Accounting for spatial dependence in random forests improves ecological understanding and enhances prediction
Ecology · 2026-11 · 107 (11)
DOI: 10.1002/ecy.70567 ↗Kurz gefasst
Fachartikel in der Zeitschrift Ecology, online erschienen am 05.11.2026. Verfasst von Matthew J Heaton und 4 weiteren Autor:innen. Beteiligte Einrichtung: Department of Statistics, Brigham Young University, Provo, Utah, USA. Hauptthemen laut MEDLINE: Rivers, Models, Biological, Ecosystem. Die Arbeit nennt 61 Literaturangaben.
Abstract (Originalsprache)
Random forests (RFs) are widely used in ecology to model complex, nonlinear relationships among environmental variables and to summarize fitted relationships through variable importance and interactions. However, many ecological datasets, particularly observational data collected across space, exhibit spatial dependence. If this spatial dependence is ignored, RFs can incorrectly attribute importance to predictors that serve at least partly as proxies for spatial location. In this study, we evaluate how accounting for spatial dependence changes both predictive performance and ecological interpretation. Using simulations with known data‐generating mechanisms and an analysis of 1061 North American river catchments, we compare standard RFs, RFs with spatial residual modeling, and a model‐agnostic spatial decorrelation transformation. The simulations show that spatially structured noise variables and raw coordinates can receive inflated importance in standard RF workflows, whereas decorrelation reduces these artifacts and recovers the relevant predictors more consistently. In the river flow application, decorrelation improves held‐out prediction and shifts interpretation from broad spatially smooth gradients toward hydrologic and soil variables with more direct mechanistic links to flow seasonality.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Statistics, Brigham Young University, Provo, Utah, USA.
- Office of Water, United States Environmental Protection Agency, Corvallis, Oregon, USA.
- Department of Mathematics & Statistics, Utah State University, Logan, Utah, USA.
ORCID-Profile: Michael Dumelle, Ryan A Hill
Zitieren
APA
Heaton, M. J., Dumelle, M., Bladen, K. K., Hill, R. A., & Rhodes, J. S. (2026). Accounting for spatial dependence in random forests improves ecological understanding and enhances prediction. Ecology, 107(11). https://doi.org/10.1002/ecy.70567
Vancouver
Heaton MJ, Dumelle M, Bladen KK, Hill RA, Rhodes JS. Accounting for spatial dependence in random forests improves ecological understanding and enhances prediction. Ecology. 2026;107(11). doi:10.1002/ecy.70567
BibTeX
@article{Heaton2026,
author = {Heaton, Matthew J and Dumelle, Michael and Bladen, Kelvyn K and Hill, Ryan A and Rhodes, Jake S},
title = {Accounting for spatial dependence in random forests improves ecological understanding and enhances prediction},
journal = {Ecology},
year = {2026},
volume = {107},
number = {11},
doi = {10.1002/ecy.70567}
}
EUVIMED