Real-scale experiment of debris flow in a natural gulley: key findings and lessons learned

Debris flows moving along steep creeks can bring serious damage to downstream regions. Most debris-flow experiments to date have been conducted in controlled, indoor settings using flumes with straight channels and constant cross-sectional areas. Such experiments cannot accurately verify the dynamic behavior of real debris flows because of their relatively small size, and they often fail to account for complex topographic shapes or the channel bed erosion that occurs during debris flows. With this in mind, a real-scale experiment was conducted in a natural outdoor gully that could credibly represent a true hazard site in South Korea. Video cameras, a load cell, pore pressure transducers, and an ultrasonic sensor were installed at the test site to capture the dynamic behavior of the debris flow. Topographic changes were analyzed before and after the experiment using terrestrial LiDAR. The results showed that the pore-fluid pressure and normal stress for the soil bed were closely related to the flow depth. In addition, the change of the frontal velocity of the debris flow decreased with the decreasing slope angle along the channel but the effect of the width of the channel on the velocity was negligible, although the velocity temporarily increased in the exposed bedrock zone despite a decrease in the slope angle. Furthermore, the erosion depth increased as the frontal velocity increased, with up to four times more erosion in the initiation zone than in the transportation zone.