ESTABLISHMENT OF PREDICTION MODEL FOR THE PACIFIC SUBTROPICAL HIGH USING NEURAL NETWORK CALCULATION METHOD

Zhang Ren; Jiang Guorong; Yu Zhihao; Jiang Quanrong

Vol.11, NO.4, 2000

Article Contents

Article Navigation > Journal of Applied Meteorological Science > 2000 > 11(4): 474-483

Zhang Ren, Jiang Guorong, Yu Zhihao, et al. Establishment of prediction model for the pacific subtropical high using neural network calculation method. J Appl Meteor Sci, 2000, 11(4): 474-483.

Citation:

Zhang Ren, Jiang Guorong, Yu Zhihao, et al. Establishment of prediction model for the pacific subtropical high using neural network calculation method. J Appl Meteor Sci, 2000, 11(4): 474-483.

Zhang Ren, Jiang Guorong, Yu Zhihao, et al. Establishment of prediction model for the pacific subtropical high using neural network calculation method. J Appl Meteor Sci, 2000, 11(4): 474-483.

Citation:

Zhang Ren, Jiang Guorong, Yu Zhihao, et al. Establishment of prediction model for the pacific subtropical high using neural network calculation method. J Appl Meteor Sci, 2000, 11(4): 474-483.

PDF( 190 KB)

ESTABLISHMENT OF PREDICTION MODEL FOR THE PACIFIC SUBTROPICAL HIGH USING NEURAL NETWORK CALCULATION METHOD

1.
PLA University of Science and Technology, Nanjing 211101
2.
Department of Atmospheric Science, Nanjing University, 210093

Received Date: 1999-08-30
Rev Recd Date: 2000-01-28
Publish Date: 2000-11-30

Abstract

Abstract

Based on the observational facts and correlative study results about the equatorial SST and Subtropical High, a prediction model between West-Pacific Subtropical High and equatorial SST is established by using the BP neural network model and its improved calculation method. By using this model, we can predict Subtropical High's basic changes and trend ahead of 3 months according to the early data of equatorial SST and subtropical high. This prediction model has better coupling precision. The forecasting accuracy and valid time of this model are meaningful in application.
- Neural network;
- Prediction model;
- Pacific SST;
- Subtropical high

FullText(HTML)

References(11)

Figures and Tables

图 1 两层BP网络模型结构, 其中S₁为隐层神经元

DownLoad: Full-Size Img PowerPoint

图 2 1963年4月~1977年12月副高面积指数网络模型拟合效果比较

DownLoad: Full-Size Img PowerPoint

图 3 1978年4月~1982年12月副高面积指数网络模型预报效果比较

(个例1) (实线为实际值, 带点线为预报值)

DownLoad: Full-Size Img PowerPoint

图 4 1986年4月~1990年12月副高面积指数网络模型预报效果比较

(个例2) (实线为实际值, 带点线为预报值)

DownLoad: Full-Size Img PowerPoint

图 5 个例1, 个例2线性回归模式的预报效果比较

(实线为实际值, 带点线为预报值)

DownLoad: Full-Size Img PowerPoint

图 6 1963年4月~1997年12月统计集成模式的拟合效果比较

(上图中实线为实际值, 带点线为拟合值, 下图为拟合误差)

DownLoad: Full-Size Img PowerPoint

图 7 个例1, 个例2副高面积指数统计集成模式的预报效果比较

(实线为实际值, 带点线为预报值)

DownLoad: Full-Size Img PowerPoint

图 8 1992年4月~1995年12月网络模型预报效果比较

(实线为实际值, 带点线为预报值)

DownLoad: Full-Size Img PowerPoint

References

[1]	叶笃正, 曾庆存, 郭裕福.当代气候研究.北京:气象出版社, 1991. 164~177.
[2]	金龙, 陈宁, 林振山.基于人工神经网络的集成预报方法研究和比较.气象学报, 1999, 57(2):198~207. http://www.cnki.com.cn/Article/CJFDTOTAL-QXXB902.007.htm
[3]	Fu C B, Zeng Z M. Ten years experioment on long range prediction of the Northwest Pacific high according to sea surface temperature anomalies. Proceedings of first WMO Conference on long-range forecasting, the practical problems and future prospects. WMO/TD. No.147, 1986. 77~86.
[4]	符淙斌, 叶笃正.热带海洋对副热带高压长期变化的影响.科学通报, 1977, 21:313~317. http://www.cnki.com.cn/Article/CJFDTOTAL-KXTB197707009.htm
[5]	陈烈庭.北太平洋副热带高压与赤道东部海温的相互作用.大气科学, 1982, 6:148~156. http://www.cnki.com.cn/Article/CJFDTOTAL-DQXK198202004.htm
[6]	藏恒范, 王绍武.赤道东太平洋水温对低纬大气环流的影响.海洋学报, 1984, 6(1):16~24. http://www.cnki.com.cn/Article/CJFDTOTAL-SEAC198401002.htm
[7]	龚道溢, 王绍武.南北半球副热带高压对赤道东太平洋海温变化的响应.海洋学报, 1998, 20(5):44~53. http://www.cnki.com.cn/Article/CJFDTOTAL-SEAC805.004.htm
[8]	黄荣辉, 李维京.夏季热带西太平洋上空的热源异常对东亚上空副热带高压的影响及物理机制.大气科学特刊, 1988, 107~116.
[9]	吴国雄, 孟文.赤道印度洋-太平洋地区海气系统的齿轮式耦合和ENSO事件.大气科学, 1998, 22(4), 470~479. http://www.cnki.com.cn/Article/CJFDTOTAL-DQXK804.008.htm
[10]	沈清, 胡德文, 时春. 神经网络应用技术. 北京: 国防科技大学出版社, 1993. 3~33, 77~86.
[11]	施阳, 李俊. MATLAB语言工具箱——TOOLBOX实用指南.西安:西北工业大学出版社, 1998. 124~209.