基于集合经验模态分解北疆降水多尺度变化特征

降水是中国西北干旱区水资源的重要组成部分,利用合理方法有效认识降水的区域变化规律对指导农业发展尤为重要。基于北疆42个气象站1961—2012年降水序列数据,从气候时间序列中提取气候信号中各个尺度的变化,对北疆过去52a来的降水进行多尺度分析,并对其空间差异进行了初步探讨。结果表明:过去52a北疆降水量存在明显的年际和年代尺度的变化,年际尺度的周期为5a和8a,年代尺度的周期为10a和30a。变化趋势上,降水量1985s之前呈减少趋势,在1985s之后呈增加趋势,且后期的增幅大于前期的减幅。突变分析表明北疆年降水量呈增加趋势。集合经验模态分解(EEMD)是一种适用于非线性、非平稳序列的信号分析方法,将EEMD应用于气候要素时间序列,可提取可靠真实的气候变化信号,同时,EEMD可以得到气候变化的固有时间尺度。

Precipitation Multi-Scale Characteristics by Ensemble Empirical Mode Decomposition in Northern Xinjiang

Multi-scale characteristics of precipitation were analyzed by ensemble empirical mode decomposition (EEMD) by using monthly precipitation data of forty-two meteorological stations from 1961 to 2012.The results showed that rainfall experienced significant multi-scale characteristics,i.e.,variability with 5 years and 8 years periodic patterns for inter-annual scale of 10-year and 30-year periodic patterns for inter-decadal scale.Precipitation tended to decline before the 1985s and increased after that.Moreover,increasing amplitude of precipitation in the later period was greater than the decreasing amplitude in the earlier period.Multiscale analysis indicated that annual precipitation presented the increasing trend in north Xinjiang.Ensemble empirical mode decomposition (EEMD) method is suitable for non-linear and non-stationary signal analysis.This method can work on nature signals (non-linear and nonstationary signals) and reduce the speckle noise.