Debris flow causes flooding and sediment deposition when it reaches alluvial fans. Many houses have been constructed on alluvial fans, and this can affect debris flow flooding and deposition. In this study, we first conducted a field survey on recent debris flow disasters in Japan; one such disaster, the Izu Oshima sediment disaster, occurred in October 2013. Houses located upstream, in lower areas, and those facing small bridges and crossroads suffered greater damage than those located in other areas. Secondly, we performed numerical simulations using debris flow simulators (KANAKO 2D and Hyper KANAKO) to determine the effect of houses on debris flow flooding and deposition. For the simulations, grid points of the locations of houses were set taking into consideration the height of the houses from the ground elevation. We simulated typical debris flow condition and real disaster cases. The simulation results showed that when houses are present, the spread of debris flow is wide in cross-direction upstream of the houses. Houses also affect the deposition area. The presence of houses increased flooding and deposition damage in some areas, whereas it reduced damage in others. When setting the houses, the areas between the houses were set lower than the houses located at the grid points. Such areas were designated as roads, and the results showed that the flow occurred along the roads, as in real disasters.
Typhoon No. 26 in 2013 attacked the Izu-Oshima Island with record heavy rainfall and caused a disaster resulting from landslides and mud flows. This extreme event is the motivation of our study on how we evaluate hazardous zones at risk of mud flows and how we design structural and non-structural measures accordingly. The present study describes sediment runout processes, mud flow control by means of a guide wall, and a method to evaluate topological conditions in which landslides and mud flows avalanche into unexpected areas. The landslides took place in the western slope of Izu-Oshima, which is only about 2500m wide. Analyses on phase shifting from solid to liquid as well as on mobility of the soil masses suggest that the soil masses released by the landslides transformed directly into mud flows, and that the mud flows developed in size through sediment erosion in their run-out processes. The predicted results by means of a numerical model based on depth-integrated governing equations of sediment-water mixture flow suggest that mud flows could be controlled well using a guide wall, which shows a high possibility of mud flow control using a storage structure with a guide wall. In addition, we propose a simple method to evaluate topological conditions to judge whether mud flows will enter unexpected areas, which will provide a key to identify hazardous zones including even those that have been missed out conventionally.
We propose a numerical simulation method for calculating the vertical distributions of the flow velocity and the sediment concentration in debris flows. Our method is based on the moving particle semi-implicit (MPS) method. We introduce the constitutive equations of Egashira et al. to the MPS method. Numerical simulations of the debris flow are performed by using an existing model based on a shallow water equation, along with our model based on the MPS method. In the condition where the riverbed gradient becomes less steep, there is good agreement with experimental results, including those involving the formation of a convex upward deposition shape in the initial deposition process. Results for the initial deposition process are not produced with existing simulation method based on a shallow water equation. Further, our model can yield clear results when the upper and lower layers have different flow directions in a numerical simulation of the collapse of a natural dam by overtopping.
Particle size-segregation can have an important feedback on the bulk flow of geophysical granular avalanches. As a polydisperse material travels downhill the larger particles rise to the surface, where they are preferentially sheared to the flow front. This coarse-rich region experiences a greater resistance to motion and the large particles are shouldered aside to form lateral levees. Wider flows may break down into a series of these lobate, ‘finger-like’ structures. In either case, the static leveed regions channelise the finer, more mobile interior, causing the resulting run-out distances to be significantly enhanced. Modelling segregation-mobility feedback effects is therefore crucial for hazard mitigation. A new class of depth-averaged continuum models is introduced that describes the transport of large particles as well as the granular rheology. The feedback arises from a basal friction law that is composition dependent, implying greater friction where there are more large particles. Numerical simulations are used to show the spontaneous formation of leveed fingers.
The first objective of the work is to test a cost-effective tool for the collection of debris flows (DF) field data such as volumes, peak flow depths and deposit depths. Secondly, we show how these data can be used for the calibration of a depth-averaged propagation model. The case study is a DF of pumiceous sediments, occurred in the Amalfi Coast (Southern Italy) in October 2013. The DF path is a steep channel, ending in a small debris fan delimitated by a gabion wall. The risk is high because DFs, having a return period of just few years, overtop the wall and hit a busy road. Both terrestrial laser scanner (TLS) and photogrammetric techniques were employed to survey the topography, before and after the event under study. The images of the channel were taken from an unmanned aerial vehicle (UAV). Digital terrain models (DTM) were obtained pre and post event while the traces left by the DF along the channel banks allowed the estimation of the peak flow depths.A finite volume two-dimensional numerical code (FLATModel), based on shallow-water equations, was used for modelling the propagation and deposition of the DF under study. Both Voellmy and pure Coulomb friction resistance laws were tested. The numerically predicted deposit was compared to the post event DTM. Such comparisons showed a good agreement in terms of both depths and shape of deposit. The calibrated model could be used to predict the DFs run-out distances in similar contexts.
Laboratory flume experiments showed the different behavior of sediment grains and wood pieces in debris flows (wood-sediment-water mixture flows) at an open check dam model. In the experiments, the debris flows were caused by putting some wood pieces on the initial movable bed and dropping the other wood pieces on the surface of the subsequent flow part. This condition resulted in the concentration of the former pieces at the flow front and the latter pieces on the subsequent flow surface. The latter pieces trapped in the check dam model were piled up on the former pieces trapped in the check dam. The sediment followed the pieces accumulating at the flow front. At the same time, sediment deposition was caused behind the trapped wood pieces.
Using a large force plate as installed in the channel bed at the Illgraben debris-flow observation station, we illustrate flow properties of three typical types of debris flow which occur there, single-surge debris flows, multiple-surge flows, and more complicated surges with many individual roll waves. In all three types of flow, the median value of the normal and shear stresses are largest at the front or fronts of each surge. Additionally, the fluctuations in normal and shear stress are also largest at the flow front, reaching values up to 1.5 times larger than the median value. The presence of large force fluctuations at the front of the surges is in agreement with previously-published results which show that the rate of erosion is largest within the first minute of the arrival of a debris-flow surge.