Statistical analysis of rainfall‐induced sediment‐related disaster events after earthquakes

Hiroaki NAKAYA, Shigetaka TAKIGUCHI, Yuki KISHIMOTO, Taku YAMADA and Hiroshi IKEDA


Sediment‐related disaster alerts are issued based on rainfall cutoffs. Rainfall cutoffs are statistically designed on past rainfall‐induced disaster events without noticeable earthquake influence. They are preset to minimize missed‐alert rates while containing false‐alarm rates for users. Rainfall cutoffs are lowered by 70 or 80% of preset levels after strong earthquakes, for fear of possible missed‐alert increases. Lowering rates were revised to current levels after 2011 earthquakes, and formalized by national government in 2013. Provisionally lowered rainfall cutoffs are raised to the preset level, based on prefectural technical studies. As a result, disaster events after earthquakes have been gathered nationally since 2013. In this study, disaster events are compiled and analyzed with characteristics of related earthquakes and of rainfall in relation to the cutoffs. The results indicate that apparent missed‐alert events can be explained by repetitive aftershock, elapsed durations, and predominant geological characteristics. Lowering rates are to be applied in focused areas to lessen adversarial impacts of false‐alarms.

Key words

sediment‐related disaster alert cutoffs, post‐earthquake response, provisional cutoffs, aftershock