Sediment disasters are often caused by rainfall. However, the occurrence of rainfall sufficient to cause a sediment disaster requires the convergence and rise of moist air in a certain area and its continuous supply and cooling. Typical sources of moist air are typhoons and fronts. In both cases, it is known that topographic factors lead to the convergence and rise of a warm and wet air flow and amplification of rainfall (orographic rainfall). This amplification effect may cause rainfall to the extent that a sediment disaster occurs. Therefore, in order to predict the rainfall that causes a sediment disaster, it is indispensable to examine weather models that reflect the detailed topography and take the influence of topography into consideration. In this research, focusing on the relationship between topography and rainfall that caused sediment disasters in the basin of the Kanna and Kabura River, the basin of the Uono River, and the upper basin of the Yoshino River from the above mentioned viewpoint, the rainfall prediction is performed by using the WRF weather model, which can express detailed topography with 1 km-mesh, using the time corresponding to seasonal rain fronts, typhoons, or other meso-α-scale phenomena that frequently caused sediment disasters as the prediction time and the GPV of MSM (every 3 hours in 39 hours (meso-α time)) published by the Japan Meteorological Agency as the initial and boundary conditions. In addition, the results are applied to CL and potential rainfalls that might cause a sediment disaster (a debris flow) are assessed. As a result, it was confirmed that the values predicted by WRF exceeded CL, half a day to 1 day (meso-α time) before the occurrence time of 4 debris flow disasters in the above 3 basins. However, it was also confirmed that any values predicted by MSM did not exceed CL within this period of time. Although this result does not mean that it is possible to assess the sediment disaster alert risk in meso-α time for all rainfalls, it does imply a way for technology to predict the sediment disaster alert risk based on rainfall prediction in meso-α time with WRF.

We previously proposed the degree of rainfall (H) as an indicator to evaluate heavy rainfall that causes sediment-related disasters. This indicator was introduced by using two factors for local rainfall : antecedent rainfall and triggering rainfall. Although the proposed degree of heavy rainfall (H) is useful to classify and evaluate extreme torrential rainfall in various places, the effectiveness as a factor to determine sediment-related disasters has yet to be evaluated. To clarify this relationship, we examined the relationship between the number of slope failure occurrence areas (n) and the degree of rainfall (H) for major heavy rain by year in Kure City from 1951 to 1972. There is a strong relationship (n=aH^b, a=3.39, b=4.3 coefficient of determination : R²=0.94), suggesting that slope failure strongly depends on the degree of heavy rainfall (H). Consequently, the degree of heavy rainfall (H) is a useful indicator to evaluate and analyze heavy rainfall causing sediment-related disasters.

Information about damages on houses due to debris flow is one of key information to plan early-warning strategies and to predict magnitude of disasters. However, most of these information was based on the field survey in 1980s and early 1990s. So, we consider that updated information should be necessary to improve strategy of countermeasures against debris flow. Here we surveyed damaged level of houses and deposited sediment depths around these houses for 34 debris flow events occurred in 2009-2018. We surveyed more than 800 damaged houses. We classified these houses into 8 levels in terms of damaged level. We quantified relationship between deposited sediment depth and damaged level of houses. Moreover, we found that effects of catchment topography, such as drainage area and riverbed gradient, were not clear. On the other hand, peak rainfall intensity gave a large impact on relationship between deposited sediment depth and damaged level of houses.

The trees in the riparian zone would fall at the time of floods and those become the driftwoods and cause disasters. The cause of falling of the trees in the riparian zones could be regarded as erosion of the base materials of the trees. In this study, a field survey was performed to clarify the falling limit of the trees by erosion of the base materials. The tractive force at the time of the recent flood was analyzed and the grain size distribution of the materials under the riparian trees was measured. The maximum diameter of the base materials of the tree (d₉₅) was smaller than diameter of stones (d_τmax) that correspond to tractive force at the time of the flood, and it was thought that the root system of the tree inhibited the erosion. As a result of the analysis, it was revealed that, at the base of survive tree, as smaller the root system interval become, the ratio of d_τmax to d₉₅ becomes larger.