5:15 PM - 7:15 PM
[MIS01-P03] Investigation of landslide dam risk using debris flow numerical simulations: a case study of the 2024 Noto Peninsula Earthquake
Keywords:Landslide, Landslide dams, Debris flow, Noto Peninsula Earthquake, Numerical simulation
When a landslide occurs, the collapsed material may continue to slide as a non-fluidized mass until it is deposited, or it may become fluidized into a debris flow. In either case, if the deposited sediment blocks the river channel, a landslide dam may form. Once a landslide dam forms, it carries the risk of an outburst flood due to breaching. Thus, it is important to investigate the conditions under which a landslide leads to a landslide dam. Previous studies have shown that a wide range of factors—such as riverbed gradient, the confluence angle between the slope and the channel, and the soil and water conditions of the collapsed material—can influence the occurrence of landslide dams. However, when assessing landslide dam risk, it is generally difficult to account for all these factors. Therefore, there is a need for risk evaluation methods that utilize relatively easily obtainable data, such as topographic information.
Here, we explore the use of debris flow numerical simulations to assess landslide dam risk. We assume that landslides can transform into debris flows immediately after they occur (i.e., at their highest fluidity) and then perform debris flow numerical simulations based on the topographic data to identify where sediment might deposit. If the sediment still deposits just downstream of the collapse site, even at this highest fluidity, it indicates that actual landslides under these topographic conditions are highly prone to forming landslide dams. In this study, we apply this approach to examine the risk of landslide dams during the 2024 Noto Peninsula Earthquake.
Our numerical simulations focused on: One collapse area at the headwaters of the Teraji River (in the Machino River system), which has turned into debris flow. Two collapsed areas, one along the downstream of the Teraji River and one along the midstream of the Ushio River (also in the Machino River system), both of which caused landslide dams. One collapse area in the Momijigawa River (Sarudani; Kawarada River system), which has partially transformed into a debris flow while the remainder caused a landslide dam. For these simulations, we used a stony debris flow model based on two-dimensional shallow-water equations, solved via the finite difference method. We employed five-meter mesh digital elevation model (DEM) data from the Geospatial Information Authority of Japan (GSI). As the initial condition, we assigned a debris flow with an average collapse depth to the mesh cells at the collapse sites.
The results showed that, in the headwaters of the Teraji River, where the riverbed gradient in contact with the slope was about 15°, and the confluence angle between the slope and the river was small, debris flowed approximately 500 meters downstream before depositing, although some sediment was deposited immediately below the collapse site. By contrast, for the collapses along the lower reaches of the Teraji River and the midstream of the Ushio River, where the channel slope was about 1° and the confluence angle with the river was large, most of the debris did not flow downstream and was deposited near the collapse area. In the Momijigawa River, with a gradient of about 5° in the channel, most of the debris was deposited immediately below the collapse site, but because the confluence angle was not too large, part of the debris reached the outlet of the valley.
Since actual landslides did not necessarily transform into debris flows immediately, our simulation results did not strictly reproduce the actual riverbed changes. Nevertheless, the findings closely matched the observed presence or absence of landslide dams. These results suggest that using debris flow numerical simulations to evaluate the risk of channel blockage can be effective.
Here, we explore the use of debris flow numerical simulations to assess landslide dam risk. We assume that landslides can transform into debris flows immediately after they occur (i.e., at their highest fluidity) and then perform debris flow numerical simulations based on the topographic data to identify where sediment might deposit. If the sediment still deposits just downstream of the collapse site, even at this highest fluidity, it indicates that actual landslides under these topographic conditions are highly prone to forming landslide dams. In this study, we apply this approach to examine the risk of landslide dams during the 2024 Noto Peninsula Earthquake.
Our numerical simulations focused on: One collapse area at the headwaters of the Teraji River (in the Machino River system), which has turned into debris flow. Two collapsed areas, one along the downstream of the Teraji River and one along the midstream of the Ushio River (also in the Machino River system), both of which caused landslide dams. One collapse area in the Momijigawa River (Sarudani; Kawarada River system), which has partially transformed into a debris flow while the remainder caused a landslide dam. For these simulations, we used a stony debris flow model based on two-dimensional shallow-water equations, solved via the finite difference method. We employed five-meter mesh digital elevation model (DEM) data from the Geospatial Information Authority of Japan (GSI). As the initial condition, we assigned a debris flow with an average collapse depth to the mesh cells at the collapse sites.
The results showed that, in the headwaters of the Teraji River, where the riverbed gradient in contact with the slope was about 15°, and the confluence angle between the slope and the river was small, debris flowed approximately 500 meters downstream before depositing, although some sediment was deposited immediately below the collapse site. By contrast, for the collapses along the lower reaches of the Teraji River and the midstream of the Ushio River, where the channel slope was about 1° and the confluence angle with the river was large, most of the debris did not flow downstream and was deposited near the collapse area. In the Momijigawa River, with a gradient of about 5° in the channel, most of the debris was deposited immediately below the collapse site, but because the confluence angle was not too large, part of the debris reached the outlet of the valley.
Since actual landslides did not necessarily transform into debris flows immediately, our simulation results did not strictly reproduce the actual riverbed changes. Nevertheless, the findings closely matched the observed presence or absence of landslide dams. These results suggest that using debris flow numerical simulations to evaluate the risk of channel blockage can be effective.