Abstract: The author used Morlet complex wavelet analysis to analyze the precipitation fluctuation from 1965 to 2011 at Saihanba in Hebei Province and to reveal the multi-time scale feature of precipitation pattern. It shows that the anomolies of monthly precipitaion, growing season precipitation (May to September) and annual precipitation in the last 47 years vary at two time-scales: one is 4 years of inter-annual oscillation, which is the low-frequency part of the inter-annual fluctuation after high-frequency filtering; and the other is about 18 years of decadal oscillation. The growing season precipitation of Saihanba accounts for 83.1 percent of the annual precipitation, thus both of them often show the same trends but with some exceptions. The wavelet coefficient curve under the scale of 18 years indicates that precipitation will keep higher than mean annual precipitation with decreased anomalies in 2012?2015.

Key words: precipitation, multi-time scale, wavelet analysis, Morlet wavelet, Saihanba

摘要： 采用Morlet复小波函数对河北省塞罕坝地区1965?2011年降水变化规律进行了小波分析, 揭示了塞罕坝地区降水变化的多时间尺度特征。结果表明, 47年来该地区月降水量、年降水量与生长季降水量距平值变化主要表现在两种尺度上: 一种是4年左右的年际尺度振荡, 这是经过高频滤波后的年际变化中的低频部分; 另一种是18年左右时间尺度的年代际振荡。由于塞罕坝地区生长季降水平均占年降水量的83.1%, 通常二者的变化趋势相同, 但也不完全一致。由典型尺度小波系数过程线图可预测, 2012?2015年塞罕坝地区的降水量仍然偏多, 但距平值会减少。

关键词: 降水量, 多时间尺度, 小波分析, Morlet小波, 塞罕坝地区