Area-to-Point Kriging of Soil Phosphorus Composite Samples

Abstract

Mapping of phosphorus (P) is based on sampling and laboratory analysis. Although laboratory analysis is costly, the number of samples is restricted in practice. In zone sampling, areas of the field are used to composite samples from sets of sampling points to reduce efforts. This study introduces area-to-point (ATP) kriging for downscaling composite samples with different sizes and shapes of the sampling areas. ATP kriging makes use of the coordinates of the sampling points of the composite samples. The applicability was tested on a simulated data set as well as on a spatially dense sample set of P measurements. Validation shows that ATP kriging outperforms point kriging with centroid interpolation. The root mean square error (RMSE) is reduced from 39.5 to 33.5 mg kg⁻¹. ATP kriging predictions were better at retaining the P value of the sampling area. The smoothing effect of interpolation and the aggregation effect of compositing the samples were reduced.

Keywords:

• Areal interpolation
• change of support
• geostatistics
• soil sampling
• spatial prediction

Acknowledgments

The authors thank the Leibniz-Institute for Agricultural Engineering, Potsdam-Bornim, for financial support; Katrin Witzke for her excellent assistance; and the editor and the reviewers for their valuable recommendations pertaining to the revision of this article.