Dynamic fragmentation characteristics of rock avalanches based on multi-source data synergism

Dynamic fragmentation is the process by which a rock mass continuously disintegrates during the high-speed movement of a large-scale rock avalanche, which significantly influencing the dynamic behavior of rock avalanche. Based on the multi-source data integration approach that combines unmanned aerial vehicle (UAV) image recognition, seismic signal analysis, and discrete element method (DEM) numerical simulations, a joint analysis was conducted to characterize the dynamic fragmentation process and the spatial distribution of the 2017 Nayong rock avalanche in Guizhou, China. The DEM results indicate that the formation of a basal shear failure zone in the source area was the key mechanism triggering large-scale toppling and collapse. The simulated maximum sliding velocity (36.5 m/s) aligns closely with the result from seismic inversion (31.6 m/s), which verifies the reliability of the numerical model. Dynamic fragmentation indicators show that the velocity, relative breakage ratio, fracture number growth rate, fractal dimension D, and the shape parameter β all change significantly within the first 20 seconds—reaching 31.6 m/s, 0.85, 2000/s, 2.14, and 3.96 respectively—suggesting that the primary fragmentation process occurs at the initial stage. The results from UAV image recognition and DEM simulations both reveal that the particle size of rock debris decreases with increasing planar transport distance. In addition, the DEM simulations show that larger rock debris predominantly accumulates in the upper part of the deposit, while smaller debris tends to concentrate in the basal layer.