省域尺度下不同时序景观指数集与粒度效应分析

ANALYSIS OF LANDSCAPE INDICE SET AND GRAIN EFFECT OF DIFFERENT TIME SERIES AT PROVINCIAL SCALE

关键词:  景观指数粒度效应代表性指数土地覆被安徽

Key words:  landscape indices  grain effect  representative indices  land cover  Anhui

摘要:

［目的］景观格局变化是景观生态学中的热点问题。研究省域尺度下不同时序景观指数的粒度效应、选取计算景观指数所需的适宜粒度范围并筛选出代表性景观指数集对分析景观格局变化具有重要意义。［方法］以安徽省土地覆被为研究对象，以2000年、2005年、2010年的Landsat TM遥感影像为主要数据源，采用土地利用转移矩阵对安徽省土地覆被的转移数量与方向进行研究；通过设置不同的栅格大小研究景观指数的粒度效应并选取合适的粒度范围；基于相关性分析、因子分析与敏感性分析，筛选出代表性景观指数集。［结果］（1）2000—2010年安徽省的土地覆被始终以耕地为主，其次是林地与人工表面；人工表面的增加是建立在耕地面积减少基础之上的，在10年间共有超过3 200km2的耕地转化为人工表面；（2）景观指数存在粒度效应，粒度变化会对景观指数的计算结果产生影响；省域尺度下计算景观指数所需的合适粒度范围为100~125m；（3）研究安徽省景观格局变化的代表性景观指数有4个，分别为最大斑块指数、景观形状指数、平均斑块面积、香农多样性指数。［结论］计算安徽省景观指数的适宜粒度范围为100~125m，安徽省的代表性景观指数集为最大斑块指数、景观形状指数、平均斑块面积、香农多样性指数，为进一步研究景观格局变化及驱动机制提供了依据。

Abstract:

The evolution of landscape pattern is one of the most hot issues in landscape ecology. It is of great significance to study the grain effect of landscape indices in different time series at provincial scale, select the appropriate grain extents for calculating landscape indices and select representative landscape indices set for analyzing landscape pattern changes. Taking the land cover in Anhui province as experimental object and the Landsat TM remote sensing images in 2000, 2005 and 2010 as main data sources, this study took the method of land use transfer matrix to intuitively analyze the quantity and direction change of land cover in Anhui province. Moreover, in order to study the grain size effect of landscape indices and determine the suitable grain extent, the grain size effect curves of landscape indices were depicted by setting different grid sizes. In addition, it used the correlation analysis, factor analysis and sensitivity analysis to select the representative minimum landscape indices set. The results were showed as follows. Firstly, the dominant landscape sequence of land cover in Anhui province had not changed. It was always dominated by cultivated land, followed by forest and artificial surface from 2000 to 2010. The increase of artificial surface was based on the decrease of cultivated land. The main change of land cover type was the conversion of cultivated land into artificial surface. From 2000 to 2010, there was more than 3 200 km2 of cultivated land converted into artificial surface. Secondly, landscape indices had obvious scale dependence. Some landscape indices had obvious grain effect, and the change of grain can affect the calculation results of landscape indices, which would affect the precise analysis of landscape pattern changes. According to the first scale domain, the grain extent for calculating the landscape indices from 100 to 125m was more suitable than the other extents at provincial scale. Finally, the representative landscape indices set reflecting landscape pattern changes from 2000 to 2010 in Anhui province contained four landscape indices, which were largest patch index, landscape shape index, mean patch area and Shannon′s diversity index. The grain extent from 100 to 125m is more suitable than the other extents at Anhui provincial scale, and the representative landscape indices are largest patch index, landscape shape index, mean patch area and Shannon′s diversity index. The conclusions of this study can provide a basis for further research on landscape pattern changes and driving mechanism.