中国降水混沌识别及空间聚类

Identification and Spatial Clustering for Chaos of Precipitation of China

0-1混沌测试方法是一种新的可直接作用于时间序列的混沌识别方法,不需要进行相空间重构,通过对Chebyshev映射的检验验证了有效性。运用0-1方法对中国640个气象站1960—2011年月降水序列进行混沌识别,并运用基于密度的顾及非空间属性的空间聚类方法(DBSC)对计算结果进行空间聚类。结果表明,各气象站月降水序列均表现出明显的混沌特性,且K值的空间分布具有明显的分区特征:从大尺度上看,全国可分为西北高值区、南方次高区、华北—东北中值区和青藏低值区;从小尺度上看,全国分出了29个空间簇。中国降水混沌空间聚类结果不仅与大尺度的气候类型分布相吻合,而且也反映了局部降水动力系统变化特征,这就为降水系统的时空规律研究提供了一条新的途径和方法。

The 0-1test is a new binary test to decide the regular or chaotic motions,which is applied directly to time series data and does not require phase space reconstruction.It is proved to be an effective test after test for chaos of Chebyshev map.The 0-1test was carried out on monthly precipitation time series from 1958to 2011of 640meteorological stations of China,and the density-based spatial clustering algorithm considering both spatial proximity and attribute similarity（DBSC）was used to perform spatial clustering.It is showed that monthly precipitation time series of all stations are chaotic,further more, the spatial distribution of Kshows significant characteristic with spatial variability.In the large-scale,the country can be divided into four sub-regions,that are northwest with the highest Kvalue,southern with the second highest,north and northeast with median and Qingzang district with the lowest.In the small-scale,29spatial clusters are found considering both spatial location and chaos degrees.The chaotic spatial clustering results not only coincide with the large-scale distributions of climate types,but also reflect the local variation of precipitation systems,which provides a new approach and methodology to study the temporal and spatial variation of precipitation systems.