Stress change prior to the major events in the 1989 earthquake swarm off the eastern Izu Peninsula, Japan

Hironori Kawakata,Shizuka Uyama,Hiroshi Ogasawara,Shoji Sekiguchi,Kazuo Mino

doi:10.1186/bf03351926

Abstract

AbstractWe investigate a temporal change in a stress parameter of earthquakes that occurred in the 1989 swarms off the eastern Izu Peninsula in Japan. We use the energy index (EI), which have been monitored in deep South African gold mines for predicting major events, as an estimate of apparent stress that is proportional to the ratio of seismic energy (E) to seismic moment (Mo).EImeasures an excess or shortage inEwith respect to the empirical relationship betweenEandMo. We check thatEIis almost proportional to the apparent stress for the ranges ofMoand frequency for our analysis, althoughEis underestimated due to the artifact of limited frequency band of monitoring. The largest events (M= 5.2 and 5.5) took place off the tip of a vertical crack, which opened associated with magma intrusion. While the opening continued to load the source areas of theM5.2 and 5.5, we find significant decrease inEIprior to the events. Based on the experimental result that the stress decreases when yielding takes place prior to final failure, we interpret this observation as yielding in the vicinity of the large earthquake hypocenters, following a rapid increase in stress caused by magma intrusion.

Highlights

Stress state is one of the most important parameters for describing the physical processes of earthquakes
If the rupture propagates toward a station, the seismic energy E estimated from the seismogram recorded at the station will be overestimated, but the seismic moment Mo is not affected
We examine stress change prior to the largest earthquakes (M = 5.2 and 5.5) in the 1989 earthquake swarm off the eastern Izu Peninsula in Japan using energy index EI, which are estimated from seismic moment and seismic energy

Summary

Introduction

Stress state is one of the most important parameters for describing the physical processes of earthquakes. We can indirectly estimate the stress state of a source region by analyzing the seismograms of preceding small earthquakes. For this to be possible, we need sufficient earthquakes (a large number of foreshocks or a high level of background seismicity). Ide and Beroza (2001) pointed out that the apparent stress tends to be underestimated for small earthquakes due to an artifact, if the effective frequency range is not wide enough. We take into account wholepath anelastic attenuation, radiation pattern, and fault orientation, and we estimate the temporal variation of the stress state from EI, prior to the two largest earthquakes (M = 5.2 and 5.5) in the swarm. We compare our results with strain and tilt data observed in a surrounding region

The 1989 Earthquake Swarm Off the Eastern Izu Peninsula

Discussion

Rapid stress accumulation and yielding prior to the largest earthquakes

Conclusions