Soil Characteristics Spatial Prediction Of Farmland Soil In Aba Prefecture

The spatial prediction of soil characteristics is not only the basic of scientific soil management, but also the basic of realizing the agricultural sustainable development. Meanwhile, the reasonable sample number is the guarantee of spatial prediction accuracy. Therefore, choosing the spatial prediction methods of farmland soil characteristics accurately and determining the sampling number reasonably have a great significance to the quantitative study on soil characteristics spatial differentiation.This study took use of geostatistics and GIS technology. Firstly, discussing the spatial prediction methods of farmland soil characteristics and simulating the spatial distribution of the farmland soil characteristics in Aba prefecture, Sichuan province. Meanwhile, the study reduced and optimized the sampling sites in the study area by using ordinary kriging and co-kriging interpolated methods. Secondly, the study analyzed the main factors which influenced the spatial distribution of soil characteristics. The main results were as follows:(1) In the farmland soil of study area, when the spatial variability of soil properties in isotropic conditions, the AN had a moderate spatial correlation. The pH and AK had a moderate to strong spatial correlation. The OM, TN and AP had a strong spatial correlation. It showed both of the structural factors and random factors influenced the soil characteristics at the same time. When the spatial variability of soil properties in anisotropic conditions, the spatial correlation is significantly stronger than in isotropic conditions. It indicated the influence of structural factors on the spatial variability of soil properties had an obvious direction.(2) From the spatial distribution pattern of several soil nutrients, it presented the high concentration mainly located the middle part, the northeast and the west of the study area. In this study, it compared the interpolation methods of the spatial distribution simulation of the soil characteristics. From the prediction results, the most appropriate method of pH was the IDW. At the same time, the best prediction method of the AP and AK content was the OK. Meanwhile, the most ideal method of OM, TN and AN was the CK.(3) In condition of single variable and interaction variable, the study analyzed the spatial variabilities of TN and AN content. When the number of sampling sites reduced, the prediction accuracy of OK and CK increased gradually. The study concluded that, when the prediction sample number decreased to 60% even 50% of the initial sample, the prediction accuracy of TN and AN content could still increased by using CK. In the comparative analysis, the ideal sampling interval was 1185 to 1423 sampling points in the study area.(4) Both of the soil parent, elevation and slope made the soil characteristics show different values, illustrated these factors had a significant influence on soil characteristics. Among the social and economic factors, the basic conditions and the structure of agricultural economy had the most obvious impact on spatial distribution of regional soil characteristics.