Occurrence of sediment disasters is concerned with plural predispositions. Although a lot of topographical features are arranged by field survey, it is difficult to find which combination of predispositions is concerned with occurrence of sediment disasters. Rough set is used as one way to solve this problem. Rough set presents the predisposition and rules that affect the landslides. However, there has been a problem in which a lot of the rules are extracted. In this study, using the concept of rules defined by rough set, propose a method of evaluating the risk of sediment disasters. We have shown how to evaluate quantitatively by adding the number of rules that are the conventional problems.

In 1926 a volcanic mudflow triggered by the eruption of Mt. Tokachi, named “Taisho mudflow”, ran through the Furano River in Hokkaido and caused destructive damage along the course. Several disaster maps presenting the path have been produced since then, while additional information of the arrival time and the levels of damage on the map was considered helpful to establish evacuation plans for the future eruption. In this study, the travelling time of the flow was estimated by interviewing locals who experienced the disaster (Nanri et al., 2004). The damage level was also classified into three groups, High (Ⅰ) : 90% of houses were completely destroyed and 30% of residents were killed, Moderate (Ⅱ) : 30-90% of houses were completely destroyed and 10-30% of residents were killed, and Low (Ⅲ) : houses were partly destroyed and less than 10% of residents were killed. Estimating fluid dynamic force produced by the flow, more than 10³ KN/m for the level Ⅰ, 10²-10³ KN/m for the level Ⅱ, and 10¹-10² KN/m for the level Ⅲ were required during the travel respectively. Applying the relationship between average river bed slope and the force given, the distribution of the damage level was estimated and demonstrated through the path. In summary, it took the flow 20 minutes to travel from the crater through the valley to 0.5 km downstream of the outlet after the eruption, giving a high level of damage. Subsequently, another 10 minutes were required to travel over a flood plain as damaging the area to a moderate degree. Finally, after 40 minutes from the eruption the flow arrived at the Kami-Furano Township, as yielding a low degree of damage.

In order to detect high risk slopes with possible deep landslide hazards in volcanic areas, it is necessary to understand the hydrogeological structure at large scales. We consider Airborne Electromagnetic (AEM) survey to be very effective for large-scale hydrogeological survey. In Aso volcanic area, we used AEM system as well as borehole and microtremor array exploration to study the hydrogeological structure. The results show : 1) An observed unique contour pattern that is the lowest resistivity point at each sample stretches horizontally. The connected low resistivity contours match the groundwater layer detected in the borehole and the location of the spring water nearby. 2) The pattern of low resistivity contour line matches the low-velocity zone of the S-wave of the PS logging in the borehole and microtremor array exploration. Therefore the low resistivity contour indicates a groundwater layer and soft layer, which can be an important factor to cause deep landslide in the studied survey area. This suggests that finding the same low contour pattern could lead to detecting possible deep landslide hazards in other high risk areas.

Recently many steel open-type Sabo dams (hereafter, steel open dams) have been constructed in Japan. These structures consist of steel pipes, and they are designed to allow most soil and small gravel to pass downstream through the gaps. However, if the debris flow has occurred, the steel open dam can capture large rocks. This paper presents experimental and computational approaches to examine the trapping mechanism whereby rocks in a debris flow are captured by a steel open dam. First, a model test was performed by using a flume to examine the mechanism whereby rocks (gravels) in a debris flow are trapped by a steel open dam. Second, a new three-dimensional distinct element method (3-D DEM) was developed in which the rocks were modeled as assembled elements representing natural, irregularly shaped rocks. The assembled elements are consisted of six equally size spherical elements symmetrically arranged in the shape of a regular octahedron. The ratio of protrusions to the diameter of the spherical elements was termed the protruding coefficient β , which was determined by performing the angle of repose test at the site. Third, the proposed 3-D DEM was applied to simulate the trap performance of the model test. Finally, the projected trap performance and shape of sediment by the proposed DEM were compared with the model test result and the DEM using a single spherical element.

Since slope failure induced by heavy rain has recently become a serious problem in Japan, detection of the relationship between the behavior of water flow vector in the slope and the premonitory phenomena of slope failure is attracting increasing attention. However, to date, research on the water flow vector in slopes has not been conducted. In the model experiment using the slope consisting of Masa-soil, simultaneous measurement of the three-dimensional water flow vector and thermal properties based on the quintuple-probe heat-pulse (QPHP) technique was conducted. A high intensity of rainfall (30 mm/h) was applied to the slope surface for 39 min by using an artificial precipitation equipment, and then, the creep of slope and behavior of water flow vector in the slope was observed. Although it was not possible to qualitatively determine the relationship between the water flow vector and timing of crack generation in the slope caused by heavy rain, the behavior of the water flow vector was detected.