Near-simultaneous great earthquakes at Tongan megathrust and outer rise in September 2009

Journal name:
Date published:

The Earth’s largest earthquakes and tsunamis are usually caused by thrust-faulting earthquakes on the shallow part of the subduction interface between two tectonic plates, where stored elastic energy due to convergence between the plates is rapidly released1, 2. The tsunami that devastated the Samoan and northern Tongan islands on 29 September 2009 was preceded by a globally recorded magnitude-8 normal-faulting earthquake in the outer-rise region, where the Pacific plate bends before entering the subduction zone. Preliminary interpretation suggested that this earthquake was the source of the tsunami3. Here we show that the outer-rise earthquake was accompanied by a nearly simultaneous rupture of the shallow subduction interface, equivalent to a magnitude-8 earthquake, that also contributed significantly to the tsunami. The subduction interface event was probably a slow earthquake with a rise time of several minutes that triggered the outer-rise event several minutes later. However, we cannot rule out the possibility that the normal fault ruptured first and dynamically triggered the subduction interface event. Our evidence comes from displacements of Global Positioning System stations and modelling of tsunami waves recorded by ocean-bottom pressure sensors, with support from seismic data and tsunami field observations. Evidence of the subduction earthquake in global seismic data is largely hidden because of the earthquake’s slow rise time or because its ground motion is disguised by that of the normal-faulting event. Earthquake doublets where subduction interface events trigger large outer-rise earthquakes have been recorded previously4, but this is the first well-documented example where the two events occur so closely in time and the triggering event might be a slow earthquake. As well as providing information on strain release mechanisms at subduction zones, earthquakes such as this provide a possible mechanism for the occasional large tsunamis generated at the Tonga subduction zone5, where slip between the plates is predominantly aseismic6.

At a glance


  1. The 2009 Samoa-Tonga earthquake as a double event, with observed and modelled GPS displacements and fault-plane solutions.
    Figure 1: The 2009 Samoa–Tonga earthquake as a double event, with observed and modelled GPS displacements and fault-plane solutions.

    The yellow arrows (with 95% confidence uncertainty ellipses) are observed displacements at GPS sites. The black arrows show displacements from an outer-rise normal-fault dislocation model fitted to observations from continuous GPS sites only (that is, not NTPT or NTRP). The predictions of this model at GPS sites NTPT and NTRP are in strong disagreement with the observations. The red arrows are displacements for a model fitted to all GPS data using two faults: a subduction interface thrust fault and an outer-rise normal fault. VAVS and NIUM, two sites located south of the region in the main figure, are well fitted by the two-fault model but NIUM is poorly fitted by the single-fault model. The outlines of the uniform-slip rectangular fault planes projected to the surface are shown by yellow rectangles with associated lower-hemisphere focal mechanisms. See Supplementary Tables 1, 3 and 5 for the displacement values and fault parameters. The red star shows the epicentre of the main shock from the US Geological Survey Preliminary Determination of Epicentres (PDE) catalogue, and the open grey circles are PDE epicentres for the first month of aftershocks. Bathymetry contours are shown from 1,000 to 6,000m at 1,000-m intervals. Inset, locations of DART buoys whose data we fitted (red dots), seismic stations (green squares) and some GPS sites including those not in the main figure (black triangles). The magenta lines are a representation of tectonic plate and microplate boundaries.

  2. Observed displacements at the two campaign GPS sites on Niuatoputapu showing their large eastward offset between 2005 and 2009.
    Figure 2: Observed displacements at the two campaign GPS sites on Niuatoputapu showing their large eastward offset between 2005 and 2009.

    A best-fit linear trend from the start of the record until December 2005 is subtracted from each time series; the dashed lines show the ±1σ uncertainty of the projected fit. The dotted lines show the predicted coseismic displacement from the two-fault model of Fig. 1. Error bars, 1σ.

  3. Observed seismic noise before onset of outer-rise earthquake arrivals at AFI compared with model subduction thrust seismograms.
    Figure 3: Observed seismic noise before onset of outer-rise earthquake arrivals at AFI compared with model subduction thrust seismograms.

    Red trace shows the observed AFI vertical-component velocity record (Vz); the onset of strong ground motion from the normal fault is at 17:48:34.5, and the record is clipped after this time. The green, blue, brown and black traces show the calculated velocity for Mw = 8 subduction interface thrust events starting 170s before the normal fault with respective rise times of 50, 100, 150 and 200s. The predicted thrust signal is at or below the noise level of the data for rise times longer than ~200s. If the thrust event started after the normal fault, its signal would be disguised within the high-amplitude waves from the normal fault. Supplementary Fig. 5 shows the horizontal components.

  4. Observed and modelled tsunami wave elevations and mis-fit contours.
    Figure 4: Observed and modelled tsunami wave elevations and mis-fit contours.

    ac, Observations (black) at the three nearest DART ocean-bottom pressure sensors, with predictions of the GPS two-fault model of Fig. 1 (red) and the preliminary US Geological Survey outer-rise normal-fault model25 (green). The latter disagrees strongly with the data, especially at DART buoys 54401 and 51246. The two-fault predictions use a thrust-fault rise time of 200s, with the normal fault rupturing 105s after the initiation of the thrust fault. Time zero is the origin time of the normal-fault event, and 1.5h of data around the first tsunami arrival time are shown for each DART record. d, Contours show root-mean-squared mis-fit (in metres) of DART sea-level data predicted by this GPS source model as a function of thrust-faulting rise time and time delay between the initiation of the two events. See Supplementary Fig. 6 for the predictions of an alternative GPS model and Supplementary Fig. 8 for the predictions at far-field DART buoys.


  1. Kanamori, H. Rupture process of subduction-zone earthquakes. Annu. Rev. Earth Planet. Sci. 14, 293322 (1986)
  2. US Geological Survey. Largest Earthquakes in the World Since 1900. left fence fence (2010)
  3. US Geological Survey. Poster of the Samoa Islands Region Earthquake of 29 September 2009 - Magnitude 8.0. left fence fence (2009)
  4. Ammon, C. J., Kanamori, H. & Lay, T. A great earthquake doublet and seismic stress transfer in the central Kuril islands. Nature 451, 561565 (2008)
  5. Okal, E. A., Borrero, J. & Synolakis, C. E. The earthquake and tsunami of 1865 November 17: evidence for far-field tsunami hazard from Tonga. Geophys. J. Int. 157, 164174 (2004)
  6. Bevis, M. et al. Geodetic observations of very rapid convergence and back-arc extension at the Tonga arc. Nature 374, 249251 (1995)
  7. Thatcher, W. Order and diversity in the modes of circum-pacific earthquake recurrence. J. Geophys. Res. 95, 26092623 (1990)
  8. Pacheco, J. F. & Sykes, L. R. Seismic moment catalog of large shallow earthquakes, 1900–1989. Bull. Seismol. Soc. Am. 82, 13061349 (1992)
  9. US Geological Survey. USGS Centroid Moment Tensor Solution. left fence fence (2009)
  10. US Geological Survey. Global CMT Project Moment Tensor Solution: Samoa Islands Region. left fence fence (2009)
  11. National Oceanographic and Atmospheric Administration. DART Buoy Comparison with Modelled Results. left fence fence (2010)
  12. Kanamori, H. Mechanism of tsunami earthquakes. Phys. Earth Planet. Inter. 6, 346359 (1972)
  13. Satake, K. & Tanioka, Y. Sources of tsunami and tsunamigenic earthquakes in subduction zones. Pure Appl. Geophys. 154, 467483 (1999)
  14. Bonnardot, M.-A., Régnier, M., Ruellan, E., Christova, C. & Tric, E. Seismicity and state of stress within the overriding plate of the Tonga-Kermadec subduction zone. Tectonics 26 doi:10.1029/2006TC002044 (2007)
  15. Aki, K. & Richards, P. G. Quantitative Seismology: Theory and Methods Vol. II 799849 (Freeman, 1980)
  16. Cifuentes, I. L. & Silver, P. G. Low-frequency source characteristics of the great 1960 Chilean earthquake. J. Geophys. Res. 94, 643663 (1989)
  17. Dominey-Howes, D. & Thaman, R. UNESCO-IOC International Tsunami Survey Team Samoa, Interim Report of Field Survey 14th-21st October 2009. Misc. Report No. 2, 2627 (Aust. Tsunami Res. Centre, Univ. New South Wales, 2009)
  18. Wilson, K. J. et al. Post-Tsunami Survey of Niuatoputapu Island, Tonga, Following the 30th September 2009 South Pacific Tsunami (GNS Science Report 2009/71, Institute of Geological and Nuclear Sciences, 2009)
  19. Freed, A. M. Earthquake triggering by static, dynamic and postseismic stress transfer. Annu. Rev. Earth Planet. Sci. 33, 335367 (2005)
  20. Linde, A. T. & Sacks, I. S. Slow earthquakes and great earthquakes along the Nankai trough. Earth Planet. Sci. Lett. 203, 265275 (2002)
  21. Frohlich, C. et al. Huge erratic boulders in Tonga deposited by a prehistoric tsunami. Geology 37, 131134 (2009)
  22. Sobolev, S. V. et al. Tsunami early-warning using GPS-shield arrays. J. Geophys. Res. 112 doi:10.1029/2006JB004640 (2007)
  23. Song, Y. T. Detecting tsunami genesis and scales directly from coastal GPS stations. Geophys. Res. Lett. 34 doi:10.1029/2007GL031681 (2007)
  24. Blewitt, G. et al. GPS for real-time earthquake source determination and tsunami warning systems. J. Geod. 83, 335343 (2009)
  25. US Geological Survey. Finite Fault Model. left fence fence (2009)
  26. Dach, R., Hugentobler, U., Fridez, P. & Meindl, M. Bernese GPS Software Version 5.0 (Astronomical Institute, Univ. Bern, 2007)
  27. Okada, M. Surface deformation due to shear and tensile faults in a half-space. Bull. Seismol. Soc. Am. 75, 11351154 (1985)
  28. Wang, X. & Liu, P. L.-F. An analysis of 2004 Sumatra earthquake fault plane mechanism and Indian Ocean tsunami. J. Hydraul. Res. 44, 147154 (2006)
  29. Coutant, O. Expression of the Green’s Functions in Cylindrical Coordinates used with a Reflectivity Method. Tech. Report (Laboratoire de Géophysique Interne et Tectonophysique, 1994)

Download references

Author information


  1. GNS Science, PO Box 30368, Lower Hutt 5040, New Zealand

    • J. Beavan,
    • X. Wang,
    • C. Holden,
    • K. Wilson,
    • W. Power &
    • G. Prasetya
  2. School of Earth Sciences, Ohio State University, Columbus, Ohio 43210, USA

    • M. Bevis
  3. Ministry of Lands, Survey, Natural Resources and Environment, PO Box 5, Nuku’alofa, Tonga

    • R. Kautoke


J.B. processed and modelled the GPS data. X.W. modelled the DART and tsunami run-up data with support from G.P. and W.P. C.H. modelled the seismic data. K.W. and W.P. made field observations of the tsunami effects in Samoa and on Niuatoputapu. R.K.’s team at the Tonga MLSNRE has been instrumental in the collection of GPS data from Niuatoputapu, including the critical post-earthquake observations, and they also made field observations of tsunami effects. M.B. analysed high-rate GPS data. J.B., X.W., C.H., M.B., K.W., G.P. and W.P. were responsible for the data interpretation and all authors discussed the results and commented on the manuscript.

Competing financial interests

The authors declare no competing financial interests.

Corresponding authors

Correspondence to:

Author details

Supplementary information

PDF files

  1. Supplementary Information (6.2M)

    This file contains Supplementary Texts 1 -7, Supplementary Tables 1-6 and Supplementary Figures 1-11.

Additional data