When debris flows occur, bridges in mountainous streams may become dangerous when blocked by woody debris. When bridges are blocked with accumulated woody debris, high flow depths can cause the flow to spread widely. However, not all bridges become blocked with woody debris. Many studies have examined bridges blocked with woody debris for gentle slopes, but few studies have been conducted on steep mountain streams with supercritical flow. To better understand the interaction of woody debris and bridges across steep streams, we conducted laboratory experiments using one-pier bridge model and considered factors for the blockage of bridges by woody debris. We used straight rectangular channel flume 7 m in length, 0.3 m width, and with a variable slope. We supplied steady water from upstream end. We supplied woody debris model to the upstream end of the flume at approx.1 second. We set the bridge model 1.5 m upstream from the downstream end of the flume. We used ABS plastic material with a specific weight of 1.05 for the experimental woody debris. We varied the flume slope, water discharge, supply of woody debris, length of wood, height of bridge piers and Froude number. When the woody debris reached the bridge model, the total time of woody debris to pass through became longer from the supplied upstream condition due to rotational motion and diffusion. Because the total time changed, the amount of woody debris per unit time at the bridge was smaller than supplied condition. When examining the blocking conditions, we applied the woody debris condition at bridge model. From the results, we proposed methods to estimate the threshold condition of woody debris blocking at bridge from dimensional analysis. We applied parameters combining the experimental conditions of the bridge model, woody debris model, and hydraulic conditions, as well as the amount of woody debris per unit time required for bridge blocking.

Five machine learning techniques-- classical nonlinear regression (NLR), multi-layer perceptrons (MLP), support vector machines (SVM) with radial-basis function (RBF) kernel, k nearest neighbour (kNN) and decision tree (DT) schemes-- were applied for regression of velocity distribution along the depth of debris flows by using experimental data of steady uniform open-channel flows. Programs coded in Python and package scikit-learn were developed for machine learning analyses. Experimental results of two cases conducted and published by Matsumura and Mizuyama (1990) were adopted for training and prediction curves of the velocity distributions using the five different machine learning techniques. Three theoretical formulas were employed for comparison and investigation, the power-law derived by Takahashi (1978) based on Bagnold dilatant flow, theory modified by Matsumura and Mizuyama (1990), and the two-region formula derived by Su et al. (1993). R-squared scores for each case were calculated to check the fitness of the machine learning results to the experimental data and then to verify the fitness of the theoretical formulas to the machine learning predictions. The quantified results revealed that machine learning schemes provide powerful approaches for building prediction models for velocity distribution of debris flows.

Rheological characteristics are important information for understanding or simulating debris-flow movement. Debris-flow movements involves complex and heterogeneous material with grain size distributions ranging from silt to large rocks. Conventional rheometers are usually limited to measure the rheological parameters of debris-flow of fine particles. Slump-tests has been used to evaluate the flow behaviour of fresh concretes which allow the tested concrete slurries to have larger particles. In this study, the relationship between the parameters obtained from rheometer measurements and slump tests for debris-flow slurries with/without big particles were investigated. At the initial stage, we used fine-sediment slurries to conduct rheological experiments to find the relationship between the parameters obtained from the rheometer measurements and slump tests. The rheological parameters of slurries were measured using the ‘Brookfield DV-III rheometer’. The rheological behavior of the slurry samples used in this study follow the Bingham fluid model. Rheological parameters (i.e., yield stress and viscosity) are affected by the concentration of slurry, indicating that the higher the concentration, the greater the value of the rheological parameters. Slump test was then conducted using the same material samples prepared for rheometer test and the slumped height and spreading diameter of the tested sample were measured. The result shows that the slump height ratio and spreading ratio of the tested slurry decrease with the increase of slurry sediment concentration. Experimental sediment slurry samples were prepared by mixing coarse sands of about 1 mm in diameter. Our results show that the parameters obtained by rheometer measurements are closely related with those by slump tests for the slurries used in this study, indicating that there is a high potential to evaluate rheological parameters of debris-flow using a slump test as an alternative method.

Mountainous regions of Brazil, especially where rural families live, need to be assessed for debris flow. Though debris flows rarely occur in this country, they have caused serious damages including human losses. Computational modeling of debris flows is an important tool to develop hazard maps and to improve the understanding of debris-flow mechanisms, since observed occurrences are rare. Therefore, the objective of the present study was to evaluate the potential for debris flows in a small rural basin (0.712 km²), in the municipality of Alto Feliz, Rio Grande do Sul state (Brazil), by using the Kanako-2D model which was calibrated with another debris flow in the same region. We simulated three scenarios by altering the debris volume and consequently the hydrograph (peak flow and peak time). All the scenarios show that debris flows would impact an existing rural house, even with the smallest potential debris volume. The modeled erosion and deposition areas along the debris flow are similar, with the magnitudes (depths) of erosion and deposition being different among the scenarios. In general, in each transversal section, the most pronounced point of erosion or deposition is almost always at the thalweg location. Along the stream channel, deposition was greatest upstream of an abrupt reduction in slope. The formation of a natural dam is observed at the channel junctions where erosion and deposition alternatively took place. Because of the investigation of the potential of debris flows, the simulation results were not compared with the actual occurrence in the present study. However, the present study could show that computational modeling of debris flow is very important for localities where debris flow occurs and that the debris-flow hazard map is useful for land-use planning.

The design of countermeasures such as barriers and filter dams needs an accurate estimation of the impact load. However, debris flows typically contain poorly sorted grains, whose size can span several orders of magnitude. Large grains can induce impulsive loads on a barrier, and potentially clog the openings designed to induce self-cleaning after an event. The current modeling techniques, mostly based on continuum-based depth-integrated approximations, cannot accurately describe these mechanisms, and analytical approaches often fail to tackle this complexity. In an effort to reproduce a realistic impact load, a sample flow composed of grains is reproduced with a three-dimensional model based on the Discrete Element Method (DEM). The mass impinges upon a barrier with a prescribed velocity. The barrier design is inspired by a monitored dam built on a catchment located in the Italian Alps, which features multiple outlets. The grains can clog the outlets, forming frictional arches. The load pattern on the barrier is analyzed in terms of single-grain impact and of collective behaviors. The impulse transferred by the granular mass to the structure is then used as input for a structural analysis of the barrier through a Finite Element analysis. The results highlight how frictional chains can induce loads that are substantially different from those determined by standard analytical approaches.

Regional modelling of rainfall-induced landslide triggering poses several technical challenges. These events can originate from a number of hydro-mechanical processes, such as soil-strength degradation, development of localized zones of pore-water pressures, liquefaction, among others. At the same time, the interplay between the spatial variability of topographic attributes, soil properties and transient infiltration can lead to a widespread distribution of distinct slope failure mechanisms across the same landscape. To this aim, this contribution describes a simulation platform for the efficient generation of storm-induced, landslide susceptibility maps in which different slope instability mechanisms can be considered. The framework relies on a vectorized finite element (FE) algorithm that performs fully-coupled simulations of transient infiltration in unsaturated soils, while input and output processing stages are linked to a Geographical Information System. To illustrate the capabilities of the proposed framework, the role of several hydro-mechanical processes on the inception of slope instability are first explored (i.e., coupled flow-deformation analyses, constitutive couplings). After this, results of regional-scale simulations are presented, where it is shown that such considerations can affect the computed spatio-temporal patterns of landslide triggering. Lastly, approaches to incorporate uncertainty of input data into landslide susceptibility zonation by using spatially-correlated random fields are discussed. The proposed framework provides an important step towards the development of robust, physically-based models for regional landslide hazard assessment.

Recent observations have shown that debris flows containing fine particles in volcanic regions exhibit greater mobility compared to stony debris flows. Recent researches have described that greater mobility occurred from fine sediment phase shift from solid phase to fluid phase in debris flow. In Japan, debris-flow research and sabo or erosion control planning widely apply the equilibrium concentration methods proposed by Takahashi. For considering fine sediment phase shift with the equilibrium method, it is proposed to set high fluid density. However, the mechanism of phase shift and behaviours of debris flows with fine sediment, are not fully understood. In this study, we conducted hydraulic experiments with sediment particles of two different diameters, defined as fine sediment and coarse sediment. We applied the equilibrium methods and took into account the increased fluid phase density due to the sediment phase shift to fluid. From the results, we found that part of the fine particles contribute to the increase in the fluid phase density. When conducting experiments, not only fine sediment, but some parts of coarse sediment behaved as a fluid. For considering the shift of sediment to fluid phase in debris flows, we presumed that the flow turbulence in debris flow affected. Regarding the sediment concentration, higher total sediment (coarse and fine) concentration increased the fluid phase density. A larger ratio of coarse sediment increased the fluid phase density more than when sediment contained only fine particles. It was speculated to occur from the flow turbulence owing to the mixture condition. Cases with smaller total sediment discharge showed higher fluid phase density though in same sediment concentration. We also found that the larger dimensionless tractive force showed a smaller ratio behaving as fluid phase, which was in contrast with the trends in recent studies.

Large-scale landslides often induce debris flows and cause serious damage to humans. These events typically have water contents in the landslide mass less than 60% and sediment concentrations more than 40%. In spite of high sediment concentrations, landslide-induced debris flows can runout long distances. For large-scale stony debris flows, many previous studies have suggested that coarse gravels behave as a solid phase, whereas fine particles with interstitial water can behave as a fluid phase. We hypothesized this fine sediment might be one of the key processes controlling the long travel distances of landslide-induced debris flows. Here we assumed that the maximum diameter of the fine sediment behave as a fluid phase should vary depending on the friction velocity of the debris flow and the settling velocity of sediments. We conducted detailed field surveys for four landslide-induced debris flows and applied our numerical simulation model to describe the travel distance of the debris flows. Our results show that, if we set the ratio of the friction velocity of debris flow to the settling velocity of sediments around 1 to 4, the simulated travel distance agreed well with our studied four debris flows. We also confirmed that, while the total volume or mean sediment diameter of debris flows varied between study cases, the variability of ratios was small. We believe that our new method and the information it provides, may be helpful for predicting the future risk from the landslide-induced debris flows.

We studied the run-up behavior of debris flows in a small-scale experimental flume using various material compositions, approach flow velocities, and geometries of the obstacle. The experiments were performed with a straight and 4 m to 6 m long flume channel with a circular cross-section of 15 cm top width. The debris flows were released from a head tank. We used three debris-flow mixtures, three channel slopes and either a vertical wall or an adverse slope of 30° as an obstacle. Additional tests were performed using water without sediment. The inclined channel was followed by a 30 cm long horizontal transition reach. Immediately upstream and along this reach we measured the shape of the approaching debris-flow surge with four laser sensors, and determined also the approach flow velocity and depth. The run-up conditions were recorded with a high-speed video camera. The measured run-up conditions were compared with four different theoretical models. The observed run-up conditions differed to some extent between debris-flow mixtures and clear water flows, and there were also some differenced among the debris flow mixtures depending on the relative proportion of coarse particles. The observed run-up heights were generally within the range predicted by the theoretical models, but none of them appears to be universally applicable to the entire range of investigated flow conditions. The commonly used energy principle is not always a conservative method to estimate run-up heights, as has been reported in previous studies.

Debris flows are mass movements that develop along drainage networks and involve generally dense fluids, compose of materials of different grain sizes, as well as woods and variable amounts of water, identified as natural processes that constitute the dynamics and the modeling of the landscape. The areas most susceptible to the occurrence of these processes in Brazil are in the foothills of the Serra do Mar, Serra da Mantiqueira and the Serra Geral, and on the north coast of São Paulo State. In 03/18/1967 there was an important landslide and debris-flows which affected the region of Caraguatatuba and São Sebastião. In this area, there is a pipeline network associated with Petrobras Treatment Units, other enterprises, structures and a large urban area in growth. The aim of this work is to show the results of the back-analysis of the debris-flow events that occurred in 1967 in a mountain area in the Serra do Mar in Caraguatatuba region (São Paulo State, Brazil) with RAMMS numerical simulation, using calibrated input parameters. The inputs were viscosity, DEM, landslide scars as release areas, the density of the debris-flow material, duration of debris-flow process and orthophoto. The modeling results were compared with the deposit area mapped in aerials photos, which was established zones of iso-thickness of the materials. The results showed a good correlation between the area and thickness of deposition modeled and observed. Moreover, the fieldwork and the retro-analysis studies revealed that the Serra do Mar debris flows have a predominantly granular rheological flow and the modeling results showed that the deposition zones are given preferably in regions with slope less than 5º.