Estimation and mapping of soil organic matter content at a national scale based on grid soil samples, a soil map and DEM data

Abstract

Soil property maps at continental and national scales provide important input for research on biogeochemical and hydrological cycles. This paper develops a methodology for producing soil property maps at a continental or national scale based on soil samples from continental or national soil surveys (especially geochemical baseline surveys) with a 2–70 km grid, soil maps, and Digital Elevation Model data. The proposed method consists of four major steps: 1) determining the overall trend of soil property distribution using the linkage method; 2) estimating the variation inside each polygon through an extended Inverse Distance Weighted interpolation based on both the environmental similarity and spatial autocorrelation with stratification; 3) integrating variations in the soil property with its overall trend; 4) re-estimating soil properties in transition zones by using a weighted average of soil properties estimated from soil samples from different neighboring soils to produce a continuous soil property map across soil boundaries. A case study using Soil Organic Matter content as an example was conducted in Jilin Province (190,000 km2) in China based on soil samples from a 8–32 km grid from the China Soil Pollution Survey, the 1:1 million soil map of China and Shutter Radar Topographic Mission Digital Elevation Model data. Independent validation indicated that the proposed method decreased the mean error, the mean absolute error and the root mean square error by 10–30% compared with two commonly used soil mapping methods, the linkage method and Inverse Distance Weighted interpolation. The SOM map produced by the proposed method improved defects in the maps produced through Inverse Distance Weighted interpolation and the linkage method such as “bull's eye” patterns, abrupt change across soil boundaries and the ignored variation inside polygons on the maps. The proposed approach generates more accurate soil property maps at a continental or national scale based on grid soil samples from continental or national soil surveys than those produced through commonly used methods.