GIS-Based Elaborate Spatial Prediction of Soil Nutrient Elements Using Ancillary Terrain Data in Chongqing Tobacco Planting Region, China

The precision agriculture hopes to manage the variation in soil nutrient status continuously, which requires reliable predictions at places between sampling sites. For the long time, ordinary kriging has been used as one prediction method when the data are spatially dependent and a suitable variogram model exists. However, even if data are spatially correlated, there are often few soil sampling sites in relation to the area to be managed. Recently, Digital elevation models(DEMs) and remotely sensed data are becoming more readily available, these data are usually far more intensive than those from soil surveys. If these ancillary data are coregionalized with the sparse soil data, they might be used to increase the accuracy of predictions of the soil properties. Under ArcGIS platform, this paper employed spatial predictions of the soil total N, P, K in Chongqing tobacco planting region, China, with cokriging and regression kriging respectively. For the both, intensive terrain data including elevation, slope and aspect were used with the soil data. Traditional ordinary kriging(OK) was investigated as comparison basis to determine which approach is appropriate for different soils properties mapping. And the results suggest that the use of intensive ancillary data can increase the accuracy of predictions of soil properties in arable fields provided that the variables are related spatially. 346 Xuan Wang , Jiake Lv , Chaofu Wei , Deti Xie