Open research questions in Landslides and related hazards

What the literature leaves open

• Traditional monitoring methods have limitations such as high acquisition costs and low temporal resolution. Computer vision-based detection methods require improvement in accuracy and robustness. The paper identifies the need for a deep learning transformer architecture that can integrate multi-source feature modeling and efficient attention mechanisms.

  A frequency-domain and context-guided deep learning transformer architecture for real-time landslide detection · 2026 · DOI

• The need for improved understanding of cliff stability. The lack of data on the mechanical properties of the cliff's rock. The complexity of simulating the cliff's behavior due to internal and external factors.

  Assessment of the stability of the Bonifacio cliff using 3D geomechanical modelling · 2026 · DOI

• The correlation between each evaluation factor and landslide occurrence is not effectively quantified. The unified expression between different factors in landslide susceptibility modeling is not realized. The existing methods for landslide susceptibility assessment have limitations and inaccuracies.

  Landslide susceptibility assessment under non-landslide sampling strategy based on a clustering algorithm · 2026 · DOI

• The mechanisms of rainfall-induced landslides are not well understood, particularly the role of pore-pressure generation. There is a lack of experimental and numerical studies on the initiation of landslide failure. The paper aims to bridge the gap between laboratory-scale mechanisms and field-scale observations.

  Pore-pressure generation initiating a rainfall-induced landslide: experimental insights and numerical modelling · 2026 · DOI

• The current methods for landslide susceptibility mapping do not fully incorporate soil moisture data. The integration of soil moisture and machine learning models can improve the reliability of susceptibility assessment. There is a need for a more physically realistic depiction of slope hydrological processes in landslide susceptibility mapping.

  Enhancing landslide susceptibility mapping with the integration of soil moisture and machine learning models in Minxian County, China · 2026 · DOI

• The paper identifies the need to understand the deformation and failure mechanisms of shallow landslides in the Qinghai-Tibet Plateau, China. The study highlights the gap in knowledge regarding the combined effects of toe excavation and cryogenic degradation on landslide triggering.

  Deformation and failure mechanisms of shallow landslides in the Qinghai-Tibet Plateau, China: a case study of Zhaxuecun landslide · 2026 · DOI

• The research identifies the need for effective mitigation measures, including drainage and supporting structures, to prevent landslide reactivation.

  Analysis of reactivation mechanisms and deformation patterns in a colluvial landslide: the role of drainage and supporting structures · 2026 · DOI

• The lack of physical interpretability and generalization ability of machine learning models. The need for detailed slope attributes and large datasets for accurate predictions.

  Earthquake-induced landslides hazard assessment with fused newmark displacement and multi-event inventories · 2026 · DOI

• Existing studies largely focus on quantifying the probability of landslide occurrence, neglecting the post-failure characteristics. The spatial variability of saturated hydraulic conductivity is not fully considered in existing studies. There is a need for a coupled hydro-mechanical framework that integrates the Random Limit Equilibrium and Material Point Methods.

  Probabilistic large deformation analysis of rainfall-induced slope failures considering spatial variability of saturated hydraulic conductivity · 2026 · DOI

• The difficulty of collecting data in the field, particularly in areas with limited access. The challenge of integrating multiple datasets and methodologies to provide a predictive framework for landslide hazards.

  Spatial heterogeneity of rock mass degradation and landslide susceptibility in the fault damage zone, Northwestern Himalayas, Pakistan · 2026 · DOI

• The literature has mainly focused on events developed in pyroclastic soils, leaving a gap in understanding other types of landslides. There is a need to understand the factors that contribute to fast landslides in calcareous debris and weathered pyroclastic soils.

  Fast landslide events in calcareous debris and weathered pyroclastic soils: a case study in Southern Italy · 2026 · DOI

• Limited access to post-landslide topographic data. Need for simplified and practical approaches for estimating peak outburst discharge and evaluating downstream channel capacity.

  Monitoring and characteristics of formation and failure of a recurrent landslide dam: A case study of the Danan River landslide dam in Taitung, Taiwan · 2026 · DOI

• The paper identifies the need for practical monitoring and reinforcement measures to prevent further destabilization of the slope. The study highlights the importance of considering the coupled effects of dynamic loads and varying water table levels on slope stability.

  Numerical Simulation and Stability Forecasting of a Critical Slope Considering Future Reservoir Impoundment under Coupled Hydrological and Seismic Scenarios · 2026 · DOI

• The model's generalization capabilities may be limited by the quality and quantity of the training data.

  A deep learning multi-branch network method for landslide susceptibility mapping in the Sichuan-Tibet transportation corridor · 2026 · DOI

• Limited size of training samples for geohazard risk assessment. Spatiotemporal heterogeneity of geohazard data, making it challenging to develop accurate models. Multicollinearity issues among influencing factors, requiring careful selection and evaluation of factors.

  Enhancing geohazard susceptibility mapping with a sample reliability algorithm: a case study in Guangzhou, China · 2026 · DOI

• Stability assessments cannot be easily conducted without access to the site, - Post-landslide topographic data could not be obtained through aerial surveys, - The dam volume was estimated using Eq.

  Monitoring and characteristics of formation and failure of a recurrent landslide dam: A case study of the Danan River landslide dam in Taitung, Taiwan · 2026 · DOI

• Extend the current architecture to a single model capable of detecting landslides, floods, fires, and storms simultaneously, - Incorporate temporal modeling to enable change-based detection for slow-onset hazards like droughts or progressive flooding, - Incorporate rigorous spatial cross-validation, - Integrate large-scale, multi-regional multi-source spatial…

  A frequency-domain and context-guided deep learning transformer architecture for real-time landslide detection · 2026 · DOI

• Further investigation of the mechanisms responsible for the failure of cut slopes in marly formations, - Evaluation of the influence of hydrogeological conditions on slope stability, - Development of more accurate design approaches for high cut…

  Geotechnical and Numerical Analysis of Slope Stability in Potentially Unstable Marly Formations · 2026 · DOI

• The LSA problem cannot be simply regarded as a binary classification problem but should be regarded as a classification problem of an unbalanced dataset - The number of non-landslide samples should be higher than the number of landslide samples in the susceptibility assessment - The applicability of this method under the unbalanced conditions between non-landslide and landslide must be further…

  Landslide susceptibility assessment under non-landslide sampling strategy based on a clustering algorithm · 2026 · DOI

• InSAR technology is particularly susceptible to geometric distortions and temporal decorrelation when monitoring landslide deformation in complex terrain and dense vegetation, - Landslides often occur in mountainous regions characterized by complex terrain and dense vegetation, - The reliability of the SBAS-InSAR monitoring results requires further…

  Multi-model landslide integrated hazard detection using SBAS-InSAR and U-Net segmentation · 2026 · DOI

• Further improvement of the reliability of the SBAS-InSAR monitoring results, - Incorporation of the geological and environmental context of landslides, - Application of deep learning techniques for landslide hazard…

  Multi-model landslide integrated hazard detection using SBAS-InSAR and U-Net segmentation · 2026 · DOI

• Further research could explore the application of the methodology to other regions, - Investigating the impact of other factors on landslide susceptibility, - Developing more advanced models that incorporate additional…

  Enhancing landslide susceptibility mapping with the integration of soil moisture and machine learning models in Minxian County, China · 2026 · DOI

• Employing advanced deep learning techniques, such as knowledge distillation and structural reparameterization, to reduce parameter counts and mitigate overfitting while preserving accuracy. Incorporating metaheuristic optimization algorithms to automatically search for optimal hyperparameters and network architectures.

  A deep learning multi-branch network method for landslide susceptibility mapping in the Sichuan-Tibet transportation corridor · 2026 · DOI

• Limited size of training samples, - Spatiotemporal heterogeneity in real-world scenarios, - Extensive data required on geotechnical and hydrological conditions for physically based…

  Enhancing geohazard susceptibility mapping with a sample reliability algorithm: a case study in Guangzhou, China · 2026 · DOI

• Preliminary geotechnical characterization, - Limited to a specific area in Southern Italy, - No detailed mechanical characterization due to operational…

  Fast landslide events in calcareous debris and weathered pyroclastic soils: a case study in Southern Italy · 2026 · DOI

Most-cited papers in Landslides and related hazards

• Machine learning methods for landslide susceptibility studies: A comparative overview of algorithm performance · Earth-Science Reviews · 2020 · 1,202 citations
• A massive rock and ice avalanche caused the 2021 disaster at Chamoli, Indian Himalaya · Science · 2021 · 707 citations
• Comparisons of heuristic, general statistical and machine learning models for landslide susceptibility prediction and mapping · CATENA · 2020 · 480 citations
• A random forest model of landslide susceptibility mapping based on hyperparameter optimization using Bayes algorithm · Geomorphology · 2020 · 449 citations
• Geomorphic Thresholds: The Concept and Its Applications · Transactions of the Institute of British Geographers · 1979 · 444 citations
• Landslide detection from an open satellite imagery and digital elevation model dataset using attention boosted convolutional neural networks · Landslides · 2020 · 439 citations
• Comparing the prediction performance of a Deep Learning Neural Network model with conventional machine learning models in landslide susceptibility assessment · CATENA · 2020 · 435 citations
• Landslide susceptibility mapping using machine learning algorithms and comparison of their performance at Abha Basin, Asir Region, Saudi Arabia · Geoscience Frontiers · 2020 · 430 citations
• Assessing the predictive capability of ensemble tree methods for landslide susceptibility mapping using XGBoost, gradient boosting machine, and random forest · SN Applied Sciences · 2020 · 429 citations
• Landslide susceptibility mapping using hybrid random forest with GeoDetector and RFE for factor optimization · Geoscience Frontiers · 2021 · 404 citations

Most recent work

• Review article: Deep learning for potential landslide identification: data, models, applications, challenges, and opportunities · Natural hazards and earth system sciences · 2026
• Three-dimensional (3D) laser scanning–based identification of rock mass discontinuities for rockfall modeling using 3D discontinuous deformation analysis · International Journal of Rock Mechanics and Mining Sciences · 2026
• Region similarity assessment for empowering physics-informed transfer learning-based landslide susceptibility mapping · Journal of Rock Mechanics and Geotechnical Engineering · 2026
• Unveiling the long-term cascading effects of the 2018 Baige landslide and subsequent outburst flood with satellite radar observations · Remote Sensing of Environment · 2026
• Insights into evolution of rockfalls on a high-steep slope using UAV photogrammetry and cone complementary-based 3D-DDA · Canadian Geotechnical Journal · 2026
• A surrogate-model-driven framework for rockfall impact assessment on high-speed railway bridge piers: Shape effects, hybrid retrofit, and fragility analysis · Engineering Failure Analysis · 2026
• Ice-rock avalanches in a warming Himalaya indicate pathways toward effective preparedness · Communications Earth & Environment · 2026
• Over 100,000 landslides triggered by typhoon-induced rainfall in North China in July 2023 · Landslides · 2026
• Spatiotemporal prediction and mapping of fractures from successive tunnel faces: an integrated SMF and ST-LSTM framework · Tunnelling and Underground Space Technology · 2026
• Landslide assessment research in the three gorges reservoir area: A review of methodological advances and future directions · Bulletin of Engineering Geology and the Environment · 2026

Related topics in Environmental Science