基于离散小波变换的水稻生育期降水量多时间尺度特征分析

Multiple Time Scale Analysis of Rice Growing Period Precipitation Based on Discrete Wavelet Transform

为深入分析水稻生育期降水多层次变化特征,运用小波分析理论,采用morlet小波函数对降水量序列进行离散小波变换,分析其多时间尺度特性,揭示降水量在不同时间尺度下的小波变换时频特征、丰枯交替变化的周期规律以及未来的发展趋势,为制定科学合理的灌溉制度提供一定的参考。研究结果表明,不同的时间尺度分辨率下,水稻生育期降水量序列会表现出不同的周期交替现象,主要存在5 a、13 a和32 a的周期。

In order to study the multi-level characteristics of rice growing precipitation, the wavelet analysis theory is adapted. This paper carries out discrete wavelet transform through morlet wavelet function to analyze the multiple time scale characteristics of precipitation, and then reveals the wavelet transform time- frequency features of precipitation at different time scales and interchanged periodic law between the rain- fall period and the rainless period. The results show that the precipitation series of rice growing period have different cycle variation under different time-scale, what＇s more, it mainly exists the cycle of 5-year, 13-year and 32-year.