In order to investigate the dynamic evolution characteristics of the normalized difference vegetation index(NDVI)time series from the perspective of topological mechanics,and to ascertain the connections and differences between different provinces(regions),the NDVI data of nine provinces(regions)in the Yellow River Basin from 2001 to 2022 are utilized to construct an NDVI network based on the visibility graph algorithm of the complex network.The networks are used to analyse the dynamic characteristics of NDVI.K-means and Spectral Clustering based on the visibility graph are used to cluster the nine provinces(regions).The results indicate that Henan is a"scale-free"network and is less susceptible to sudden events compared to the other provinces.However,its NDVI is more susceptible to recent changes in the vegetation cover.Additionally,the NDVI exhibits a certain trend and correlation over time,and different seasons may result in varying patterns of vegetation cover.The NDVI of the nine provinces(regions)exhibits the characteristics of a"small world".The nine provinces(regions)are divided into three groups.Spectral clustering based on the visibility graph is more effective in studying the fluctuations of NDVI in these provinces(regions).
关键词
复杂网络/可视图/归一化植被指数/黄河流域/聚类/时间序列
Key words
complex network/visibility graph/normalized difference vegetation index/Yellow River Basin/clustering/time series
维普
万方数据