基于景观指数聚类的粤北典型区景观生态区划

Landscape ecological cluster analysis of typical area in north Guangdong

通过选取不同层次的景观指数,对粤北典型区域的景观生态子区进行分区研究,探讨了景观指数在景观生态区划方面的应用,为景观生态区划提供了一种新的思路和方法。采用2009年广东省南雄市DOM数字正射影像为数据源,应用ENVI 4.3和ARCGIS 9.3对遥感影像进行解译和分类,把研究区景观类型分为耕地、林地、疏林地、城镇建设用地、水体、未利用地6大类;根据镇级行政界线把研究区分为18个景观生态子区;选取不同层次的景观指数,应用Fragstats 3.3计算研究区的景观指数;以经过筛选后的景观指数为变量因子,18个景观生态子区为聚类对象,在SPSS 17.0中选择Q型聚类欧式距离平方法进行聚类分析,在聚类图谱中选择水平距离7.5为标准线,对景观子区进行分类,把18个景观生态子区分为4个区域。聚类结果显示,各分区内子区的景观结构特征高度一致。把景观指数应用到景观生态区划,通过聚类分析,分区结果能够较好的反映景观生态结构的显著特征与空间分异状况。

landscape ecological regionalization play an important role in landscape ecological planning,but the apply of landscape index is still rare to see in landscape ecological regionalization.By selected different levels of landscape index,this paper make a zoning for Landscape Ecological sub-areas in typical area of Northern Guangdong,which aimed to discuss the apply of landscape index in landscape ecological regionalization and provide a new concept and method for Landscape Ecological Regionalization.this paper take 2009 Digital Orthophoto map（DOM） of Nanxiong,Guangdong Province as a data source.First of all,RS images were interpreted and classified through ENVI 4.3 and ARCGIS 9.3,then the landscape type are divided into 6 classes include：Farmland,Woodland,open woodland,urban construction land,water,unused land;secondly,the study area are divided into 18 sub-areas by Township administrative boundaries;thirdly,different of levels landscape indexes are calculated by Fragstats 3.3;Finally,uses choose landscape index as variable factors,18 landscape ecology sub-areas as clustering object,select the SPSS 17.0 Q Cluster Analysis method calculate the landscape indexes of study area.After these works,The area divided into 4 partitions by select the horizontal distance of 7.5 as the standard line.The result shows that the landscape structures are highly consistent in each partition.Through the research above all,the author argued that the apply of landscape metrics in landscape ecological regionalization is better reflect the prominent landscape features and spatial variation conditions of landscape ecological structure through clustering analysis,it provided a new zoning method and ideas for landscape ecology regionalization.