Spatial Heterogeneity of Soil Nutrients in a Small Catchment of the Loess Plateau
黄土高原小流域土壤养分的空间异质性

An understanding of the distribution of soil nutrients at the field and catchment scale is important for improving agricultural management and for assessing the effects of agriculture on environmental quality and the influence degree of random factors such as human disturbance on soil properties. However, soil nutrients are highly heterogeneous whether large scale or small scale, and their heterogeneity results from many processes acting and interacting across a continuum of spatial and temporal scales. Due to serious soil erosion and nutrient loss in hilly area of the Loess Plateau, some researchers have studied the mechanism of soil nutrient loss and how to control soil nutrient loss, and others have explored the variability of soil nutrient described by classical statistical methods. However, there is little information on spatial variability and scales of soil nutrients in this area. With the wide applications of GIS and the development of soil science and landscape ecology, the relationships between spatial heterogeneity and ecological processes have aroused more and more attention in recent years. Geostatistical methods originally deriving for the mining industry have extended to assess spatial and temporal heterogeneity in soil science, ecology and water resources because of its advantages in analyzing the variability at a wide range of scales. Based on spatial analysis function of GIS, spatial heterogeneity of soil nutrients in a small catchment of the Loess Plateau was studied using semivariogram of geostatistics in this paper. Theoretical semivariogram models of soil organic matter, total N, available N, total P, and available P were spherical model with a sill. The ratio of random variance (nugget) to total variance (sill) was low. These values were 13.333% for soil organic matter, 10.938% for total N, 22.000% for available N, 9.091% for total P and 27.536% for available P, respectively. This indicated that the five soil nutrients had relatively strong spatial correlation patterns. However, the significant differences in ranges of the five soil nutrients were found. Total P and available P had the highest values in the same ranges of 160m. Soil organic matter in range of 120m followed them, while total N and available N were the lowest values, 90m and 110m. These differences in ranges can be explained by particular properties of the five soil nutrients, soil erosion, land use patterns and topography. The information obtained from this study may be helpful for improving the sampling design of soil nutrients and for interpolation and map of soil nutrients.