Open Access

Complexity in the recurrence of large earthquakes in southwestern Japan: A simulation with an interacting fault system model

Earth, Planets and Space201453:BF03352382

DOI: 10.1186/BF03352382

Received: 1 September 1999

Accepted: 30 May 2000

Published: 18 June 2014


Activity of large earthquakes in southwestern Japan is simulated with a model that incorporates mechanical interactions between faults, including both interplate and intraplate faults. In this simulation, each fault element is assumed to accumulate stress with a constant slip deficit rate and redistribute its accumulated stress to surrounding faults by making a forward (coseismic) slip when the cumulative stress reaches an assumed threshold. The results from the inversion of geodetic data by Hashimoto and Jackson (1993) were used to specify slip deficit rates for these faults. Each fault in this model is divided into four equal-sized elements, two in the length direction and two in the width direction, so that this model can simulate events as small as M6. A complex pattern of seismicity arises from a 10,000-year run of the simulation. The rate of stress accumulation is not necessarily constant for all faults, which may be attributed to the interaction between faults. It is interesting that fluctuations in the amplitude of stress changes with periods of 1,500 years or longer are seen for some inland faults. A variety of sizes of events occur according to the number of simultaneously rupturing elements. Smaller events in which only one element on a fault ruptures frequently occur, but large events with three or more rupturing elements are rarely seen. This implies that the difference between geodetic and geological/seismological strain rates might be made up by smaller events. Simulations indicate that two models with 1 initial conditions may separate by a factor of about 2030 in the state space after hundreds of years. The increase of this distance in the state space slows down or is linear in tome depending on initial conditions.