Influence of fault friction heterogeneities on strain accumulation and release on Megathrust
Over the last few years, our view of how Megathrust faults behave over the seismic cycle has evolved significantly because of advances in geodetic monitoring techniques, and due to the occurrence of a number of exceptionally large Megathrust earthquakes. The sequence on the Sumatra Megathrust, which started with the giant Mw 9.15 earthquake of 2004, has been particularly instructive. These earthquakes occurred within the area monitored by the Sumatra Geodetic Array (SuGAr), which provided exceptional records of near field co- seismic and postseismic ground displacements. Ground displacement in the interseismic period has also been documented from geodetic and paleogeodetic data, making it possible to describe the full process and stress accumulation and release over the seismic cycle. The Mw 8.0, 2007 Pisco earthquake offshore Peru is another instructive example where co- post and interseismic deformation can be observed. The emerging picture is that 1- the pattern of locking of the Megathrust in the interseismic period is highly heterogeneous showing both downdip and along strike variations; 2- the large earthquakes rupture patches which lie within areas that remain locked in the interseismic period and don't reach the trench; -3 postseismic deformation is dominated by afterslip on patches surrounding the seismically ruptured areas in particular at shallow depth along the trench. These observations suggest that friction properties of faults are heterogeneous with interfingering of patches with velocity-weakening frictional properties (mostly undergoing stick-slip motion) or velocity-strengthening frictional properties (mostly creeping aseismically). The topology of this patchwork probably exerts a fundamental control on the seismic behavior of Megathrust. Numerical simulations show that depending on the characteristics of intervening rate-strengthening areas, rate- weakening patches might tend to break in isolation or rupture jointly to produce larger events. This model can explain the kind of systematic and non systematic patterns observed in reality. Geodetic observations of interseismic strain can therefore help assess spatial variations of frictional properties and their eventual effect on seismic ruptures. For example, a locally low interseismic coupling can reveal a systematic barrier to seismic rupture propagation due to a rate-strengthening patch.
Measuring surface deformation in subduction zones with InSAR: Examples from South America and Cascadia
Ground displacements in subduction zones provide several important constraints: the location and magnitude of fault slip on the megathrust during large earthquakes and slow slip events, as well as the nature and extent of inter-seismic, post-seismic, and even pre-seismic surface deformation. The development of satellite Interferometric Synthetic Aperture Radar (InSAR) has allowed such ground displacements to be measured in many areas without dense arrays of continuously recording instruments on the ground, such as South America and Indonesia. Even where such arrays exist (like in Cascadia), InSAR observations can complement ground observations with increased spatial coverage or by helping to reconstruct the full three-dimensional deformation field. We will demonstrate the capabilities of InSAR in subduction zones with examples of co- seismic, post-seismic, and inter-seismic ground displacement in the Peru-Chile subduction zone including nine subduction zone earthquakes (6.7 < Mw < 8.5). Because of the arid climate in southern Peru and northern Chile, conventional C-band (5.6 cm radar wavelength) InSAR is successful, although like all geodetic methods, InSAR can be effected by changes in the refractive properties of the troposphere and ionosphere. We demonstrate that these atmospheric effects and orbital errors must be removed or accounted for when measuring small amplitude deformation over large spatial scales. In other subduction zones with more vegetation ground cover, L-band (23 cm wavelength) InSAR is successful, but observations are infrequent and the available data starts in 2006. To develop a longer time series (starting in 1992), we will demonstrate the potential of persistent scatterer C-band InSAR to reveal surface deformation in vegetated areas like Cascadia.
Subducting Seamounts and the Rupturing Process of Great Subduction Zone Earthquakes
It was suggested in the 1970's that subducting ocean floor features may delimit the along-strike rupture lengths of large subduction zone earthquakes. With the dramatic improvement in data quality, both for seismic and ocean floor bathmetry data, we can now see how the actual rupturing process of great earthquakes is also influenced by such subducting features. Here we present three great (Mw > 8) subduction zone earthquakes, in very different parts of the world, for which a relation between the ocean floor and the earthquake source process is seen. These include the 1986 Andreanof Islands, Alaska and the 1996 Biak, Indonesia earthquakes, in which the regions of large slip concentrate in patches, reminiscent of the "asperity model" of earthquakes, and appear to be related to subducted seamounts. For the 2001 Peru earthquake, a subducting fracture zone, with its associated bathymetric peak and trough, seems to have been the cause of the rupture being stalled for ~30s, before producing an earthquake of Mw 8.4, the third largest earthquake worldwide since 1965. Similarities and differences in the earthquake rupturing properties for these two different types of subducting features will be discussed. An outstanding question is what controls whether a seamount obducts or subducts.
Structure and Deformation Across Epicentral Region of 26 December 2004 Great Sumatra Earthquake
Joint interpretation of deep seismic reflection and wide angle seismic data reveals detailed structure and deformation of the subduction zone across the epicentral region of the 26 December 2004 Great Sumatra earthquake. This seismic line starts from the oceanic plate through the accretionary prism, outer-arc high (Simeulue plateau) and Simeuleu fore-arc basin. Top and bottom of subducting Indo-Australia plate, can be followed on the seismic section from 8.5 s (in the oceanic part) downto about 14 s TWT beneath the Simuelue fore-arc basin and are segmented by several landward dipping faults. Velocity structure derived from wide angle data estimates the thickness of this subducting oceanic plate is about 4.5 km. The accretionary prism is characterized by massive active faulting and folding, in which several of the faults may splay downto the subducting plate. Simeuleu plateau is formed by very compacted sediments characterized by low seismic penetration and may reflect the dynamic backstop to the accretionary prism. Major structural feature on the Simeulue plateau is the push up ridge which is interpreted as the continuation of West Andaman Fault (WAF). Further landward, Tuba ridge borders the outer arc high and Simeuleu fore-arc basin. Underneath Simeuleu basin, a strong landward verging reflector (backthrust) is marked and was inactive since Early/Mid Miocene as indicated by structural reconstruction based on seismic stratigraphy. Reactivation of this fault during late Miocene/Pliocene may have built up Tuba ridge as well as the piggy back basin on its seaward direction. The backthrust may demarcate the static (continental) backstop to the entire accretionary prism. This continental backstop was formed by accumulation of different arc/continental materials and is likely to be most heterogeneous along the margin.
Backthrusting And Forearc Evolution At Northern Sumatra Subduction Zone
Sumatra megathrust system is an obliquely convergent boundary where slip-partitioning has been variously inferred from slip-vectors of recorded earthquakes and geodetic models. The submarine West Andaman fault (WAF) has been variously interpreted as the dextral twin of the on-land great Sumatra fault, accommodating the strike parallel convergence. However the deeper structure, activity and role of the WAF in inner forearc dynamics remain unclear. Using high-resolution seismic tomography results we demarcate the presence of a seaward dipping backstop buttress structure beneath the inner forearc high. The velocity structure of the backstop buttress indicates that it is composed of continental crust. On coincident deep seismic reflection profile the presence of active backthrusts is imaged down to at least 15 km depth, which can be traced to the surface near the western edge of the Aceh forearc basin. The relationship of the forearc backthrusting with the dextral strike-slip movement along WAF sheds light on the complex deformation of the forearc at this obliquely convergent margin. Sharp bathymetric features at the seafloor suggest that the imaged backthrust branches are active, which can also be related to aftershock patterns in the region. Uplifting along the backthrust branches may explain the presence of forearc islands observed all along Sumatran margin and their evolution. Moreover, if these backthrusts slip coseismically during megathrust events they would contribute to tsunami and seismic hazard in the region.
Tsunami risk due to Active Backthrusting and Landslides at the NE Margin of Mentawai Islands, SW Sumatra
The Indo-Australian plate subducts obliquely beneath the Sunda plate leading to a slip partitioning into pure thrust and strike-slip motion. Just in the last three and half years, three pure thrust earthquakes of Mw>8.4 have occurred along this subduction interface. The great Sumatra Fault, traversing the Sumatra continental block along the volcanic arc, takes up a significant part of the trench parallel strike-slip motion. Moreover the submarine Mentawai Fault bounding the NE margin of Mentawai forearc Islands has also been suggested to accommodate a part of the strike-slip motion. Although the great Sumatra Fault is active, no significant seismicity has been observed along the Mentawai fault. Based on combined interpretation of high- resolution seismic reflection and bathymetry dataset, we infer that the Mentawai Fault is characterised by active SW dipping backthrusts instead. The presence of recent steeply dipping thrust earthquakes confirms that these faults are active. We also observed several large landslides, which could be earthquake triggered. Large localised uplift along the steeply dipping backthrusts and associated landslides at the NE margin of Mentawai Islands in the fully locked region pose serious seismic and tsunami risks to the SW coast of Sumatra in the coming future.