of ability and learning correlations[J]. Advances in the Psychology of Human Intelligence,1984,2:47-103.
[28] Kuroda K,Hashiguchi H,Fujiwara K,et al. Reconstruction of network structures from marked point processes using multidimensional scaling[J]. Physica A:Statistical Mechanics and Its Applications,2014,415(1):194-204.
[29] Natalla Jaworska,Angelina Chupetlovska-Anastasova. A review of multidimensional scaling(MDS) and its utility in various psychological domains[J]. Tutorials in Quantitative Methods for Psychology,2009.5(1):1-10.
[30] CoxT.E.Cox M.A.A. Multidimensional Scaling[M]. London:Chapman Hall,1994.
[31] Darrell Whitley. A genetic algorithm tutorial[J]. Statistics and Computing,1994,4(2):65-85.
[32] M Mitchell. An Introduction to Genetic Algorithms[M]. Cambridge, 1998,5-15.
[33] 张文修,梁怡. 遗传算法的数学基础[M]. 西安:西安交通大学出版社,2000,4-20. ZHANG Wenxiu,LIANG Yi. The Mathematical Basis of Genetic Algorithm[M]. Xian:Xi’an Jiaotong University Press,2000,4-20. (in Chinese)
[34] 刑文训,谢金星. 现代优化计算方法[M]. 北京:清华大学出版社,1999,140-245. XING Wenxun, XIE Jinxing. Modern Optimization Method[M]. Beijing:Tsinghua University Press,1999,140-245. (in Chinese)
[35] 金龙,况雪源,黄海洪,等. 人工神经网络预报模型的过拟合研究[J]. 气象学报,2004,62(1):62-70. JIN Long,KUANG Xueyuan,WANG Haihong,et al. Study on the overfitting of the artifical neural network forecasting model[J]. 2004,62(1):62-70. (in Chinese)
[36] North G R,Bell T L,Cahalan R F,et al. Sampling errors in the estimation of empirical orthogonal function[J]. Mon Wea Rev,1982,110(2):699-706.

相似文献/References:

[1]刘彤,闫天池.我国的主要气象灾害及其经济损失[J].自然灾害学报,2011,(02):090.
 LIU Tong,YAN Tian-chi.Main meteorological disasters in China and their economic losses[J].,2011,(06):090.
[2]尹洁,杜惠良,吴静.3个深入内陆热带气旋产生大暴雨的成因——对比分析[J].自然灾害学报,2011,(03):079.
 YIN Jie,DU Hui-liang,WU Jing.Formation cause of three heavy rainstorms induced by tropical cyclones invading interior territory:comparative analysis[J].,2011,(06):079.
[3]陈惠芬.热带气旋灾害等级预评估方法初探[J].自然灾害学报,2011,(05):136.
 CHEN Hui-fen.Tentative study on Pre-evaluation method of tropical cyclone disaster grade[J].,2011,(06):136.
[4]吴晓京,陈云浩,郑新江,等.洋面风场的QuikSCAT/SeaWinds遥感探测[J].自然灾害学报,2007,(02):007.
 WU Xiao-jing,CHEN Yun-hao,ZHENG Xin-jiang,et al.Remote sensing survey of sea surface wind field by QuikSCAT/SeaWinds[J].,2007,(06):007.
[5]梁军,陈联寿,吴士杰,等.影响黄渤海域热带气旋的灾害分析[J].自然灾害学报,2007,(02):027.
 LIANG Jun,CHEN Lian-shou,WU Shi-jie,et al.Analysis of disaster of tropical cyclone affecting Yellow Sea and Bohai Sea areas[J].,2007,(06):027.
[6]袁金南,郑彬.广东热带气旋及其降水的年际变化特征[J].自然灾害学报,2008,(03):140.
 YUAN Jin-nan,ZHEN Bin.Yearly variation features of tropical cyclone and its precipitation in Guangdong Province of China[J].,2008,(06):140.
[7]王毅,石汉青,黄思训.西北太平洋热带气旋的时空分布特征分析[J].自然灾害学报,2009,(06):166.
 WANG Yi,SHI Han-qing,HUANG Si-xun.Spatiotemporal distribution of tropic cyclones in north-west Pacific Ocean[J].,2009,(06):166.
[8]张京红,刘少军,蔡大鑫.基于GIS的海南岛橡胶林风害评估技术及应用[J].自然灾害学报,2013,(04):175.
 ZHANG Jinghong,LIU Shaojun,CAI Daxin.GIS-based wind disaster assessment technology of rubber forest in Hainan Island and its application[J].,2013,(06):175.
[9]田荣湘.南半球气候变暖对西北太平洋热带气旋的影响[J].自然灾害学报,2005,(4):025.
 TIAN Rong-xiang.Influence of climatic warming in Southern Hemisphere on tropical cyclone of northwestern Pacific Ocean[J].,2005,(06):025.
[10]魏应植,张长安,林秀斌,等.“艾利”台风异常路径与登陆地点分析[J].自然灾害学报,2006,(05):054.
 WEI Ying-zhi,ZHANG Chang-an,LIN Xiu-bin,et al.Analyses of abnormal path and landing site of Typhoon AERE[J].,2006,(06):054.

备注/Memo

备注/Memo:
收稿日期:2017-01-04;改回日期:2017-02-24。
基金项目:国家自然科学基金(41575051,61562008);广西区气象局气象科研计划重点项目(桂气科2016Z03);广西青年基金(2014GXNSFBA118211)
作者简介:黄小燕(1978-),女,高级工程师,博士,主要从事非线性智能预报技术在天气预报中的应用.E-mail:Gx_huangxy@163.com
通讯作者:林开平(1960-),男,正研高工,博士,主要从事天气分析、预报工作.E-mail:linkp0305@aliyun.com
更新日期/Last Update: 1900-01-01