Do Vegetation Root Systems Affect Landslide Mobility? A Flume Experiment

Landslides are common natural hazards that greatly impact lives and property worldwide. The magnitude of landslide impacts depends strongly on how far landslide sediments travel, widely known as landslide mobility. Numerous studies showed that landslide mobility is complex, but largely affected by initial water content during landslide initiation. Here, water acts as a medium that carries the collapsed landslide mass downslope. Vegetation root systems may alter the initial water content by modifying the flow path within the soil. The mechanical reinforcement of root systems may also limit the spatial propagation of the landslide mass. Thus, vegetation root systems may exert significant effects on landslide mobility. Nevertheless, effects of root systems on landslide mobility have rarely been discussed in landslide studies. The objective of this study is to evaluate the effect of rooting systems on landslide mobility.

A flume constructed at a 1:70 scale was used to evaluate the effect of root systems on landslide mobility. The flume consisted of two segments representing landslide initiation (120 cm long, 35° inclination) and deposition (150 cm long, 35° inclination). All segments were 80 cm wide, 15 cm high, and constructed with 1-cm thick acrylic material. Sand (density=1.4 g/cm³, D₅₀=0.23 mm) was placed in the initiation segment to a depth of 10 cm. For conditions with vegetation (V), we grew pea (Pisum sativum L.) bean sprouts in the sand to simulate the root system. Sprouts were grown at 3 cm intervals for two weeks to simulate the root system on 2200 stem/ha of Japanese cedar forest. To initiate landslides, 90 mm/h of rainfall was applied via nozzles installed at 2 m above the flume. Timing of landslide initiation was then measured. Water content was also measured by TDR sensors installed at 3 and 7 cm depths below the soil surface. The L/H ratio was estimated based on total travel distance and total descent height of the landslide mass.

Vegetated conditions (V; n=3) were more stable than non-vegetated conditions (NV; n=3). Indeed, landslides initiated at 889-959 s (SD=41 s) on V, while on NV was 510-519 s (SD=5 s). Mean volumetric water content during landslide initiation was 0.2-0.22 (SD=0.01) on V, while on NV was 0.16-0.2 (SD=0.02). Because V had higher water content than NV, V was 1.2-1.4 times more mobile than NV. The L/H was 2.2-2.4 (SD=0.09) on V, while on NV it was 1.7-1.8 (SD=0.06). In general, vegetation root systems maintain slope stability by adding more cohesion to soils. Due to this reinforcement, greater gravitational forces and pore water pressure are needed to destabilize the slope. This consequently elevates the threshold of water content for landslide initiation. Since water content greatly influences mobility, wetter conditions enhance the mobility of the collapsed landslide mass. Our findings concur with previous studies that root reinforcement can mitigate slope instability. However, we highlight that such reinforcement can also enhance the mobility, which may elevate the potential impacts of landslides. We further investigate the effect of various stem densities on landslide mobility.