Real-time monitoring of rainfalls associated with shallow landslides : Development and verification of SWING system based on two rainfall patterns

: A proto type of real-time monitoring and predicting system for a shallow landslide haz-ard during heavy rainfall in Japan, the system with Soil Water Index Normalized by Greatest-value (SWING system), is being developed. The system is based on empirical two rainfall patterns associated with shallow landslide initiation: short-duration, high-intensity (SH) and long-duration, low-intensity (LL) types. The system classifies present rainfall into SH and LL types using Radar/Rain-gauge-Analyzed Precipitation and Soil Water Index provided by the Japan Meteorological Agency. We verified the system by monitoring the major rainfall-induced shallow landslide disasters of July 2010 in the southern part of Kyushu, Yaotsu town in Gifu pref., and Shobara city in Hiroshima pref. Results showed that we succeeded in monitoring and predicting SH and LL types as the conditions likely to initiate shallow landslide in the place. Verification also underscored the need for some ad-ditional development. The system will be useful for real-time shallow landslide warnings, which should be modified and verified using more case studies.