Work

Geostatistical active learning for expanding monitoring networks for environmental decision making

Year: 2025

Type: article

Abstract: This is the R software documentation to our research about an Active Learning Geostatistical approach to densify the existing monitoring of nitrate concentrations in ground water across Germany. All constructed criteria to select candidate locations are based on Kriging with External Drift but can be adopted to any other Geostatistical or Machine Learning Model which provides a measure of uncertainty. In addition, our approach is not specific to nitrate concentrations in ground water but can be adapted to any other monitoring network for a given environmental variable. Furthermore we provided additional criteria, which can account for a second potential threshold. As dataset we used nitrate concentrations in the ground water provided by the Umweltbundesamt (German Federal Environment Agency), which we weren’t allowed to publish. Therefore, we provided a demo script with the meuse data set from the R Package sp to demonstrate all constructed criteria in addition to our real code. You can find our published preprint on SSRN: http://dx.doi.org/10.2139/ssrn.5743632 Our research was funded by the Umweltbundesamt (German Federal Environment Agency), which had no role in study design, data analysis, decision to publish, or preparation of the manuscript.

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.17493159

Authors: Henkel, Felix, Frank, Jonathan Karl, Suesse, Thomas, Brenning, Alexander

Institution: Friedrich Schiller University Jena

Language: English

Topic: Soil Geostatistics and Mapping

Subfield: Environmental Engineering

Field: Environmental Science

Domain: Physical Sciences

SDG: Clean water and sanitation

Open Access status: green

Funder: Umweltbundesamt