Boundary Condition Effect Of Cultivated Soil Characteristics Evaluation On Spatial Interpolation In A Mountain Area

Farmland soil characteristics evaluation is the most important process for the technique of soil formulation to applying fertilizer which is being expanded strongly in our country these years. The current way used to evaluate soil nutrient by interpolating soil characteristics in the range of all research area without considering the different soil types of boundary problem. This research studied the soil samples that divided into two parts randomly, a total of 4189 samples of experimental data points, including:red soil for the 527 sampling points, limestone soil for the 157 sampling points, paddy soil for the 3442 sample points, purple soil for the 9 sampling points, yellow soil for the 50 sampling points, yellow brown soil for the 4 sampling points. As another part of a total of 1216 test data samples, including:red for the 171 sampling points, limestone soil for the 65sampling points, paddy soil for the 958 sampling points, purple soil for the 5 sampling points, yellow soil for the 15 sampling points, yellow brown soil for the 2 sampling points. Because the number of sampling points of the sub-soil types of some soils is less than normal, so the red soil and paddy soil are studied only.Case studies of pH and soil organic matter (SOM), the interpolation accuracy of none sub-soil types and sub-soil types are contrasted by using Kriging interpolation method, and the interpolation accuracy of different interpolation method also be studied under sub-soil types condition. The results are as follows:1. The accuracy of pH value of soil and organic matter is found that the accuracy of sub-soil types is higher than the precision of none sub-soil types using Kriging interpolation method, and the accuracy of PH is increased by 1.01 percentage points. The accuracy of organic matter is increased by 1.00 percentage points.2. The accuracy of interpolation is found that inverse distance weighting interpolation accuracy is increased under sub-soil types condition. The accuracy of PH is increased by 0.84 percentage points. The accuracy of organic matter is increased by 2.72 percentage points.3. The combination of GIS and Geo-statistics shows a distinctive advantage in the study the spatial variability of soil nutrients. In mountainous conditions, due to the soil unit fragmented, the accuracy of sub-soil type inverse distance weighting the spatial interpolation is the highest, for futune variability of soil nutrients offering some guidance.