基于时间序列与小波分析耦合模型区域降雨量预测研究  

Study on regional rainfall forecast based of time series and wavelet analysis coupled model

作　　者：徐淑琴[1] 雷兴元 刘宇佳[1] 苏鑫[1] 王莉莉[1] 路豪杰 

机构地区：[1]东北农业大学水利与建筑学院,哈尔滨150030

出　　处：《东北农业大学学报》2015年第11期63-69,共7页Journal of Northeast Agricultural University

基　　金：黑龙江省教育厅科研项目(11551044);东北农业大学农业水土资源创新团队

摘　　要：为提高区降雨量预测精度,提出小波分析与时间序列组合模型。以黑龙江省虎林市858农场区为例,采用小波变换A Trous算法将小波信号分解,对分解后各小波变换序列互相关分析,随后运用时间序列分析分别对该农场周期项W 1（t）、随机项W2（t）、趋势项C2（t）模型求解,采用小波重构算法得到组合模型。运用该模型模拟该区1999~2014年降雨量,结果与实际测量值接近。预测未来3年短期降雨量,为指导区域农业生产及管理提供科学依据。A combination model of wavelet analysis and time series was proposed to improve the accuracy of regional rainfall prediction. The 858 farm of Hulin City in Heilongjiang Province was selected as an example, wavelet signal was decomposed by wavelet transformed A Trous algorithm, and crosscorrelation analysis of decomposed wavelet transform sequences was performed, periodic term W1（t）,random item W2（t） and trend term C2（t） of the model worked out by time series analysis, the combination model was obtained by wavelet reconstruction algorithm. The model simulated rainfall from 1999 to 2014 of this region, the results showed that it was close to actual measured values, and the model also forecasted the rainfall in the next three years to provide scientific basis for the guidance of regional agricultural production and management.

关 键 词：小波分析 时间序列 A Trous算法 互相关分析 降雨量 

分 类 号：TV125[水利工程—水文学及水资源]