Debris flows pose severe threats to ecologically sensitive mountain areas, especially UNESCO World Heritage sites, where traditional mitigation measures often fail to balance risk reduction with environmental conservation. This study develops an integrated framework that combines high-resolution dynamic simulation with adaptive eco-engineering strategies. Using the Massflow model calibrated with empirical hydrodynamic parameters and 1m-resolution DEM data, we simulated debris flow scenarios under 20- and 50-year rainfall return periods. Under the 50-year event, debris flows reached peak velocities of 6.49 m/s and discharges of 38.33 m³/s. Hazard zoning revealed that high-, medium-, and low-risk zones accounted for 1.16%, 8.07%, and 90.77% of the study area, respectively, threatening 11,745.33 m² of infrastructure in Shuzheng Village. To mitigate these risks while preserving the natural landscape, we designed a novel pine pile–gabion composite dam (PPGD) system integrated with terrain-adaptive ecological restoration. The PPGD raises the erosion base level to dissipate impact energy and features a stepped gabion structure that promotes progressive solid-liquid separation, thereby reducing sediment concentration and enhancing ecological resilience. Simulation results show that three cascaded PPGDs reduced the total inundation extent by 45.78%. At the same time, peak flow velocities downstream of Dams No. 1, No. 2, and No. 3 were decreased by 45.34%, 40.34%, and 37.14%, respectively, compared to upstream values. These findings highlight the effectiveness of coupling dynamic process modeling with eco-engineering interventions for quantitative debris flow risk governance and sustainable landscape protection in ecologically sensitive mountain areas.

GSD (grain size distribution), constitutes a paramount parameter for comprehending the behavior and dynamic mechanics of mass movements, such as debris flows, rock avalanches, sediment transport etc. Alongside traditional sieving methodologies, the past few decades have witnessed a growing interest in photo-sieving, the technique of deducing GSD directly from photographic data. Photo-sieving holds promise for augmenting the spatial and temporal resolution of superficial GSD analysis by virtue of its accessibility, reduced labor intensity, and non-invasive nature. Moreover, the integration of aerial photography within the discipline enables to include the coarse-grained fraction, expanding the scope of particle size analysis beyond the capabilities of traditional sieving. This study introduces a novel algorithm for extracting the coarse-grained fraction using UAV (unmanned aerial vehicle) photography. This novel approach enables us to analyze hectare-scale extents, probing tens of thousands of clasts - surpassing previous similar techniques by two orders of magnitude - and generates a detailed map of the position and dimensions of each particle within the sedimentary system. Furthermore, the algorithm exhibits remarkable resilience in navigating real-world complexities, effectively discerning clasts from vegetation, anthropogenic artifacts, and handling exceptionally large boulders, rendering it suitable for application in diverse field settings. We anticipate that this technique could become a valuable tool for advancing our understanding of debris flow and rock avalanche dynamics, sediment transport processes, and the stability of landslide dams.

Located in the north of the Great Bend of Yarlung Zangbo River, the Purlung Zangbo basin is the largest distribution area of marine glaciers in China. Here is one of the most serious sections for geological disasters in the Sichuan-Tibet transportation corridor, which poses a serious threat to the planning and construction of the Sichuan-Tibet Railway and Expressway and other important infrastructures. In order to assess the risk of geological disasters along the proposed Sichuan-Tibet Railway, more than 80 typical moraine landslides, glacier collapsed debris flows, and glacial lake breakouts in the basin were selected through remote sensing interpretation and field verification, which were carried out ground investigation and geotechnical properties tests in detail. The main types, material sources and key influencing factors of the moraine geological hazards are analyzed by in-situ tests of large-scale straight shear, double ring penetration, matrix suction, dynamic penetration and indoor sample testing methods such as consolidation compression, triaxial shear, and OSL dating, combined with regional temperature, rainfall, multi-stage remote sensing images and geological mapping data. Finally, the starting mechanism and disaster scale of loose moraine are preliminarily simulated by combining the water and thermal product index method and FLOW 3D numerical analysis software. The results show that the physical and mechanical properties of moraine have good statistical relations with its water content, void ratio and clay content. The natural density and compression modulus are negatively linear related to the void ratio, while the vertical permeability coefficient and the free expansion ratio show a quadratic relationship with clay content, and also the shear strength and compressive strength are binary relationship with void ratio and moisture content. The permeability, consolidation degree and free surface characteristics are the key parameters affecting the stability of moraine slope. Moraine located in three depositional positions (the front of modern glacier tongue, the middle reaches of glacial valley and the bank slope of main river ) has different starting modes, which are "shoveled-scraped and migration type", "erosion and blocking-burst type", and "unloaded and permeated type". When it intersects with the main traffic lines, different countermeasures and application should be adopted: avoidance by bridges or tunnels, slope reinforcement, and cutting or subgrade.