Elsevier

Tectonophysics

Volume 721, 28 November 2017, Pages 143-150
Tectonophysics

Rupture evolution of the 2006 Java tsunami earthquake and the possible role of splay faults

https://doi.org/10.1016/j.tecto.2017.10.003Get rights and content

Highlights

  • The 2006 Java earthquake produced frequency-dependent seismic radiation.

  • A transition in high-frequency radiation suggests a two stage rupture.

  • The second-stage seismic radiation may be hosted by reactivated splay faults in the outer wedge.

Abstract

The 2006 Mw 7.8 Java earthquake was a tsunami earthquake, exhibiting frequency-dependent seismic radiation along strike. High-frequency global back-projection results suggest two distinct rupture stages. The first stage lasted ∼65 s with a rupture speed of ∼1.2 km/s, while the second stage lasted from ∼65 to 150 s with a rupture speed of ∼2.7 km/s. High-frequency radiators resolved with back-projection during the second stage spatially correlate with splay fault traces mapped from residual free-air gravity anomalies. These splay faults also colocate with a major tsunami source associated with the earthquake inferred from tsunami first-crest back-propagation simulation. These correlations suggest that the splay faults may have been reactivated during the Java earthquake, as has been proposed for other tsunamigenic earthquakes, such as the 1944 Mw 8.1 Tonankai earthquake in the Nankai Trough.

Introduction

Tsunami earthquakes are characterized by a disproportionately large tsunami for their size, and often exhibit a disparity between estimates of moment magnitude derived from long and short period seismic radiation Kanamori, 1972, Kanamori and Kikuchi, 1993. The July 17, 2006 Java earthquake was a classic tsunami earthquake with body-wave magnitude mb = 6.1, surface-wave magnitude Ms = 7.1, and moment magnitude Mw = 7.7 Ekström et al., 2012, International Seismological Centre, 2013. Such a large variation in magnitude estimates is atypical and may indicate a deficiency in high-frequency radiation compared to low-frequency radiation Ammon et al., 2006, Newman and Okal, 1998. The 2006 Java earthquake initiated at shallow depth (20 km, (International Seismological Centre, 2013); Fig. 1) and ruptured eastward along the trench axis for ∼200 km Ammon et al., 2006, Bilek and Engdahl, 2007. Given the source dimension, the unusually long source duration (∼185 s) indicates anomalously slow rupture propagation for the event Ammon et al., 2006, Bilek and Engdahl, 2007. The earthquake generated a large tsunami (∼8 m) resulting in over 800 fatalities Fritz et al., 2007, Fujii and Satake, 2006, Mori et al., 2007. This was the second tsunami earthquake that struck the Java region since instrumental records began, and a Mw 7.8 earthquake in June 1994 produced an even larger tsunami (∼13 m), resulting in 250 fatalities Abercrombie et al., 2001, Mori et al., 2007. These two earthquakes are only 600 km apart, highlighting the major tsunami hazard along the south coast of Indonesia (Mori et al., 2007). Is the Java trench prone to more tsunami earthquakes and if so, what properties of the margin promote this type of rupture?

Finite-fault slip models of the 2006 Java earthquake suggest a smooth slip distribution with an unusually slow (∼1 km/s) rupture propagation (Fig. 2b). Finite-fault slip models obtained from body waves (P and SH waves, ∼0.001–0.2 Hz) have similar slip distributions, with the largest slip concentrated near the hypocenter (Fig. 2b) Ammon et al., 2006, Bilek and Engdahl, 2007, Yagi and Fukahata, 2011, Ye et al., 2016, Ye et al., 2016. In contrast, finite-fault slip models obtained from both body and surface waves (both Rayleigh and Love waves) suggest that the largest slip is close to the trench and is up-dip and ∼50 km east of the hypocenter (Fig. 2) Hayes, 2011, Shao et al., 2011. Surface waves have been shown to be effective at resolving near-trench slip distributions, which are difficult to resolve just with body waves (Shao et al., 2011).

The 2006 Java earthquake was one of the best-recorded tsunami earthquakes with modern instruments. Combining the wealth of data with new observational approaches enables us to investigate the earthquake in great detail. We first analyze bathymetry and gravity anomalies in conjunction with active-source seismic profiles to constrain margin structure and the location of splay faults. We then build on published kinematic slip models of the 2006 Java earthquake source by performing global P-wave back-projection using two different frequency bands to examine the earthquake kinematics. In addition, we back-propagate first-crest arrivals in tsunami waveforms of five nearby tide gauges at various azimuths to locate tsunami sources. Our high-frequency back-projection results suggest a unilateral rupture extending ∼200 km with a slow first-stage rupture (∼1.2 km/s) from west to east until ∼65 s and a fast second-stage rupture (∼ 2.7 km/s) from ∼65 to 150 s. The second-stage rupture colocates with a major tsunami source located by first-crest tsunami back-propagation. The spatial correlation between the stage-two rupture imaged by back-projection and splay fault traces delineated by gravity data suggests that splay faults may have been reactivated during the 2006 Java earthquake and possibly contributed to tsunamigenesis. This mechanism of enhanced tsunami excitation due to splay faulting has been proposed for the 1944 Mw 8.1 Tonankai earthquake in the Nankai Trough (Moore et al., 2007).

Section snippets

Tectonic setting and residual gravity anomaly

The Java subduction zone accommodates underthrusting of the Indo-Australian plate beneath Eurasia at approximately 67 mm/yr (Tregoning et al., 1994). The incoming plate in offshore western Java is structurally complex, hosting a dense population of seamounts and the Roo Rise oceanic plateau (Shulgin et al., 2011). The forearc is characterized by an outer-arc high, which typically extends 100 km from the trench-axis with water-depths of 2–3 km Kopp et al., 2002, Planert et al., 2010. Landward of

Seismic P-wave back-projection

We perform P-wave back-projection using the procedure described in Fan and Shearer (2015), using vertical-component velocity records from the International Federation of Digital Seismograph Networks (FDSN) seismic stations that are available and distributed by the Data Management Center (DMC) of the Incorporated Research Institutions for Seismology (IRIS). Because back-projection techniques do not make assumptions about fault geometry or rupture velocity, they are able to resolve complex

Tsunami tide gauge back-propagation

To constrain tsunami source locations, we perform tsunami back-propagation with five nearby tide gauges recording the tsunami of the 2006 Java earthquake (Fig. 5a). The tsunami waveforms are high-pass filtered at 2 h to remove tidal signals, from which the initial and first-crest arrivals are estimated (Table 1). With the first-crest arrivals, back-propagation of tsunami waves from the tide gauges is used to delineate possible source locations of sea surface displacements (Fig. 5b–f). We

Discussion

Tsunami waveform inversion suggests that the tsunami source of the 2006 Java earthquake was about 200 km long with the largest slip (∼2.5 m) stably located about 150 km east of the epicenter, regardless of the assumed earthquake rupture velocity (Fujii and Satake, 2006). This tsunami-derived slip model is significantly different from the seismic slip models Ammon et al., 2006, Bilek and Engdahl, 2007, Fujii and Satake, 2006, Yagi and Fukahata, 2011, Ye et al., 2016, which suggest the largest slip

Conclusions

The 2006 Mw 7.8 Java earthquake ruptured more than 200 km from west to east, lasting for more than ∼180 s. Finite-fault slip models suggest a smooth and slow rupture with the largest slip patch within ∼50 km away from the hypocenter Ammon et al., 2006, Bilek and Engdahl, 2007, Yagi and Fukahata, 2011, Ye et al., 2016, Ye et al., 2016, which is supported by our low-frequency back-projection results. In contrast, high-frequency global back-projection results suggest a two-stage rupture. The first

Acknowledgments

We thank the editor Dr. Kelin Wang and two reviewers for their constructive suggestions, which led to improvements in our paper. We would also like to thank Lingling Ye for sharing her finite-fault slip model. Finite-fault slip model of Yagi and Fukahata (2011) is downloaded from the Source Inversion Validation (SIV) database (SRCMOD, http://equake-rc.info/) (Mai et al., 2016). Finite-fault slip model obtained with both body and surface waves is downloaded from the U.S. Geological Survey

References (75)

  • BilekS.L. et al.

    Rupture characterization and aftershock relocations for the 1994 and 2006 tsunami earthquakes in the Java subduction zone

    Geophys. Res. Lett.

    (2007)
  • DeDontneyN. et al.

    Applying wedge theory to dynamic rupture modeling of fault junctions

    Bull. Seismol. Soc. Am.

    (2012)
  • DeDontneyN. et al.

    Tsunami wave analysis and possibility of splay fault rupture during the 2004 Indian ocean earthquake

    Pure Appl. Geophys.

    (2012)
  • DenolleM.A. et al.

    Dynamics of the 2015 M7.8 Nepal earthquake

    Geophys. Res. Lett.

    (2015)
  • DuputelZ. et al.

    Real-time W-phase inversion during the 2011 off the Pacific coast of Tohoku earthquake

    Earth. Planets. Space.

    (2011)
  • EfronB. et al.

    An Introduction to the Bootstrap

    (1994)
  • FanW. et al.

    Detailed rupture imaging of the 25 April 2015 Nepal earthquake using teleseismic P waves

    Geophys. Res. Lett.

    (2015)
  • FanW. et al.

    Local near instantaneously dynamically triggered aftershocks of large earthquakes

    Science

    (2016)
  • FritzH.M. et al.

    Extreme runup from the 17 July 2006 Java tsunami

    Geophys. Res. Lett.

    (2007)
  • FujiiY. et al.

    Source of the July 2006 west Java tsunami estimated from tide gauge records

    Geophys. Res. Lett.

    (2006)
  • FukahataY. et al.

    Theoretical relationship between back-projection imaging and classical linear inverse solutions,

    Geophys. J. Int.

    (2014)
  • FukaoY.

    Tsunami earthquakes and subduction processes near deep-sea trenches

    (1979)
  • GarciaE.S. et al.

    Retracking Cryosat-2, Envisat and Jason-1 radar altimetry waveforms for improved gravity field recovery

    Geophys. J. Int.

    (2014)
  • HayesG.P.

    Rapid source characterization of the 2011 Mw 9.0 off the Pacific coast of Tohoku earthquake

    Earth Planets Space

    (2011)
  • HayesG.P. et al.

    88 hours: The U.S. Geological Survey National Earthquake Information Center response to the 11 March 2011 Mw 9.0 Tohoku earthquake

    Seismol. Res. Lett.

    (2011)
  • HayesG.P. et al.

    Slab1.0: A three-dimensional model of global subduction zone geometries

    J. geophys. Res.

    (2012)
  • HouserC. et al.

    Shear and compressional velocity models of the mantle from cluster analysis of long-period waveforms

    Geophys. J. Int.

    (2008)
  • On-Line Bulletin

    (2013)
  • IshiiM. et al.

    Extent, duration and speed of the 2004 Sumatra-Andaman earthquake imaged by the Hi-net array

    Nature

    (2005)
  • LaskeG. et al.

    Update on CRUST1.0-A 1-degree global model of Earth's crust

  • JiangJ. et al.

    Probabilistic imaging of tsunamigenic seafloor deformation during the 2011 Tohoku-oki earthquake

    J. Geophys. Res.

    (2016)
  • KameN. et al.

    Effects of prestress state and rupture velocity on dynamic fault branching

    J. Geophys. Res

    (2003)
  • KanamoriH. et al.

    The 1992 Nicaragua earthquake: a slow tsunami earthquake

    Nature

    (1993)
  • KennettB.L.N. et al.

    Traveltimes for global earthquake location and phase identification

    Geophys. J. Int.

    (1991)
  • KiserE. et al.

    The 2010 Mw 8.8 Chile earthquake: triggering on multiple segments and frequency-dependent rupture behavior

    Geophys. Res. Lett

    (2011)
  • KiserE. et al.

    Hidden aftershocks of the 2011 Mw 9.0 Tohoku, Japan earthquake imaged with the backprojection method

    J. Geophys. Res.

    (2013)
  • KoperK.D. et al.

    Frequency-dependent rupture process of the 2011 Mw 9.0 Tohoku earthquake: comparison of short-period P wave backprojection images and broadband seismic rupture models

    Earth Planets Space

    (2011)
  • Cited by (26)

    • Raised potential earthquake and tsunami hazards at the North Sulawesi subduction zone after a flurry of major seismicity

      2023, Marine and Petroleum Geology
      Citation Excerpt :

      Consequently, if an earthquake enucleates in the downdip segment of the megathrust and propagates into this shallow segment underneath the frontal prism, it could rupture continuously and easily to the mega-splay fault, inducing damaging tsunami waves. Indeed, examples are observed in 1975 and 1963 tsunami earthquakes in Kuril subduction zone (Fukao, 1979) and more recently in 2006 Mw 7.8 Java tsunami earthquake (Fan et al., 2017), and also trench-breaking mega earthquakes e.g., 1946 Mw 8.2 Nankai earthquake (Cummins and Kaneda, 2000), 2004 Mw 9.2 Sumatra earthquake (Singh et al., 2008). The existence of the splay fault posts a large tsunamigenic potential that is highly necessary to be considered in future hazard assessment at the NSSZ.

    • Great earthquake and tsunami potential in the eastern Makran subduction zone: New insights from geodetic and structural constraints

      2022, Tectonophysics
      Citation Excerpt :

      Among these data sets the seismic structural analysis from seismic images shows the MSZ has the world's thickest sedimentary accretionary wedge (∼7 km), in which it has developed numerous complex imbricate thrusts (Kopp et al., 2000; Mokhtari, 2015; Mokhtari et al., 2019; Mokhtari et al., 2008). Although it is challenging to assess the activities of these thrusts due to the lack of seafloor measurements, global observations from tsunami earthquakes and trench-breaking giant ruptures demonstrate these faults may have ruptured and played an important role in tsunami excitation along coastlines in the rupture region (Cummins and Kaneda, 2000; DeDontney and Rice, 2012; Fan et al., 2017; Fukao, 1979; Hananto et al., 2020; Hubbard et al., 2015; Lieser et al., 2014; Park et al., 2002; Singh et al., 2008; Suleimani and Freymueller, 2020). Therefore, in addition to the megathrust, we also incorporate the source mechanism from these imbricate thrusts of the outer wedge to achieve an improved understanding of the tsunami wave dynamics and hazards.

    • Tsunami excitation in the outer wedge of global subduction zones

      2022, Earth-Science Reviews
      Citation Excerpt :

      The later Ms. 7 event to the north was responsible for the strong shaking on the Nias island [Martin et al., 2019]. Similar slow and fast two-stages ruptures have also been reported in the 2006 Java [Fan et al., 2017] and the 1963 and 1975 Kuril tsunami earthquakes [Fukao, 1979]. Combining the near- and far-field tsunami observations with forward tsunami modeling, Martin et al. [2019] estimated a ~220 km rupture extending along the shallow Sunda megathrust with ~21 m of maximum slip offshore the Simeulue and Nias islands (Fig. 3).

    • Seafloor morphology along the active margin in Guerrero, Mexico: Probable earthquake implications

      2020, Journal of South American Earth Sciences
      Citation Excerpt :

      It is the first study of its kind in the GSG. This is a new important piece of puzzle in understanding seismic and especially tsunami hazard in the offshore area (e.g. Simons et al., 2011; Fan et al., 2017), alongside with previous works focused on slab geometry and/or seismicity (e.g. Suárez et al., 1990; Graham et al., 2016), and paleoseismological studies onshore (e.g. Ramírez-Herrera et al., 2007, 2009, 2012). The GSG lies in an area of subduction-related tectonics where the Cocos plate subducts beneath the North American plate slightly oblique to the trench normal (9–14°; Gaidzik et al., 2016) and with a speed of subduction increasing from northwest (64 mm/year) to southeast (66 mm; DeMets et al., 2010).

    • Sediment-starved trenches and rough subducting plates are conducive to tsunami earthquakes

      2019, Tectonophysics
      Citation Excerpt :

      It has been speculated that the 2006 earthquake may have been laterally bounded by excess topography on the subducting plate (Bilek and Engdahl, 2007; Bassett and Watts, 2015a) but the lack of high-resolution marine geophysical data in the area complicates the analyses. Furthermore, Fan et al. (2017) showed evidence that upper-plate faults may have been involved in the second stage of the rupture therewith contributing to the tsunami. Off Java, Cretaceous oceanic basement of the Australian Plate subducts under the Sunda Plate.

    View all citing articles on Scopus
    View full text