A new spatial interpolation approach based on inverse distance weighting: Case study from interpolating soil properties

Abstract

This paper analyzed and discussed the effect of three similarity measure modes, i.e., spatial distance measure (SDM), non-spatial attribute similarity measure (NSASM) and their hybrid (SDM+NSASM), on interpolation accuracy by inverse distance weighting algorithm (IDW). The newly proposed approach differed from the traditional IDW one in that it also took use of non-spatial information (or environmental co-variables, either NSASM or SDM+NSASM mode) in the weighting vector computation for known spatial samples which are used to make interpolation for unknown locations. The proposed approach was tested on a case study in soil property (soil pH) interpolation. The result from cross-validation demonstrated that the NSASM and SDM+NSASM helped to decrease the prediction error compared to traditional SDM method, with good correlation between its environmental variables and the interpolation variable. Meanwhile, the IDW modeling result by NSASM and SDM+NSASM revealed closer local spatial relationships between soil pH and its environmental factors in comparison with the result from the traditional SDM. © Springer-Verlag Berlin Heidelberg 2013.