2.3.2. Geostatistical Analysis

Geostatistical methods have been widely employed to describe the spatial variability of various soil physicochemical properties [61,62]. Previous studies have confirmed that the kriging interpolation method should be preferentially considered to predict the values of soil physicochemical properties at unsampled locations [23,48,63]. In this study, the ordinary kriging (OK) was adopted as the interpolation method to present the spatial distribution characteristics of the K factor and Ks, the optimal input parameters for which were obtained by the application of a semivariogram. The formula of the semivariogram is shown as follows:

where γ(h) refers to the semivariance for the given lag distance h, N(h) represents the total data pair number between sample sites separated by h, and F(x_i) is the value of the variable F at x_i.