Abstract:

In this paper, taking Tazi Gou forest farm as the research area, taking the two phase Landsat TM images of 1997 and 2009 as the main data source, with the help of ENVI 5.1, ArcGIS 10.3 and Geoda095i, the uniformity theory analysis method is introduced into the spatial correlation analysis of landscape pattern to analyze the global autocorrelation of the study area landscape pattern. The conclusion is compared with the analysis method of Moran's I and General G-statistic. In 1997, the uniformity of landscape pattern is 0.320 7, which is between the lower limit 0.297 2 and the upper limit 0.339 7, the coefficient of Moran's I is -0.11 and the value of Z Score is 0.92, it shows the pattern of the study area is random pattern. In 2009, the uniformity of landscape pattern is 0.288 3, which is less than the lower limit value 0.296 9, the coefficient of Moran's I is 0.42 and the value of Z Score is 2.06, it shows the pattern of the study area is aggregated pattern. The same conclusion is drawn from the three methods, that is, from 1997 to 2009, the pattern of the study area changed from random pattern to aggregated pattern in the 12 years.

Key words: landscape pattern, uniformity theory, Moran's I, General G-statistic, global autocorrelation