3.2. Spatial Autocorrelation Analyses

Spatial autocorrelation analysis is a commonly used method of spatial statistics and can quantitatively describe the spatial distribution characteristics of statistics [67]. Xue et al. (2019) and Hu et al. (2018) revealed the distribution patterns of regional landscape ecological risk and ecological security, respectively, by using spatial autocorrelation analysis [15,68]. In this study, the global Moran’s I index and local indicators of spatial association (LISA) were employed to analyse the spatial statistics of the ERI.

The global Moran’s I provides an overall measure of spatial autocorrelation, with Moran’s I ranging from +1 to −1 [69]. Moran’s I values >0 would indicate positive spatial autocorrelation of the ERI; values closer to +1, indicate strong spatial autocorrelation of the ERI. Moran’s I = 0 indicates no spatial autocorrelation, and the ERI exhibits a random spatial distribution; Moran’s I values <0 would indicate negative spatial autocorrelation of the ERI. LISA represents a set of local statistical methods that can measure the spatial autocorrelation of each observed object in a spatially adjacent area to reveal the spatial clustering pattern of the research object. The spatial clustering can be defined using the local Moran’s I index [67]. A positive LISA indicates that the object value is similar to the neighbouring values and exhibits a high-high cluster (H-H, high value in a high-value neighbourhood) or low-low cluster (L-L, low-value in a low-value neighbourhood) spatial clustering mode. A negative LISA indicates potential spatial outliers of target values, including high and low outliers (H-L, high values in low-value neighbourhoods) and low-high outliers (L-H, low values in high-value neighbourhoods), that are significantly different from the values in the surrounding area [67].