Megathrust earthquakes are responsible for some of the most devastating natural disasters1. To better understand the physical mechanisms of earthquake generation, subduction zones worldwide are continuously monitored with geophysical instrumentation. One key strategy is to install stations that record signals from Global Navigation Satellite Systems2,3 (GNSS), enabling us to track the non-steady surface motion of the subducting and overriding plates before, during and after the largest events4,5,6. Here we use a recently developed trajectory modelling approach7 that is designed to isolate secular tectonic motions from the daily GNSS time series to show that the 2010 Maule, Chile (moment magnitude 8.8) and 2011 Tohoku-oki, Japan (moment magnitude 9.0) earthquakes were preceded by reversals of 4–8 millimetres in surface displacement that lasted several months and spanned thousands of kilometres. Modelling of the surface displacement reversal that occurred before the Tohoku-oki earthquake suggests an initial slow slip followed by a sudden pulldown of the Philippine Sea slab so rapid that it caused a viscoelastic rebound across the whole of Japan. Therefore, to understand better when large earthquakes are imminent, we must consider not only the evolution of plate interface frictional processes but also the dynamic boundary conditions from deeper subduction processes, such as sudden densification of metastable slab.
Subscribe to Journal
Get full journal access for 1 year
only $3.90 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
The daily GNSS displacement time series and the predicted displacements from fluid loading models are available in a data supplement to this paper66.
Kajitani, Y., Chang, S. E. & Tatano, H. Economic impacts of the 2011 Tohoku-Oki earthquake and tsunami. Earthq. Spectra 29, 457–478 (2013).
Sagiya, T., Miyazaki, S. I. & Tada, T. Continuous GPS array and present-day crustal deformation of Japan. Pure Appl. Geophys. 157, 2303–2322 (2000).
Báez, J. C. et al. The Chilean GNSS network: current status and progress toward early warning applications. Seismol. Res. Lett. 89, 1546–1554 (2018).
Heki, K. & Mitsui, Y. Accelerated Pacific plate subduction following interplate thrust earthquakes at the Japan trench. Earth Planet. Sci. Lett. 363, 44–49 (2013).
Loveless, J. P. & Meade, B. J. Two decades of spatiotemporal variations in subduction zone coupling offshore Japan. Earth Planet. Sci. Lett. 436, 19–30 (2016).
Melnick, D. et al. The super-interseismic phase of the megathrust earthquake cycle in Chile. Geophys. Res. Lett. 44, 784–791 (2017).
Bedford, J. & Bevis, M. Greedy automatic signal decomposition and its application to daily GPS time series. J. Geophys. Res. Solid Earth 123, 6992–7003 (2018).
Reid, H. F. The Mechanics of the Earthquake, The California Earthquake of April 18, 1906, Report of the State Investigation Commission Vol. 2, 16–28 (Carnegie Institution of Washington, 1910).
Schurr, B. et al. Gradual unlocking of plate boundary controlled initiation of the 2014 Iquique earthquake. Nature 512, 299–302 (2014).
Mavrommatis, A. P., Segall, P. & Johnson, K. M. A decadal-scale deformation transient prior to the 2011 Mw 9.0 Tohoku-oki earthquake. Geophys. Res. Lett. 41, 4486–4494 (2014).
Wang, K. et al. Learning from crustal deformation associated with the M9 2011 Tohoku-oki earthquake. Geosphere 14, 552–571 (2018).
Kato, A. et al. Propagation of slow slip leading up to the 2011 Mw 9.0 Tohoku-Oki earthquake. Science 335, 705–708 (2012).
Johnson, P. A. et al. Acoustic emission and microslip precursors to stick-slip failure in sheared granular material. Geophys. Res. Lett. 40, 5627–5631 (2013).
Kaproth, B. M. & Marone, C. Slow earthquakes, preseismic velocity changes, and the origin of slow frictional stick-slip. Science 341, 1229–1232 (2013).
Lohman, R. B. & Murray, J. R. The SCEC geodetic transient detection validation exercise. Seismol. Res. Lett. 84, 419–425 (2013).
Dill, R. & Dobslaw, H. Numerical simulations of global-scale high-resolution hydrological crustal deformations. J. Geophys. Res. Solid Earth 118, 5008–5017 (2013).
van Dam, T. et al. Crustal displacements due to continental water loading. Geophys. Res. Lett. 28, 651–654 (2001).
Heki, K. Snow load and seasonal variation of earthquake occurrence in Japan. Earth Planet. Sci. Lett. 207, 159–164 (2003).
Wdowinski, S., Bock, Y., Zhang, J., Fang, P. & Genrich, J. Southern California permanent GPS geodetic array: spatial filtering of daily positions for estimating coseismic and postseismic displacements induced by the 1992 Landers earthquake. J. Geophys. Res. Solid Earth 102, 18057–18070 (1997).
Bouchon, M. et al. Potential slab deformation and plunge prior to the Tohoku, Iquique and Maule earthquakes. Nat. Geosci. 9, 380–383 (2016).
Gardonio, B. et al. Seismic activity preceding the 2011 M w 9.0 Tohoku earthquake, Japan, analyzed with multidimensional template matching. J. Geophys. Res. Solid Earth 124, 6815–6831 (2019).
Ito, Y. et al. Episodic slow slip events in the Japan subduction zone before the 2011 Tohoku-Oki earthquake. Tectonophysics 600, 14–26 (2013).
Ito, Y., Hino, R., Suzuki, S. & Kaneda, Y. Episodic tremor and slip near the Japan Trench prior to the 2011 Tohoku-Oki earthquake. Geophys. Res. Lett. 42, 1725–1731 (2015).
Katakami, S. et al. Spatiotemporal variation of tectonic tremor activity before the Tohoku-Oki earthquake. J. Geophys. Res. Solid Earth 123, 9676–9688 (2018).
Becker, T. W., Hashima, A., Freed, A. M. & Sato, H. Stress change before and after the 2011 M9 Tohoku-oki earthquake. Earth Planet. Sci. Lett. 504, 174–184 (2018).
Yokota, Y. & Koketsu, K. A very long-term transient event preceding the 2011 Tohoku earthquake. Nat. Commun. 6, 5934 (2015).
Pollitz, F. F. Coseismic deformation from earthquake faulting on a layered spherical Earth. Geophys. J. Int. 125, 1–14 (1996).
Hayes, G. P. et al. Slab2, a comprehensive subduction zone geometry model. Science 362, 58–61 (2018).
Rogers, G. & Dragert, H. Episodic tremor and slip on the Cascadia subduction zone: the chatter of silent slip. Science 300, 1942–1943 (2003).
Yamashita, Y. et al. Migrating tremor off southern Kyushu as evidence for slow slip of a shallow subduction interface. Science 348, 676–679 (2015).
Ide, S., Beroza, G. C., Shelly, D. R. & Uchide, T. A scaling law for slow earthquakes. Nature 447, 76–79 (2007).
Savage, J. C. A dislocation model of strain accumulation and release at a subduction zone. J. Geophys. Res. Solid Earth 88, 4984–4996 (1983).
Kodaira, S. et al. High pore fluid pressure may cause silent slip in the Nankai Trough. Science 304, 1295–1298 (2004).
Tsuji, Y., Nakajima, J. & Hasegawa, A. Tomographic evidence for hydrated oceanic crust of the Pacific slab beneath northeastern Japan: implications for water transportation in subduction zones. Geophys. Res. Lett. 35, L14308 (2008).
Liu, X., Zhao, D. & Li, S. Seismic attenuation tomography of the Northeast Japan arc: insight into the 2011 Tohoku earthquake (Mw 9.0) and subduction dynamics. J. Geophys. Res. Solid Earth 119, 1094–1118 (2014).
Taetz, S., John, T., Bröcker, M., Spandler, C. & Stracke, A. Fast intraslab fluid-flow events linked to pulses of high pore fluid pressure at the subducted plate interface. Earth Planet. Sci. Lett. 482, 33–43 (2018).
Ujiie, K. Chemical origin of tectonic tremor. Nat. Geosci. 12, 962–963 (2019).
Incel, S. et al. Reaction-induced embrittlement of the lower continental crust. Geology 47, 235–238 (2019).
Austrheim, H. Eclogitization of lower crustal granulites by fluid migration through shear zones. Earth Planet. Sci. Lett. 81, 221–232 (1987).
John, T. & Schenk, V. Partial eclogitisation of gabbroic rocks in a late Precambrian subduction zone (Zambia): prograde metamorphism triggered by fluid infiltration. Contrib. Mineral. Petrol. 146, 174–191 (2003).
Seno, T. & Maruyama, S. Paleogeographic reconstruction and origin of the Philippine Sea. Tectonophysics 102, 53–84 (1984).
Bach, W. & Früh-Green, G. L. Alteration of the oceanic lithosphere and implications for seafloor processes. Elements 6, 173–178 (2010).
Burnley, P. C., Green, H. W. & Prior, D. J. Faulting associated with the olivine to spinel transformation in Mg2GeO4 and its implications for deep-focus earthquakes. J. Geophys. Res. Solid Earth 96, 425–443 (1991).
Thielmann, M., Rozel, A., Kaus, B. J. P. & Ricard, Y. Intermediate-depth earthquake generation and shear zone formation caused by grain size reduction and shear heating. Geology 43, 791–794 (2015).
Fukao, Y. & Obayashi, M. Subducted slabs stagnant above, penetrating through, and trapped below the 660 km discontinuity. J. Geophys. Res. Solid Earth 118, 5920–5938 (2013).
Kawakatsu, H. & Yoshioka, S. Metastable olivine wedge and deep dry cold slab beneath southwest Japan. Earth Planet. Sci. Lett. 303, 1–10 (2011).
Panet, I., Bonvalot, S., Narteau, C., Remy, D. & Lemoine, J. M. Migrating pattern of deformation prior to the Tohoku-Oki earthquake revealed by GRACE data. Nat. Geosci. 11, 367–373 (2018).
Wang, L. & Burgmann, R. Statistical significance of precursory gravity changes before the 2011 M w 9.0 Tohoku-Oki earthquake. Geophys. Res. Lett. 46, 7323–7332 (2019).
Blewitt, G., Hammond, W. C. & Kreemer, C. Harnessing the GPS data explosion for interdisciplinary science. Eos 99, https://doi.org/10.1029/2018EO104623 (2018).
Gendt, G. et al. GFZ Analysis Center of IGS. Annual Report for 2013 1–10 (GFZ German Research Centre for Geosciences, 2013); ftp://ftp.gfz-potsdam.de/GNSS/DOCS/IGS_repro2/GFZ_igs_annual_report_2013_iss1-0.pdf.
Deng, Z., Fritsche, M., Nischan, T. & Bradke, M. Multi-GNSS Ultra Rapid Orbit-, Clock- & EOP-Product Series (GFZ Data Services, 2016); https://doi.org/10.5880/GFZ.1.1.2016.003.
Rebischung, P., Altamimi, Z., Ray, J. & Garayt, B. The IGS contribution to ITRF2014. J. Geodyn. 90, 611–630 (2016).
Altamimi, Z., Rebischung, P., Métivier, L. & Collilieux, X. ITRF2014: A new release of the International Terrestrial Reference Frame modeling nonlinear station motions. J. Geophys. Res. Solid Earth 121, 6109–6131 (2016).
Nakagawa, H., et al. Development and validation of GEONET New Analysis Strategy (Version 4) [in Japanese] Annual Report of the Geographical Survey Institute Vol. 118, 1–8 (GSI, 2009); https://www.gsi.go.jp/common/000054716.pdf.
Needell, D., Tropp, J. & Vershynin, R. Greedy signal recovery review. In 2008 42nd Asilomar Conf. on Signals, Systems and Computers 1048–1050, https://core.ac.uk/reader/23798095 (IEEE, 2008).
Bevis, M. & Brown, A. Trajectory models and reference frames for crustal motion geodesy. J. Geod. 88, 283–311 (2014).
Rebischung, P. et al. IGS08: the IGS realization of ITRF2008. GPS Solutions 16(4), 483–494 (2012).
Zumberge, J. F., Heflin, M. B., Jefferson, D. C., Watkins, M. M. & Webb, F. H. Precise point positioning for the efficient and robust analysis of GPS data from large networks. J. Geophys. Res. Solid Earth 102, 5005–5017 (1997).
Liu, L., Khan, S. A., van Dam, T., Ma, J. H. Y. & Bevis, M. Annual variations in GPS-measured vertical displacements near Upernavik Isstrøm (Greenland) and contributions from surface mass loading. J. Geophys. Res. Solid Earth 122, 677–691 (2017).
Kusche, J. E. J. O. & Schrama, E. J. O. Surface mass redistribution inversion from global GPS deformation and Gravity Recovery and Climate Experiment (GRACE) gravity data. J. Geophys. Res. Solid Earth 110, B09409 (2005).
Borsa, A. A., Agnew, D. C. & Cayan, D. R. Ongoing drought-induced uplift in the western United States. Science 345, 1587–1590 (2014).
Bevis, M. et al. Accelerating changes in ice mass within Greenland, and the ice sheet’s sensitivity to atmospheric forcing. Proc. Natl Acad. Sci. USA 116, 1934–1939 (2019).
Grant, M. & Boyd, S. CVX: Matlab software for disciplined convex programming. Version 2.1, http://cvxr.com/cvx/citing/ (2014).
Moreno, M. et al. Locking of the Chile subduction zone controlled by fluid pressure before the 2010 earthquake. Nat. Geosci. 7, 292–296 (2014).
Chen, Y. W., Wu, J. & Suppe, J. Southward propagation of Nazca subduction along the Andes. Nature 565, 441–447 (2019).
Bedford, J. et al. Trajectory models for daily displacement time series in the five years preceding the 2010 Maule Mw 8.8, Chile, and 2011 Tohoku-oki Mw 9.0, Japan earthquakes (GFZ Data Services, 2020); https://doi.org/10.5880/GFZ.4.1.2020.001.
Hunter, J. D. Matplotlib: a 2D graphics environment. Comput. Sci. Eng. 9, 90–95 (2007).
Wessel, P., Smith, W. H., Scharroo, R., Luis, J. & Wobbe, F. Generic mapping tools: improved version released. Eos 94, 409–410 (2013).
We thank the Geospatial Information Authority of Japan (GSI) and the Nevada Geodetic Laboratory (NGL), University of Nevada, for their assistance and for providing time series for this study. We thank Y. Bock and K. Heki for comments. J.R.B. thanks S. Sobolev for his comments. J.R.B. thanks the German Science Foundation (DFG) for grant MO-2310/3. M.M. acknowledges support from FONDECYT 1181479, the Millennium Nucleus “The Seismic Cycle Along Subduction Zones” grant NC160025, and the Research Center for Integrated Disaster Risk Management (CIGIDEN), CONICYT/FONDAP 15110017. J.C.B. acknowledges support from FONDECYT projects 1170430 and 1181479.
The authors declare no competing interests.
Peer review information Nature thanks Yehuda Bock, Kosuke Heki and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended data figures and tables
Extended Data Fig. 1 Stations used in the processing and analysis of the pre-Maule-earthquake GNSS data.
Locations of IGS stations used to define the reference frame of the network solutions are shown in gold. Black dots are the stations where network solutions are used in the time series analysis of the pre-Maule signals. Coloured triangles indicate locations of time series shown in Fig. 2 and Extended Data Fig. 3. There are some stations used to define the reference frame that are not used in the time series analysis owing to lack of data in the desired window (1 January 2005 until 25 February 2010). Source Data
Extended Data Fig. 2 Time series before the Tohoku-oki earthquake and the effect of noise removal with GrAtSiD in all three directional components.
Left panels show the pre-Tohoku-oki F3 time series. Right panels show these time series after the removal of background seasonal and common-mode noise (with the GrAtSiD routine). The transient behaviour in the months before Tohoku-oki is heavily obscured by seasonal and common-mode noise. Colours correspond to locations on Fig. 1. For clarity, steps have been removed from all time series. Time series in these plots extend until three days before the mainshock. Source Data
Extended Data Fig. 3 Time series before the Maule earthquake and the effect of noise removal with GrAtSiD in all three directional components.
Left panels show the pre-Maule time series. Right panels show these time series after the removal of background seasonal and common mode noise (with the GrAtSiD routine). The transient behaviour in the months before Maule event is heavily obscured by seasonal and common-mode noise. Colours correspond to locations on Fig. 1. For clarity, steps have been removed from all time series. Time series in these plots extend until two days before the mainshock. Source Data
Extended Data Fig. 4 Visualizing the along-strike signal migration and reversal of Japan in the years and months preceding the Tohoku-oki earthquake for all three directional components.
Velocities within non-overlapping rectangular regions before the Tohoku-oki earthquake. The velocity for each region is detrended relative to the median velocity of that region between 1 January 2006 and 8 March 2011. Green lines indicate the along-strike locations and times of earthquakes of moment magnitude exceeding 6. Panels d–f are zoom-ins of panels a–c between the beginning of September 2010 and the beginning of February 2011. The dashed line on panel d indicates the velocity front that migrates across Japan from the southwest (shown in Supplementary Videos 3 and 4 and Fig. 3). Source Data
Extended Data Fig. 5 The corrections to the time series made possible after application of the regression model solved by GrAtSiD.
a, The example time series is for the East component of station Ooamishirasato in Japan. Blue dots show the time series input into the GrAtSiD routine. This time series has been corrected for the common-mode error. The red line shows the complete fit of the regression model solved by GrAtSiD. This includes steps, oscillation terms, the first-order polynomial and the multi-transients. The time series has been optimally tilted (detrended) for clarity of presentation. b, The time series (blue dots) and trajectory model (red line) after removal of the modelled step offsets. c, The time series and the regression model following the removal of the modelled seasonal and step terms. The remaining terms in the model are the first-order polynomial and multi-transients. It is these detrended modelled trajectories following seasonal and step removal (shown in panel c in red) that are represented in Supplementary Videos 3–8 and Figs. 3 and 4 and Extended Data Fig. 4. Source Data
Extended Data Fig. 6 Investigating spatial extents of the pre-Maule-earthquake wobbling in the network solutions.
The map shows the locations of two groups of stations used in the investigation into spatial extent of the unstable period observed before the Maule earthquake. The time series show the average (median) deviation from median velocity at each station of the two groups in the above map for each directional component, where the median velocity of each station is determined between 1 January 2005 and 25 February 2010. Source Data
Extended Data Fig. 7 Investigating spatial extents of the pre-Tohoku-oki-earthquake wobbling using the Nevada Geodetic Laboratory’s IGS08 PPP solutions.
The map shows the locations of two groups of stations used in the investigation into spatial extent of the unstable period observed before the Tohoku-oki earthquake. The time series show the average (median) deviation from median velocity at each station of the two groups in the above map for each directional component, where the median velocity of each station is determined between 1 January 2006 and 8 March 2011. Source Data
Extended Data Fig. 8 Comparison of GSI’s F3 solutions with NGL’s IGS08 PPP solutions for selected stations across Japan before the Tohoku-oki earthquake.
All time series shown are in the East component. Circles show the F3 and crosses show the PPP solutions. Colours correspond to the stations located on the inset map. Source Data
Extended Data Fig. 9 Comparing the transient surface motions recorded by GNSS and predicted by fluid-loading models before the Tohoku-oki earthquake.
The map shows the GNSS station locations used in the analysis comparing fluid-loading displacement predictions to GNSS displacement measurements for the pre-Tohoku-oki case. The time series show a comparison of the median velocity variations for GNSS-measured (GSI’s F3 solutions) and fluid-loading-predicted displacements at the locations in the map. Velocities are taken from the trends estimated by GrAtSiD. In the horizontal components, the prediction from fluid loading produces much lower velocities than those observed. In the vertical component, there is considerable deviation from steady-state velocity in both the GNSS observation and fluid-loading prediction but with visibly low agreement in sense of motion. Source Data
Extended Data Fig. 10 Comparing the transient surface motions recorded by GNSS and predicted by fluid-loading models before the Maule earthquake.
The map shows the GNSS station locations used in the analysis comparing fluid-loading displacement predictions to GNSS displacement measurements for the pre-Maule case. The time series show a comparison of the median velocity variations for GNSS-measured and fluid-loading-predicted displacements at the locations in the map. Velocities are taken from the trends estimated by GrAtSiD. In the East component (in which the pre-Maule unstable motion is most pronounced) the prediction from fluid loading produces much lower deviation from steady-state velocities than those observed by GNSS. Source Data
This file contains detailed descriptions of the supplementary videos.
About this article
Cite this article
Bedford, J.R., Moreno, M., Deng, Z. et al. Months-long thousand-kilometre-scale wobbling before great subduction earthquakes. Nature 580, 628–635 (2020). https://doi.org/10.1038/s41586-020-2212-1