The 25 October 2010 Mentawai tsunami earthquake (Mw 7.8) andthe tsunami hazard presented by shallow megathrust ruptures

T. Lay,1 C. J. Ammon,2 H. Kanamori,3 Y. Yamazaki,4 K. F. Cheung,4 and A. R. Hutko5

Received 20 December 2010; revised 9 February 2011; accepted 15 February 2011; published 22 March 2011.

[1] The 25 October 2010 Mentawai, Indonesia earthquake(Mw 7.8) ruptured the shallow portion of the subduction zoneseaward of the Mentawai islands, off‐shore of Sumatra,generating 3 to 9 m tsunami run‐up along southwesterncoasts of the Pagai Islands that took at least 431 lives. Analysesof teleseismic P, SH and Rayleigh waves for finite‐faultsource rupture characteristics indicate ∼90 s rupture durationwith a low rupture velocity of ∼1.5 km/s on the 10° dippingmegathrust, with total slip of 2–4 m over an ∼100 km longsource region. The seismic moment‐scaled energy releaseis 1.4 × 10−6, lower than 2.4 × 10−6 found for the 17 July2006 Java tsunami earthquake (Mw 7.8). The Mentawaievent ruptured up‐dip of the slip region of the 12 September2007 Kepulauan earthquake (Mw 7.9), and together with the4 January 1907 (M 7.6) tsunami earthquake located seawardof Simeulue Island to the northwest along the arc, demonstratesthe significant tsunami generation potential for shallowmegathrust ruptures in regions up‐dip of great underthrustingevents in Indonesia and elsewhere. Citation: Lay, T., C. J.Ammon, H. Kanamori, Y. Yamazaki, K. F. Cheung, and A. R. Hutko(2011), The 25 October 2010Mentawai tsunami earthquake (Mw 7.8)and the tsunami hazard presented by shallow megathrust ruptures,Geophys. Res. Lett., 38, L06302, doi:10.1029/2010GL046552.

1. Introduction

[2] Coseismic slip zones for great underthrusting earth-quakes on subduction zone megathrusts often do not extendseaward to the trench, leaving uncertainty regarding whetherthe up‐dip region is a zone of stable sliding that slips aseis-mically before or after great events [e.g.,Hsu et al., 2006] or astill coupled region with potential for future large earth-quakes. Large earthquakes that rupture the shallow, near‐trench region of megathrusts are relatively rare, but suchevents can be exceptionally tsunamigenic and are oftencharacterized by low rupture velocity and large slip in rela-tively low rigidity material [e.g., Polet and Kanamori, 2000;Bilek and Lay, 2002]. When the rupture velocity for a tsu-namigenic faulting event is low enough that the seismicwaves have unusually weak short‐period seismic wave

energy relative to the long‐period signals levels, the event iscalled a tsunami earthquake [Kanamori, 1972;Kanamori andKikuchi, 1993].[3] A large tsunamigenic earthquake struck the shallow

Sumatra subduction zone on 25 October 2010 (3.484°S,100.114°E, 14:42:22 UTC, mb = 6.5, MS = 7.3; U.S. Geo-logical Survey, Magnitude 7.7 – Kepulauan Mentawai Region,Indonesia, 2010, http://earthquake.usgs.gov/earthquakes/eqinthenews/2010/usa00043nx/#details.). The Global Cen-troid Moment Tensor solution (GCMT; G. Ekström, GlobalCMT Web Page, 2010, http://www.globalcmt.org/) for theevent indicates a shallow dipping (7°) underthrustingmechanism with a best double couple moment Mo = 6.7 ×1020 Nm (Mw 7.8). Our inversion of 40 three‐component0.001–0.005 Hz W‐phase [Kanamori and Rivera, 2008]observations from 28 global broadband network stationshas a predominantly double‐couple point‐source solution12 km deep with Mo = 5 × 1020 Nm (Mw 7.7) and a nodalplane dipping 10° toward the northeast (Figure 1).[4] The rupture area is seaward of the Pagai islands of the

Mentawai group offshore of Sumatra. This stretch of thesubduction zone likely failed in the great Sumatran events of1833 (M ∼ 9) and 1797 (M ∼ 8.8), although the up‐dipextent of these ruptures is not well constrained [Natawidjajaet al., 2006]. The 2010 event ruptured the megathrust up‐dipof the 12 September 2007 Kepulauan earthquake (Mw 7.9),and northwest of the great 12 September 2007 Sumatranearthquake (Mw 8.4) (Figure 1) [Konca et al., 2008]. Finite‐source rupture models for the 2007 Kepulauan event do notextend to the trench, although GPS and paleogeodetic dataindicate strong coupling (>50%) of the megathrust wellseaward of Pagai prior to the 2007 events [Chlieh et al.,2008].[5] The 2010 Mentawai event struck at 9:42 PM local time,

producing ground shaking on Pagai less intense than for the2007 Kepulauan event, and although tsunami warnings werebroadcast on local television many of the isolated coastal areaswere unprepared for the large tsunami waves that swept ontoPagai’s southwestern coast, with peak run‐ups ranging from2.5 to 9 m (K. Satake, personal communication, 2010). Theofficial loss of life reported by the IndonesianNationalDisasterManagement Agency is currently 431, with 88 missing.

2. Rupture Process of 25 October 2010 Event

[6] We develop finite‐source models for the rupture pro-cess of the Mentawai earthquake using teleseismic record-ings of P, SH, and short‐arc Rayleigh wave (R1) recordingsfrom global seismic network stations. Inversion for finite‐fault models requires many assumptions and constraints onthe solution. We adopt a 10° dip for the fault plane based onthe W‐phase solution, a strike of 324° based on the local

1Department of Earth and Planetary Sciences,University of California,Santa Cruz, California, USA.

2Department of Geosciences, Pennsylvania State University, UniversityPark, Pennsylvania, USA.

3Seismological Laboratory,California Institute of Technology, Pasadena,California, USA.

4Department of Ocean andResource Engineering,University of Hawaiiat Manoa, Honolulu, Hawaii, USA.

5Data Management Center, Incorporated Research Institutions forSeismology, Seattle, Washington, USA.

Copyright 2011 by the American Geophysical Union.0094‐8276/11/2010GL046552

GEOPHYSICAL RESEARCH LETTERS, VOL. 38, L06302, doi:10.1029/2010GL046552, 2011

L06302 1 of 5

trend of the trench, and the rupture velocity, Vr, is constrainedusing R1 source time functions (STF) in finite‐fault inver-sions [Ammon et al., 2006a, 2006b] and back‐projection ofseismic network recordings of P‐waves [Ishii et al., 2005;Xu et al., 2009].[7] Broadband P wave recordings from the F‐Net in Japan

were used in the back‐projection imaging to bound theoverall rupture finiteness. The back‐projections suggest thatthe source region radiated short‐period energy for about90 s, with slow propagation toward the northwest over adistance of about 100 km with Vr ∼ 1.0–1.5 km/s (Figure 2and Animation S1 of the auxiliary material), with slipconcentrated seaward and beneath the Pagai Islands.1

[8] Finite‐fault inversions of teleseismic P, SH, and R1STFs assuming fixed Vr values ranging from 1.0 to 2.5 km/sindicate a slight preference for Vr = 1.5 km/s (Figure S1),and the corresponding slip distribution covers a 100 km longregion of 2–3.5 m slip located seaward of Pagai (Figure 1,left). For this model the source velocity structure was a half‐space with P wave velocity of 5.0 km/s, S wave velocity of2.9 km/s, and density of 2300 kg/m3, with each subfaulthaving a half‐cosine source time function with duration of8, 12, 16, 20 or 24 s. The seismic moment for this inversionis Mo = 6.2 × 1020 Nm (Mw 7.8) and about 83% of theweighted data signal was fit (Figure S2). The subfaultdetails are shown in Figure S3.[9] Linear inversion of 55 broadband P and SHwaves using

Vr = 1.5 km/s and a layered velocity structure (Table S1)appropriate for the low velocities found in the shallow wedgeregion [Collings et al., 2010] gave a similar single patch ofseismic moment release, with the slip being enhanced in theregion near the toe of the sedimentary wedge (Figure 1, right).The inversion used 16 s long subfault source functionscomprised of 7 overlapping 2‐s rise‐time triangles. The tel-

eseismic body wave data set alone has very limited resolutionof source finiteness; inversions with different rupture veloc-ities all fit the data waveforms at about the 83% level (e.g.,Figure S4). The subfault details are shown in Figure S5. Theinversion including R1 STFs indicates about 50 km morenorthwestward expansion of the source than the body wavesolution, while the more compact body wave slip distributionis shifted slightly up‐dip. The models differ in data andparameterization, with smoother features in the slip distri-bution being better resolved when R1 is included, but rougherfeatures better resolved when just body waves are used. Thebody waves have somewhat better depth‐resolution, whilethe inversion with long‐periods better constrains seismicmoment and along‐strike directivity. Both solutions placethe large slip region up‐dip of the primary slip zone for the2007 Kepulauan event and on the northwestern margin ofthe 2007 great Sumatran event rupture zone (Figure 1).The two slip models in Figure 1 indicate the overall uncer-tainty in the teleseismic solutions, and are similar to modelsproduced by other research groups [Newman et al., 2011; e.g.,Y. Yamanaka, 2010, http://www.seis.nagoya‐u.ac.jp/sanchu/Seismo_Note/2010/NGY31.html]. It may be viable to improveconstraints on the slip distribution when GPS data on Pagai areeventually analyzed. Aftershocks fringe the strong slip regions,with a concentration up‐dip of the hypocenter (Figure S6 andAnimation S2). Animations S3 and S4 depict the ruptureexpansion for both models.[10] The spectrum of the source time function from the

P and SH inversion (Figure 3, inset) was combined withazimuthally‐averaged P wave spectra to characterize theoverall source spectrum (Figure 3), which is significantlydepleted in short‐period amplitudes relative to expectationsfor a reference w‐squared source spectrum, consistent withother tsunami earthquakes (Figure S7). The P wave groundvelocities were used to compute the seismic energy releaseover the duration indicated by the source time function,

Figure 1. Maps showing the source region bathymetry and topography near the 25 October 2010 Mentawai tsunami earth-quake (Mw 7.8). The base maps show the large‐slip regions for the 12 September 2007 Kepulauan (pink areas) and Suma-tran earthquakes (blue region) [Konca et al., 2008]. The focal mechanism is the W‐phase moment tensor solution describedin the text. (left) The slip distribution obtained by inversion of P, SH and R1 source time functions. The rake was allowed tovary but is almost uniform with a mean value of 98° for this model. (right) The slip distribution obtained by the linearinversion of only teleseismic P and SH waves. The arrows indicate the slight variations in rake obtained for this model.

1Auxiliary materials are available in the HTML. doi:10.1029/2010GL046552.

LAY ET AL.: OCTOBER 2010 MENTAWAI TSUNAMI EARTHQUAKE L06302L06302

2 of 5

giving a radiated seismic energy estimate [Venkataramanand Kanamori, 2004], of ER = 9.2 × 1014 J. Using theGCMT seismic moment, the moment‐scaled energy ratioER/Mo is 1.4 × 10−6 (Figure S8). Similarly computed scaled‐energy values for the 2 June 1994 (Mw 7.8) and 17 July2006 (Mw 7.8) Java tsunami earthquakes were found to be0.37 × 10−6 and 3.5 × 10−6, respectively. The large dis-crepancy between mb and Mw, low moment‐scaled energyvalue, and compact source area and low rupture velocity ofthe Mentawai event establish that it shares common char-acteristics with other tsunami earthquakes [Polet andKanamori, 2000; Bilek and Lay, 2002].

3. Tsunami of 25 October 2010 Event

[11] The tsunami generated by the Mentawai event wasrecorded by DART buoy 56001, which is located far from thecoastline and in deep water (13.961°S 110.004°E; Figure S9).We computed the surface elevation at DART 56001 usingthe non‐linear dispersive wave model NEOWAVE [Yamazaki

et al., 2009, 2011] for the two finite‐sourcemodels in Figure 1.NEOWAVE is a staggered finite difference model, whichincludes a vertical momentum equation and a non‐hydrostaticpressure term in the nonlinear shallow‐water equations todescribe tsunami generation from seafloor deformationand propagation of weakly dispersive tsunami waves. Thefault parameters at each subfault allow computation ofthe seafloor deformation (Figures S10 and S11) using theplanar fault model of Okada [1985]. Superposition of thedeformation from the subfaults with consideration of theirrupture initiation time and rise time reconstructs the timehistory of seafloor vertical displacement and velocity forthe input to NEOWAVE. We use the British OceanographicData Centre General Bathymetric Chart of the Oceans(GEBCO), re‐sampled on a 1‐arcmin grid (≈1800‐m reso-lution) for the computations.[12] The observed and computed tsunami waveforms and

spectra at DART 56001 for the two finite source models areshown in Figure 4. The source model obtained when the R1

Figure 2. Back‐projections of 0.5–2.0 Hz bandpass filtered P‐waves recorded by the Japanese F‐Net. These are framesfrom the continuous animation presented in the auxiliary material. The colors indicate the amplitude of the cube‐rootstack of the back‐projected signals at each position for each time step. The lower trace indicates the peak amplitude in theimage area as a function of time, with the vertical lines indicating the image time steps.

LAY ET AL.: OCTOBER 2010 MENTAWAI TSUNAMI EARTHQUAKE L06302L06302

3 of 5

STFs are included predicts the primary tsunami amplitudewell, but has a slight phase lag resulting from lower pre-dictions of waves shorter than 22 min. The tsunami calcu-lation for the body wave model slightly underestimates thefirst wave’s amplitude due to smaller amplitude predictedfor the 30‐min and 73‐min wave components, but showsvery good agreement with the overall phase. The first waveis directly from the source, while the subsequent waves are acombination of waves from the source, reflection from is-lands, and leakage of trapping waves over the shelf. High

resolution bathymetry is required to reproduce the reflectedand leaked trapped waves at DART 56001. Considering thegird resolution, the general agreement indicates that theshallow seismic slip models are compatible with the far‐field tsunami excitation.

4. Discussion

[13] The 25 October 2010 Mentawai earthquake qualifiesas a tsunami earthquake, and shares attributes with the M ∼7.6 1907 Sumatra earthquake [Kanamori et al., 2010] thatlikely ruptured off‐shore of Simeulue, Nias and Batu Islands(Figure S9). Both are tsunamigenic events that ruptured up‐dip regions of strongly coupled megathrust zones that havedeeper great interplate events. The 1994 and 2006 Java[Ammon et al., 2006b; Abercrombie et al., 2001] tsunamiearthquakes have similar seismic radiation characteristics tothe Mentawai event, but the Java events ruptured up‐dip ofweakly coupled megathrust zones that appear not to fail ingreat interplate events. The low levels of short‐period seis-mic wave radiation for these events, and the resulting weaklocal ground shaking can undermine public response to thepossibility of imminent tsunami arrival. Technical methodsfor very rapid (<15 min) local event detection and momentcharacterization appear to be essential for developing anyeffective tsunami warning capability for tsunami earth-quakes, and this must include effective societal implemen-tation and public education.[14] The seaward position of the Mentawai islands

enabled geodetic inversion for locked portions of the shal-low megathrust, and the Mentawai earthquake occurred in aregion identified as having been partially to fully locked.For most regions the resolution of geodetic inversions foroffshore locking of the megathrust is very limited, and whena great event like the 27 February 2010 Chile (Mw 8.8)earthquake does not rupture all the way to the trench it isunclear whether there remains potential for shallow tsunami

Figure 3. The source spectrum for the 25 October 2010Mentawai earthquake estimated from a combination of thespectrum of the moment‐rate function (inset) determined byinversion of teleseismic P and SH waves (solid black line)scaled to the GCMT seismic moment and the stacked spectraof teleseismic P–waves corrected for propagation and radia-tion pattern (dots). A reference w‐squared spectrum for thecorresponding seismic moment is shown in blue.

Figure 4. Comparisons of the observed (black lines) tsunami (left) time series and (right) spectrum recorded by DART56001 for the 25 October 2010 Mentawai earthquake with calculations (red lines) for (a) joint inversion of P, SH, andR1 STFs and (b) inversion of the P and SH body waves alone.

LAY ET AL.: OCTOBER 2010 MENTAWAI TSUNAMI EARTHQUAKE L06302L06302

4 of 5

earthquakes to occur. The most promising approach tothis problem is to improve imaging of coseismic and post‐seismic slip for great events along with geodetic inversionfor megathrust locked regions in the interseismic interval[e.g., Moreno et al., 2010]. Indonesia is evidently exposedto hazard from conventional great megathrust events as wellas up‐dip tsunami earthquakes in both strongly and weaklycoupled portions of the subduction zone. It is reasonable toexpect that similar hazard of shallow megathrust tsunamiearthquakes exists in other subduction zones in both regionswhere great underthrusting events occur as well as in weaklycoupled environments.

[15] Acknowledgments. This work made use of GMT and SAC soft-ware. The IRIS DMS and the F‐Net data centers were used to access thedata. We thank Kenji Satake for information about the tsunami run‐upobservations on Mentawai. E. Geist and two anonymous reviewers pro-vided very helpful comments on the manuscript. This work was supportedby NSF grant EAR0635570 and USGS Award Number 05HQGR0174.[16] The Editor thanks Eric Geist and two anonymous reviewers for

their assistance in evaluating this paper.

ReferencesAbercrombie, R. E., M. Antolik, K. Felzer, and G. Ekström (2001), The1994 Java tsunami earthquake: Slip over a subducting seamount, J. Geo-phys. Res., 106, 6595–6607, doi:10.1029/2000JB900403.

Ammon, C. J., A. A. Velasco, and T. Lay (2006a), Rapid estimation offirst‐order rupture characteristics for large earthquakes using surfacewaves: 2004 Sumatra‐Andaman earthquake, Geophys. Res. Lett., 33,L14314, doi:10.1029/2006GL026303.

Ammon, C. J., H. Kanamori, T. Lay, and A. A. Velasco (2006b), The17 July 2006 Java tsunami earthquake, Geophys. Res. Lett., 33,L24308, doi:10.1029/2006GL028005.

Bilek, S. L., and T. Lay (2002), Tsunami earthquakes possibly widespreadman-ifestations of frictional conditional stability, Geophys. Res. Lett., 29(14),1673, doi:10.1029/2002GL015215.

Chlieh, M., J. P. Avouac, K. Sieh, D. H. Natawidjaja, and J. Galetzka(2008), Heterogeneous coupling of the Sumatran megathrust constrainedby geodetic and paleogeodetic measurements, J. Geophys. Res., 113,B05305, doi:10.1029/2007JB004981.

Collings, R., A. Rietbrock, D. Lange, F. Tilmann, D. H. Natawidjaja, andB. W. Suwargadi (2010), The structure of the Mentawai segment of theSumatra subduction zone revealed by local earthquake travel timetomography, Abstract T11D‐2127 presented at Fall Meeting, AGU,San Francisco, Calif., 13–17 Dec.

Hsu, Y.‐J., M. Simons, J.‐P. Avouac, J. Galetzka, K. Sieh, M. Chlieh,D. Natawidjaja, L. Prawirodirdjo, and Y. Bock (2006), Frictional afterslipfollowing the 2005 Nias‐Simeulue earthquake, Sumatra, Science, 312,1921–1926, doi:10.1126/science.1126960.

Ishii, M., P. M. Shearer, H. Houston, and J. E. Vidale (2005), Extent, dura-tion and speed of the 2004 Sumatra‐Andaman earthquake imaged by theHi‐net array, Nature, 435, 933–936.

Kanamori, H. (1972), Mechanism of tsunami earthquakes, Phys. EarthPlanet. Inter., 6, 346–359, doi:10.1016/0031-9201(72)90058-1.

Kanamori, H., and M. Kikuchi (1993), The 1992 Nicaragua earthquake: Aslow tsunami earthquake associated with subducted sediments, Nature,361, 714–716, doi:10.1038/361714a0.

Kanamori, H., and L. Rivera (2008), Source inversion of W‐phase: Speed-ing up seismic tsunami warning, Geophys. J. Int., 175, 222–238,doi:10.1111/j.1365-246X.2008.03887.x.

Kanamori, H., L. Rivera, and W. H. K. Lee (2010), Historical seismogramsfor unraveling a mysterious earthquake: The 1907 Sumatra earthquake,Geophys. J. Int., 183, 358–374, doi:10.1111/j.1365-246X.2010.04731.x.

Konca, A. O., et al. (2008), Partial rupture of a locked patch of the Sumatramegathrust during the 2007 earthquake sequence, Nature, 456, 631–635,doi:10.1038/nature07572.

Moreno, M., M. Rosenau, and O. Oncken (2010), Maule earthquake slipcorrelates with pre‐seismic locking of Andean subduction zone, Nature,467, 198–202, doi:10.1038/nature09349.

Natawidjaja, D. H., K. Sieh, M. Chlieh, J. Galetzka, B. W. Suwargadi,H. Cheng, R. L. Edwards, J.‐P. Avouac, and S. N. Ward (2006), Sourceparameters of the great Sumatran megathrust earthquakes of 1797 and1833 inferred from coral microatolls, J. Geophys. Res., 111, B06403,doi:10.1029/2005JB004025.

Newman, A. V., G. Hayes, Y. Wei, and J. Convers (2011), The 5 October2010 Mentawai tsunami earthquake, from real‐time discriminants, finite‐fault rupture, and tsunami excitation, Geophys. Res. Lett., 38, L05302,doi:10.1029/2010GL046498.

Okada, Y. (1985), Surface deformation due to shear and tensile faults in ahalf space, Bull. Seismol. Soc. Am., 75(4), 1135–1154.

Polet, J., and H. Kanamori (2000), Shallow subduction zone earthquakesand their tsunamigenic potential, Geophys. J. Int., 142, 684–702,doi:10.1046/j.1365-246x.2000.00205.x.

Venkataraman, A., and H. Kanamori (2004), Observational constraints onthe fracture energy of subduction zone earthquakes, J. Geophys. Res.,109, B05302, doi:10.1029/2003JB002549.

Xu, Y., K. D. Koper, O. Sufri, L. Zhu, and A. R. Hutko (2009), Ruptureimaging of the Mw 7.9 12 May 2008 Wenchuan earthquake from backprojection of teleseismic P waves, Geochem. Geophys. Geosyst., 10,Q04006, doi:10.1029/2008GC002335.

Yamazaki, Y., Z. Kowalik, and K. F. Cheung (2009), Depth‐integrated,non‐hydrostatic model for wave breaking and run‐up, Int. J. Numer.Methods Fluids, 61(5), 473–497, doi:10.1002/fld.1952.

Yamazaki, Y., K. F. Cheung, and Z. Kowalik (2011), Depth‐integrated,non‐hydrostatic model with grid nesting for tsunami generation, propa-gation, and run‐up, Int. J. Numer. Methods Fluids, doi:10.1002/fld.2485, in press.

C. J. Ammon, Department of Geosciences, Pennsylvania State University,440 Deike Bldg., University Park, PA 16802, USA.K. F. Cheung and Y. Yamazaki, Department of Ocean and Resource

Engineering, University of Hawaii at Manoa, Holmes Hall 402, 2540Dole St., Honolulu, HI 96822, USA.A. R. Hutko, Data Management Center, Incorporated Research

Institutions for Seismology, 1408 NE 45 St., Ste. 201, Seattle, WA 98105,USA.H. Kanamori, Seismological Laboratory, California Institute of

Technology, 1200 E. California Blvd., Pasadena, CA 91125, USA.T. Lay, Department of Earth and Planetary Sciences, University of

California, 1156 High St., Santa Cruz, CA 95064, USA. (tlay@ucsc.edu)

LAY ET AL.: OCTOBER 2010 MENTAWAI TSUNAMI EARTHQUAKE L06302L06302

5 of 5