Seismic characterization of a landslide dam failure hazard chain: Insights into flow dynamics and implications for warning

Abstract

Landslide dam failure hazard chains (i.e., landslide dam breaches → outburst debris flow → river blocking) are common in alpine valley regions worldwide, and they pose a great threat to human safety and infrastructure. However, the formation and dynamic processes of cascading hazards are still poorly understood, which limits our ability to recognize warning signals for this multi-hazard chain. We conducted 12 large-scale field tests on the multi-hazard chain induced by landslide dam failure, measuring the flow dynamics and the resulting seismic signals. The results showed that river blockage caused by a dam-breaching debris-flow underwent stages of injection and initial deposition, river damming and equilibration, during which the evolution of seismic signals can help interpret the physical processes and mechanisms involved. The radiated energy level resulting from the dam breaching and the debris flow increased with the diameter, speed and volume of the fluid-particles mixture; while the frequency of the seismic data increased with the speed of the fluid-mixture particles but decreased with the diameter of the particles. Greater tributary inflow discharge or dam height generally resulted in larger and more destructive outburst debris flows, characterized by greater seismic energy and a wider frequency domain, resulting in a greater degree of obstruction within the receiving channel. We also show that the seismic technique employed is an efficient means of detecting and characterizing landslide dam failure hazard chains, providing downstream warnings of a river blockage by estimating the critical blockage conditions, such as the tributary flow velocity, sediment concentration, and kinetic energy. This study provides a scientific perspective for understanding the dynamic signatures of the entire hazard chain evolutionary process and it highlights opportunities to use seismic data for multi-hazard chain research and warning.