Abstract
Large-scale sediment-related disasters have been increasing in recent years [Uchida et al., 2009; Inoue and
Doshida, 2012], such as the deep-seated catastrophic landslides caused by Typhoon Talas (T1112) in the
Kii Peninsula, southwest Japan, in 2012. The formation and collapse processes of landslide dams strongly
influence water and sediment runoff. When a large-scale landslide dam collapses, the peak discharge of
downstream flooding can sometimes become several times as large as the inflow discharge upstream. Such
abrupt increases in flow discharge often cause serious disasters in downstream areas. The objective of this
study was to support prediction systems for hazards caused by debris flows. We developed a crisis
management system to contribute to evacuation decisions and actions by providing information to residents.
We adopted Hyper KANAKO as the basic tool of the crisis management system. This system can also
provide results of temporal changes such as flow depth and sedimentation depth for the administrator. The
time-series images are associated with a world file that responds to geographic information from a 2D
topographic model. It is possible to check the simulation results on a GIS as soon as calculations have been
performed. In addition to the existing system, we developed functions to reproduce the side-bank erosion
caused by landslide dam overtopping. This new system requires setting the height of the landslide dam, the
width of the river channel, the initial overflow channel width of the landslide dam, and the numerical
constant for side-bank erosion velocity as calculation conditions. The system can display the widening of
the overflow channel of the landslide dam caused by overtopping, and analyze a series of phenomena
related to flooding from the overtopping erosion of the landslide dam. Moreover this system can assist in
developing more effective countermeasures and hazard maps for disaster mitigation.
