Figure 8 Vegetation diversity indices at two scales: Margalef, Pielou, Simpson and Shannon-Wiener indices.
The spatial distribution of diversity indices shows a gradual decrease from east to west (Figure 8), influenced by species composition. The eastern region with mixed shrub-grassland has high diversity values, the middle region dominated by trees and shrubs has moderate values, and the western region with shrubs like Artemisia and Salix has low values. Salix possibly has a strong ability to occupy resources, which suppresses the growth of other species. Additionally, its large canopy limits the amount of sunlight reaching vegetation underneath its branches and leaves, resulting in little nearby vegetation (Alvarez et al., 2011; Pierce et al., 2019).

3.2.3 Spatial Autocorrelation Analysis

Spatial autocorrelation analysis was conducted to examine the distribution of vegetation diversity and to determine the presence of clustering. The analysis showed that global Moran’s I index for each diversity index was greater than 0, indicating clustering. Shannon index had the highest global Moran’s I value among all indices. For example, at two scales, the Moran scatter plots were mainly distributed in the first and third quadrants with global Moran’s I indices of 0.469 and 0.463, respectively (Figure 9). Both were statistically significant at the significance level of α=0.05 (z>1.96), indicating strong positive spatial autocorrelation and significant spatial clustering rather than random distribution in space. High-high clustering was mainly located in the east of the study area, corresponding to areas of high diversity. In contrast, low-low clustering was mainly located in the west, corresponding to the low-diversity region. High-low and low-high clustering were sporadically distributed throughout the study area. These results demonstrate the spatial distribution characteristics of clustering and differentiation using the Shannon index.