Search

Recent megathrust events in Tohoku (Japan), Maule (Chile), and Sumatra (Indonesia) were well recorded. Much has been learned about the dominant physical processes in megathrust zones: (partial) locking of the plate interface, detailed coseismic slip, relocking, afterslip, viscoelastic mantle relaxation, and interseismic loading. These and older observations show complex spatial and temporal patterns in crustal deformation and displacement, and significant differences among different margins. A key question is whether these differences reflect variations in the underlying processes, like differences in locking, or the margin geometry, or whether they are a consequence of the stage in the earthquake cycle of the margin. Quantitative models can connect these plate boundary processes to surficial and far-field observations. We use relatively simple, cyclic geodynamic models to isolate the first-order geodetic signature of the megathrust cycle. Coseismic and subsequent slip on the subduction interface is dynamically (and consistently) driven. A review of global preseismic, coseismic, and postseismic geodetic observations, and of their fit to the model predictions, indicates that similar physical processes are active at different margins. Most of the observed variability between the individual margins appears to be controlled by their different stages in the earthquake cycle. The modeling results also provide a possible explanation for observations of tensile faulting aftershocks and tensile cracking of the overriding plate, which are puzzling in the context of convergence/compression. From the inversion of our synthetic GNSS velocities we find that geodetic observations may incorrectly suggest weak locking of some margins, for example, the west Aleutian margin.

Subduction is substantially multi-scale process where the stresses are built by long-term tectonic motions, modified by sudden jerky deformations during earthquakes, and then restored by following multiple relaxation processes. Here, we develop a cross-scale thermomechanical model aimed to simulate the subduction process from 1 minute to million years' time scale. The model employs elasticity, nonlinear transient viscous rheology, and rate-and-state friction. It generates spontaneous earthquake sequences and by using an adaptive time-step algorithm, recreates the deformation process as observed naturally during the seismic cycle and multiple seismic cycles. The model predicts that viscosity in the mantle wedge drops by more than three orders of magnitude during the great earthquake with a magnitude above 9. As a result, the surface velocities just an hour or day after the earthquake are controlled by viscoelastic relaxation in the several hundred km of mantle landward of the trench and not by the afterslip localized at the fault as is currently believed. Our model replicates centuries-long seismic cycles exhibited by the greatest earthquakes and is consistent with the postseismic surface displacements recorded after the Great Tohoku Earthquake. We demonstrate that there is no contradiction between extremely low mechanical coupling at the subduction megathrust in South Chile inferred from long-term geodynamic models and appearance of the largest earthquakes, like the Great Chile 1960 Earthquake.

The relationship between individual earthquakes and the longer-term growth of topography and of geological structures is not fully understood, but is key to our ability to make use of topographic and geological datasets in the contexts of seismic hazard and wider-scale tectonics. Here we investigate those relationships at an active fold-and-thrust belt in the southwest Tarim Basin, Central Asia. We use seismic waveforms and interferometric synthetic aperture radar (InSAR) to determine the fault parameters and slip distribution of the 2015 Mw 6.3 Pishan earthquake - a blind, reverse-faulting event dipping towards the Tibetan Plateau. Our earthquake mechanism and location correspond closely to a fault mapped independently by seismic reflection, indicating that the earthquake was on a pre-existing ramp fault over a depth range of ˜9–13 km. However, the geometry of folding in the overlying fluvial terraces cannot be fully explained by repeated coseismic slip in events such as the 2015 earthquake nor by the early postseismic motion shown in our interferograms; a key role in growth of the topography must be played by other mechanisms. The earthquake occurred at the Tarim-Tibet boundary, with the unusually low dip of 21° . We use our source models from Pishan and a 2012 event to argue that the Tarim Basin crust deforms only by brittle failure on faults whose effective coefficient of friction is ≤0.05±0.025. In contrast, most of the Tibetan crust undergoes ductile deformation, with a viscosity of order 10²⁰–10²² Pa s. This contrast in rheologies provides an explanation for the low dip of the earthquake fault plane.

The Queen Charlotte plate boundary, near Haida Gwaii, B.C., includes the dextral, strike-slip, Queen Charlotte Fault (QCF) and the subduction interface between the downgoing Pacific and overriding North American plates. In this study, we present a comprehensive repeating earthquake catalogue that represents an effective slip meter for both structures. The catalogue comprises 712 individual earthquakes (0.3≤MW≤3.5) arranged into 224 repeating earthquake families on the basis of waveform similarity and source-separation estimates from coda wave interferometry. We employ and extend existing relationships for repeating earthquake magnitudes and slips to provide cumulative slip histories for the QCF and subduction interface in 6 adjacent zones within the study area between 52.3° N and 53.8° N. We find evidence for creep on both faults; however, creep rates are significantly less than plate motion rates, which suggests partial locking of both faults. The QCF exhibits the highest degrees of locking south of 52.8° N, which indicates that the seismic hazard for a major strike-slip earthquake is highest in the southern part of the study area. The October 28, 2012, MW 7.8 Haida Gwaii thrust earthquake occurred in our study area and altered the slip dynamics of the plate boundary. The QCF is observed to undergo accelerated, right-lateral slip for 1-2 months following the earthquake. The subduction interface exhibits afterslip thrust motion that persists for the duration of the study period (i.e., 3 years and 2 months after the Haida Gwaii earthquake). Afterslip is greatest (5.7 - 8.4 cm/yr) on the periphery of the main rupture zone of the Haida Gwaii event.

This study analyzes surface displacements generated by a low magnitude crustal earthquake in the Ecuadorian Andes by combining analysis of SAR Interferometry, geological field investigations and seismological data. In March 2010, a significant surface faulting event occurred in the Pisayambo area (Eastern cordillera), along the major dextral fault zone bounding the North Andean Sliver and the South-America Plate. Interferograms were inverted to determine fault plane geometry and slip displacement distribution. The event affected a 9 km-long previously unknown fault, referred as the Laguna Pisayambo Fault (LPF), with purely dextral movement reaching 45 cm and concentrated in the top 3 km of the crust. Geological investigations confirm both the fault mechanism and the amplitude of displacements. While these large displacements would be related to an event with a magnitude of 5.44 if using a standard crustal rigidity, we show that they can be convincingly associated with an Mw5.0 earthquake, that occurred on 2010/03/26. Reconciling the apparent differences in magnitude requires the existence of a low rigidity medium at shallow depths and/or postseismic activity of the fault. However, considering only the latter hypothesis would imply an unusually active postseismic process, in which 400-500% of the coseismic moment is released in the 6 days following the earthquake. Our observations highlight that the scaling laws relating surface observations to earthquake magnitude, classically used for seismic hazard assessment, should be carefully used. This study also illustrates how systematic InSAR analysis, even in places where no clues of ground deformation are present, can reveal tectonic processes.

We document interactions between intermediate depth and interplate seismicity in the North Chile subduction zone, over a 25-year period (1990 - 2015). We show that the 2005 Mw 7.8 Tarapaca slab-pull earthquake was followed by 9 years of enhanced deep and shallow seismicity, together with the decrease of eastward average GPS velocities and associated interplate coupling, eventually leading to the 2014 Mw 8.1 Iquique megathrust earthquake. In contrast, megathrust ruptures (e.g. Mw 8.0 Antofagasta in 1995, or Mw 8.1 Iquique in 2014) initiate several years of silent background seismicity in the studied area, both at shallow and intermediate depths. The plunge of a rigid slab into a viscous asthenospheric mantle could explain the observed synchronization between deep and shallow seismicity, and their long-term interactions.

In April 2017, a sequence of earthquakes offshore Valparaiso, Chile, raised concerns of a potential megathrust earthquake in the near future. The largest event in the 2017 sequence was a M6.9 on April 24th, seemingly co-located with the last great-sized earthquake in the region - a M8.0 in March 1985. The history of large earthquakes in this region shows significant variation in rupture size and extent, typically highlighted by a juxtaposition of large ruptures interspersed with smaller magnitude sequences. We show that the 2017 sequence ruptured an area between the two main slip patches during the 1985 earthquake, re-rupturing a patch that had previously slipped during the October 1973 M6.5 earthquake sequence. A significant gap in historic ruptures exists directly to the south of the 2017 sequence, with large enough moment deficit to host a great-sized earthquake in the near future, if it is locked.

In July 2013, a sequence of more than 340 earthquakes was induced by reservoir stimulations and well‐control procedures following a gas kick at a deep geothermal drilling project close to the city of St. Gallen, Switzerland. The sequence culminated in an ML 3.5 earthquake, which was felt within 10–15 km from the epicenter. High‐quality earthquake locations and 3‐D reflection seismic data acquired in the St. Gallen project provide a unique data set, which allows high‐resolution studies of earthquake triggering related to the injection of fluids into macroscopic fault zones. In this study, we present a high‐precision earthquake catalog of the induced sequence. Absolute locations are constrained by a coupled hypocenter‐velocity inversion, and subsequent double‐difference relocations image the geometry of the ML 3.5 rupture and resolve the spatiotemporal evolution of seismicity. A joint interpretation of earthquake and seismic data shows that the majority of the seismicity occurred in the pre‐Mesozoic basement, hundreds of meters below the borehole and the targeted Mesozoic sequence. We propose a hydraulic connectivity between the reactivated fault and the borehole, likely through faults mapped by seismic data. Despite the excellent quality of the seismic data, the association of seismicity with mapped faults remains ambiguous. In summary, our results document that the actual hydraulic properties of a fault system and hydraulic connections between its fault segments are complex and may not be predictable upfront. Incomplete knowledge of fault structures and stress heterogeneities within highly complex fault systems additionally challenge the degree of predictability of induced seismicity related to underground fluid injections.

Following a sequence of three Slow Slip Events (SSEs) on the northern Hikurangi Margin, between June 2015 and August 2016, a Mw 7.1 earthquake struck ~30 km offshore of the East Cape region in the North Island of New Zealand on the 2nd September 2016 (NZ local time). The earthquake was also followed by a transient deformation event (SSE or afterslip) northeast of the North Island, closer to the earthquake source area. We use data from New Zealand's continuous Global Positioning System (cGPS) networks to invert for the SSE slip distribution and evolution on the Hikurangi subduction interface. Our slip inversion results show an increasing amplitude of the slow slip towards the Te Araroa earthquake foreshock and mainshock area, suggesting a possible triggering of the Mw 7.1 earthquake by the later stage of the slow slip sequence. We also show that the transient deformation following the Te Araroa earthquake ruptured a portion of the Hikurangi Trench northeast of the North Island, further north than any previously observed Hikurangi margin SSEs. Our slip inversion and the coulomb stress calculation suggest that this transient may have been induced as a response to the increase in the static coulomb stress change downdip of the rupture plane on the megathrust. These observations show the importance of considering the interaction between slow slip events, seismic and aseismic events, not only on the same megathrust interface, but also on faults within the surrounding crust.

The JFAST drilling project endeavored to establish the stress state on the shallow subduction megathrust that slipped during the M9 Tohoku earthquake. Borehole breakout data from the drillhole can constrain both the orientation and magnitude of the principal stresses. Here we reanalyze that data to refine our understanding of the stress state on the fault. In particular, we: (1) Improve the identification of breakouts, (2) Consider a fuller range of stress states consistent with the data, and (3) Incorporate new and more robust laboratory constraints on rock strength. The original conclusion that the region is in a normal faulting regime after the earthquake is strengthened by the new analysis. The combined analysis suggests the earthquake released sufficient elastic strain energy to reset the local stress field.

Following the 2016 Mw 7.8 Kaikoura earthquake, uncertainty over the nature of the coseismic rupture developed. Seismological evidence pointed to significant involvement of the subduction megathrust, while geodetic and field observations pointed to a shallow set of intracrustal faults as the main participants during the earthquake. The addition of tsunami observations and modeling as reported in Bai et al. (2017) places additional constraints on the specific location of coseismic slip, which when combined with other observations indicates the simultaneous occurrence of shallow slip on the subduction interface and slip on overlying, upper crustal fault structures. This Kaikoura-style earthquake, involving synchronous ruptures on multiple components of the plate boundary, is an important mode of plate boundary deformation affecting seismic hazard along subduction zones.

On 25 December 2016, the Mw 7.6 Chiloe earthquake broke a plate-boundary asperity in South- Central Chile near the center of the rupture zone of the Mw 9.5 Valdivia earthquake of 1960. To gain insight on decadal-scale deformation trends and their relation with the Chiloe earthquake, we combine geodetic, teleseismic and regional seismological data. GPS velocities increased at continental scale after the 2010 Maule earthquake, probably due to a readjustment in the mantle flow and an apparently abrupt end of the viscoelastic mantle relaxation following the 1960 Valdivia earthquake. It also produced an increase in the degree of plate locking. The Chiloe earthquake occurred within the region of increased locking, breaking a circular patch of ~15 km radius at ~30 km depth, located near the bottom of the seismogenic zone. We propose that the Chiloe earthquake is a first sign of the seismic reawakening of the Valdivia segment, in response to the interaction between postseismic viscoelastic relaxation and changes of interseismic locking between Nazca and South-America.

We investigated spatiotemporal change in the interplate seismic activity following the 2011 Tohoku-oki earthquake (M9.0) in the region where interseismic interplate coupling was relatively weak and large postseismic slip was observed. We classified earthquakes by their focal mechanisms to identify the interplate events and conducted hypocenter relocation to examine the detailed spatiotemporal distribution of interplate earthquakes in the mostly creeping area. The results show that many interplate earthquakes, including M~6 events, emerged immediately after the Tohoku-oki earthquake in areas where very few interplate earthquakes had been observed in the 88 previous years. The emergent earthquakes include repeating sequences, and the extremely long quiescence of small to moderate earthquakes before the Tohoku-oki earthquake suggests that the source areas for the post-M9 events slipped aseismically during the quiescence. The repeaters’ magnitudes decayed over time following the Tohoku-oki earthquake and some sequences disappeared within a year. The emergence of interplate earthquakes suggests that areas where aseismic slip had been dominant before the Tohoku-oki earthquake, started to cause seismic slip after the earthquake, probably due to the increased loading rate from the afterslip. The magnitude decrease and disappearance of repeaters can be interpreted as shrinkage in seismic areas around the repeaters’ sources as the loading rate decreased due to the afterslip decay over time. These observations suggest that changes in the loading rate can cause slip behavior transition between seismic and aseismic. This indicates that such loading-rate-dependent slip behavior plays an important role in the spatiotemporal distribution of earthquakes in interplate seismogenic zones.

Earthquake swarms are characterized by an increase in seismicity rate that lacks a distinguished mainshock and does not obey Omori's law. At subduction zones, they are thought to be related to slow-slip events (SSEs) on the plate interface. Earthquake swarms in subduction zones can therefore be used as potential indicators of slow-slip events. However, the global distribution of earthquake swarms at subduction zones remains unclear. Here, we present a method for detecting such earthquake sequences using the space–time epidemic-type aftershock-sequence (ETAS) model. We applied this method to seismicity (M ≥ 4.5) recorded in the ANSS catalog at subduction zones during the period 1995–2009. We detected 453 swarms, which is about 6.7 times the number observed in a previous catalog. Foreshocks of some large earthquakes are also detected as earthquake swarms. In some subduction zones, such as at Ibaraki-Oki, Japan, swarm-like foreshocks and ordinary swarms repeatedly occur at the same location. Given that both foreshocks and swarms are related to SSEs on the plate interface, these regions may have experienced recurring SSEs. We then compare the swarm activity and tectonic properties of subduction zones, finding that swarm activity is positively correlated with curvature of the incoming plate before subduction. This result implies that swarm activity is controlled either by hydration of the incoming plate, or by heterogeneity on the plate interface due to fracturing related to slab bending.

Surface rupturing earthquakes produced by intraplate active faults generally have long recurrence intervals of a few thousands to tens of thousands of years. We here report the first evidence for an extremely short recurrence interval of 300 years for surface rupturing earthquakes on an intraplate system in Japan. The Kamishiro fault of the Itoigawa-Shizuoka Tectonic Line (ISTL) active fault system generated a Mw 6.2 earthquake in 2014. A paleoseismic trench excavation across the 2014 surface rupture showed the evidence for the 2014 event and two prior paleoearthquakes. The slip of the penultimate earthquake was similar to that of 2014 earthquake, and its timing was constrained to be after A.D. 1645. Judging from the timing, the damaged area, and the amount of slip, the penultimate earthquake most probably corresponds to a historical earthquake in A.D. 1714. The recurrence interval of the two most recent earthquakes is thus extremely short compared with intervals on other active faults known globally. Furthermore, the slip repetition during the last three earthquakes is in accordance with the time-predictable recurrence model rather than the characteristic earthquake model. In addition, the spatial extent of the 2014 surface rupture accords with the distribution of a serpentinite block, suggesting that the relatively low coefficient of friction may account for the unusually frequent earthquakes. These findings would affect long-term forecast of earthquake probability and seismic hazard assessment on active faults.

Myanmar is drawing rapidly increasing attention from the world for its seismic hazard. The Sagaing Fault (SF), an active right-lateral strike-slip fault passing through Myanmar, has been being the source of serious seismic damage of the country. Thus, awareness of seismic hazard assessment of this region is of pivotal significance by taking into account the interaction and migration of earthquakes with respect to time and space. We investigated a seismic series comprising10 earthquakes with M > 6.5 that occurred along the SF since 1906. The Coulomb failure stress modeling exhibits significant interactions among the earthquakes. After the 1906 earthquake, eight out of nine earthquakes occurred in the positively stress-enhanced zone of the preceding earthquakes, verifying that the hypothesis of earthquake triggering is applicable on the SF. Moreover, we identified three visible positively stressed earthquake gaps on the central and southern SF, on which seismic hazard is increased.

The 2016 M7.8 Kaikōura (New Zealand) earthquake struck the east coast of the northern South Island, resulting in strong ground shaking and large surface fault slip. Since the earthquake was well recorded by a local strong-motion seismic network, near-fault data may provide direct measurements of dynamic parameters associated with the fault-weakening process. Here we estimate a proxy for slip-weakening distance Dc′′, defined as double the fault-parallel displacement at the time of peak ground velocity, from accelerograms recorded at a near-fault station. Three-component ground displacements were recovered from the double numerical integration of accelerograms, and the corresponding final displacements are validated against coseismic displacement from geodetic data. The estimated Dc′′ is 4.9 m at seismic station KEKS located ∼2.7 km from a segment of the Kekerengu fault where large surface fault slip (∼12 m) has been observed. The inferred Dc′′ is the largest value ever estimated from near-fault strong motion data, yet it appears to follow the scaling of Dc′′ with final slip for several large strike-slip earthquakes. The energy budget of the M7.8 Kaikōura earthquake inferred from the scaling of Dc′′ with final slip indicates that a large amount of energy was dissipated by on- and off-fault inelastic deformation during the propagation of the earthquake rupture, resulting in a slower average rupture speed ( ≲2.0 km/s).

The seismicity in the Korean Peninsula has increased since the 2011 Mw9.0 Tohoku-Oki megathrust earthquake. Two strike-slip earthquakes with magnitudes of ML5.1 and 5.8 occurred in the southeastern Korean Peninsula on September 12, 2016. The two events occurred within 48 minutes. The ML5.8 earthquake was the largest event in the Korean Peninsula since 1978 when national seismic monitoring began. Both events produced strong high-frequency ground motions. More than 500 aftershocks with local magnitudes greater than or equal to 1.5 followed the events for two months. An unreported subsurface strike-slip fault with a dip of 65∘ to the east and a strike of N27∘ E was responsible for the earthquakes. The fault ruptured at depths of 11-16 km, resulting in a rupture plane of ∼26 km2. The aftershock distribution displayed horizontal streaks at a depth of ∼14 km, which was consistent with the focal mechanism solutions from long-period waveform inversion. The number of aftershocks decreased exponentially with time. The two ML5.1 and 5.8 earthquakes produced regional Coulomb stress changes of -4.9 to 2.5 bar. The spatial distribution of the aftershocks correlated with the Coulomb stress changes. The peak dynamic stress induced by strong ground motions reached 14.2 bar. The groundwater levels changed coseismically in some regions of decreased static stresses. The earthquakes on previously unidentified faults raised attention for the potential seismic hazards by earthquakes with long recurrence intervals.

The 17 December 2016 Solomon Islands earthquake (M_w 7.9) initiated ~103 km deep in the subducting Solomon Sea slab near the junction of the Solomon Islands and New Britain trenches. Most aftershocks are located near the Solomon Islands plate boundary megathrust west of Bougainville, where previous large interplate thrust faulting earthquakes occurred in 1995 (M_w 7.7) and 1971 (M_w 8.0). Teleseismic body wave modeling and aftershock relocations indicate that the initial 30 s of the 2016 rupture occurred over depths of 90 to 120 km on an intraslab fault dipping ~30° to the southwest, almost perpendicular to the dipping slab interface. The next 50 s of rupture took place at depths of 32 to 47 km in the deeper (Domain C) portion of the overlying megathrust fault dipping ~35° to the northeast. High susceptibility to triggering in the region accounts for this compound rupture of two separate fault planes.

Rupture history of the 2016 Mw 7.0 Kumamoto earthquake is constrained by using the waveforms of strong motion observations, teleseismic broadband body waves, and long-period surface waves. Its fault geometry is modeled with Hinagu (orienting 205° and dipping 73°) and Futagawa (orienting 235° and dipping 60°), two segments. The result reconciles the difference between moment tensor solutions and the surface fault trace. It reveals a complex rupture process that initiated on the Hinagu segment in dextral motion, propagated northeastward unilaterally, and after 15 s ceased near Aso volcano with normal fault motion. The average slip, rise time, and slip rate are 1.8 m, 2.0 s, and 1.2 m/s, respectively. The rupture broke through an ~30° fault intersection without notable delay, which can be a result of dynamic “unclamping.” The northeast boundary of the largest asperity might mark the bottom of the seismogenic zone, which becomes shallower gradually near Aso volcano.

We present an integrated approach of the seismic structure and activity along the offshore SW Hellenic subduction from combined observations of marine and land seismic stations. Our imaging of the slab top topography from teleseismic receiver function analysis at ocean bottom seismometers supports a trenchward continuation of the along-dip slab faults beneath the Peloponnesus. We further show that their morphostructural control accounts for the backstepping of the thrust contact of the Mediterranean Ridge accretionary wedge over the upper plate. Local seismic activity offshore SW Peloponnesus constrained by ocean bottom seismometer observations reveals a correlation with specific features of the forearc: the Matapan Troughs. We study the Mw6.8 14.02.2008 interplate earthquake offshore SW Peloponnesus and show that its nucleation, rupture zone, and aftershocks sequence are confined to one slab panel between two adjacent along-dip faults and are thus controlled by not only the offshore slab top segmentation but also the upper plate sea-bottom morphology.

The Mw7.8 Gorkha Earthquake struck Nepal and ruptured the boundary between the Indian and Eurasian plates. We conducted 2D Pg-wave tomography to clarify the seismogenic structure and try to understand causal mechanisms for this large earthquake, using the aftershock data recorded by 15 broadband seismic stations located near the China-Nepal border. Our high-resolution results show that coseismic slip area of the main shock is consistent with the high P-wave velocity anomaly, and the region of maximum slip has a larger area with higher velocity than the region of initial slip, possibly resulting in the dominant low-frequency radiation of energy observed after the dominant high-frequency radiation of energy in the source rupture process. The boundary between these regions of contrasting high and low seismic velocity anomalies suggests a potential crustal tearing at the southern end of the Tangra Yum Co Rift, possibly resulting from different thrust speeds in the Greater Himalaya.

Two oceanic plates (the Pacific (PA) and Philippine Sea (PH)) subduct beneath the land plate and they forms deeper (PH-PA) and shallower (land-PH) plate boundaries beneath Kanto. Remarkably enhanced seismicity was observed in the densely populated area after the 2011 Tohoku-oki earthquake which cannot simply be explained by the southern expansion of the postseismic slip of the earthquake. We examine interplate repeating earthquakes to constrain the relative plate motion across aseismically slipping faults. The repeater slip rates show creep rates in the deeper and shallower boundaries respectively increased to 2.4-6.6 times and 1.3-6.2 times the pre-Tohoku-oki rates. In addition, the repeater slip directions show no change larger than 4 degree in the deeper boundary. The interplate creep rates and slip directions suggest regional movements of both the two plates had accelerated. They probably caused the seismicity increase, frequent slow slips on the shallower boundary, and enhanced probability of larger earthquakes.

The relationship (scaling) between scalar moment, M0, and duration, T, potentially provides key constraints on the physics governing fault slip. The prevailing interpretation of M0-T observations proposes different scaling for fast (earthquakes) and slow (mostly aseismic) slip populations and thus fundamentally different driving mechanisms. We show that a single model of slip events within bounded slip zones may explain nearly all fast and slow slip M0-T observations, and both slip populations have a change in scaling, where the slip area growth changes from 2-D when too small to sense the boundaries to 1-D when large enough to be bounded. We present new fast and slow slip M0-T observations that sample the change in scaling in each population, which are consistent with our interpretation. We suggest that a continuous but bimodal distribution of slip modes exists and M0-T observations alone may not imply a fundamental difference between fast and slow slip.

The 2015 Mw 7.1 earthquake on the Charlie-Gibbs transform fault along the Mid-Atlantic Ridge is the latest in a series of seven large earthquakes since 1923. We propose that these earthquakes form a pair of quasi-repeating sequences with the largest magnitudes and longest repeat times for such sequences observed to date. We model teleseismic body waves and find that the 2015 earthquake ruptured a distinct segment of the transform from the previous 1998 earthquake. The two events display similarities to earthquakes in 1974 and 1967, respectively. We observe large oceanic transform earthquakes to exhibit characteristic slip behavior, initiating with small slip near the ridge, and propagating unilaterally to significant slip asperities nearer the center of the transform. These slip distributions combined with apparent segmentation support multimode slip behavior with fault slip accommodated both seismically during large earthquakes and aseismically in between.

The Mw8.4 Illapel earthquake occurred on 16 September was the largest global event in 2015. This earthquake was not unexpected because the hypocenter was located in a seismic gap of the Peru-Chile subduction zone. However, the source model derived from 3-D spectral-element inversion of teleseismic waves reveals a distinct two-stage rupture process with completely different slip characteristics as a composite megathrust event. The two stages were temporally separated. Rupture in the first stage, with a moment magnitude of Mw8.32, built up energetically from the deeper locked zone and propagated in the updip direction toward the trench. Subsequently, the rupture of the second stage, with a magnitude of Mw8.08, mainly occurred in the shallow subduction zone with atypical repeating slip behavior. The unique spatial-temporal rupture evolution presented in this source model is key to further in-depth studies of earthquake physics and source dynamics in subduction systems.

The 30 May 2015 Chichi-jima M8 earthquake is one of the largest deep focus earthquakes ever recorded and its depth of 682 km puts it near the base of the mantle transition zone. Before source mechanisms and slip models of this earthquake can be reliably assessed, a better understanding of the tectonic setting and structures of the region near the origin is required. Here we present evidence from receiver functions, a method of isolating subsurface material contrast with converted seismic waves, that the earthquake initiated within the upper mantle transition zone, above a significantly depressed 660 km phase boundary. Additionally, we observe multiple conversions within and below the transition zone, which we associate with seismic waves passing into and out of segments of the subducting Pacific plate. From this, we infer slab material is piling up at the base of the transition zone and segments are penetrating into the lower mantle.

Earthquake slip distributions are asymmetric along strike, but the reasons for the asymmetry are unknown. We address this question by establishing empirical relations between earthquake slip profiles and fault properties. We analyze the slip distributions of 27 large continental earthquakes in the context of available information on their causative faults, in particular on the directions of their long-term lengthening. We find that the largest slips during each earthquake systematically occurred on that half of the ruptured fault sections most distant from the long-term fault propagating tips, i.e., on the most mature half of the broken fault sections. Meanwhile, slip decreased linearly over most of the rupture length in the direction of long-term fault propagation, i.e., of decreasing structural maturity along strike. We suggest that this earthquake slip asymmetry is governed by along-strike changes in fault properties, including fault zone compliance and fault strength, induced by the evolution of off-fault damage, fault segmentation, and fault planarity with increasing structural maturity. We also find higher rupture speeds in more mature rupture sections, consistent with predicted effects of low-velocity damage zones on rupture dynamics. Since the direction(s) of long-term fault propagation can be determined from geological evidence, it might be possible to anticipate in which direction earthquake slip, once nucleated, may increase, accelerate, and possibly lead to a large earthquake. Our results could thus contribute to earthquake hazard assessment and Earthquake Early Warning.

We developed a new method to classify the faulting types of small earthquakes using interevent waveform similarity and applied it to earthquakes in the northeast Japan subduction zone. In the method, we used separate time windows for P and S waves and established a relationship between waveform similarity and differences in focal mechanisms from event pairs whose focal mechanisms we know. Then we applied the relationship to many pairs of such focal-mechanism-known events (5607 from the moment tensor catalogue and 3623 events from the interplate repeating earthquakes catalogue) and focal-mechanism-unknown events for the period from 1984 to 2013. As a result, 8984 earthquakes were newly classified into interplate (N = 5401), noninterplate thrust (N = 631), normal (N = 1070), and strike-slip faulting earthquakes (N = 165). From the new data set, which doubles the number of mechanism types, we confirmed that there have been almost no interplate earthquakes in the area of large coseismic slip of the Tohoku-oki earthquake since that event. We also saw that this trend continued until at least the end of 2013, suggesting a nearly complete stress release and slow interplate stress recovery. The abundant interplate aftershocks also indicate the precise spatial extent of postsesimic slip, which is usually difficult to obtain from land-based geodetic data. The postseismic slip also suggests stress concentration at the asperities of the 1968 Tokachi-oki (M7.9) and 1994 Sanriku-oki (M7.6) earthquakes. The present-day faulting types offshore Tohoku correlated well with the static-stress change from the Tohoku-oki earthquake, suggesting a stress state change during the earthquake cycle of megathrust earthquakes.

The 10 January 2012 Mw 7.2 Sumatra earthquake in the Wharton basin occurred 3 months before the great Mw 8.6 and Mw 8.2 earthquakes in the same region, which had complex ruptures and are the largest strike-slip earthquakes ever recorded. Teleseismic P wave back projection of the Mw 7.2 earthquake images a unilateral rupture lasting ∼40 s without observable frequency dependency (low frequency, 0.05–0.3 Hz, high frequency, 0.3–1 Hz). In addition to radiation bursts during the Mw 7.2 main shock, coherent energy releases from 50 to 75 s and from 100 to 125 s are observed about 143 km northeast of the main shock rupture and landward of the trench. Analysis of globally recorded P waves, in both 0.02–0.05 Hz velocity records and 1–5 Hz stacked envelope functions, confirms the presence of coherent sources during the time windows. The observed energy bursts are likely to be large early aftershocks occurring on or near the subduction interface. Both dynamic and static triggering could have induced these early aftershocks, as they initiated after the surface wave passed by, and the Coulomb stress perturbations from the Mw 7.2 main shock promote earthquakes in the observed locations. The earthquake sequence is a clear example of a seaward-intraplate strike-slip earthquake triggering landward-intraplate earthquakes in the same region, in contrast to previously reported normal-reverse or reverse-normal interactions at subduction zones.

On 16 September 2015, an Mw 8.3 earthquake struck middle Chile due to the subduction of the Nazca plate beneath the South America plate. This earthquake is the consequence of 72 years of strain accumulation in the region since the 1943 M 8.3 event. In this study, we apply the compressive sensing method (CS) to invert for the spatiotemporal distribution of the coseismic radiation at different frequencies of this event. The results show clear frequency-dependent feature of earthquake rupture with low-frequency (LF) radiation located in the updip region while high-frequency (HF) radiation concentrated in the downdip region of the megathrust. We also compare the CS results with three coseismic slip models as well as the stress drop distributions inferred from these slip models. The comparison confirms our understanding of coseismic radiation that energy sources are mostly located in the margin of large coseismic slip regions. Furthermore, we find that the LF radiation sources are mainly within the stress-decreasing (releasing) regions while the HF radiation sources are mainly located in the stress-increasing (loading) regions due to rupturing of relatively large asperities nearby (stress decreasing and releasing). These results help to better understand the physics of the rupture process during megathrust earthquakes. Moreover, our results do not show radiation sources south of the epicenter, suggesting that the subducting Juan Fernandez Ridge probably stopped the rupture of this earthquake toward the south.

A decadal-scale deformation transient preceding the 2011 Mw 9 Tohoku-oki, Japan, earthquake was reported from continuous GPS data and interpreted as accelerating aseismic slip on the Japan Trench megathrust. Given the unprecedented nature of this transient, independent confirmation of accelerating slip is required. Here we show that changes in the recurrence intervals of repeating earthquakes on the Japan Trench megathrust in the period 1996 to 2011 are consistent with accelerating slip preceding the Tohoku-oki earthquake. All sequences of repeating earthquakes with statistically significant trends in recurrence interval (at 95% confidence) offshore south central Tohoku occurred at an accelerating rate. Furthermore, estimates of the magnitude of slip acceleration from repeating earthquakes are consistent with the completely independent geodetic estimates. From a joint inversion of the GPS and seismicity data, we infer that a substantial portion of the megathrust experienced accelerating slip, partly surrounding the eventual rupture zone of the Mw 9 earthquake.

The subduction rate of an oceanic plate may accelerate after large earthquakes rupture the interplate coupling between the oceanic and overriding continental plates. To better understand postseismic deformation processes in an incoming oceanic plate, we directly measured the displacement rate of the Pacific Plate near the Japan Trench after the 2011 Tohoku-Oki earthquake using a GPS/acoustic technique over a period of 2 years (September 2012 to September 2014). The displacement rate was measured to be 18.0 ± 4.5 cm yr−1 (N302.0°E) relative to the North American Plate, which is almost twice as fast as the predicted interseismic plate motion. Because the sum of steady plate motion and viscoelastic response to the Tohoku-Oki earthquake roughly accounts for the observed displacement rate, we conclude that viscoelastic relaxation is the primary mechanism responsible for postseismic deformation of the Pacific Plate and that significant subduction acceleration did not occur at least not during the observation period.

The 2015 M7.8 Nepal earthquake ruptured part of the Main Himalayan Thrust beneath Kathmandu. To study the dynamics of this event, we compute P wave spectra of the main shock and of two large aftershocks to estimate stress drop and radiated energy. We find that surface reflections (depth phases) of these shallow earthquakes produce interference that severely biases spectral measurements unless corrections are applied. Measures of earthquake dynamics for the main shock are within the range of estimates from global and regional earthquakes. We explore the azimuthal and temporal variations of radiated energy and highlight unique aspects of the M7.8 rupture. The beginning of the earthquake likely experienced a dynamic weakening mechanism immediately followed by an abrupt change in fault geometry. Correlation of backprojection results with frequency-dependent variations in the radiated energy rate and with the suggested geometry of the Main Himalayan Thrust yields new constraints on dynamic ruptures through geometrical barriers.

Many efforts have been made to quickly estimate the maximum runup height of tsunamis associated with large earthquakes. This is a difficult task because of the time it takes to construct an accurate tsunami model using real-time data from the source. It is possible to construct a database of potential seismic sources and their corresponding tsunami a priori. However, such models are generally based on uniform slip distributions and thus oversimplify the knowledge of the earthquake source. Here we show how to predict tsunami runup from any seismic source model using an analytic solution that is specifically designed for subduction zones with a well-defined geometry, i.e., Chile, Japan, Nicaragua, and Alaska. The main idea of this work is to provide a tool for emergency response, trading off accuracy for speed. The solutions we present for large earthquakes appear promising. Here runup models are computed for the following: the 1992 Mw 7.7 Nicaragua earthquake, the 2001 Mw 8.4 Peru earthquake, the 2003 Mw 8.3 Hokkaido earthquake, the 2007 Mw 8.1 Peru earthquake, the 2010 Mw 8.8 Maule earthquake, the 2011 Mw 9.0 Tohoku earthquake, and the recent 2014 Mw 8.2 Iquique earthquake. The maximum runup estimations are consistent with measurements made inland after each event, with a peak of 9 m for Nicaragua, 8 m for Peru (2001), 32 m for Maule, 41 m for Tohoku, and 4.1 m for Iquique. Considering recent advances made in the analysis of real-time GPS data and the ability to rapidly resolve the finiteness of a large earthquake close to existing GPS networks, it will be possible in the near future to perform these calculations within the first minutes after the occurrence of similar events. Thus, such calculations will provide faster runup information than is available from existing uniform-slip seismic source databases or past events of premodeled seismic sources.

Large earthquakes that occur in subduction zones are likely to generate long-period ground motions that can cause severe damage even at great distances from the epicenter. We extracted surface-to-surface impulse response functions from the ambient seismic field recorded by offshore ocean bottom seismometers located atop the Nankai subduction zone and onshore stations. We showed that these offshore-onshore impulse response functions can be used to accurately simulate the long-period ground motions generated by an offshore moderate subduction earthquake. Moreover, we also found that the distributions of the earthquake and impulse response function pseudovelocity response spectra have similar maximum amplifications in the same area close to the earthquake epicenter. This suggests that the ambient seismic field recorded by the increasing number of ocean bottom seismometers around the world can be used to assess seismic hazard related to offshore subduction earthquakes without prior knowledge of the velocity structure.

We found very low frequency earthquakes (VLFEs) at a shallow subduction zone close to the Japan Trench off the Pacific coast of Tohoku, Japan. Centroid moment tensor solutions of VLFEs showed reverse fault mechanisms with a compression axis in the east-west direction. A cross-correlation analysis of seismograms with template events between 2005 and 2013 revealed three major VLFE clusters and their temporal evolution. A VLFE cluster in the central off-Tohoku region located in the large slip area of the 2011 Tohoku earthquake was detectable only before the Tohoku earthquake. However, VLFEs in the northern and southern off-Tohoku regions at the rim of the large slip area were activated after the Tohoku earthquake. The change in the activity may reflect the stress redistribution by the coseismic and/or afterslip processes of the Tohoku earthquake.

On 22 April 1997 the largest earthquake recorded in the Trinidad-Tobago segment of the Caribbean-South American plate boundary zone (Mw 6.7) ruptured a shallow (~9 km), ENE striking (~250° azimuth), shallowly dipping (~28°) dextral-normal fault ~10 km south of Tobago. In this study, we describe this earthquake and related foreshock and aftershock seismicity, derive coseismic offsets using GPS data, and model the fault plane and magnitude of slip for this earthquake. Coseismic slip estimated at our episodic GPS sites indicates movement of Tobago 135 ± 6 to 68 ± 6 mm NNE and subsidence of 7 ± 9 to 0 mm. This earthquake was anomalous and is of interest because (1) its large component of normal slip and ENE strike are unexpected given the active E-W dextral shearing across the Caribbean-South American plate boundary zone, (2) it ruptured a normal fault plane with a low (~28°) dip angle, and (3) it reactivated and inverted the preexisting Tobago terrrane-South America ocean-continent (thrust) boundary that formed during early Tertiary oblique plate convergence.

Time series analysis of interferometric synthetic aperture radar data reveals coseismic and postseismic surface displacements associated with the 2007 M5.5 earthquake along the southern Ghazaband fault, a major but little studied fault in Pakistan. Modeling indicates that the coseismic surface deformation was caused by ~9 cm of strike-slip displacement along a shallow subvertical fault. The earthquake was followed by at least 1 year of afterslip, releasing ~70% of the moment of the main event, equivalent to a M5.4 earthquake. This high aseismic relative to the seismic moment release is consistent with previous observations for moderate earthquakes (M

The characteristic earthquake hypothesis is the basis of time-dependent modeling of earthquake recurrence on major faults. However, the characteristic earthquake hypothesis is not strongly supported by observational data. Few fault segments have long historical or paleoseismic records of individually dated ruptures, and when data and parameter uncertainties are allowed for, the form of the recurrence distribution is difficult to establish. This is the case, for instance, of the Corinth Gulf Fault System (CGFS), for which documents about strong earthquakes exist for at least 2000 years, although they can be considered complete for M ≥ 6.0 only for the latest 300 years, during which only few characteristic earthquakes are reported for individual fault segments. The use of a physics-based earthquake simulator has allowed the production of catalogs lasting 100,000 years and containing more than 500,000 events of magnitudes ≥ 4.0. The main features of our simulation algorithm are (1) an average slip rate released by earthquakes for every single segment in the investigated fault system, (2) heuristic procedures for rupture growth and stop, leading to a self-organized earthquake magnitude distribution, (3) the interaction between earthquake sources, and (4) the effect of minor earthquakes in redistributing stress. The application of our simulation algorithm to the CGFS has shown realistic features in time, space, and magnitude behavior of the seismicity. These features include long-term periodicity of strong earthquakes, short-term clustering of both strong and smaller events, and a realistic earthquake magnitude distribution departing from the Gutenberg-Richter distribution in the higher-magnitude range.

We examined the temporal variation of the size of repeating earthquakes related to the 2011 Tohoku-oki earthquake (M9.0) in the northeastern Japan subduction zone for the period from July 1984 to the end of 2011. The repeaters (M2.5–6.1) show postseismic magnitude increases for most sequences located in the area of large postseismic slip at the downdip extension of the M9 source region. The magnitudes of the first events after the M9 earthquake increased by an average of about 0.3 for sequences having three or more earthquakes over the 9 months following it. We also examined the slip area in detail for Kamaishi repeaters whose magnitudes had been M4.9 ± 0.2 but which increased by about 1 after the M9 earthquake. Waveform modeling shows that the slip area for the post-M9 Kamaishi earthquakes overlaps with that before the Tohoku-oki earthquake but enlarged by about 6 times. Until the occurrence time of the last event (September 2011) in the analysis period, the rupture area remained larger than before but appeared to shrink over time. The enlargement of the rupture area suggests that an aseismic-to-seismic transition occurred in the region surrounding the pre-M9 repeaters and is most likely related to fast loading of the repeaters due to rapid postseismic slip estimated to have occurred in the area. The existence of conditionally stable regions around the repeating earthquakes and/or patches slightly larger than the earthquake nucleation sizes may explain such behavior. The temporal change of loading rate is an important factor in determining earthquake size in this case.

The 2011 Mw 9.0 Tohoku-Oki megathrust earthquake accompanied coseismic and postseismic displacements around the eastern Eurasian continental plate. Noise cross correlations produced transient seismic waveforms along interstation paths in the Korean Peninsula. We measured the traveltime changes of the fundamental mode Rayleigh waves over the range of 0.03–0.08 Hz after the megathrust earthquake. The temporal seismic velocity changes in the lower crust were assessed from the traveltime changes. The traveltimes increased instantly after the megathrust earthquake and were gradually recovered over several hundreds to thousands of days. The instant shear wave velocity decreases ranged between 0.731 (±0.057)% and 4.068 (±0.173)%. The temporal medium perturbation might be caused by the transient uniaxial tensional stress due to the coseismic and postseismic displacements. The medium properties may be recovered by progressive stress field reconstruction.

Motivated by the fact that temporal earthquake aperiodicity was observed off Kamaishi just after the 2011 Tohoku earthquake, we performed numerical simulations of chain reactions due to the postseismic slip of large earthquakes by applying rate- and state-dependent friction laws. If the repeater is composed of single asperity, our results show that, (i) a mixture of partial and whole rupturing of a single asperity can explain some of the observed variability in timing and size of the repeating earthquakes off Kamaishi; (ii) the partial rupturing can be reproduced with moderate frictional instability with the aging-law and not the slip or Nagata laws; (iii) the perturbation of the activated earthquake hypocenters observed mostly in the ESE-WNW direction may reflect the fact that the large postseismic slip of the 2011 Tohoku earthquake propagated from ESE to WNW off Kamaishi; (iv) the observed region of repeating earthquake quiescence may reflect the strong plate coupling of megathrust earthquakes.

The systematic behavior of earthquake rupture as a function of earthquake magnitude and/or tectonic setting is a key in our understanding of the physical mechanisms involved during earthquake rupture. Geophysical evidence suggests that although deep earthquakes—including intermediate-depth and deep—are similar to shallow ones, the mechanism involved during deep earthquakes is different from that of shallow ones. In particular, the magnitude and depth dependence of scaled duration, a measure of earthquake rupture duration, has led to controversy of what controls deep earthquake behavior. Here we estimate scaled source durations for 600 intermediate-depth and deep-focus earthquakes recorded at teleseismic distances and show deviation from self-similar scaling. No depth dependence is observed which we interpret as due to little differences between intermediate-depth and deep-focus earthquake mechanisms. The data show no correlation between durations and plate age or thermal parameters, suggesting that the thermal properties of the plate have little effect on source durations. We nevertheless report differences in average source duration and scaling between subduction zones and along-strike variations of source durations that more closely resemble the geometry of subduction (flat or steep subduction) rather than plate age.

The northern Vanuatu (formerly New Hebrides) subduction zone (11°S to 14°S) has experienced large shallow thrust earthquakes with Mw > 7 in 1966 (MS 7.9, 7.3), 1980 (Mw 7.5, 7.7), 1997 (Mw 7.7), 2009 (Mw 7.7, 7.8, 7.4), and 2013 (Mw 8.0). We analyze seismic data from the latter four earthquake sequences to quantify the rupture processes of these large earthquakes. The 7 October 2009 earthquakes occurred in close spatial proximity over about 1 h in the same region as the July 1980 doublet. Both sequences activated widespread seismicity along the northern Vanuatu subduction zone. The focal mechanisms indicate interplate thrusting, but there are differences in waveforms that establish that the events are not exact repeats. With an epicenter near the 1980 and 2009 events, the 1997 earthquake appears to have been a shallow intraslab rupture below the megathrust, with strong southward directivity favoring a steeply dipping plane. Some triggered interplate thrusting events occurred as part of this sequence. The 1966 doublet ruptured north of the 1980 and 2009 events and also produced widespread aftershock activity. The 2013 earthquake rupture propagated southward from the northern corner of the trench with shallow slip that generated a substantial tsunami. The repeated occurrence of large earthquake doublets along the northern Vanuatu subduction zone is remarkable considering the doublets likely involved overlapping, yet different combinations of asperities. The frequent occurrence of large doublet events and rapid aftershock expansion in this region indicate the presence of small, irregularly spaced asperities along the plate interface.

We propose a stochastic model for characteristically repeating earthquake sequences to estimate the spatiotemporal change in static stress loading rate. These earthquakes recur by a cyclic mechanism where stress at a hypocenter is accumulated by tectonic forces until an earthquake occurs that releases the accumulated stress to a basal level. Renewal processes are frequently used to describe this repeating earthquake mechanism. Variations in the rate of tectonic loading due to large earthquakes and aseismic slip transients, however, introduce nonstationary effects into the repeating mechanism that result in nonstationary trends in interevent times, particularly for smaller-magnitude repeating events which have shorter interevent times. These trends are also similar among repeating earthquake sites having similar hypocenters. Therefore, we incorporate space-time structure represented by cubic B-spline functions into the renewal model and estimate their coefficient parameters by maximizing the integrated likelihood in a Bayesian framework. We apply our model to 31 repeating earthquake sequences including 824 events on the Parkfield segment of the San Andreas Fault and estimate the spatiotemporal transition of the loading rate on this segment. The result gives us details of the change in tectonic loading caused by an aseismic slip transient in 1993, the 2004 Parkfield M6 earthquake, and other nearby or remote seismic activities. The degree of periodicity of repeating event recurrence intervals also shows spatial trends that are preserved in time even after the 2004 Parkfield earthquake when time scales are normalized with respect to the estimated loading rate.

We observed seafloor crustal deformation at two observation sites on opposite sides of the Suruga Trough off Japan from 2005 to 2011 to investigate the interplate locking condition at the source region of the anticipated great subduction earthquake, named Tokai earthquake. We estimated the displacement velocity vectors relative to the Amurian Plate on the basis of repeated observations. Our results at the two points, Suruga northeast and Suruga northwest (SNW) were 42 ± 8 mm/yr toward N94 ± 3°W and 39 ± 11 mm/yr toward N84 ± 9°W, respectively. These directions are the same as those measured at on-land GPS stations. The magnitudes of the velocity vectors indicate a significant shortening of approximately 4 mm/yr between SNW and on-land GPS stations located to the west of the Suruga Trough. The results show that the plate interface is strongly locked (no slip) shallower than the source region of the anticipated Tokai earthquake.

16 pages, 9 figures, We investigate potential relations between variations in seafloor relief and age of the incoming plate and interplate seismicity. Westward from Osa Peninsula in Costa Rica, a major change in the character of the incoming Cocos Plate is displayed by abrupt lateral variations in seafloor depth and thermal structure. Here a Mw 6.4 thrust earthquake was followed by three aftershock clusters in June 2002. Initial relocations indicate that the main shock occurred fairly trenchward of most large earthquakes along the Middle America Trench off central Costa Rica. The earthquake sequence occurred while a temporary network of OBH and land stations ∼80 km to the northwest were deployed. By adding readings from permanent local stations, we obtain uncommon P wave coverage of a large subduction zone earthquake. We relocate this catalog using a nonlinear probabilistic approach within both, a 1-D and a 3-D P wave velocity models. The main shock occurred ∼25 km from the trench and probably along the plate interface at 5-10 km depth. We analyze teleseismic data to further constrain the rupture process of the main shock. The best depth estimates indicate that most of the seismic energy was radiated at shallow depth below the continental slope, supporting the nucleation of the Osa earthquake at ∼6 km depth. The location and depth coincide with the plate boundary imaged in prestack depth-migrated reflection lines shot near the nucleation area. Aftershocks propagated downdip to the area of a 1999 Mw 6.9 sequence and partially overlapped it. The results indicate that underthrusting of the young and buoyant Cocos Ridge has created conditions for interplate seismogenesis shallower and closer to the trench axis than elsewhere along the central Costa Rica margin. © 2014. American Geophysical Union. All Rights Reserved, This work was supported by DFG grant GR1964/17-1 and by former project SFB574