对云贵高原的降水量序列进行多尺度分析，可为喀斯特地区流域的综合治理、水资源的合理利用及地区的可持续发展提供保证。利用互补集合经验模态分解算法（complete ensemble empirical mode decomposition, CEEMD）结合快速傅里叶变换分析（fast Fourier transform, FFT）的特征提取方法。选取云贵高原60个站点20年、时间间隔为12 h的降水量观测数据进行周期识别。结果表明：CEEMD算法能够有效提取序列中的月、半周年、周年等特征分量，结合各气象站点位置，发现地理位置上临近站点的周期变化趋势数相似。与云贵高原的柯本-盖革气候分类指数图相对比，符合地区的气候特征，证明该分解方法可有效应用于气候统计；探讨在不同周期性特征影响下，实际降水序列如何响应于各种气候驱动因素，利用自相关函数（autocorrelation function, ACF）分析降水序列与准两年振荡（quasi biennial oscillation, QBO）、Niño3.4指数、热带气旋之间的时滞和相关性。发现区域的降水序列受厄尔尼诺南方涛动（El Niño-Southern Oscillation, ENSO）以及准两年期振荡的影响。区域48.3%站点固有模态函数（intrinsic mode functions, IMFs）呈现双峰年周期，使用极端降水事件与热带气旋降水事件对比，发现云贵高原降水序列双峰值周期的产生受热带气旋影响。

Multiscale analysis of precipitation series in the Yunnan-Guizhou Plateau is essential for the comprehensive management of karst region watersheds, rational utilization of water resources, and regional sustainable development. The complete ensemble empirical mode decomposition(CEEMD) algorithm combined with fast Fourier transform (FFT) was utilized for feature extraction. Precipitation observation data, collected from 60 stations over 20 years with a 12-hour interval, were used to identify periodicity in the precipitation series of the Yunnan-Guizhou Plateau. The results show that the CEEMD algorithm effectively extracts characteristic components, such as monthly, semi-annual, and annual cycles from the precipitation series. Analyzing the positions of various meteorological stations reveals that stations in close geographical proximity exhibit similar periodic variation trends. When compared with the Köppen-Geiger climate classification map of the Yunnan-Guizhou Plateau, the results align with regional climatic characteristics, validating the applicability of this decomposition method in climate statistics. How actual precipitation series respond to various climatic driving factors under different periodic characteristics was further explored. Using the autocorrelation function(ACF), the lag and correlation was examined between precipitation series and climatic oscillations such as the quasi biennial oscillation(QBO), the Niño3.4 index, and tropical cyclones. It is found that the regional precipitation series are influenced by both El Niño-Southern Oscillation(ENSO) and QBO. Specifically, 48.3% of the station intrinsic mode functions (IMFs) exhibit a bimodal annual cycle. By comparing extreme precipitation events with tropical cyclone precipitation events, it is discovered that the bimodal annual cycle in the Yunnan-Guizhou Plateau's precipitation series is influenced by tropical cyclones.