Here, Zi is the deviation of an attribute for county i from its mean
ðXi XÞ, wij is the spatial weight between feature i and j, a and b
mean change rates of different land use types, n is equal to the total
number of features. The value of I represents the global spatial
association of change rates of two land use types, and the greater
the absolute value means the stronger spatial association. And the
positive value indicates positive spatial association; negative ones
represent negative spatial association.
As illustrated in Table 3, the reduction of arable land and construction
land expansion had the highest Moran's I index, which
indicated the spatial agglomeration of the two main types of land
use change. And the fact that there exists a negative correlation
between the change rate of arable land and the change rates of
orchard land or water area could be interpreted as the arable land
supplement, due to the reduction of orchard land and water area in
Jiangsu. By contrast, changes in the use of construction land could
better explain the arable land change and comprehensive land use
change, which can be interpreted as the construction land expansion
leading to the decrease in arable land, and dominating the
regional land use change. This spatial autocorrelation was gradually
strengthened through the process of economic transition. Thus we
can conclude that the dominant land use changes in Jiangsu are the
decrease in arable land and the expansion of construction land.