The Comparative Analysis Between2D And3D Landscape Partern Indices Based On TM Images

As an important element of environmental change, LUCC is also causing ecosystem changes. It is important for defining ecological processes that accurate determinating it. Landscape indices is an important means of land cover changes in quantitative research. But the traditional two-dimensional landscape indices use two-dimensional remote sensing images as a data source, ignoring the effects of topography on landscape pattern. This has resulted in lack of ecological significance of the three-dimensional structure. The height difference of Jiuzhaigou National Nature Reserve is2768m,it can be a typical representatives. Reserving Jiuzhaigou National Nature Reserve’s landscape pattern is important to comprehensive understanding of the status features of Nature Reserve and reasonable using its landscape resources. In this paper, select Jiuzhaigou Nature Reserve as a case.Using1997,2010TM image as the main data source, combining the DEM data, comparative analysis of the difference between the traditional2D landscape indices and3D landscape indices in land cover change during the measurement. According to the research needs, selecting21indexs from the patch level, class level, landscape level to comparative analysing. Obtaining the following conclusions:Comparing the3D landscape pattern analysis and2Dlandscape pattern analysis in Jiuzhaigou during1997and2010has concluded that:(1)Advantage of a single landscape types in the study area decreased, different landscape types dominance was tending to balance; Landscape discrete levels is declining, the overall complexity of patch shape slight is increasing fragmentation declined, unicom degree rised; landscape pattern general stability. (2) The patch levels, this study selected PA (plaque area), PP (plate perimeter), SI (shape index) values of3D and2Dare significantly differently; Some values of SI (shape index) and FD (fractal dimension index) are1in2D landscape pattern index, which does not match with the fact.So the3D values are more realistic. Overall, In the level patches of2D and3D landscape indices were significantly different.(3) The type level, CA (class area), AREA-MN (mean patch area), AREA_CV (average coefficient of variation patch), LPI (largest patch index), ED (edge density)have significant difference between2D and3D landscape indices; FRAC-MN (fractal dimension), LSI (landscape shape index) difference are not obvious; AI (polymerization degree index) has no difference. The type level, compare with2D,3D landscape indices show the overall area of the study area increased, the average size of patch increased, the degree of dispersion narrow patch shape tends to be complex.(4)The landscape level, AREA-MN (mean patch area), AREA_CV (average coefficient of variation patch), LSI (landscape shape index), ED (edge density) have significant difference between2D and3D. LPI (the largest patch index), FRAC-MN (fractal dimension) difference are not obvious; PRD and SHDI slightly decreased, which is related to the overall size increases,the difference was not significant. CONTAG (spread index) there is no difference; The landscape level, compared with2D,3D landscape indices showing the study area average patch area increased, the degree of dispersion increased; patch shape tends to be complex.Overall, the traditional2D landscape indices in the quantitative analysis of alpine gorge area of land cover change is not accurate enough in the measuring of the underlying index of area and perimeterto, The3D landscape indices are relatively accurate.