Rock-ice avalanches (RIAs) are extreme geological hazards characterized by high hazard potential, severe destructiveness, and sudden onset, predominantly occurring in high- altitude glacial regions. Their propagation involves interactions with multiphase media and complex, dynamic evolutionary processes. Amid ongoing climate change, global warming continues to undermine the geological stability of glacierized areas, which is projected to increase the frequency of RIA events. Moreover, the cascading disaster effects of RIAs often trigger large-scale secondary hazards, such as glacial debris flows, glacial lake outburst floods, dammed lake outbursts, and subsequent floods. These pose severe threats to regional ecological environments and engineering infrastructure. It is foreseeable that research on RIA dynamics will become a core topic in the field of glacial geohazards. Against the dual backdrop of significant potential RIA risks in high-altitude engineering zones and the urgent need to advance the currently underdeveloped research framework, this thesis adopts a combination of methods, including field geological surveys, laboratory mechanical strength tests, physical model experiments, Smoothed Particle Hydrodynamics (SPH) numerical simulations, and theoretical analysis to investigate the initiation, propagation, and deposition mechanisms of RIAs. This study reveals multiple dynamic mechanisms underlying RIA evolution, thereby contributing to the refinement of future research frameworks for RIA dynamics and the development of effective disaster risk management systems.
1. Statistical analysis of 50 global RIAs shows they concentrate in warm-glacier regions and are temperature-sensitive. Two geomechanical initiation models are proposed: creep-tensile cracking-sliding for 30–50° slopes with a basal moraine layer, and creep-hanging fracture-collapse for steeper (>50°) slopes without it.
2. Temperature-controlled triaxial tests reveal ice-rock mixtures (IRM) weaken sharply from −8 °C to 2 °C, with rapid loss of cohesion and friction angle near −2 to 0.5 °C; residual strength drops by 69.4% due to melted ice cementation. Numerical analysis confirms the frontal basal moraine layer fails first under warming, breaking ice-bedrock bonding to trigger RIA.
3. Flume tests show dry rock-ice flows with high ice-rock size ratios segregate easily; ice-rich/coarse-ice flows travel farther (↑32%) and form gentler deposits (↓24%). For wet flows, meltwater content <10% reduces mobility, while higher meltwater increases pore pressure and basal lubrication, enhancing mobility.
4. SPH simulations with the μ(i) model reproduce flow stratification and depositional structures (faults, folds). A modified μ(i) model for wet flows reconstructs the 2002 Kolka RIA, showing meltwater exceeding 2.2% drastically accelerates RIA dynamics.
.
 

Rock-ice avalanche, Rock-ice mixtures, Basal moraine layer, Initial mechanism, Dynamic behavior, Depositional mode