青海湖流域景观格局空间粒度效应分析

An analysis of landscape pattern spatial grain size effects in Qinghai Lake watershed

流域是生态系统的基本单元和重要空间尺度单元,流域尺度景观格局分析的科学性很大程度上依赖于最佳空间粒度的选取是否准确。选择青海湖流域为研究区,基于卫星遥感数据资料,利用面向对象分类方法,解译获得流域景观数据,通过重采样得到不同空间粒度的流域景观数据,计算得到不同景观格局指数并绘制各指数与空间粒度之间的关系曲线,识别景观格局指数的尺度效应,评估空间粒度增加而引起的景观格局指数信息损失量,确定流域尺度景观格局分析的最佳空间粒度。结果表明:流域景观格局指数随着空间粒度增大变化显著,但规律不同;综合景观指数的空间粒度效应和信息损失量变化特征,流域景观格局分析的最佳空间粒度选择以90 m为最佳。

Watershed is the basic and important spatial scale unit of ecosystem.The scientific analysis of watershed scale landscape pattern depends heavily on the accuracy of the selection of the optimal grain size.In this paper,Qinghai Lake watershed was selected as the study area.Based on the object-oriented classification method,the authors interpreted satellite remote sensing data to generate watershed landscape data.Resampling method was used to obtain different spatial grain size watershed landscape data.Then the landscape pattern indexes were calculated and statistical relation curve was drawn between each landscape pattern index and grain size.Thus the scale effect of landscape pattern index could be identified.Finally,landscape pattern index information loss caused by increasing grain size was used to determine the optimum spatial grain size of watershed scale landscape pattern analysis.The result showed watershed landscape pattern index changed significantly with the increase of spatial grain size,but laws of change were different.Taking into account the landscape pattern spatial grain size effect and the change characteristics of information loss,the authors hold that the best spatial grain size choice of watershed landscape pattern analysis was 90 m.