Effects of active fault types on earthquake-induced deep-seated landslides: A study of historical cases in Japan

We investigated the relationship between the distribution of deep-seated landslides (DSLs; landslide volume > 105 m3) induced by inland earthquakes as well as the distribution of corresponding active faults by compiling preexisting documents on historical DSL occurrence records. The following points are found: (1) The DSLs induced by reverse fault earthquakes tend to occur equally within a wide range of about 20 km from the faults, whilst > 80% of DSLs induced by strike-slip fault earthquakes are concentrated within a small range of about 5 km from the faults. (2) Most of the DSLs are distributed on the hanging wall side of the active faults. (3) The distribution of some historical DSLs may reflect the directivity of the seismic waves of the historical earthquakes. The minimum peak ground velocity (PGV) and peak ground acceleration (PGA) during earthquakes that can induce DSLs are estimated to be 15-20 cm s− 1 and 300-400 cm s− 2, although most of the DSLs examined were induced by strike-slip fault earthquakes with PGV > 60 cm s− 1 and PGA > 900 cm s− 2. This discrepancy may be attributed to a possible limitation of the proposed equation, which was established mainly for cases of reverse fault earthquakes. It is implied that the type of fault, the side of the epicenter location (hanging wall/footwall side), and the directivity of seismic waves should be considered for assessing the distribution of ground motion in terms of DSL occurrence, and that these factors may reflect the level of risk for earthquake-induced landslides around active faults.