应用R/S方法分析NDVI时间序列  

Application of R/S Method in Analyzing NDVI Time Series

作　　者：王新明[1] 王长耀[1] 牛铮[1] 

机构地区：[1]中国科学院遥感应用研究所遥感科学国家重点实验室,北京100101

出　　处：《地理与地理信息科学》2005年第5期20-23,48,共5页Geography and Geo-Information Science

基　　金：中国科学院知识创新工程重大项目(KZCX1-SW-01-02);国家重点基础研究发展规划项目(G2000077902)

摘　　要：针对1982—2001年月合成PAL NDVI时间序列,应用R/S方法计算中国陆地范围内的赫斯特指数。结果表明:全国范围内的赫斯特指数普遍大于0.5,意味着NDVI时间序列不是马尔科夫序列,而是存在明显的趋势性;中国陆地范围内的赫斯特指数分布存在“西高东低”的特点,说明西部NDVI时间序列表现出更强的持久性;通过分析赫斯特指数分布与陆地覆盖类型的关系发现,在同一生态地区内,裸地的赫斯特指数最高,与其它覆盖类型差异明显。陆地覆盖类型之外的自然条件和人为因素对赫斯特指数空间分布的影响有待进一步研究。NDVI time series data are valuable in researching various geo - phenomena on the earth surface. In this paper, the R/S method was the first time adopted to calculate the Hurst exponents of NOAA/NASA PAL monthly NDVI time series from 1982 to 2001 on Chinese terrestrial regions. Since 99.5 percent of Hurst exponents of the NDVI time series are greater than 0.5, it is obvious that NDVI time series are generally persistent,and a rising tendency in the past will lead to some increase in the future. Mostly higher Hurst exponents distribute in west China,which means that there is stronger persistence for NDVI time series in the west China than in the east. When land cover types are considered,bating land shows much higher Hurst exponents than other land- cover types. As NDVI is affected by various natural and human factors,the relations between such factors and Hurst exponents of the NDVI time seties are to he quantified in the coming research.

关 键 词：遥感 NDVI 时间序列 R/S方法 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]