This paper reviews previous studies of volcanic sector collapse, with particular attention to the relationship of the collapse direction to the regional stress field in Japan. The large number of examples of Quaternary age in Japan has allowed analysis of the factors that control the direction of collapse, but researchers have not reached a consensus about those controls. Early researchers concluded that most collapses have been perpendicular to the alignment of parasitic cones and dikes on the volcanic edifice. Subsequent research identified no clearly preferred direction of collapse. More recent research has proposed the following relationship between sector collapse and the regional tectonic stress field. Under regional compressional stress, collapse is subparallel to the axis of maximum horizontal principal stress in response to a local extensional stress field developed along mountain ridges. Under regional extensional stress, collapse is subparallel to the axis of maximum horizontal principal stress. Furthermore, if in an extensional regime the difference between intermediate and minimum principal stresses is small, the relationship of the two can be temporarily reversed in response to stress released due to volcanic activity. Another model, based on analogue experiments, has indicated that slope failures occur at an angle to the axis of regional maximum horizontal principal stress. A strike-slip model for generation of crater breaches can explain this phenomenon, although the model is not applicable to Japanese volcanoes yet. Much of the previous research has suffered from the inclusion of too few volcanoes to provide statistically sound results. Further consideration of the strike-slip model and research using larger datasets published during the past decade may overcome these shortcomings.

This study examined geomorphological location conditions suitable for the slope failures triggered by heavy rainfall in 19 to 20 August 2014 in Hiroshima using random forest (RF) method, a kind of machine learning. We focused on three geomorphic units with different spatial scales: 1) neighborhood of failure heads, 2) first-order basins, and 3) large basins whose low end reach the base of the mountains. The failure heads mainly located at around the steepest slope near the ridge top. The places where slope type changes from convergent (concave) to straight tend to turn into failure head, possibly reflecting the effect of so-called zero-order shallow valley. The first-order basins elevated near the main divide with steep slopes and with rough textures tend to have slope failures, probably reflecting the susceptibility to the rainfall event. In large basins, the number of surface failures normally increases with the increase of the basins area and the basin roughness. Basin area is one of the most important factors to control disaster risk along the foot of the steep mountains where debris flows originated from failures can reach the entrance of the valley.

A probability model of landslide risk evaluated with both triggering and resistance factors is proposed in order to compare spatiotemporally the tendency of landslide occurrence. Assuming a landslide tendency is the function of the intensity of triggering factors (such as rainfall or earthquake), we calculate a risk level as the summation of the product of the occurrence probability and the landslide tendency at an arbitrary intensity of triggering factors. This model enables us to discuss about the landslide risk together with the effects of immunity to rainfall triggering and increase in heavy rain due to climate change. In terms of immunity to rain fall, there are many arguments about relationships among the landslide tendency, critical rainfall to trigger landslides and regional rainfall characteristics. We applied this model to two cases of rainfall triggered landslides reported by previous articles, and compared the calculated risk levels in a rainy and a less-rainy region. The result in the first case shows that, the risk level in a rainy region is much higher than that in a less-rainy region. It is consistent with the landslide density map in Japan, On the contrary, the result in the second case demonstrates that neither the occurrence probability nor the landslide tendency are affected by the regional rainfall characteristics.

In late August 2016, Typhoon Lionrock took an unusual course across Japan and induced heavy rainfall over the Tohoku Region, causing a serious disaster in the Omoto River basin, which left 20 dead and 1 missing in Iwaizumi Town. The disaster included slope failures and debris flows in upper mountainous parts, sediment flooding with driftwoods along tributary channels, and a widespread flood with abundant sediment and countless driftwoods along the main channel in the downstream. This paper firstly overviews the disaster situation and societal characteristics of the Iwaizumi area, and secondly discusses the relationship between its topographical characteristics and damaged areas along the main and tributary channels by analyzing aerial photographs and 1:25,000 topographic maps as well as the results from field investigations. In addition, temporal changes of house distribution and land use since 1947 at three downstream sites were analyzed to investigate relationship between houses and their damage. Our investigation has revealed that the occurrence of flood expansion, overflow and sediment f looding are related to sudden change of local topography, such as valley width, channel sinuosity, and longitudinal profiles from steep to mild. Although most of the houses in the area are historically settled at some distance from the river in the downstream, some of the houses built after 1977 have been found to stand where channel bars used to be and to have suffered severe damage in this flood disaster. It has also found that the f lood expanded over the valley in the downstream and inflicted some damage on most of the houses there, which suggests that this flood disaster was unprecedented in scale and unpreventable even for the residents who had been practicing traditional land use adjusted to local land conditions.

This study revealed wetland development in the Sugenuma Landslide, a huge translational landslide on the Funagata volcano in the northeastern Japan, based on sediment core analysis obtained from 4 wetlands, ¹⁴C dating and tephrochronology. The Sugenuma landslide was divided into several large-scale masses based on their motion direction and surface roughness. Large-scale wetlands are distributed in between the masses, and small-scale ones are formed in linear depressions on the masses. The three wetlands were formed after 11-12 ka probably associated with a large-scale landslide activity and the other was formed by a secondary small-scale landslide a few centuries before the present. This suggests that wetlands have been formed one after another by subsequent landslide activities which sustains distinctive landscape diversity, including wetland, forest and scarp in a landslide.