Segmentation and Holocene surface faulting on the Median Tectonic Line, southwest Japan

The Median Tectonic Line (MTL) active fault system is a 300-km-long, arc-parallel, right-lateral strike-slip fault related to oblique subduction of the Philippine Sea plate beneath the Eurasian plate at the Nankai trough, southwest Japan. The fault on Shikoku Island has a slip rate of 5-10 mm/yr, one of the highest in onshore Japan. We identified 12 geometric segments along the MTL separated by discontinuities such as en echelon steps, bends, changes in strike, and gaps in the surface trace. A chronology of latest Holocene surface faulting on the MTL was constructed based on previously published and new work including trenching, geomorphic mapping, and acoustic profiling and piston coring of the submarine extension of the fault zone. The MTL probably does not rupture along its entire length in a single earthquake but instead consists of multiple ''earthquake segments'' that rupture independently of one another. The recurrence interval and surficial offset for surface-rupturing earthquakes at four individual sites on the MTL in Shikoku Island are 1000-3000 years and 5-8 m, respectively. The relatively long recurrence interval and large surficial offset per earthquake suggest large seismic moment release during surface-rupturing earthquakes on the MTL, as commonly observed for Japanese moderate to large intraplate earthquakes. Because the predicted rupture lengths based on empirical relations between surface displacement and rupture length are considerably larger than those of individual segments, large earthquakes on the MTL probably rupture multiple geometrically defined segments. The fault was previously thought to be unruptured at least in the past 1000 years based on the absence of historical evidence for destructive earthquakes. However, trench excavations show that part of the MTL ruptured most recently during or after the 16th century A.D. with at least 6.9 +/- 0.7 m of slip along the main fault trace. This rupture may be correlated to the 1596 Keicho-Kinki earthquake, but further geological and historical investigations are needed to confirm this hypothesis.