汛期雨量长期预测建模分析

南京大学学报(自然科学版) ›› 2017, Vol. 53 ›› Issue (5): 887–.

汛期雨量长期预测建模分析

周雨婷1，陈元芳2，王　栋1*，邹　鹰3，贺瑞敏4

出版日期:2017-09-25 发布日期:2017-09-25
作者简介: 1.南京大学地球科学与工程学院，南京，210023；2.河海大学水文水资源学院，南京，210098；3.南京水利科学研究院，南京，210029；4.南京水利科学研究院水文水资源与水利工程科学国家重点实验室，南京，210029
基金资助:
基金项目：国家重点研发计划(2016YFC0401501)
收稿日期：2017－01－16
*通讯联系人，E－mail：wangdong@nju.edu.cn

Model establishment for the analysis of long－term forecast of flood season precipitation

Zhou Yuting1，Chen Yuanfang2，Wang Dong1*，Zou Ying3，He Ruimin4

Online:2017-09-25 Published:2017-09-25
About author:1.School of Earth Sciences and Engineering，Nanjing University，Nanjing，210023，China；
2．College of Hydrology and Water Resources，Hohai University，Nanjing，210098，China；
3．Nanjing Hydraulic Research Institute，Nanjing，210029，China；
4．State Key Laboratory of Hydrology－Water Resources and Hydraulic Engineering，Nanjing Hydraulic
Research Institute，Nanjing，210029，China

摘要/Abstract

摘要： 通过趋势分析、跳跃分析和周期分析对上海徐家汇站汛期雨量时间序列进行组成成分分析．在此基础之上，为了探究最适用于上海徐家汇站汛期雨量的长期预测模型，选择改进的周期均值叠加模型，自回归模型AR(p)，门限自回归模型TAR，灰色模型GM(1，1)，误差反向传播(BP)神经网络和径向基函数(RBF)神经网络建立长期预测模型．采用平均相对误差、均方根误差、确定性系数、合格率和特大值预测合格率这5项评价指标对比分析各模型预测效果．研究结果表明，徐家汇站汛期雨量时间序列的周期成分明显，不存在趋势和跳跃成分；综合5项指标，TAR模型对汛期雨量一般值的预测效果最优，对特大值的预测需要同时参考AR(p)模型和RBF神经网络．

Abstract: The study has selected Xujiahui Station in Shanghai as a case of the flood season precipitation series to establish the most suitable model for the long term forecast in the area.Firstly，the component analysis is carried out to separate deterministic components，including the trend，the mutation and the periodic component.Then，the article takes time series analysis，gray theory and artificial neural network into consideration.So，different kinds of models，covering the improved periodic mean superposition model，auto regressive model，threshold auto regressive model，gray model(1，1)，back－propagation(BP) neural network and radial basis function(RBF) neural network，are built for the forecast of the series.In the end，the performance of different models will be analyzed through the evaluation indexes such as mean absolute percentage error(MAPE)，root mean square errors(RMSE)，deterministic coefficient(Dc)，qualified rate(T)and qualified rate of super large value prediction(S)．The result of component analysis shows that the periodic component of flood season precipitation series of Xujiahui Station is obvious，but the trend and mutation are not pronounced.The analysis of evaluation indexes illustrates that in the long term forecast of flood season precipitation series of Xujiahui Station，the threshold auto regressive model and two kinds of artificial neural network are better than the others for the general value prediction，and the former is the most suitable.But for the super large value prediction，the auto regressive model and RBF neural network perform best.Thus，the most suitable model for general value forecast is threshold auto regressive model，and the super large value forecast needs to combine the auto regressive model and RBF neural network.Long term forecast plays an important role in the different types of hydraulic engineering and urban flood defence.On the basis of the fact，this study provides important reference value for the flood prevention in Shanghai.

周雨婷1，陈元芳2，王　栋1*，邹　鹰3，贺瑞敏4. 汛期雨量长期预测建模分析[J]. 南京大学学报(自然科学版), 2017, 53(5): 887–.

Zhou Yuting1，Chen Yuanfang2，Wang Dong1*，Zou Ying3，He Ruimin4. Model establishment for the analysis of long－term forecast of flood season precipitation[J]. Journal of Nanjing University(Natural Sciences), 2017, 53(5): 887–.

参考文献

[1]　桑燕芳，王中根，刘昌明．水文时间序列分析方法研究进展．地理科学进展，2013，32(1)：20－30.(Sang Y F，Wang Z G，Liu C M．Research progress on the time series analysis methods in hydrology.Progress in Geography，2013，32(1)：20－30.)
[2] Yule G U．On a method of investigating periodicities in disturbed series，with special reference to Wofer’s sunspot numbers.Philosophical Transactions of the Royal Society of London，1927，226(226)：267－298.
[3] Walker G．On periodicity in series of related terms.Proceedings of the Royal Society of London，1931，131(818)：518－532.
[4] 陈守煜．从研究汛期描述论水文系统模糊集分析的方法论．水科学进展，1995，6(2)：133－138.(Chen S Y．Methodology of fuzzy sets analysis to hydrologic system from research on flood period description.Advances in Water Science，1995，6(2)：133－138.)
[5] 傅　军，丁　晶，邓育仁．洪水混沌特性初步研究．水科学进展，1996，7(3)：226－230.(Fu J，Ding J，Deng Y R．Preliminary study on the chaotic behavior of flood flows.Advances in Water Science，1996，7(3)：226－230.)
[6] Singh V P．The use of entropy in hydrology and water resources.Hydrological Processes，1997，11(6)：587－626.
[7] Kumar P，Foufoula－Georgiou E．Wavelet analysis for geophysical applications.Reviews of Geophysics，1997，35(4)：385－412.
[8] 邓聚龙．灰理论基础．武汉：华中科技大学出版社，2002，361－392.(Deng J L．Theoretic basis of grey system.Wuhan：Huazhong University of Science and Technology Press，2002，361－392.)
[9] 覃光华，丁　晶，李眉眉等．敏感型人工神经网络及其在水文预报中的应用．水科学进展，2003，14(2)：163－166.(Qin G H，Ding J，Li M M，et al.Application of an ANNs with sensitive ability to hydrologic forecast.Advances in Water Science，2003，14(2)：163－166.)
[10] 马富明，陈元芳，林元润．晋江流域年径流统计特性分析与预测．水　文，2006，26(1)：87－88，54.(Ma F M，Chen Y F，Lin Y R．Statistic characteristics analysis and prediction for the annual runoff of the Jinjiang River in Fujian Province.Journal of China Hydrology，2006，26(1)：87－88，54.)
[11] 丁　晶，刘权授．随机水文学．北京：中国水利水电出版社，1997，17－27.(Ding J，Liu Q S．Stochastic hydrology.Beijing：China Water and Power Press，1997，17－27.)
[12] 范钟秀．中长期水文预报．南京：河海大学出版社，1999，6－87.(Fan Z X．Mid and long term hydrologic forecasting.Nanjing：Hohai University Press，1999，6－87.)
[13] 谭巧巧，杨吉云．改进的自回归AR(p)预测模型在WSN中的应用．计算机工程与应用，2015，51(12)：83－87.(Tan Q Q，Yang J Y．Application of improved AR(p)prediction model in WSN.Computer Engineering and Applications，2015，51(12)：83－87.)
[14] 郝永红，黄登宇，张文忠等．山西神头泉流量的灰色预测模型研究．水利学报，2004，35(2)：111－114.(Hao Y H，Huang D Y，Zhang W Z，et al.Gray prediction model for forecasting spring discharge.Journal of Hydraulic Engineering，2004，35(2)：111－114.)
[15] 罗　党，刘思锋，党耀国．灰色模型GM(1，1)优化．中国工程科学，2003，5(8)：50－53.(Luo D，Liu S F，Dang Y G．The optimization of grey model GM(1，1)．Engineering Science，2003，5(8)：50－53.)
[16] 朱星明，卢长娜，王如云等．基于人工神经网络的洪水水位预报模型．水利学报，2005，36(7)：806－811.(Zhu X M，Lu C N，Wang R Y，et al.Artificial neural network model for flood water level forecasting.Journal of Hydraulic Engineering，2005，36(7)：806－811.)
[17] 杨晓华，杨志峰，沈珍瑶等．水环境非线性时序预测的高精度RBF网络模型．水科学进展，2005，16(6)：788－791.(Yang X H，Yang Z F，Shen Z Y，et al.High precision radial basis function network model for predicting water environment nonlinear time series.Advances in Water Science，2005，16(6)：788－791.)
[18] 中化人民共和国国家标准．水文情报预报规范．GB/T 22482－2008.(State Standard of the People’s Republic of China.Standard for hydrological information and hydrological forecasting.GB/T 22482－2008.)

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed