Study On Spatial Variability Expression Of Soil Fertility Factors In Hills In Medium And Large Scales

According to the sampling data in the Lianshui Valley, Xingguo County, Jiangxi Province in 2002, comparing to the influences of the sampling methods including random, grid, stratified, synthetical optimization and the synthetical sampling on the spatial variability of soil organic matter (SOM), total phosphorus (STP), available phosphorus(SAP), available potassium (SAK), bulk density(BD) by GPS, GIS and geostatistics, the better sampling methods which adapted to predict the spatial variability of soil fertility in given condition was obtained. Besides, the paper compared to interpolation methods including the inverse distance weighting(IDW), radial basis functions(RBF), ordinary kriging(OK), global polynomial(GPI), local polynomial(LPI) in the condition of the better sampling methods in order to obtain the optimal interpolation method for soil fertility in different scale. The main results were as follows.In the different scales, random, grid, stratified, synthetical optimization and the synthetical sampling methods made the difference in expressing spatial variability of soil fertility in hills. Based on the average, prediction error, spatial pattern and area of grade for soil fertility factors, the result indicated that no one approaches synthetical sampling method fully except for some aspects, but the relative better sampling methods of soil fertility factors were obtained in different scale synthesizing different indexes. In medium scale, grid sampling method was the best sampling method; random sampling method was superior to synthetical optimization and stratified sampling methods and the better sampling method to adapt to predict soil fertility factors. In large scale, synthetical optimization and grid sampling methods were closer to synthesis sampling method than the others because that could reflect the influence of the difference between vegetation, relief, parent material, land use and distribution of soil types on spatial variability of soil properties. Besides, with the expansion of area, random sampling method was getting worse to predict spatial variability of soil fertility factors. The results of comparing the several sampling methods in medium and large scales showed that grid sampling method was the better sampling method in medium and large scale.