Evaluation of Soil Data Interpolation Methods

Abstract

Applying geostatistical interpolation methods that predict soil properties of the surrounding area may present an efficient solution to create interpolation maps for site-specific field management. These statistical methods produce different distributions of interpolated data. According to the data type, different methods are appropriate. In this study, we compared Kriging and Inverse Distance Weighting to determine the potential application within management zones. Therefore, we interpolated soil mineral nitrogen content (Nmin) data. We evaluated the accuracy of real vs. predicted data by bootstrapping and considering the standard error. Both interpolation methods were able to predict the Nmin content with a root mean square error of below 0.025.