An adaptive uncertainty-guided sampling method for geospatial prediction and its application in digital soil mapping

Abstract

Sampling design can significantly reduce the uncertainty in geospatial predictions. In this paper, we developed an adaptive uncertainty-guided stepwise sampling (AUGSS) method to select sampling locations to supplement existing legacy sample points whose representation should be improved. The proposed method selects supplemental samples in a stepwise manner as guided by an objective function with two weighted sub-objectives. One reduces the area with high prediction uncertainty, and the other minimizes the overall prediction uncertainty for the entire area. The method takes an adaptive approach to adjust weights for the two sub-objectives and to tune an uncertainty threshold controlling whether a location can be reliably predicted during the sampling procedure. A case study on soil property prediction shows that AUGSS outperforms the stratified random sampling (SRS) and the non-adaptive uncertainty guided sampling method (UGSS) in terms of RMSE and Lin’s concordance correlation coefficient with different sample sizes. This study shows that the AUGSS method offers a potential for effectively adding supplemental samples to existing samples which are insufficient for spatial prediction. The adaptive strategy guided by predicted uncertainty provides an efficient support to improve the spatial pattern of samples, which plays a key role in the result accuracy of geospatial predictive mapping.

Keywords:

• Spatial sampling
• prediction uncertainty
• adaptive approach
• spatial prediction
• digital soil mapping

Acknowledgments

Lei Zhang thanks to the support from Postgraduate Research and Practice Innovation Program of Jiangsu Province (KYCX22_0109). Supports to A-Xing Zhu through the NSFC Project (41871300), PAPD, the Vilas Associate Award, the Hammel Faculty Fellow Award, and the Manasse Chair Professorship from the University of Wisconsin-Madison are greatly appreciated. The authors express sincere gratitude to Editor May Yuan, Editor Jennifer Miller and anonymous reviewers, whose valuable comments and suggestions have greatly improved the quality of the paper.

Disclosure statement

No potential competing interest was reported by the author(s).

Data and codes availability statement

The data and codes that support the findings of this study are available with a digital object identifier (DOI) at: https://doi.org/10.5281/zenodo.7070697.

Additional information

Funding

The study was supported by National Natural Science Foundation of China [Project No.: 41871300, 41971054, 41901062], the 111 Program of China [Approved Number: D19002], Postgraduate Research and Practice Innovation Program of Jiangsu Province (KYCX22_0109), and PAPD.

Notes on contributors

Lei Zhang

Lei Zhang is currently a Ph.D. candidate at the School of Geography and Ocean Science, Nanjing University, and a visiting researcher at Wageningen University. His research interests include vegetation growth and soil carbon dynamics under the impacts of climate change and human activities, spatial predictive mapping, efficient spatial sampling strategy, remote sensing, and machine learning. L.Z.'s homepage: https://leizhang-geo.github.io.

A-Xing Zhu

A-Xing Zhu is a full professor at the Department of Geography and the Manasse Chair Professor, the University of Wisconsin-Madison. He currently serves as the Editor-in-Chief of Annals of GIS. His research interests focus on theoretical and methodological developments in GIS, including artificial intelligence, fuzzy logic, intelligent geocomputing, and their applications in environmental modeling and scenario analysis. His signature work includes the Third Law of Geography and similarity-based spatial prediction.

Junzhi Liu

Junzhi Liu is a professor in the Center for the Pan-Third Pole Environment, Lanzhou University. His research interests mainly include land surface modeling and spatio-temporal data mining. He developed the SEIMS (Spatially Explicit Integrated Modeling System) watershed modeling framework (https://github.com/lreis2415/SEIMS), which has been widely used in different types of watersheds.

Tianwu Ma

Tianwu Ma is a Ph.D. candidate at the School of Geography, Nanjing Normal University. His research interests include spatial prediction and sampling design.

Lin Yang

Lin Yang is a professor in the School of Geography and Ocean Science, Nanjing University. Her research interests focus on spatial sampling, digital soil mapping, soil carbon cycling.

Chenghu Zhou

Chenghu Zhou is an academician of the Chinese Academy of Sciences, and currently a professor in the Institute of Geographical Sciences and Natural Resources Research, Chinese Academy of Sciences. His research interests are broadly situated in research on the application of GIS and remote sensing.