Objective Landslide dams are natural blockages formed by the rapid accumulation of slope failure materials, typically resulting from mass movements such as rockslides or debris avalanches. Due to gravitational sorting during downslope transport and complex topographic constraints in narrow valleys, these dams generally exhibit inherently three-dimensional (3D), spatially heterogeneous internal structures. However, most existing breach simulation models assume material homogeneity for computational convenience, neglecting the influence of real-world structural non-uniformity. This simplification results in significant deviations in breach morphology and peak discharge predictions, ultimately impairing the reliability of hazard assessments and early warning systems for downstream communities. This study clarifies the fundamental influence of 3D material heterogeneity on the breach process of landslide dams, enhancing the predictive capabilities of numerical models in geohazard risk ma-nagement. Methods A novel 3D landslide dam breach model was developed by coupling Large Eddy Simulation (LES) with sediment mass conservation equations that incorporated phase transitions between solid and suspended states, accurately capturing the erosion dynamics in heterogeneous dams. The model considered the non-directional transport and deposition behavior of sediment under complex flow conditions. A Volume-of-Fluid (VOF) method was employed to simulate the evolution of the free surface and interface, enabling the precise tracking of water-sediment interactions. The model was validated through physical model experiments simulating three structural types: homogeneous, vertically heterogeneous, and laterally heterogeneous dams. Key breach parameters, including incision rate, breach geometry, and outflow hydrographs, were compared to experimental data to verify the model's reliability and accuracy. Results For homogeneous dams, breach behavior varied significantly with the material grain size. Fine-grained dams exhibited rapid, layered erosion, forming triangular longitudinal profiles, with breach durations under 60 seconds and peak discharges reaching 3.32 L/s. In contrast, coarse-grained dams underwent multistage headcut erosion, requiring higher flow shear stress for sediment entrainment. This delayed breach development resulted in longer times to peak (up to 120 s) and reduced peak discharge (2.1 L/s). Reverse vortices formed at the headcut bases enhanced local scouring but exerted limited influence on overall erosion rates. Medium-grained dams exhibited intermediate characteristics, characterized by single-stage headcuts and corresponding breach metrics that fell between those of fine- and coarse-grained dams. As the median grain size increased, breach morphology transitioned from uniform scouring to progressively complex headcut erosion patterns, with delayed peak times and reduced peak flows. For vertically heterogeneous dams, the breach process was susceptible to the configuration of layered materials. The upper layer (V1) influenced breach initiation: in tests where it consisted of fine particles, breach formation was accelerated; in contrast, coarse-grained V1 layers delayed erosion onset and increased upstream impoundment volumes by up to 40%. The middle layer (V2) governed vertical incision rates: fine-grained layers accelerated downward erosion, whereas coarse layers formed headcuts that restricted further deepening. The lower layer (V3) controlled basal stability: coarse-grained foundations enhanced dam resistance to scouring, while fine-grained ones raised undercutting and subsequent collapse of the overlying mass. In addition, an interactive effect was observed between the various layer combinations. For example, "inverse grading" structures (coarse-over-fine) tended to form armor layers on the surface, delaying breach initiation and leading to sudden failure modes with elevated peak discharges. In contrast, "normal grading" (fine-over-coarse) favored headcut erosion. These structural patterns critically influenced both the erosion mechanisms and the hydraulic response. In laterally heterogeneous dams, spatial variability in material properties across dam zones significantly altered breach morphology and flow dynamics. The breach core zone (C1) determined incision and discharge characteristics, while adjacent zones (C2 and D1) influenced breach asymmetry and propagation direction. The D1 zone played a major role in controlling breach widening, whereas D2 exhibited minimal influence. Comparative simulations (Tests 7~9) revealed differences: Test 7 yielded a wide breach, Test 8 produced a narrow and deep breach, and Test 9 exhibited slow incision but extensive lateral expansion. In addition, the deposition of coarse particles within the breach resulted in a reduction of the slope angle and flow velocity, which suppressed the initiation of further coarse particle entrainment and impeded upstream headward erosion. These feedback mechanisms resulted in a lower peak discharge and a more gradual evolution of the breach. Conclusion This study demonstrates that the spatial heterogeneity of dam materials is a primary factor governing the evolution of breaches in landslide dams. Vertically inverse grading structures (coarse-over-fine) contribute to delayed breach initiation and energy storage, often resulting in abrupt failure events with peak discharges exceeding those of homogeneous dams by more than 30%. In laterally heterogeneous dams, the sorting of materials on the overtopping side directly influences the breach depth-to-width ratio (ranging from 0.8 to 2.5), controlling the discharge capacity. A positive feedback mechanism associated with coarse sediment deposition is identified: the accumulation of sediment reduces the breach slope, which in turn decreases flow velocity and further enhances deposition, transforming the breach mode from rapid incision to a slow-release erosion regime. The developed 3D VOF-LES-based hydro-sediment coupled model overcomes the limitations of the traditional homogeneous assumption and, for the first time, enables high-resolution simulation of breach morphology evolution under realistic heterogeneous conditions. With prediction errors maintained within 10% under complex experimental conditions, this model provides a robust tool for enhancing risk assessments and emergency planning in regions prone to landslide dam breaches.
CasagliN, ErminiL, RosatiG.Determining grain size distribution of the material composing landslide dams in the Northern Apennines:Sampling and processing methods[J].Engineering Geology,2003,69(1/2):83‒97. doi:10.1016/s0013-7952(02)00249-1
FanXuanmei, DufresneA, WhiteleyJ,et al.Recent technological and methodological advances for the investigation of landslide dams[J].Earth‒Science Reviews,2021,218:103646. doi:10.1016/j.earscirev.2021.103646
[4]
PengM, ZhangL M.Analysis of human risks due to dam-break floods:Part 1:A new model based on Bayesian networks[J].Natural Hazards,2012,64(1):903‒933. doi:10.1007/s11069-012-0275-5
[5]
ZhouYuanyuan, ShiZhenming, QiuTong,et al.Experimental study on morphological characteristics of landslide dams in different shaped valleys[J].Geomorphology,2022,400:108081. doi:10.1007/s10346-022-01907-6
[6]
XuFugang, YangXingguo, ZhouJiawen,et al.Experimental research on the dam‒break mechanisms of the Jiadanwan landslide dam triggered by the Wenchuan earthquake in China[J].The Scientific World Journal,2013,2013:272363. doi:10.1155/2013/272363
[7]
HuangWei, QiZijie, DuanWengang,et al.Large‒scale experiment on impact of downstream slope on breaching mechanisms of landslide dams[J].Advanced Engineering Sciences,2022,54(3):14‒24.
CaoZhixian, YueZhiyuan, PenderG.Landslide dam failure and flood hydraulics.Part I:Experimental investigation[J].Natural Hazards,2011,59(2):1003‒1019. doi:10.1007/s11069-011-9814-8
[10]
YangYang, CaoShuyou, YangKejun,et al.Experimental study of breach process of landslide dams by overtopping and its initiation mechanisms[J].Journal of Hydrodynamics,Ser B,2015,27(6):872‒883. doi:10.1016/s1001-6058(15)60550-9
[11]
ChenS C, LinT W, ChenC Y.Modeling of natural dam failure modes and downstream riverbed morphological changes with different dam materials in a flume test[J].Engineering Geology,2015,188:148‒158. doi:10.1016/j.enggeo.2015.01.016
[12]
JiangXiangang, WeiYunwei, WuLei,et al.Experimental investigation of failure modes and breaching characteristics of natural dams[J].Geomatics,Natural Hazards and Risk,2018,9(1):33‒48. doi:10.1080/19475705.2017.1407367
[13]
JiangXiangang, HuangJiahua, WeiYunwei,et al.The influence of materials on the breaching process of natural dams[J].Landslides,2018,15(2):243‒255. doi:10.1007/s10346-017-0877-9
[14]
ShiZhenming, WangYouquan, PengMing,et al.Landslide dam deformation analysis under aftershocks using large-scale shaking table tests measured by videogrammetric technique[J].Engineering Geology,2015,186:68‒78. doi:10.1016/j.enggeo.2014.09.008
[15]
ShiZhenming, WangYouquan, PengMing,et al.Characteristics of the landslide dams induced by the 2008 Wenchuan earthquake and dynamic behavior analysis using large‒scale shaking table tests[J].Engineering Geology,2015,194:25‒37. doi:10.1007/978-3-319-09057-3_202
[16]
ItohT, IkedaA, NagayamaT,et al.Hydraulic model tests for propagation of flow and sediment in floods due to breaking of a natural landslide dam during a mountainous torrent[J].International Journal of Sediment Research,2018,33(2):107‒116. doi:10.1016/j.ijsrc.2017.10.001
[17]
PengMing, JiangQiulu, ZhangQingzhao,et al.Stability an-alysis of landslide dams under surge action based on large-scale flume experiments[J].Engineering Geology,2019,259:105191. doi:10.1016/j.enggeo.2019.105191
[18]
PengMing, MaChenyi, ChenHongxin,et al.Experimental study on breaching mechanisms of landslide dams composed of different materials under surge waves[J].Engineering Geology,2021,291:106242. doi:10.1016/j.enggeo.2021.106242
[19]
ShiZhenming, ZhangGongding, PengMing,et al.Experimental investigation on the breaching mechanisms of landslide dams with heterogeneous structures[J].Advanced Engineering Sciences,2023,55(1):129‒140.
WuWeiming.Simplified physically based model of earthen embankment breaching[J].Journal of Hydraulic Engineering,2013,139(8):837‒851. doi:10.1061/(asce)hy.1943-7900.0000741
[22]
ShanYibo, ChenShengshui, ZhongQiming,et al.Development of an empirical model for predicting peak breach flow of landslide dams considering material composition[J].Landslides,2022,19(6):1491‒1518. doi:10.1007/s10346-022-01863-1
[23]
ShenDanyi, ShiZhenming, PengMing,et al.Longevity analysis of landslide dams[J].Landslides,2020,17(8):1797‒1821. doi:10.1007/s10346-020-01386-7
[24]
PengM, ZhangL M, ChangD S,et al.Engineering risk mitigation measures for the landslide dams induced by the 2008 Wenchuan earthquake[J].Engineering Geology,2014,180:68‒84. doi:10.1016/j.enggeo.2014.03.016
[25]
ChenChen, ZhangLimin, XiaoTe,et al.Barrier lake bursting and flood routing in the Yarlung Tsangpo Grand Canyon in October 2018[J].Journal of Hydrology,2020,583:124603. doi:10.1016/j.jhydrol.2020.124603
[26]
CaiYaojun, ChengHaiyun, WuShuaifeng,et al.Breaches of the baige barrier lake:Emergency response and dam breach flood[J].Science China Technological Sciences,2020,63(7):1164‒1176. doi:10.1007/s11431-019-1475-y
[27]
MeiShengyao, ChenShengshui, ZhongQiming,et al.Effects of grain size distribution on landslide dam breaching—Insights from recent cases in China[J].Frontiers in Earth Science,2021,9:658578. doi:10.3389/feart.2021.658578
[28]
LiuWei, HeSiming.Dynamic simulation of a mountain disaster chain:Landslides,barrier lakes,and outburst floods[J].Natural Hazards,2018,90(2):757‒775. doi:10.1007/s11069-017-3073-2
[29]
ZhouG G D, LuXueqiang, XieYunxu,et al.Mechanisms of the non‒uniform breach morphology evolution of landslide dams composed of unconsolidated sediments during overtopping failure[J].Journal of Geophysical Research (Earth Surface),2022,127(10):e2022JF006664. doi:10.1029/2022jf006664
[30]
BuiH H, NguyenG D.Smoothed particle hydrodynamics (SPH) and its applications in geomechanics:From solid fracture to granular behaviour and multiphase flows in porous media[J].Computers and Geotechnics,2021,138:104315. doi:10.1016/j.compgeo.2021.104315
[31]
FengRuofeng, FourtakasG, RogersB D,et al.Two‒phase fully-coupled smoothed particle hydrodynamics(SPH) model for unsaturated soils and its application to rainfall-induced slope collapse[J].Computers and Geotechnics,2022,151:104964. doi:10.1016/j.compgeo.2022.104964
[32]
IssakhovA, BorsikbayevaA, IssakhovA.Dam-break flow on mobile bed through an idealized city:Numerical study[J].Water Resources Management,2022,36(11):4425‒4446. doi:10.1007/s11269-022-03253-7
[33]
MeiShengyao, ChenShengshui, ZhongQiming,et al.Detailed numerical modeling for breach hydrograph and morphology evolution during landslide dam breaching[J].Landslides,2022,19(12):2925‒2949. doi:10.1007/s10346-022-01952-1
[34]
ZhangLucheng, ZhongQiming, YangMeng,et al.Centrifugal model tests and numerical modeling on overtopping-induced breach processes of landslide dams[J].Frontiers in Earth Science,2023,10:1062981. doi:10.3389/feart.2022.1062981
MastbergenD R, Van Den BergJ H.Breaching in fine sands and the generation of sustained turbidity currents in submarine canyons[J].Sedimentology,2003,50(4):625‒637. doi:10.1046/j.1365-3091.2003.00554.x
[37]
Meyer‒PeterE, MüllerR.Formulas for bed‒load transport[C]//Proceedings of the 2nd Meeting of the International Association for Hydraulic Structures Research,1948.