Detection and Monitoring of Ongoing Aseismic Slip in the Tokai Region, Central Japan

See allHide authors and affiliations

Science  01 Nov 2002:
Vol. 298, Issue 5595, pp. 1009-1012
DOI: 10.1126/science.1076780

This article has a correction. Please see:


Analysis of global positioning system data shows that the rate of crustal deformations in the Tokai region of Japan, a seismic gap area, changed over the past 18 months. Kalman filtering analysis shows aseismic slip on the plate boundary in the western Tokai region centered on Lake Hamana, adjacent to the anticipated Tokai earthquake source area. The cumulative moment magnitude reaches 6.7 in June 2002 with a relative slip increase northeast of Lake Haman from January 2002. An existence of aseismic slip in the western Tokai supports the hypothesis of a silent event as the cause of uplifting several days before the 1944 Tonankai earthquake.

The Tokai region is located along the Pacific coast of central Japan about 200 km to the southwest of Tokyo. The Suruga trough, a subduction plate boundary between the Eurasian plate and the Philippine Sea plate, runs just off the shore of this area (Fig. 1). In this tectonic setting, the Tokai area has experienced large offshore earthquakes with time intervals of around 150 years. Since the 1854 Tokai earthquake, Richter magnitudeM = 8.4, the Tokai region has been loaded by the Philippine Sea plate and did not rupture at the time of the 1944 Tonankai earthquake (moment magnitude M w = 8.1) (1–3). Continuous global p system (GPS) data since 1994 and historical geodetic survey data by the Geographical Survey Institute of Japan (GSI) for about 100 years indicate a steady strain accumulation in this region (3, 4). On the basis of these pieces of evidence, the Tokai region has been regarded as a seismic gap with a potential of M w = 8 earthquake.

Figure 1

(A) Tectonic setting in and around Japan. The solid lines indicate plate boundaries. Solid circle shows the location of Ohgata GPS site. (B) Magnified map of a rectangular area in (A). The Philippine Sea plate subducts under the Eurasian plate from the Suruga and Sagami troughs. The black arrow at bottom represents the observed ground displacement rate in cm/year for the period between 1997 and 1999 relative to Ohgata station (Fig. 1A). Northwestward motions (small arrows) in the Kanto and Tokai regions are mainly due to the coupling effect between the subducting Philippine Sea plate and the overriding Eurasian plate. This crustal deformation is referred as the steady deformation. Dashed line represents the source area of the future Tokai earthquake adopted by the Central Disaster Management Council of the Japanese government on the basis of many proposed models that were constrained using crustal deformation, seismicity, or other available data [e.g. (4,14)]. The open circles show the location of the selected GPS stations whose time series are shown in Fig. 2.

A large seismo-volcanic event occurred in the Izu islands from July to September 2000 (5, 6), causing southeastward horizontal displacements in the Tokai and northeastward displacements in the Kanto region, central Japan, in a range of up to 3 cm from the dike opening and fault creeping caused by magma intrusion in the Izu Islands (5). After October 2000, the effect of the Izu islands activity faded and the Tokai and Kanto regions returned approximately to the previous normal state of crustal deformation for the period between 1997 and 1999, when displacement rates were almost constant (Fig. 1B). However, another stage of surface displacements gradually became evident in the Tokai region at the beginning of 2001 in the GPS time series.

By subtracting the estimated linear and annual components (7), we detrended the raw time series for the period between April 1996 and June 2002. The detrended time series of selected GPS sites indicates that ground motion started to change in January 2001 (Fig. 2). The spatial pattern of the observed anomalous ground motion for the period between January 2001 and June 2002 (Fig. 3A) indicates southeastward motion of around 2 cm at stations 054, 096, 097, and 098 in the western Tokai region. Because of the steady northwestward motion observed during 1997–1999 (Fig. 1B), the post-2001 data suggests the possibility of aseismic slip between the Philippine Sea plate and the overriding Eurasian plate in the western Tokai region during 2001–2002. Maximum uplift of 2 cm is observed at station 054 east of Lake Hamana where the maximum horizontal motions occur. Crustal deformation on the Izu islands (Fig. 3A) shows a continuation of the 2000 Izu islands event with much less intensity than the peak period between July and August 2000. Southeastward surface displacements on the Izu peninsula are mostly attributed to the Izu islands activity because displacements in the same direction were observed during the 2000 Izu islands event (5).

Figure 2

Detrended time series of east-west (top), north-south (middle), and up-down (bottom) displacements at selected GPS stations [(A) 303, (B) 054, (C) 096] denoted by open circles in Fig. 1B. East, north, and up are positive. Zero slant in these figures means that ground motion is reduced to the steady state deformation in Fig. 1B. The selected stations show clear deviated motion from 2001 onwards. The blue lines indicate computed values from the estimated model.

Figure 3

(A) Solid arrows represent detrended crustal deformation or anomaly from the steady deformation inFig. 1B for the period between January 2001 and June 2002. Open circles numbered 1, 2, 3, and 4 represent 054, 098, 097, and 096 stations, respectively. Southeastward motion in the Tokai region suggests occurrence of a slow earthquake. Color represents observed vertical anomaly motion from the steady state for the same period. Northeast-southwest trending uplift area is observed east of Lake Hamana with the maximum of about 2 cm. (B) Computed crustal deformation for the same period from the estimated model in this study.

On the assumption that aseismic slip is causing the abnormal crustal deformation in the Tokai, we estimated the slip history between the Philippine Sea plate and the overriding plate by a Kalman filter–based analytical method (8, 9) analogous to the time-dependent inversion method (10, 11) using the data in Figs. 2 and 3, with the 2000 Izu islands activity taken into account (12). We used east-west, north-south, and up-down displacement data at 99 selected GPS sites (Fig. 3B) on the Izu island chain and in the Kanto and Tokai regions. As a model region, we use the plate boundary estimated by Ishida (13) and the 2000 Izu islands model (12) (Fig. 4). Adopting these model regions represented by spline surfaces (8), we estimated the slip history between September 2000 and June 2002. We set slip components at zero at the edge of the fault patch as a boundary condition. Furthermore we adopted the condition that the slip motion is southward and eastward for the Tokai aseismic slip.

Figure 4

Estimated time evolution of interplate slip in the Tokai region. Solid arrows represent aseismic slip of the Eurasian plate against the Philippine Sea plate. 1 standard deviation is estimated at about 1 cm. Dashed line indicates the estimated source area of the anticipated Tokai earthquake. Gray solid lines represent isodepth contours of the plate boundary estimated by Ishida (13) in km. Data from (A) 2000.69– 2001.25 (Universal time–Japan time), (B) 2001.25–2001.77, and (C) 2001.77–2002.41 years (in universal time). Solid dots numbered 1 and 2 [in (A)] show locations of two representative points where modeled interplate slip is plotted in fig. S1. Solid lines denoted by “right-lateral,” “dike,” and “left-lateral” in the Izu Islands [in (A)] represent right-lateral, dike, and left-lateral faults of the adopted Izu Islands model (12).

Our analysis for the Tokai aseismic slip gives an area of slip around the Lake Hamana, western Tokai region, close to the estimated Tokai seismic gap (4, 14) (Fig. 1B) from October 2000, which spread slightly through July 2001 (Fig. 4). We think that aseismic slip infiltrating the Tokai earthquake source region is very small and within uncertainties of the identified source area and our model. From November 2001, the estimated slip magnitude decreased near Lake Hamana and a relative slip increase is observed northeast of Lake Hamana (Fig. 4). Slip motion occurs from October 2000 and slows down from November 2001 at point 1, whereas point 2 shows aseismic slip starting after point 1 and the amount of motion nears that at point 1 in 2002 (Fig. 4 and fig. S1). The time evolution of moment released from the estimated Tokai interplate aseismic slip starts in October 2000, increases linearly until July 2001, and continues with a slightly decreased release rate from July 2001 (fig. S2). Cumulative moment amounts to the equivalent of an M w = 6.7 earthquake in June 2002.

The computed ground displacements (Figs. 2 and 3) from the estimated slip model are consistent with the observed displacements. A notable discrepancy is observed in the Boso peninsula, where observed eastward crustal motion is not consistent with the model (Fig. 3).

Previous studies of silent earthquakes suggest a tendency for aseismic slips to occur in weakly coupled regions (8, 15,16), such as the western Tokai region. Re-examination of the past baseline measurements by electromagnetic distance meter suggests a possibility of at least two similar silent events near the area of the event of 2001 during the periods 1978–83 and 1987–91 (17,18). Though the past two silent events in the western Tokai region did not lead to a catastrophic rupture, we do not know whether the 2001 silent earthquake will eventually subside or lead to a catastrophic event. However, one thing is clear: the current aseismic slip in the western Tokai is changing stress state such that it is favorable for the Tokai earthquake with the Coulomb failure stress change (ΔCFS), estimated at about 5 kilopascal (kPa) near Cape Omaezaki (Fig. 4) in the Tokai earthquake source area in June 2002, assuming rigidity of 30 gigapascal (GPa), Poisson's ratio of 0.25, and friction coefficient of 0.2 (19).

The detection of the 2001 silent event supports the hypothesis of a silent event as the cause of the upheavals in the Kakegawa (Fig. 3) (inland of the Tokai region), detected by leveling several days before the 1944 Tonankai earthquake (3), because the present aseismic slip area is close to the assumed silent event in 1944 (20) and because both events caused uplift in Kakegawa (Fig. 3). The 2000 Izu Islands event may have triggered the aseismic slip in the Tokai region, though estimated ΔCFS of ∼0.5 kPa near Lake Hamana in the western Tokai is very small ordinarily (19).

Supporting Online Material

Materials and Methods

Figs. S1 and S2

  • * To whom correspondence should be addressed. E-mail: ozawa{at}


View Abstract

Stay Connected to Science

Navigate This Article