arxiv: 2604.25196 · v1 · submitted 2026-04-28 · 💻 cs.LG

Recognition: unknown

Knowledge-Data Dually Driven Paradigm for Accurate Landslide Susceptibility Prediction under Data-Scarce Conditions Using Geomorphic Priors and Tabular Foundation Model

Yuting Yang , Gang Mei , Feng Chen , Yongshuang Zhang , Jianbing Peng

Authors on Pith no claims yet

Pith reviewed 2026-05-07 16:54 UTC · model grok-4.3

classification 💻 cs.LG

keywords landslide susceptibility predictiondata-scarce conditionsgeomorphic priorstabular foundation modelknowledge-data dual paradigmgeohazard risk assessmentmountainous regionspermafrost terrain

0 comments

The pith

Integrating geomorphic prior knowledge with scarce landslide data inside a tabular foundation model produces accurate susceptibility predictions even when inventories cover only a fraction of events.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes a knowledge-data dually driven paradigm that merges physical landscape rules with whatever limited landslide records are available. Traditional data-driven methods fail in many mountainous and plateau settings because they need dense conditioning factors and large inventories that rarely exist. In a controlled test on a data-rich Italian region the new approach matched the accuracy of a full-data baseline while using only thirty percent of the landslide points. The same framework then delivered usable maps in a genuinely data-poor permafrost zone on the Tibetan Plateau. This matters because it turns partial field observations plus established geomorphic understanding into reliable hazard assessments where conventional techniques cannot operate.

Core claim

The central claim is that a knowledge-data dually driven paradigm, formed by encoding geomorphic prior knowledge and combining it with scarce landslide inventory data inside a tabular foundation model, achieves predictive accuracy comparable to a conventional data-driven model trained on the entire inventory when only thirty percent of the landslide data is supplied, and yields reliable susceptibility maps in a real data-scarce application region.

What carries the argument

The knowledge-data dually driven paradigm that encodes geomorphic prior knowledge and merges it with scarce landslide data inside a tabular foundation model to drive predictions.

If this is right

Landslide susceptibility maps become feasible in remote mountainous and plateau regions where full inventories cannot be assembled.
The data-collection effort for geohazard risk assessment can be reduced to roughly one-third of conventional requirements while retaining accuracy.
Predictions remain stable across both test regions with artificially reduced data and genuinely data-poor field sites such as permafrost terrain.
The approach extends the reach of tabular foundation models to geoscience tasks by letting physical priors offset missing observations.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same dual-driving structure could be tested on other geohazards such as debris flows or rockfalls where inventories are also incomplete.
Systematic reduction of the training fraction below thirty percent would show the minimum inventory size at which accuracy begins to degrade.
Incorporating additional priors from climate or vegetation layers might allow the paradigm to track changing susceptibility under warming conditions.
Operational use would ultimately need validation against future landslide occurrences rather than historical hold-out sets.

Load-bearing premise

Geomorphic prior knowledge can be encoded and integrated into the model without introducing systematic bias that would require region-specific tuning data unavailable under scarcity.

What would settle it

Independent mapping of actual landslides in a new data-scarce region over multiple seasons that shows the model's high-susceptibility zones contain far fewer events than its low-susceptibility zones would falsify reliable performance.

Figures

Figures reproduced from arXiv: 2604.25196 by Feng Chen, Gang Mei, Jianbing Peng, Yongshuang Zhang, Yuting Yang.

**Figure 1.** Figure 1: The workflow of the proposed paradigm. 2.2. Step 1: Estimation of the geomorphic prior knowledge via morphometric modeling 2.2.1. Calculation of drainage area and average drainage area slope The morphometric model requires two primary topographic inputs: upstream drainage area (A) and average drainage area slope (β). Both factors were derived within ArcGIS using the D8 flow-direction algo4 view at source ↗

**Figure 2.** Figure 2: Regional location of the Central Italy study area. view at source ↗

**Figure 3.** Figure 3: delineation of Slope unit and distribution of labels in the Central Italy study area. view at source ↗

**Figure 4.** Figure 4: Landslide susceptibility maps of the central Italy study area via the proposed paradigm: (a) using 10% available landslide view at source ↗

**Figure 5.** Figure 5: Landslide susceptibility maps of the central Italy study area via conventional data-driven paradigm: (a) conventional view at source ↗

**Figure 6.** Figure 6: Comparison of predictive performance between the proposed paradigm and conventional data-driven paradigm under view at source ↗

**Figure 7.** Figure 7: Regional location of the Qilian Permafrost Region, Tibetan Plateau study area. view at source ↗

**Figure 8.** Figure 8: delineation of Slope unit and distribution of labels in the Qilian Permafrost Region, Tibetan Plateau study area. view at source ↗

**Figure 9.** Figure 9: Landslide susceptibility maps of the Qilian Permafrost Region, Tibetan Plateau study area under different paradigms: (a) view at source ↗

read the original abstract

Landslide susceptibility prediction is critical for geohazard risk assessment and mitigation. Conventional data-driven paradigm achieves high predictive accuracy but require sufficient conditioning factors and large-scale landslide inventories. However, in practical engineering applications across mountainous and plateau regions, data-scarce conditions are commonly observed, where such data requirements are rarely satisfied, rendering conventional data-driven paradigm inapplicable. To address this issue, we propose a knowledge-data dually driven paradigm for accurate landslide susceptibility prediction under data-scarce conditions. The essential idea behind the proposed novel paradigm is the integration of the geomorphic prior knowledge with scarce landslide data. To validate the proposed paradigm, we first applied it to a data-rich region in central Italy, where a conventional data-driven paradigm trained on the full dataset served as the baseline. By utilizing only 30% of the available landslide data, the proposed paradigm achieved comparable predictive accuracy to the baseline, demonstrating its effectiveness under data-scarce conditions. The paradigm was further evaluated in a genuinely data-scarce environment for application, the Qilian Permafrost Region of the Tibetan Plateau, where it also yielded reliable susceptibility predictions, confirming its applicability under data-scarce conditions.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The 30% data result in Italy is the only quantified win; the Qilian claim rests on an unbacked assertion of 'reliable predictions.'

read the letter

The paper's clearest contribution is showing that geomorphic priors plus a tabular foundation model can match a full-data baseline while using only 30% of the landslide inventory in central Italy. That single controlled comparison is the part worth paying attention to, because it directly tests the data-scarcity claim on a region where a strong baseline exists. The rest of the work is an attempt to carry the same setup into the Qilian Permafrost Region, where inventories are genuinely thin. They report that the method produced usable susceptibility maps there, but supply no AUC, precision-recall, spatial cross-validation, or even a simple inventory overlap check. Without those numbers it is impossible to know whether the priors are adding signal or simply producing maps that look reasonable for that terrain. The integration step itself is described at a high level; there is no ablation on how the priors are encoded, no sensitivity test on prior strength, and no discussion of whether the tabular foundation model requires region-specific fine-tuning that would be unavailable in true data-scarce settings. The Italy experiment is reproducible in principle and addresses a real engineering constraint, but the Qilian application functions more as a demonstration than as evidence. Readers who need a method they can apply tomorrow in another data-poor mountain range will still have to do their own validation work. The paper is worth sending to referees because the Italy result is concrete and the underlying problem matters, but any acceptance should require the authors to add quantitative metrics and at least one independent check for the target region before the claims can be treated as reliable.

Referee Report

2 major / 2 minor

Summary. The paper proposes a knowledge-data dually driven paradigm that integrates geomorphic prior knowledge with limited landslide inventories via a tabular foundation model to enable accurate susceptibility prediction when conventional data-driven methods cannot be applied due to insufficient data. It reports that training on only 30% of the landslide inventory in central Italy produces predictive accuracy comparable to a full-data baseline, and states that the same paradigm yields reliable susceptibility predictions when applied to the genuinely data-scarce Qilian Permafrost Region of the Tibetan Plateau.

Significance. If the quantitative claims are supported by detailed metrics, ablations, and independent validation, the work could offer a practical route to susceptibility mapping in data-poor mountainous and plateau settings where large inventories are unavailable. The dual-knowledge approach addresses a recurring limitation in applied geohazard modeling, but the current absence of reported performance numbers for the Qilian case prevents assessment of whether the priors add genuine predictive value or merely produce plausible maps.

major comments (2)

[Qilian Permafrost Region evaluation] Qilian Permafrost Region evaluation (abstract and application section): the manuscript asserts that the paradigm 'yielded reliable susceptibility predictions' yet reports no AUC, F1, precision-recall, spatial cross-validation, or inventory-comparison metrics. This omission is load-bearing for the central claim of effectiveness under data-scarce conditions, as it leaves open whether the output reflects successful prior-data fusion or simply terrain-consistent but untested maps.
[Central Italy validation] Central Italy 30% data-reduction experiment (abstract and validation section): the claim of 'comparable predictive accuracy' to the full-data baseline is presented without accompanying numerical values, confidence intervals, or ablation results that isolate the contribution of the geomorphic priors versus the tabular foundation model alone. Without these, the strength of the data-scarcity demonstration cannot be evaluated.

minor comments (2)

[Methods] The abstract and available text do not define the tabular foundation model architecture, the precise encoding of geomorphic priors, or the loss function used for dual training; these details are needed for reproducibility.
[Results] No comparison is shown against pure knowledge-driven or pure data-driven baselines in either study area; adding such controls would clarify the incremental benefit of the proposed dual paradigm.

Simulated Author's Rebuttal

2 responses · 1 unresolved

We thank the referee for the insightful comments on our manuscript. We provide point-by-point responses to the major comments below, and we will incorporate clarifications and additional details in the revised version where appropriate.

read point-by-point responses

Referee: [Qilian Permafrost Region evaluation] Qilian Permafrost Region evaluation (abstract and application section): the manuscript asserts that the paradigm 'yielded reliable susceptibility predictions' yet reports no AUC, F1, precision-recall, spatial cross-validation, or inventory-comparison metrics. This omission is load-bearing for the central claim of effectiveness under data-scarce conditions, as it leaves open whether the output reflects successful prior-data fusion or simply terrain-consistent but untested maps.

Authors: We agree that quantitative metrics would strengthen the claim, but they are not feasible in this case. The Qilian Permafrost Region lacks a comprehensive landslide inventory, which is the very reason it exemplifies data-scarce conditions. Our assessment of 'reliable susceptibility predictions' is based on qualitative validation: alignment with geomorphic priors, expert review by local geologists, and consistency with known permafrost and terrain characteristics. We will revise the application section to explicitly describe this validation methodology and any supporting evidence. revision: partial
Referee: [Central Italy validation] Central Italy 30% data-reduction experiment (abstract and validation section): the claim of 'comparable predictive accuracy' to the full-data baseline is presented without accompanying numerical values, confidence intervals, or ablation results that isolate the contribution of the geomorphic priors versus the tabular foundation model alone. Without these, the strength of the data-scarcity demonstration cannot be evaluated.

Authors: We acknowledge the need for explicit numerical reporting. Although the full manuscript contains the detailed results in the validation section, we will ensure the abstract and main text prominently feature the specific performance metrics (e.g., AUC values for 30% vs. full data), confidence intervals, and ablation experiments separating the effects of geomorphic priors from the tabular foundation model. This will be added in the revised manuscript. revision: yes

standing simulated objections not resolved

The lack of quantitative performance metrics for the Qilian Permafrost Region evaluation, since no independent landslide inventory exists to compute AUC, F1, or similar metrics.

Circularity Check

0 steps flagged

No circularity: paradigm integrates external geomorphic priors with empirical validation on independent test regions

full rationale

The paper proposes integrating geomorphic prior knowledge with scarce landslide inventories inside a tabular foundation model. Validation proceeds by training on 30% of Italy inventory and comparing AUC/F1-style metrics to a full-data baseline, then applying the same pipeline to the separate Qilian region. No equations, parameter-fitting steps, or derivations are described that reduce any output to a re-labeling of the input data or to a self-citation chain. The geomorphic priors are treated as external domain knowledge rather than quantities fitted from the target landslide labels. Consequently the claimed performance gains rest on reported cross-region empirical results rather than on any definitional or self-referential reduction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Abstract provides insufficient technical detail to enumerate free parameters or invented entities; the central claim rests on the domain assumption that geomorphic priors meaningfully compensate for missing data.

axioms (1)

domain assumption Geomorphic prior knowledge can be effectively encoded and fused with scarce landslide data inside a tabular foundation model
Invoked as the essential idea of the proposed paradigm

pith-pipeline@v0.9.0 · 5524 in / 1235 out tokens · 38321 ms · 2026-05-07T16:54:47.416297+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

31 extracted references · 31 canonical work pages

[1]

, author Cs\'amer, A

author Abdelkader, M.M. , author Cs\'amer, A. , year 2025 . title Comparative assessment of machine learning models for landslide susceptibility mapping: a focus on validation and accuracy . journal Natural Hazards volume 121 , pages 10299--10321 . https://doi.org/10.1007/s11069-025-07197-0, :10.1007/s11069-025-07197-0

work page doi:10.1007/s11069-025-07197-0 2025
[2]

, author Aju, C.D

author Achu, A.L. , author Aju, C.D. , author Di Napoli, M. , author Prakash, P. , author Gopinath, G. , author Shaji, E. , author Chandra, V. , year 2023 . title Machine-learning based landslide susceptibility modelling with emphasis on uncertainty analysis . journal Geoscience Frontiers volume 14 , pages 101657 . https://www.sciencedirect.com/science/ar...

work page doi:10.1016/j.gsf.2023.101657 2023
[3]

, author Amitab, K

author Ado, M. , author Amitab, K. , author Maji, A.K. , author Jasinska, E. , author Gono, R. , author Leonowicz, Z. , author Jasinski, M. , year 2022 . title Landslide susceptibility mapping using machine learning: A literature survey . journal Remote Sensing volume 14 , pages 3029 . https://doi.org/10.3390/rs14133029, :10.3390/rs14133029

work page doi:10.3390/rs14133029 2022
[4]

, author Loche, M

author Alvioli, M. , author Loche, M. , author Jacobs, L. , author Grohmann, C.H. , author Abraham, M.T. , author Gupta, K. , author Satyam, N. , author Scaringi, G. , author Bornaetxea, T. , author Rossi, M. , author Marchesini, I. , author Lombardo, L. , author Moreno, M. , author Steger, S. , author Camera, C.A.S. , author Bajni, G. , author Samodra, G...

work page doi:10.1016/j.earscirev.2024.104927 2024
[5]

Machine Learning , author =

author Breiman, L. , year 2001 . title Random forests . journal Machine Learning volume 45 , pages 5--32 . https://doi.org/10.1023/A:1010933404324, :10.1023/A:1010933404324

work page doi:10.1023/a:1010933404324 2001
[6]

Khyathi Raghavi Chandu, Linjie Li, Anas Awadalla, Ximing Lu, Jae Sung Park, Jack Hessel, Lijuan Wang, and Yejin Choi

author Brier, G.W. , year 1950 . title Verification of forecasts expressed in terms of probability . journal Monthly Weather Review volume 78 , pages 1--3 . https://journals.ametsoc.org/view/journals/mwre/78/1/1520-0493_1950_078_0001_vofeit_2_0_co_2.xml, :10.1175/1520-0493(1950)078<0001:VOFEIT>2.0.CO;2

work page doi:10.1175/1520-0493(1950)078 1950
[7]

XGBoost: A Scalable Tree Boosting System

author Chen, T. , author Guestrin, C. , year 2016 . title XGBoost : A scalable tree boosting system , in: booktitle 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD) , address San Francisco, CA, USA . pp. pages 785--794 . https://doi.org/10.1145/2939672.2939785, :10.1145/2939672.2939785

work page doi:10.1145/2939672.2939785 2016
[8]

, author Xie, X

author Chen, W. , author Xie, X. , author Wang, J. , author Pradhan, B. , author Hong, H. , author Bui, D.T. , author Duan, Z. , author Ma, J. , year 2017 . title A comparative study of logistic model tree, random forest, and classification and regression tree models for spatial prediction of landslide susceptibility . journal Catena volume 151 , pages 14...

work page doi:10.1016/j.catena.2016.11.032 2017
[9]

, author Lombardo, L

author Dahal, A. , author Lombardo, L. , year 2025 . title Towards physics-informed neural networks for landslide prediction . journal Engineering Geology volume 344 , pages 107852 . https://www.sciencedirect.com/science/article/pii/S0013795224004526, :10.1016/j.enggeo.2024.107852

work page doi:10.1016/j.enggeo.2024.107852 2025
[10]

, author Whipple, K.X

author DiBiase, R.A. , author Whipple, K.X. , author Heimsath, A.M. , author Ouimet, W.B. , year 2010 . title Landscape form and millennial erosion rates in the san gabriel mountains, ca . journal Earth and Planetary Science Letters volume 289 , pages 134--144 . https://www.sciencedirect.com/science/article/pii/S0012821X09006451, :10.1016/j.epsl.2009.10.036

work page doi:10.1016/j.epsl.2009.10.036 2010
[11]

, author Petley, D.N

author Froude, M.J. , author Petley, D.N. , year 2018 . title Global fatal landslide occurrence from 2004 to 2016 . journal Natural Hazards and Earth System Sciences volume 18 , pages 2161--2181 . https://doi.org/10.5194/nhess-18-2161-2018, :10.5194/nhess-18-2161-2018

work page doi:10.5194/nhess-18-2161-2018 2018
[12]

, author Guzzetti, F

author Gariano, S.L. , author Guzzetti, F. , year 2016 . title Landslides in a changing climate . journal Earth-Science Reviews volume 162 , pages 227--252 . https://doi.org/10.1016/j.earscirev.2016.08.011, :10.1016/j.earscirev.2016.08.011

work page doi:10.1016/j.earscirev.2016.08.011 2016
[13]

Accurate predictions on small data with a tabular foundation model.Nature, 637(8045):319–326, 2025

author Hollmann, N. , author Müller, S. , author Purucker, L. , author Krishnakumar, A. , author Körfer, M. , author Hoo, S.B. , author Schirrmeister, R.T. , author Hutter, F. , year 2025 . title Accurate predictions on small data with a tabular foundation model . journal Nature volume 637 , pages 319--326 . https://doi.org/10.1038/s41586-024-08328-6, :10...

work page doi:10.1038/s41586-024-08328-6 2025
[14]

, author Xiong, H

author Huang, F. , author Xiong, H. , author Jiang, S.H. , author Yao, C. , author Fan, X. , author Catani, F. , author Chang, Z. , author Zhou, X. , author Huang, J. , author Liu, K. , year 2024 . title Modelling landslide susceptibility prediction: A review and construction of semi-supervised imbalanced theory . journal Earth-Science Reviews volume 250 ...

work page doi:10.1016/j.earscirev.2024.104700 2024
[15]

, author Kapnick, S.B

author Kirschbaum, D. , author Kapnick, S.B. , author Stanley, T. , author Pascale, S. , year 2020 . title Changes in extreme precipitation and landslides over high mountain asia . journal Geophysical Research Letters volume 47 , pages e2019GL085347 . https://doi.org/10.1029/2019GL085347, :10.1029/2019gl085347

work page doi:10.1029/2019gl085347 2020
[16]

, author Feng, W

author Kong, L. , author Feng, W. , author Yi, X. , author Xue, Z. , author Bai, L. , year 2025 . title Enhanced landslide susceptibility mapping in data-scarce regions via unsupervised few-shot learning . journal Gondwana Research volume 138 , pages 31--46 . https://www.sciencedirect.com/science/article/pii/S1342937X2400306X, :10.1016/j.gr.2024.10.011

work page doi:10.1016/j.gr.2024.10.011 2025
[17]

, author Wang, L

author Liu, S. , author Wang, L. , author Zhang, W. , author He, Y. , author Pijush, S. , year 2023 . title A comprehensive review of machine learning-based methods in landslide susceptibility mapping . journal GEOLOGICAL JOURNAL volume 58 , pages 2283--2301 . https://doi.org/10.1002/gj.4666, :10.1002/gj.4666

work page doi:10.1002/gj.4666 2023
[18]

, author Yunus, A.P

author Merghadi, A. , author Yunus, A.P. , author Dou, J. , author Whiteley, J. , author Binh, T. , author Dieu Tien, B. , author Avtar, R. , author Abderrahmane, B. , year 2020 . title Machine learning methods for landslide susceptibility studies: A comparative overview of algorithm performance . journal Earth-Science Reviews volume 207 , pages 103225 . ...

work page doi:10.1016/j.earscirev.2020.103225 2020
[19]

, author Bozzolan, E

author Ozturk, U. , author Bozzolan, E. , author Holcombe, E.A. , author Shukla, R. , author Pianosi, F. , author Wagener, T. , year 2022 . title How climate change and unplanned urban sprawl bring more landslides . journal Nature volume 608 , pages 262--265 . https://doi.org/10.1038/d41586-022-02141-9, :10.1038/d41586-022-02141-9

work page doi:10.1038/d41586-022-02141-9 2022
[20]

, year 2012

author Petley, D. , year 2012 . title Global patterns of loss of life from landslides . journal Geology volume 40 , pages 927--930 . https://doi.org/10.1130/G33217.1, :10.1130/g33217.1

work page doi:10.1130/g33217.1 2012
[21]

, author Ozturk, U

author Rana, K. , author Ozturk, U. , author Malik, N. , year 2021 . title Landslide geometry reveals its trigger . journal Geophysical Research Letters volume 48 , pages e2020GL090848 . https://doi.org/10.1029/2020GL090848, :10.1029/2020gl090848

work page doi:10.1029/2020gl090848 2021
[22]

, author Rossi, M

author Reichenbach, P. , author Rossi, M. , author Malamud, B.D. , author Mihir, M. , author Guzzetti, F. , year 2018 . title A review of statistically-based landslide susceptibility models . journal Earth-Science Reviews volume 180 , pages 60--91 . https://www.sciencedirect.com/science/article/pii/S0012825217305652, :10.1016/j.earscirev.2018.03.001

work page doi:10.1016/j.earscirev.2018.03.001 2018
[23]

, author Lee, M.L

author Sim, K.B. , author Lee, M.L. , author Wong, S.Y. , year 2022 . title A review of landslide acceptable risk and tolerable risk . journal Geoenvironmental Disasters volume 9 , pages 3 . https://doi.org/10.1186/s40677-022-00205-6, :10.1186/s40677-022-00205-6

work page doi:10.1186/s40677-022-00205-6 2022
[24]

, author van Zadelhoff, F

author Spiekermann, R.I. , author van Zadelhoff, F. , author Schindler, J. , author Smith, H. , author Phillips, C. , author Schwarz, M. , year 2023 . title Comparing physical and statistical landslide susceptibility models at the scale of individual trees . journal Geomorphology volume 440 , pages 108870 . https://www.sciencedirect.com/science/article/pi...

work page doi:10.1016/j.geomorph.2023.108870 2023
[25]

, author Yang, C

author Tan, J. , author Yang, C. , author Wang, Y. , author Xiong, H. , author Ma, C. , year 2024 . title A hybrid model to overcome landslide inventory incompleteness issue for landslide susceptibility prediction . journal Geocarto International volume 39 . https://doi.org/10.1080/10106049.2024.2322066, :10.1080/10106049.2024.2322066

work page doi:10.1080/10106049.2024.2322066 2024
[26]

, author Wen, Z

author Wang, F. , author Wen, Z. , author Gao, Q. , author Yu, Q. , author Li, D. , author Chen, L. , year 2024 . title Thermokarst landslides susceptibility evaluation across the permafrost region of the central qinghai-tibet plateau: Integrating a machine learning model with insar technology . journal Journal of Hydrology volume 642 , pages 131800 . htt...

work page doi:10.1016/j.jhydrol.2024.131800 2024
[27]

, author Mirus, B.B

author Woodard, J.B. , author Mirus, B.B. , year 2025 . title Overcoming the data limitations in landslide susceptibility modeling . journal Science Advances volume 11 , pages eadt1541 . https://doi.org/10.1126/sciadv.adt1541, :10.1126/sciadv.adt1541

work page doi:10.1126/sciadv.adt1541 2025
[28]

, author Mirus, B.B

author Woodard, J.B. , author Mirus, B.B. , author Crawford, M. , author Or, D. , author Leshchinsky, B.A. , author Allstadt, K.E. , author Wood, N.J. , year 2023 . title Mapping landslide susceptibility over large regions with limited data . journal Journal of Geophysical Research: Earth Surface volume 128 , pages e2022JF006810 . https://doi.org/10.1029/...

work page doi:10.1029/2022jf006810 2023
[29]

, author Mei, G

author Yang, Y. , author Mei, G. , author Ma, Z. , author Xu, N. , author Peng, J. , year 2026 . title Impact of freeze-thaw on landslide activity under diverse topographies using geospatial data mining: Insights from qilian permafrost region, tibetan plateau . journal Catena volume 268 , pages 110024 . https://www.sciencedirect.com/science/article/pii/S0...

work page doi:10.1016/j.catena.2026.110024 2026
[30]

, author Pourghasemi, H.R

author Youssef, A.M. , author Pourghasemi, H.R. , author Pourtaghi, Z.S. , author Al-Katheeri, M.M. , year 2016 . title Landslide susceptibility mapping using random forest, boosted regression tree, classification and regression tree, and general linear models and comparison of their performance at wadi tayyah basin, asir region, saudi arabia . journal La...

work page doi:10.1007/s10346-015-0614-1 2016
[31]

, author Li, Y

author Zhang, T. , author Li, Y. , author Wang, T. , author Wang, H. , author Chen, T. , author Sun, Z. , author Luo, D. , author Li, C. , author Han, L. , year 2022 . title Evaluation of different machine learning models and novel deep learning-based algorithm for landslide susceptibility mapping . journal Geoscience Letters volume 9 , pages 26 . https:/...

work page doi:10.1186/s40562-022-00236-9 2022