Spatial Heterogeneity of Soil Moisture in Red Beds Area and the Associated Driving Factors

Abstract Soil moisture is one of key factors which influence crop growth and ecological environment construction. Knowledge on the spatial heterogeneity of soil moisture and the associate driving factors is of great importance for water resources allocation and ecological environment management. The present study investigated the spatial heterogeneity of soil moisture under four land-use types (cropland, forestland, grassland and bare land) in the ecologically degraded of a red beds area which located in the Nanxiong basin, China. Geostatistical and classic statistical methods were used to analyze the spatial variance of soil moisture. In addition, the driving factors which affect the spatial variance were discussed. The geostatistical analysis showed that the Gaussian model is the optimal regression model of semi-variance functions for all these four land-use types, among which the model of cropland and forestland had better fitting results comparing with other land-use types. Moreover, the soil moisture in bare land exhibited intermediate spatial autocorrelation, which was affected by structural and random factors. Using a cross validation method, we found that the precision of the ordinary kriging interpolation (OK) method was better than the inverse distance weighting interpolation (IDW) method, and the OK method more accurately reflected the spatial distribution of soil moisture in all these four land-use types. Furthermore, the spatial heterogeneity of soil moisture was mainly affected by the natural and anthropogenic factors such as structural characteristics of the soil, soil texture, vegetation covered on the surface of the land, local climate, land-use types, and human intervention, among which human intervention is the most important factor.