Shallow landslide is one of the major causes of sediment related disasters in Japan. To mitigate damages caused by sediment disasters, predicting shallow landslides might be an effective countermeasure. Numerous studies have been conducted to predict shallow landslides by physical, empirical and statistical approaches. The authors have also tried to predict shallow landslide occurrence employing numerical simulation combining computations of rainwater infiltration and slope stability by critical slip surface analysis method, which can precisely specify a shape of slip surface. However, the critical slip surface method has been able to be applied only to a single slope area hitherto. In the present study, to extent the applicable area of the critical slip surface method from a single slope to a watershed, we employed stream tubes instead of rectangular grid to divide the watershed area into individual tubes and elements. The Okanazawa watershed in Izu-Oshima, where shallow landslides and debris flows took place in 2013, are divided by the stream tubes and the critical slip surface method are applied to all individual stream tubes to reproduce the shape, location and timing of the shallow landslides. The results demonstrate that the present method simulates the location of the shallow landslides quite accurately comparing to the conventional analysis by assuming infinite slope, although the timing of landslide occurrence is not well simulated. The present method can be potentially improved employing more reliable rainwater infiltration analysis method. Further investigation is needed to evaluate the applicability of the method to watersheds with different geological and topographical characteristics to use the present method as a practical countermeasure for mitigation of the sediment related disasters.

A large number of slope failure and landslide events occurred inside the Aso caldera following the Kumamoto Earthquake (Mj 7.3) in April 2016. Subsequent rainfalls also caused new slope failures as well as slope failure expansion and promoted secondary sediment movement. This study quantitatively analyzed and summarized the condition and characteristics of these sediment movements. LiDAR data used for analysis are those of three periods―before the earthquake, immediately after the earthquake, and after the subsequent rainfalls. Deviation of coordinates attributable to ground displacement by the earthquake was corrected using Classification and Combined Iterative Closest Point (CCICP) method. After that, altitude difference analysis was performed. The results show that (1) the amount of sediment load from a stream catchment basin by post-earthquake rainfalls was roughly from 10,000 to 200,000 m³/km². (2) In addition, the sediment load was greater at the central cone than at the caldera wall for both the earthquake and rainfall events. (3) However, this trend is more apparent during the rainfall event than at the time of the earthquake. The results also show the sediment load by the rainfalls was 5.9 times greater at the central cone than at the caldera wall. This is possibly attributed to the difference in grain size of sediments deposited by the earthquake. (4) Taking into consideration the sediment balance of the entire caldera, about half of the sediments deposited by the earthquake was underwent secondary erosion during the rainfalls. Furthermore, failure expansion and re-erosion inside the failure site led to new generation of the sediments on the same scale as secondary erosion. As a result, it was considered that the amount of sediments close to the sediment yield by the earthquake was produced by the rainfalls. While the sediment delivery ratio by the earthquake was about 10%, it increased significantly to about 60% due to subsequent heavy rainfalls. Approximately 11.04 million m³ were produced jointly by the earthquake and rainfalls, of which about 6.64 million m³ were deposited inside the caldera.

The sediment discharge volume is observed with hydrophones and turbid meters at 100 observation stations by Sabo work offices of MLIT. At one of those stations, located in Daiya River, Nikko city, the sediment discharge affected by debris flow was observed during the flooding, 2011. The result of the data analysis seemed that the changing of bedload discharge volume had characteristic trend in comparison with other flooding when debris flow didn't occurred. It is supposed that these trend were caused by the difference of the type of sediment supplying, such as debris flow or riverbed material movement and the riverbed condition especially grain size changing. Therefore these trend can be applied to the river basin monitoring for example occurrence of dangerous sediment discharge caused by debris flow and landslide.