Editorial: Monitoring, early warning and mitigation of natural and engineered slopes - volume V

Natural and engineered slopes, along with flash flood-prone mountainous catchments, are highly dynamic geomorphic systems that have become more susceptible to compounded risks with climate change, precipitation extremes, and expanding human activities (Dar, T., Baofeng et al.2026;Yao et al. 2026). The risks include landslides, deformation, subsidence, and flash floods in infrastructure corridors, mining areas, energy sectors, and urbanizing areas, which threaten human safety and asset resilience (Alcántara-Ayala, I. 2025; Shrestha et al. 2025). Recent developments in satellite remote sensing, multisensor monitoring systems, machine learning, hydro-mechanical modelling, and sustainable geotechnical materials have greatly enhanced our capabilities for understanding instability processes, improving prediction and forecasting, and designing mitigation measures (Jiang et al.2026;i Vijaylakshm, S. 2025;Alam et al. 2024). However, critical challenges remain, including multi-scale process coupling, monitoring reliability in complex terrain and dense vegetation, false alarm reduction in warning systems, and the shift to low-carbon, resource-efficient technologies for stabilization. In recognition of the progress achieved in Volume IV of this Research Topic (Qiu et al., 2025), Volume V comprises 16 new papers that contribute to the following key directions: Slope failure mechanisms and rainfall infiltration processes; multi-source monitoring technologies and risk assessment; flash flood warning and susceptibility mapping; sustainable mitigation and the use of solid wastes for reinforcement; and intelligent optimization and prediction by the use of artificial intelligence. Together, these studies reflect an increasingly integrated framework that combines physical process understanding with data-driven intelligence, supporting more reliable early warning and more resilient infrastructure development.Slope hazards: mechanisms, failure evolution, and rainfall-infiltration processes Four of 16 papers in this volume focus on advanced understanding of how hydrological forcing and anthropogenic disturbances drive progressive slope failure. Baofeng et al. investigate the mechanism of group-occurring loess falls due to the sustained discharge of domestic sewage. By employing multi-temporal remote sensing, investigation, and flume experiments, the authors demonstrate the mechanism of progressive basal saturation, collapsible settlement, and the consequent staged slope failure. In the context of engineered slopes, the paper presented by Daocheng et al. describes the development of a numerical model that combines the Richards equation, Van Genuchten model, and the strength reduction method with the FLAC3D code for the investigation of the rainfall-induced instability of photovoltaic (PV) slopes subjected to non-uniform infiltration. Their results demonstrate that panel shading and runoff concentration create hydrological inhomogeneity, leading to asynchronous suction dissipation, localized pore-pressure accumulation, and early instability near the slope toe. The study highlights the dominant role of infiltration heterogeneity in slope failure and emphasizes that neglecting spatial rainfall variability may result in inaccurate stability assessments and underestimated localized risk. Complementing these modelling efforts, Jinwen et al. carried out experimental studies to investigate rainfall-induced instabilities in homogeneous loess slopes subjected to intermittent rainfall. By physically monitoring water content, pore pressures, wetting front, and displacement, the authors found periodic hydrological changes in the loess slopes, indicating heterogeneous deformation responses in the soil mass. The results show that repeated saturationunsaturation cycles, combined with runoff erosion and seepage forces, progressively reduce shear strength and lead to staged failure characterized by toe erosion, fissure expansion, and gully development. Their findings provide valuable insight into the mechanisms of rainfall-induced loess landslides and inform stagespecific mitigation strategies. Finally, beyond rainfall-driven processes, Quan et al. investigated the impact of mining-induced geomorphic changes on soil structure and anti-erodibility of coal mining-induced subsidence in coal mines. Their results highlight the "slope top + 0-20 cm soil layer" as the most susceptible zone for erosion, and the research provides a scientific basis for soil conservation and ecological restoration in the loess mining region.In terms of the development and application of early warning and risk assessment systems, some researchers focus on the development and improvement of monitoring technologies. Four studies have improved the accuracy of deformation detection, data fusion, and the feasibility of monitoring technologies. For instance, Di et al. have proposed a multi-sensor monitoring technology that integrates satellite InSAR for regional detection, GB-InSAR for dense dynamic monitoring, GNSS for 3D deformation constraints, and rain-gauge forcing. The findings have demonstrated that there is a clear link between rainfall and acceleration in deformation, and that both short-duration and long-duration rainfall contribute to landslide activity and early warning. Similarly, Liu Yi et al. developed a multi-source InSAR monitoring and spatial aggregation assessment technology for ultra-high voltage (UHV) transmission corridors, which utilizes Sentinel-1 and RADARSAT-2 time-series analysis (SBAS-InSAR) and Getis Ord Gi* hotspot statistics. The study shows the capability of the multi-source SAR data sets, which can be used for complementary capabilities, broad trend detection, and high-resolution identification of localized subsidence, enabling the rapid screening of high-risk sections for maintenance scheduling and foundation stability evaluation. In order to overcome the difficulty of detecting cracks caused by landslides, Kailun et al. presented an automatic crack extraction algorithm that combines the SBAS-InSAR deformation zoning with UAV LiDAR point cloud scanning. By employing the methods of cross-scale registration (control point matching and ICP), morphological feature construction, probabilistic neural network classification, and optimized edge extraction, the algorithm can be used for robust crack extraction, which is effective for various types of landslides. For scalable and lowbudget monitoring, Daniele et al. introduce a low-cost Arduino-based 3-axis inclinometer designed for structural health monitoring and landslide-related infrastructure safety assessment. With compact size, low power consumption, full-range tilt sensing, and IoT-based transmission and storage, the instrument supports dense sensor deployments where professional-grade networks are cost-prohibitive.In advancing from monitoring to decision-making, two studies enhance flash flood warning systems as well as flash flood susceptibility assessment systems. Xiangning et al. presented a flash flood warning method based on rainfall information, which incorporates spatial stratification of stations as well as a multi-mode discrimination strategy consisting of single station, multi-station, and areal rainfall approaches. The proposed method reduces false alarms without compromising detection efficiency, thus proving the effectiveness of a combination of point sensitivity and regional stability for enhancing flash flood warning robustness in mountainous flash flood-prone villages. For the Tibetan Plateau region, Hongxuan et al. developed a flash flood susceptibility mapping framework based on the combination of logistic regression and a geographical detector method. Using long-term flood records and environmental variables, they quantify nonlinear interaction effects among drivers (notably precipitation and river proximity) and produce sensitivity maps aligned with plateau-scale hydro-climatic and geomorphic controls, supporting dynamic risk zoning and resilience planning.Sustainable mitigation and green reinforcement: solid-waste utilization and stabilized slope performance Although improved monitoring and warning systems are beneficial for the anticipation of hazards, the longterm risk mitigation depends on the development of strong and sustainable engineering techniques. In this context, three studies are proposed for green reinforcement techniques of slopes and subgrade stabilization using strong techniques of solid wastes. For instance, Shujian et al. focus on the stability of fine-grained tailings dams by exploring curing-based improvement strategies under mixed tailings conditions. The study reveals optimal proportions of cementitious materials for improved mechanical strength to ensure the stability of fine tailings dam engineering. Similarly, Li Jun et al. explore recycling techniques of basalt dust wastes through alkali activation for loess improvement in subgrade construction engineering, revealing improved feasibility and reduced costs compared to cement-based techniques. Lei Wang et al. further investigate the stabilization of geopolymer materials using alkali-activated basalt powder and slag, with significant improvements observed in compressive and shear strength, proving that an increase in the proportion of geopolymer materials can lead to reduced displacement of roadbed and slope, while keeping stress distribution more favorable under loading.In addition to the traditional approaches to the analysis and stabilization of landslides, the development of AI prediction and optimization models presents new opportunities to further enhance the efficiency of design and the forecasting of landslide stages. Chaofei et al. integrate a stacking ensemble learning susceptibility model (RF + XGBoost via stacking) with SBAS-InSAR deformation rates to construct a comprehensive susceptibility evaluation matrix for the Lexi Highway. Their approach improves susceptibility accuracy and strengthens discrimination of high-deformation zones, providing more reliable spatial support for risk management along engineering corridors. Tao ma et al. propose a deep transfer learning approach based on CNN and BiLSTM models combined and optimized by Bayesian hyperparameter optimization to effectively predict the evolutionary stages of step-like landslides in the Three Gorges Reservoir Area. The transfer learning approach from a well-monitored source landslide to a data-scarce target landslide improves the accuracy of landslide evolutionary stage prediction and achieves good performance for landslide evolution prediction tasks. Additionally, Tao Deng et al. developed an intelligent optimization strategy for open pit slopes with poor interbedded layers by using a machine learning model prediction (SVM) and multi-object optimization algorithms (NSGA-II and related strategies). The optimization strategy improves the efficiency and accuracy of optimization results by considering multiple performance indicators of slope stability and economic indicators, and the strategy is implemented by developing a Python-based GUI system.The studies included in the present Research Topic collectively reveal progress in (i) mechanism-based understanding of progressive slope instability under rainfall and anthropogenic infiltration, (ii) multi-source monitoring and data fusion for deformation detection, (iii) warning reliability improvements through explainable and operational models, and (iv) sustainable mitigation strategies incorporating solid-waste utilization. Future work is encouraged in the following directions: 1) Developing multi-scale mechanisms of integration, including monitoring, physical tests, and hydromechanical modelling, to reveal mechanisms of progressive failure of slopes under heterogeneous infiltration and complex boundary conditions. 2) Improving operational early warning, with special emphasis on quantification of uncertainty, elimination of false alarms, and interpretable AI that supports decision-making. 3) Developing low-cost dense monitoring networks with IoT sensors and remote sensing technologies to enhance community-level resilience. 4) Developing green stabilization and nature-based mitigation, integrating sustainable materials, ecological restoration, and reduced carbon footprint solutions for slopes and infrastructure corridors.