Ground subsidence prediction: a review from traditional to deep learning techniques
Geotechnical Research · 14.12.2026 · 13 (1) · 138-168
DOI: 10.1680/jgere.25.00057 ↗Kurz gefasst
Fachartikel in der Zeitschrift Geotechnical Research, online erschienen am 14.07.2026. Verfasst von Kangkang Mo und 3 weiteren Autor:innen. Die Arbeit nennt 119 Literaturangaben. Gefördert von: National Natural Science Foundation of China Youth Fund.
Abstract (Originalsprache)
Ground subsidence – the gradual or sudden lowering of the Earth’s surface – poses serious threats to infrastructure, environmental safety, and sustainable land use. Accurate prediction is therefore crucial for effective risk management. This review provides a comprehensive synthesis of subsidence prediction methods developed over the past two decades, spanning traditional empirical, analytical, and numerical approaches, as well as recent advances in deep learning. This review summarises the theoretical foundations and practical performance of major model classes, with emphasis on their ability to capture complex spatiotemporal patterns. Comparative analysis highlights the superior flexibility and accuracy of deep learning methods, while underscoring their challenges in data dependency and interpretability. Future research directions are discussed, focusing on physics-informed learning, multi-source data fusion, and uncertainty quantification. Integrating physical understanding with data-driven intelligence represents a promising pathway towards more accurate, interpretable, and robust subsidence prediction frameworks.
Förderung
- National Natural Science Foundation of China Youth Fund
Zitieren
APA
Mo, K., Li, Y., Zhang, J., & Min, B. (2026). Ground subsidence prediction: a review from traditional to deep learning techniques. Geotechnical Research, 13(1), 138-168. https://doi.org/10.1680/jgere.25.00057
Vancouver
Mo K, Li Y, Zhang J, Min B. Ground subsidence prediction: a review from traditional to deep learning techniques. Geotechnical Research. 2026;13(1):138-168. doi:10.1680/jgere.25.00057
BibTeX
@article{Mo2026,
author = {Mo, Kangkang and Li, Yongbo and Zhang, Jian and Min, Bo},
title = {Ground subsidence prediction: a review from traditional to deep learning techniques},
journal = {Geotechnical Research},
year = {2026},
volume = {13},
number = {1},
pages = {138--168},
doi = {10.1680/jgere.25.00057}
}
EUVIMED