Search

Analogue models are commonly used to model long-term geological processes such as mountain building or basin inversion. The majority of these models use granular materials like sand or glass beads to simulate the brittle behaviour of the crust. In granular materials deformation is localized into shear bands that act as analogues to natural fault zones and detachments. Shear bands aka faults are persistent anomalies in the granular package and are frequently reactivated during an experimental run. This is due to their lower strength in comparison to the undeformed bulk material. When fault motion stops, time dependent healing immediately starts to increase the strength of the fault. Therefore, older faults show a higher strength in comparison to younger faults. This time dependent healing, also called time consolidation, can therefore affect the structural style of an analogue model due to evolution of fault strength over time. Time consolidation is a well known mechanism in granular mechanics but remains poorly characterized for analogue materials and on the timescales of typical analogue models. In this study, we estimate the healing rate of several analogue materials and evaluate the consequences on the reactivation potential of analogue faults. We find that the healing rates are generally below 3 % per tenfold increase in hold time which is comparable to natural fault zones. We qualitatively compare the frictional properties of the materials with grain characteristics and find a weak correlation of healing rates with sphericity and friction with an average quality score. In models where predefined faults exist or reactivation is forced by blocks, the stability region of fault angles that can be reactivated can accordingly decrease by up to 7° over the duration of 12 hours. The stress required to reactivate a preexisting fault can double in the same time which may favor the creation of new faults. In a basin inversion scenario, normal faults can not be inverted due to severe misorientation and therefore time consolidation plays only a minor additional role for such models.

Located on the overriding plate of the Hellenic subduction margin, the 250 km-long island of Crete offers a unique opportunity to study curved-forearc deformation. The African-Eurasian plate-convergence of ~40 mm/yr (~80 %) is primarily accommodated aseismically, but intense seismicity is recorded at the plate-interface and a reverse splay faults along the Hellenic trough; frequent M6+ earthquakes and (at least one) tsunami-genic event, causing up to 10 m of paleoshoreline uplift in western Crete, are reported. Global Navigation Satellite System (GNSS) data revealed N-S shortening of ~2 mm/yr within western Crete due to pure plate convergence. Further east, the curved subduction trench accommodates increased oblique slip, causing E-W extension of ~2 mm/yr in eastern Crete.Recently, the European Ground Motion Service published dense InSAR surface deformation data in East and Up direction of whole Europe. The InSAR time-series comprise positioning samples every six days, respectively, every ~50 m, and, in Crete, exhibit long-wavelength deformation signals caused by deep-rooted, tectonic sources that are overlaid by (often seasonally-modulated) signals originating in shallow aquifers. We analyze these time-series in space and time and validate the results using available GNSS rates, a seismic catalog and an active fault data base. Preliminary results suggest a slight eastward tilt of Crete, which is not confirmed by published GNSS rates, and has to be investigated further. Spatially-confined uplift of up to ~5 mm/yr are observed at the karstic Omalos plateau, and up to ~30 mm/yr subsidence in the Messara basin, both probably related to groundwater replenishment/abstraction. Relative eastward motion increases towards eastern Crete, particularly in the fault zones embracing Mirabello bay and east of it, thus confirming the aforementioned E-W extension, and towards the southern coast.

Aftershocks are a time-dependent (exponential decay) phenomenon in the aftermath of large earthquakes. In subduction zones, those occurring in the upper plate are of special concern given their potential seismic hazard, as they may produce substantial surface shaking close to highly populated cities. Therefore, the understanding of the mechanisms that drive upper-plate aftershocks is of utmost importance to improving seismic hazard assessment. Transfer of static coseismic stresses has been commonly proposed to explain this; however, they fail to explain their exponential decay over time. This time-dependency is observed in postseismic geodetic measurements, suggesting that the processes that control the postseismic surface deformation also govern or at least are involved in the generation of upper-plate aftershocks. Here, the postseismic surface deformation is dominated by aseismic slip along the fault interface (afterslip), non-linear viscoelastic relaxation in the lower crust and upper mantle, and pore-pressure diffusion in the crust. Despite great research efforts, however, the key driver remains elusive.In this study, we investigate which postseismic mechanism mainly controls the occurrence of aftershocks in the upper plate in subduction zones using the 2014 Mw=8.1 Iquique earthquake, northern Chile, as a study case. We employ a 4D numerical forward model to simulate the transient poroelastic and non-linear viscoelastic relaxation, whose contributions are subtracted from the cumulative Global Navigation Satellite System (GNSS) measurements to then invert for afterslip. Using realistic rock material properties, we first show that this approach explains the surface displacements during the first nine months of postseismic deformation recorded by continuous GNSS. For the same period, we then compute the spatiotemporal Coulomb Failure Stress changes (ΔCFS) that result from individual postseismic processes and compare them with the upper-plate aftershocks using a high-resolution seismicity catalog and focal mechanisms. We show for the first time that the ΔCFS produced by pore-pressure diffusion induced by the mainshock are unambiguously better correlated in space and time with the increase in upper-plate aftershocks than those from afterslip or non-linear viscous relaxation. In addition, pore-pressure diffusion lowers the effective normal stress of the stress tensor more effectively, while its resulting ΔCFS are relatively independent of the fault orientation. The latter would also explain the diversity of faulting styles in the upper plate exhibited by focal mechanisms following the 2014 Iquique earthquake and other subduction zone earthquakes. Our findings provide new insights into the link between pore-pressure diffusion and upper-plate deformation in subduction zones with implications for time-dependent seismic hazard.

The eastern Mediterranean island of Crete is located on the overriding plate of the Hellenic subduction thrust which is curved and changes strike from ~170° to ~50° in a west to east direction. Crete is located in the zone of maximum curvature of the subduction thrust. Basin and range topography together with prominent limestone scarps indicate that Quaternary deformation at the ground surface on Crete is dominated by normal faults with slip rates of up to ~1 mm/yr. These active faults comprise two primary sets that strike N-NNE (0-30°) and E-ESE (90-120°), with the more easterly faults dominating in southern Crete. Each fault set is characterised by dip slip and together they accommodate coeval W-WNW and N-NNE crustal extension. The E-ESE normal faults are approximately parallel to the strike of the subducting North African plate and form part of a regional fault system that swings in strike in sympathy with depth contours on the top of the concave northwards plate. By contrast, N-NNE normal faults are sub-parallel to the line of maximum curvature on the subduction thrust. These geometric relationships support the view that normal faulting on Crete formed, at least partly, in response to Cenozoic slab retreat (e.g., Jolivet et al., 2013), which continued into the Quaternary. In this model contemporaneous multi-directional crustal extension on Crete is driven by geologically simultaneous westward and southward retreat of the slab. Jolivet, L., Faccenna, C., Huet, B., Labrousse, L., Le Pourhiet, L., Lacombe, O., et al. (2013). Aegean tectonics: Strain localisation, slab tearing and trenchretreat. Tectonophysics, 597–598, 1–33. https://doi.org/10.1016/j.tecto.2012.06.

Stress in active subduction forearcs is controlled primarily by friction along the megathrust and the gravitational force. The competing deviatoric compressive and tensional stresses generated by megathrust friction and gravity, respectively, are of the same order of magnitude and result in very low deviatoric stress in the forearc. The near neutral stress state in subduction forearcs is supported by observations of stress reversal, that is a change from deviatoric compression to deviatoric tension, caused by very small megathrust shear stress drops (

Large megathrust earthquakes like the 2011 Mw 9.0 Tohoku earthquake (Japan) are followed by numerous aftershocks in the subduction zone forearc overlying the seismogenic fault. The aftershocks in the forearc can include normal-faulting events despite the thrust mechanism of the main shock. Postseismic normal faulting has been explained by stress changes induced by the coseismic stress drop along the megathrust. However, details of stress changes in the forearc and aftershock triggering mechanisms remain poorly constrained. Here we use numerical force-balance models combined with Coulomb failure analysis to show that the megathrust stress drop indeed supports normal faulting, but that forearc-wide triggering of aftershocks is feasible within a narrow range of megathrust stress-drop values and forearc stress states only. We determine this range for the Tohoku earthquake and show that the associated stress changes explain the aftershock seismicity in unprecedented detail. In particular, our analysis reveals that ~78% of the aftershocks and ~92% of the seismic moment release occurred in areas where the Tohoku earthquake caused a stress increase, and that the detailed aftershock distribution was also governed by spatial variability in fault strength and forearc topography. Our findings provide new insights into aftershock triggering and help to understand where aftershocks occur after great earthquakes at subduction zones.

Large magnitude (Mw ∼ ≥6) earthquakes in extensional settings are often associated with simultaneous rupture of multiple normal faults as a result of static and/or dynamic stress transfer. Here, we report details of the coseismic breaching of a previously unrecognized large-scale fault relay zone in central Greece, through three successive normal fault earthquakes of moderate magnitude (Mw 5.7–6.3) that occurred over a period of ∼10 days in March 2021. Specifically, joint analysis of InSAR, GNSS and seismological data, coupled with detailed field and digital fault mapping, reveals that the Tyrnavos Earthquake Sequence (TES) was accommodated at the northern end of a ∼100 km wide transfer structure, by faults largely unbroken during the Holocene. By contrast, the southern section of this relay zone appears to have accrued significant slip during Holocene. InSAR-derived displacements agree with the loci of eight subtle, previously undetected, faults that accommodated coseismic and/or syn-seismic normal fault slip during the TES. Kinematic modeling coupled with fault mapping suggests that all involved faults are interconnected at depth, with their conjugate fault-intersections acting largely as barriers to coseismic rupture propagation. We also find that the TES mainshocks were characterized by unusually high (>6 MPa) stress-drop values that scale inversely with rupture length and earthquake magnitude. These findings, collectively suggest that the TES propagated north-westward to rupture increasingly stronger asperities at fault intersections, transferring slip between the tips of a well-established, but previously unrecognized, relay structure. Fault relay zones may be prone to high stress-drop earthquakes and associated elevated seismic hazard.

The Kaikōura Earthquake uplifted Kaikōura Peninsula by ≤∼1 m. Uplift in 2016 mainly resulted from slip on an offshore thrust fault (OSTF), modelled to splay from the plate-interface, and was further influenced by slip on two newly identified faults (Armers Beach Fault, ABF; Te Taumanu Fault, TTF) mapped onshore from differential lidar (D-lidar). Forward dislocation modelling indicates that 2016 peninsula uplift can be reproduced by mean slip of ∼2.3 m on the OSTF and 0.25–0.5 m on the ABF and TTF. The variable co-seismic uplift recorded during the 2016 earthquake differs from the near-uniform (1.2 ± 0.2°) northwest tilting of MIS5c (96 ± 5 ka) and MIS5e (123 ± 5 ka) marine terraces; these ages are constrained by Optically Stimulated Luminescence (OSL) dating and correlation to sea-level curves. Tilting of Late Quaternary marine terraces can be primarily reproduced by slip rates of ∼0.8–2.7 mm/yr on the OSTF and 0.3–0.6 mm/yr on the ABF. Slip on the TTF is not required to produce tilting of the marine terraces, suggesting that it may have ruptured less frequently than the OSTF and ABF in the Late Quaternary. The OSTF links 2016 ruptures north and south of Kaikōura, with the earthquake rupturing an interconnected network of faults.

The interaction of the northern Nazca and southwestern Caribbean oceanic plates with South America, and the collision of the Panama-Choco arc have significant implications on the evolution of the northern Andes. We integrate an alternative interpretation of the Nazca and Caribbean kinematics with the magmatic and deformation history in the region. The northeastward migration of the Caribbean plate caused a progressive change in the geometry of the subducting Farallon plate, causing flat-slab subduction throughout the late Eocene-late Oligocene, inhibition of magmatism and eastward migration of the Andean deformation. Meanwhile, the Paleocene-Eocene highly oblique convergence of the Caribbean plate against South America changed by the mid-Eocene, when the Caribbean plate began to migrate in an easterly direction. These events and the late Oligocene breakup of the Farallon plate, prompted a Miocene plate reorganization, with further plate fragmentation, changes in convergence obliquity, steepening of the subducting slabs and renewal of magmatism. This tectonics was complicated by the accretion of the Panama-Choco arc to South America, which was characterized by early Miocene subduction erosion of the forearc and trench advance, followed by breakoff of the subducting slab east of Panama and collisional tectonics from the middle Miocene. By 9 Ma the Coiba and Malpelo microplates were attached to the Nazca plate, resulting in an abrupt change in convergence directions, that correlates with the main pulse of Andean orogeny. During the late Pliocene, the Nazca slab broke, triggering the modern volcanism south of 5.5º N. Seismicity data and tomography support the proposed reconstruction.

Large magnitude (Mw ∼ ≥6) earthquakes in extensional settings are often associated with simultaneous rupture of multiple normal faults as a result of static and/or dynamic stress transfer. Here, we report details of the coseismic breaching of a previously unrecognized large-scale fault relay zone in central Greece, through three successive normal fault earthquakes of moderate magnitude (Mw 5.7–6.3) that occurred over a period of ∼10 days in March 2021. Specifically, joint analysis of InSAR, GNSS and seismological data, coupled with detailed field and digital fault mapping, reveals that the Tyrnavos Earthquake Sequence (TES) was accommodated at the northern end of a ∼100 km wide transfer structure, by faults largely unbroken during the Holocene. By contrast, the southern section of this relay zone appears to have accrued significant slip during Holocene. InSAR-derived displacements agree with the loci of eight subtle, previously undetected, faults that accommodated coseismic and/or syn-seismic normal fault slip during the TES. Kinematic modeling coupled with fault mapping suggests that all involved faults are interconnected at depth, with their conjugate fault-intersections acting largely as barriers to coseismic rupture propagation. We also find that the TES mainshocks were characterized by unusually high (>6 MPa) stress-drop values that scale inversely with rupture length and earthquake magnitude. These findings, collectively suggest that the TES propagated north-westward to rupture increasingly stronger asperities at fault intersections, transferring slip between the tips of a well-established, but previously unrecognized, relay structure. Fault relay zones may be prone to high stress-drop earthquakes and associated elevated seismic hazard.

Several decades of field, geophysical, analogue, and numerical modeling investigations have enabled documentation of the wide range of tectonic transport processes in accretionary wedges, which constitute some of the most dynamic plate boundary environments on Earth. Active convergent margins can exhibit basal accretion (via underplating) leading to the formation of variably thick duplex structures or tectonic erosion, the latter known to lead to the consumption of the previously accreted material and eventually the forearc continental crust. We herein review natural examples of actively underplating systems (with a focus on circum-Pacific settings) as well as field examples highlighting internal wedge dynamics recorded by fossil accretionary systems. Duplex formation in deep paleo-accretionary systems is known to leave in the rock record (1) diagnostic macro- and microscopic deformation patterns as well as (2) large-scale geochronological characteristics such as the downstepping of deformation and metamorphic ages. Zircon detrital ages have also proved to be a powerful approach to deciphering tectonic transport in ancient active margins. Yet, fundamental questions remain in order to understand the interplay of forces at the origin of mass transfer and crustal recycling in deep accretionary systems. We address these questions by presenting a suite of two-dimensional thermo-mechanical experiments that enable unravelling the mass-flow pathways and the long-term distribution of stresses along and above the subduction interface as well as investigating the importance of parameters such as fluids and slab roughness. These results suggest the dynamical instability of fluid-bearing accretionary systems causes either an episodic or a periodic character of subduction erosion and accretion processes as well as their topographic expression. The instability can be partly deciphered through metamorphic and strain records, thus explaining the relative scarcity of paleo-accretionary systems worldwide despite the tremendous amounts of material buried by the subduction process over time scales of tens or hundreds of millions of years. We finally stress that the understanding of the physical processes at the origin of underplating processes as well as the forearc topographic response paves the way for refining our vision of long-term plate-interface coupling as well as the rheological behavior of the seismogenic zone in active subduction settings., Geosphere, 18 (1), ISSN:1553-040X

Megathrust earthquakes are responsible for some of the most devastating natural disasters1. To better understand the physical mechanisms of earthquake generation, subduction zones worldwide are continuously monitored with geophysical instrumentation. One key strategy is to install stations that record signals from Global Navigation Satellite Systems2,3 (GNSS), enabling us to track the non-steady surface motion of the subducting and overriding plates before, during and after the largest events4–6. Here we use a recently developed trajectory modelling approach7 that is designed to isolate secular tectonic motions from the daily GNSS time series to show that the 2010 Maule, Chile (moment magnitude 8.8) and 2011 Tohoku-oki, Japan (moment magnitude 9.0) earthquakes were preceded by reversals of 4–8 millimetres in surface displacement that lasted several months and spanned thousands of kilometres. Modelling of the surface displacement reversal that occurred before the Tohoku-oki earthquake suggests an initial slow slip followed by a sudden pulldown of the Philippine Sea slab so rapid that it caused a viscoelastic rebound across the whole of Japan. Therefore, to understand better when large earthquakes are imminent, we must consider not only the evolution of plate interface frictional processes but also the dynamic boundary conditions from deeper subduction processes, such as sudden densification of metastable slab. Observed reversals in GNSS surface motions suggests greatly enhanced slab pull in the months preceding the great subduction earthquakes in Maule (Chile, 2010) and Tohoku-oki (Japan, 2011) of moment magnitudes 8.8 and 9.0.

Karpathos is a roughly north-south oriented island that emerges between Crete and Rhodes in the forearc of the eastern Hellenic subduction system. It extends for ~60 km to the north of the 40 km contour of the plate interface depth. Further, the northern part of the island is confined to a N-S trending Horst bounded by two large normal faults that shape the seafloor off both, the eastern and western shore. Furthermore, many normal faults, mainly in the north, strike parallel to the Horst and shape the topography onshore. Given the location and the structural configuration of the island, we expect that multiple processes are reflected in both the sedimentary and morphological record of vertical movement. Marine terraces and paleo-cliffs are observed all around the island recording its vertical movements over the last ~1 Ma. Moreover, sedimentary basins in the southern and central parts of the island are excellent archives of long-term uplift interrupted by subsidence over the last ~4.5 Ma. Twenty-five samples were collected at elevations between 1 and ~310 masl. We have gathered six (n=6) age/elevation data-points obtained by Sr-isotope dating, and nineteen (n=19) age/elevation data-points by radiocarbon dating. We explored the likelihood of different hypotheses on what drives the uplift: whether it is driven by upper-crust normal faults, megathrust earthquakes, underplating, or a combination of these phenomena. We present preliminary results on both the temporal and spatial fluctuations of the vertical movement of Karpathos.

Inverted structures are some of the economically most important geological features worldwide. Besides their most common manifestation as traps for hydrocarbons, they are also interesting for the storage of CO2 and extraction of other resources such as heat, minerals or hydrogen. Analogue modelling is frequently used to understand the long-term geological evolution of basins and basin inversion as an addition to numerical and mathematical models. Most analogue models use granular materials, like sands and glass beads, to simulate the brittle-plastic rheology of the crust. The main driving mechanism for basin inversion, both in nature and analogue models is the reactivation of pre-existing structures. This is due to strain-dependent weakening which leads to a reduced strength of a fault or shear zone in comparison with the surrounding bulk material. If the structure comes to a rest, several mechanisms lead to a time-dependent restrengthening of the structure. Therefore, older structures are usually more resistant to reactivation than younger ones, in the same material. In this study we use an annular shear tester to quantify the healing of granular materials commonly used for analogue models. We take advantage of a large collection of analogue material samples at the Helmholtz Laboratory for Tectonic Modelling, coming from many laboratories worldwide. To estimate granular healing, we employ slide-hold-slide tests with hold times comparable to typical analogue models of basin inversion. We show that all materials tested exhibit healing which follows a power-law relation quantified by with a healing rate. For example, fused glass microbeads showed a healing rate of 0.025 per decade in hold time. This means that for a tenfold increase in hold time the strength required to reactivate the given fault increases by 2.5%. Consequently, if a fault is inactive for a longer period of time, it is slightly stronger in comparison with a fault with shorter inactivity. Comparing the healing exponent for several materials reveals that some materials show a stronger healing than others. Glass beads have a stronger healing than sands, with quartz sands having lower healing rates than garnet or feldspar sands. Geomechanical tests on natural materials (quartz and gypsum fault gouges) and measurements of seismic velocities across fault zones suggest that healing obeys a similar power law. The healing rates in real rocks are roughly equal or higher depending on the temperature and water saturation of the fault. Albeit small, this change in reactivation strength for analogue materials might have a strong influence on the structural style of inversion if the models are run with different timespans between extensional phase and compressional phase. With a typical range of experimental time-spans of a view seconds to several hours this may result in up to 10% difference in reactivation strength similar to the difference between static and dynamic friction. This becomes especially relevant, if the angles of the formed pre-existing structures are close to the angle of internal friction of the bulk material which is the default in models where reactivated structures have been formed self-consistently in a pre-inversion phase.

Slow slip events (SSEs) in subduction zones can precede large-magnitude earthquakes and may therefore serve as precursor indicators, but the triggering of earthquakes by slow slip remains poorly understood. Here we report on a multidisciplinary dataset that captures a synergy of slow slip events, earthquake swarms and fault-interactions during the ~5 years leading up to the 2018 Mw 6.9 Zakynthos Earthquake at the western termination of the Hellenic Subduction System (HSS). We find that this long-lasting preparatory phase was initiated by a slow-slip event that released, over a period of 4-months, aseismic slip equivalent to a ~Mw 6.4 earthquake on the Hellenic plate-interface. This SSE, which is the first to be reported in the HSS, was associated with mild Coulomb failure stress changes (≤3 kPa) that were nevertheless sufficient to destabilize faults in the overriding plate. Tectonic instability was evidenced by a prolonged (~4 years) period of suppressed b-values (

Tectonic faults display a range of slip behaviours including continuous and episodic slip covering rates of more than 10 orders of magnitude. To gain insight into the slip behaviour of brittle faults, we performed laboratory stick-slip experiments at low pressure using dry “sticky” rice as rock analogue material. Rice has been shown to be a valuable material obeying the rate and state friction laws qualitatively and quantitatively and mimicking the full spectrum of seismogenic fault behaviour (“ricemic cycles”) depending on boundary conditions. The deformation mechanism is granular flow and as such transient hardening and weakening phenomena such as strain localization or stick-slip are accompanied by dilation and compaction, respectively. Such a rheology might be similar in rocks at various scales (grain scale to regional tectonic scale).We here report on ring shear test experiments on a range of rice varieties, including full-grain and crushed sorts. We imposed boundary conditions (i.e., normal load, shear velocity) scaled down from nature under which our fault analogue shows a variety of slip behaviours ranging from slow and quasi-continuous creep to episodic slow slip to dynamic rupture. The experiments demonstrate that significant interseismic creep (up to far-field loading rate) and earthquakes may not be mutually exclusive phenomena for a given location along a fault. Moreover, creep signals vary systematically with the fault’s seismic potential. Accordingly, the transience of interseismic creep scales with fault strength and seismic coupling as well as with the maturity of the seismic cycle. Loading rate independence of creep signals suggests that the long-term stationary mechanical properties of faults (e.g. seismic coupling) can be inferred from short-term observations (e.g. aftershock sequences). Moreover, we observe the number and size of small episodic slip events to systematically increase towards the end of the seismic cycle providing an observable proxy of the relative shear stress state on seismogenic faults. Importantly, very weak faults (with low effective normal loads) in a late stage of their seismic cycle might creep at rates very close to far-field loading for extended periods of the interseismic stage (~decades before failure). Given that we typically observe only a fraction of seismic cycles with high resolution (with geodetic methods) in nature, this might lead to the false belief of the fault being aseismic and not hosting large earthquakes. We thus demonstrate that seismic and aseismic behaviour might not necessarily be mutually exclusive.

Geodetic, seismological, gravimetric, and geomorphic proxies have widely been used to understand the behavior of the shallow portion of subduction megathrusts and answer questions related to seismic asperities: Where are they located, and how large are they? How close are they to failure, and how strong are they coupled? Our current knowledge of the kinematics and dynamics of megathrust earthquakes is limited due to their offshore location, and that our observations only cover a fraction of one megathrust earthquake cycle. The frictional-elastoplastic interaction between the interface and its overriding wedge causes variable surface strain signals such that the wedge strain pattern may reveal the mechanical state of the interface. We here contribute to this discussion using observations and interpretations of controlled analog megathrust experiments highlighting the variability of deformation signals in subduction zones. To examine the interaction, we investigate seismotectonic scale models representing a seismically heterogenous interface and capture the model’s surface displacements by employing a “laboratory-geodetic” method with high spatio-temporal resolution. Our experiments generate physically self‐consistent, analog megathrust earthquake ruptures over multiple seismic cycles at laboratory scale to study the interplay between short-term elastic and long-term permanent deformation. Our results demonstrate that frictional-elastoplastic interaction partitions the upper plate into a trench-parallel and -perpendicular strain domain, experiencing opposite strain (contraction vs. extension) during the co- and interseismic phase of the seismic cycle. Moreover, the pattern differs in the off- and onshore segments of the upper plate. This implies that the seismic potential of the shallow (offshore) portion of the megathrust may be underrepresented if only onshore observations are included in the estimate. However, our models suggest that, in the case of strong frictional contrast (velocity weakening vs. strengthening) on the interface, the short-term, onshore strain pattern (dominated by elastic deformation) may suffice to map the frictional heterogeneity of the shallow interface along strike. Finally, the frictional heterogeneity of the shallow interface is well reflected by the permanent surface strain observed offshore and partially in the strain observed at the coastal region. The observed along-trench segmentation predicted by our models is reasonably compatible with short-term, elastic geodetic observations and permanent geomorphic features in nature.

The growth of faults is well studied with field methods, experiments and theoretical models. Fault evolution is largely established from a geometrical and kinematic point of view with respect to the growth of isolated faults and their mutual interaction. However, the dynamics of fault growth (e.g. stress shadowing, damage zone evolution, energy budgets) and the emergence of interactions over various spatial and temporal scales in larger fault networks is a topic of recent interest less illuminated so far. We here introduce a new experimental setup allowing to study “large-n” fault networks evolving in crustal-scale brittle and brittle-ductile analogue models. We document preliminary results helping to demonstrate and verify the capability of the approach.The setup, called “The Expander”, builds on a traditional extensional setup with a basal rubber sheet expanded in one direction. The aspect ratio of the rubber sheet controls its lateral contraction (“Poisson’s effect”) and thus the bulk strain ratio under pure shear conditions. We can thus realize constrictional (prolate) to plane to flattening (oblate) kinematic basal boundary conditions depending on the sheet’s aspect ratio and whether we expand or relax the sheet. Evolving fault networks vary from anastomosing fold-and-thrust belts to conjugate sets of strike-slip fault networks to quasi-parallel normal fault populations, respectively. We apply digital image correlation (DIC) to track the kinematic surface evolution and photogrammetry (structure from motion, SFM) for topography evolution.First observations suggest that strike-slip fault networks in a purely brittle crust under basal pure shear conditions evolve into compartments of synthetic faults, the size of which scale with brittle layer thickness similar to fault spacing. The scaling seems to be controlled by slip partitioned onto the individual faults and mediated by stress shadows. Numerical simulation of the experiment suggests that the compartmentalization might evolve further through sequential de-activation of smaller faults and collapse of deformation into a single regional scale master fault with or without prescribing a zone of crustal weakness (a “seed”). Further experiments are planned to test the fault pattern evolution for different mechanical stratigraphy (brittle-viscous layers, seeds) and kinematic boundary conditions.

Large earthquakes impose differential stresses in the crust and upper mantle that are transiently relaxed during the postseismic phase mostly due to afterslip on the fault interface, viscoelastic relaxation in the lower crust and upper mantle, and poroelastic rebound in the upper crust. During the last years, the wealth of geophysical and geodetic observations, as well as great effort in forward and inverse modelling have allowed a better comprehension of the role of these mechanisms during the postseismic period. However, it is still an open question to what extent postseismic processes contribute to the surface deformation signal, especially during the early postseismic period. In this study, we use GNSS and InSAR observations collected in the first 48 days following the 2010 Maule earthquake in Chile along with a model approach that integrates afterslip, poroelasticity, and temperature-controlled power-law (non-linear viscosity) rheology. The afterslip distribution is obtained from a geodetic data inversion after removing the poro-viscoelastic component by forward modelling to the geodetic data. We find that our model approach explains the geodetic cumulative signal 14% better than a pure elastic model inverting for afterslip. This improvement is mainly produced by the better fit to the geodetic signal at the volcanic and back-arc regions due to the inclusion of non-linear viscoelastic processes, which can explain > 60% of the observed surface displacements in these regions. We also show that poroelastic processes play a significant role locally, specifically near the region where the coseismic slip was largest. Here, poroelastic processes explain most of the cumulative observed GNSS uplift signal and produce surface landward patterns that affect the horizontal GNSS component by up to 15% in the opposite direction. If poroelastic processes are ignored, our results reveal that the resulting afterslip amplitude is both amplified and suppressed by up to 40% in regions of ~50 x 50 km2. Our findings have implications for the calculation of the postseismic slip budget, and therefore the seismic hazard assessment of future earthquakes.

We used data from >100 permanent and temporary seismic stations to investigate seismicity patterns related to the 1 April 2014 M8.1 Iquique earthquake in northern Chile. Applying a multistage automatic event location procedure to the seismic data, we detected and located ~19,000 foreshocks, aftershocks, and background seismicity for 1 month preceding and 9 months following the mainshock. Foreshocks skirt around the updip limit of the mainshock asperity; aftershocks occur mainly in two belts updip and downdip of it. The updip seismicity primarily locates in a zone of transitional friction on the megathrust and can be explained by preseismic stress loading due to slow‐slip processes and afterslip driven by increased Coulomb failure stress due to the mainshock and its largest aftershock. Afterslip further south also triggered aftershocks and repeating earthquakes in several EW striking streaks. We interpret the streaks as markers of surrounding creep that could indicate a change in fault mechanics and may have structural origin, caused by fluid‐induced failure along presumed megathrust corrugations. Megathrust aftershocks terminate updip below the seaward frontal prism in the outer continental wedge that probably behaves aseismically under velocity‐strengthening conditions. The inner wedge locates further landward overlying the megathrust's seismogenic zone. Further downdip, aftershocks anticorrelate with the two major afterslip patches resolved geodetically and partially correlate with increased Coulomb failure stress, overall indicating heterogeneous frictional behavior. A region of sparse seismicity at ~40‐ to 50‐km depth is followed by the deepest plate interface aftershocks at ~55‐ to 65‐km depth, which occur in two clusters of significantly different dip.

An earthquake-induced stress drop on a megathrust instigates different responses on the upper plate and slab. We mimic homogenous and heterogeneous megathrust interfaces at the laboratory scale to ...

Subduction margins are the loci of a wide range of deformation processes occurring at different timescales along the plate interface and in the overriding forearc crust. Whereas long-term deformation is usually considered as stable over Myr-long periods, this vision is challenged by an increasing number of observations suggesting a long-term pulsing evolution of active margins. To appraise this emerging view of a highly dynamic subduction system and identify the driving mechanisms, detailed studies on high pressure-low temperature (HP-LT) exhumed accretionary complexes are crucial as they open a window on the deformation history affecting the whole forearc region.In this study, we combine structural and petrological observations, Raman spectroscopy on carbonaceous material, Rb/Sr multi-mineral geochronology and thermo-mechanical numerical models to unravel with an unprecedented resolution the tectono-metamorphic evolution of the Late-Cenozoic HP-LT nappe stack cropping out in western Crete (Hellenic subduction zone). A consistent decrease of peak temperatures and deformation ages toward the base of the nappe pile allows us to identify a minimum of three basal accretion episodes between ca. 24 Ma and ca. 15 Ma. On the basis of structural evidences and pressure-temperature-time-strain predictions from numerical modeling, we argue that each of these mass-flux events triggered a pulse in the strain rate, sometimes associated with a switch of the stress regime (i.e., compressional/extensional). Such accretion-controlled transient deformation episodes last at most ca. 1-2 Myr and may explain the poly-phased structural records of exhumed rocks without involving changes in far-field stress conditions. This long-term background tectonic signal controlled by deep accretionary processes plays a part in active deformations monitored at subduction margins, though it may remain blind to most of geodetic methods because of superimposed shorter-timescale transients, such as seismic-cycle-related events.

A complex seismic sequence took place in 2014 at the Juan Fernández microplate, a small microplate located between Pacific, Nazca and Antarctica plates. Despite the remoteness of the study region and the lack of local data, we were able to resolve earthquake source parameters and to reconstruct the complex seismic sequence, by using modern waveform-based seismological techniques. The sequence started with an exceptional Mw 7.1-6.7 thrust – strike slip earthquake doublet, the first subevent being the largest earthquake ever recorded in the region and one of the few rare thrust earthquakes in a region otherwise characterized by normal faulting and strike slip earthquakes. The joint analysis of seismicity and focal mechanisms suggest the activation of E-W and NE-SW faults or of an internal curved pseudofault, which is formed in response to the microplate rotation, with alternation of thrust and strike-slip earthquakes. Seismicity migrated Northward in its final phase, towards the microplate edge, where a second doublet with uneven focal mechanisms occurred. The sequence rupture kinematics is well explained by Coulomb stress changes imparted by the first subevent. Our analysis show that compressional stresses, which have been mapped at the northern boundary of the microplate, but never accompanied by large thrust earthquakes, can be accommodated by the rare occurrence of large, impulsive, shallow thrust earthquakes, with a considerable tsunamigenic potential.

The release of elastic energy along an active fault is accommodated by a wide range of slip modes. It ranges from long-term slow slip events (SSEs) and creep to short-term tremors and earthquakes. They vary not only in their characteristic duration but also in their magnitude, spatial extent and slip velocities. The exact relationship is unclear, as in some regions many slip modes occur simultaneously (e.g. Tohoku-Oki) and in others certain slip modes are completely absent (e.g. Cascadia).One of the driving factors in the generation of this large variety of slip modes is the interplay of fault heterogeneity and geometrical complexity of the fault system. We test various settings in terms of fault heterogeneity and geometrical complexity with a scaled physical model. The experimental results are then validated and benchmarked through multi-scale numerical simulations. We describe the system using a rate-and-state frictional framework and introduce on-fault heterogeneity with variable frictional properties. All properties are the same for analogue and numerical simulation as far as they can be determined or realized experimentally (a-b, vload, Shmax, Shmin, etc...). As analogue material we use segmented, decimetre sized neoprene foam blocks in multiple configurations (e.g. biaxial shear at forces We are able to reliably generate frequent stick-slip events of variable size and recurrence intervals. The slip characteristics, such as slip distribution, are in good agreement with analytical solutions of fault slip in elastic media. In a geometrically simple strike-slip model the recurrence behaviour and magnitude follows straightforward scaling relations in accordance with existing studies. If geometrical complexity is added to the model we observe clustering and variable recurrence that differ from the simpler geometry. Additionally, we are going to give an outlook on the interaction behaviour of multiple faults in dependence of their geometric configuration and the generation of power-law type magnitude scaling relations.

International audience; The long-lived subduction zone in the eastern Mediterranean region has been the locus of successive basal-accretion periods responsible for the formation of two main high pressure-low temperature (HP-LT) metamorphic belts recognized for a long time in the central Aegean domain (the Cycladic Blueschist Unit) and on Crete and Peloponnese (the Phyllites-Quartzites and Plattenkalk units). However, the characterization of individual tectonic slices within these metamorphic belts is only emerging and their precise timing of accretion remains enigmatic, although it is of paramount importance for constraining the protracted deformation history of the Hellenic margin.In this study, we combine structural and petrological observations, RSCM geothermometry and Rb/Sr multi-mineral geochronology to unravel with an unprecedented resolution the tectono-metamorphic evolution of the Late-Cenozoic HP-LT nappe stack cropping out in western Crete. We thus identify a minimum of six tectono-metamorphic slivers (i) bounded by major shear zones commonly reworked during exhumation and (ii) characterized by well-resolved down-stepping of the maximum recorded temperatures toward the lowermost structural levels. In addition, a consistent decrease of the accretion- and exhumation-related deformation ages, allows us to identify three slicing episodes between ca. 26 Ma and ca. 15 Ma. The subsequent fast exhumation of the nappe stack has been achieved by two detachment faults accommodating a N-S-directed, bivergent extension, which later evolved in an E-W-directed, trench-parallel extension as a consequence of the accelerating slab roll-back and the arc curvature during the middle-late Miocene.Through this multi-disciplinary approach, we finally stress that the Myr-scale cadence of slicing events resolved for the Hellenic margin sheds light on the spatial and temporal scale of deep-accretion dynamics, with implications for understanding mass fluxes in active forearc margins worldwide.

Currently, it is unknown how seismic and aseismic slip influences the recurrence and magnitude of earthquakes. Modern seismic hazard assessment is therefore based on statistics combined with numerical simulations of fault slip and stress transfer. To improve the underlying statistical models we conduct low velocity shear experiments with glass micro‐beads as fault gouge analogue at confining stresses of 5–20 kPa. As a result, we show that characteristic slip events emerge, ranging from fast and large slip to small scale oscillating creep and stable sliding. In particular, we observe small scale slip events that occur immediately before large scale slip events for a specific set of experiments. Similar to natural faults we find a separation of scales by several orders of magnitude for slow events and fast events. Enhanced creep and transient dilatational events pinpoint that the granular analogue is close to failure. From slide‐hold‐slide tests, we find that the rate‐and‐state properties are in the same range as estimates for natural faults and fault rocks. The fault shows velocity weakening characteristics with a reduction of frictional strength between 0.8% and 1.3% per e‐fold increase in sliding velocity. Furthermore, the slip modes that are observed in the normal shear experiments are in good agreement with analytical solutions. Our findings highlight the influence of micromechanical processes on macroscopic fault behavior. The comprehensive data set associated with this study can act as a benchmark for numerical simulations and improve the understanding of observations of natural faults., Plain Language Summary: Earthquakes occur when two continental plates slide past each other. The motion is concentrated at the interface of the two plates which is called a fault. In many cases the fault is filled with granular material, called gouge, that supports the pressure between the plates. Therefore, the properties of this gouge determine how fast and how large an earthquake can be. It also has an influence on the time between earthquakes. In our study, we examine a simplified version of a fault gouge in a simple small‐scale model. Instead of rock material we use glass beads and measure how different conditions affect the motion of the model. We find that our model reproduces features of fault gouge because it shows similar behavior. When there is no motion our model fault becomes stronger with a rate equal to fault gouge. Also, the type of strengthening is analogous to fault gouge. During slip, the glass beads become weaker as the slip velocity increases in a similar manner as in natural faults. These results improve the understanding of computer simulations and natural observations., Key Points: Slip modes in granular gouge are akin to natural fault slip. Glass beads are a suitable granular analogue for fault gouge and show rate‐and‐state dependent friction. Enhanced creep and small scale events are signals for imminent failure and indicate fault criticality., Deutsche Forschungsgemeinschaft (DFG) http://dx.doi.org/10.13039/501100001659, 亥姆霍兹联合会致力, Helmholtz‐Zentrum Potsdam ‐ Deutsches GeoForschungsZentrum GFZ (GFZ) http://dx.doi.org/10.13039/501100010956

SUMMARY The Northern Chilean subduction zone is characterized by long-term subduction erosion with very little sediment input at the trench and the lack of an accretionary prism. Here, multichannel seismic reflection (MCS) data were acquired as part of the CINCA (Crustal Investigations off- and onshore Nazca Plate/Central Andes) project in 1995. These lines cover among others the central part of the MW 8.1 Iquique earthquake rupture zone before the earthquake occurred on 1 April 2014. We have re-processed one of the lines crossing the updip parts of this earthquake at 19°40′S, close to its hypocentre. After careful data processing and data enhancement, we applied a coherency-based pre-stack depth migration algorithm, yielding a detailed depth image. The resulting depth image shows the subduction interface prior to the Iquique megathrust earthquake down to a depth of approximately 16 km and gives detailed insight into the characteristics of the seismogenic coupling zone. We found significantly varying interplate reflectivity along the plate interface which we interpret to be caused by the comparably strong reflectivity of subducted fluid-rich sediments within the grabens and half-grabens that are predominant in this area due to the subduction-related bending of the oceanic plate. No evidence was found for a subducted seamount associated to the Iquique Ridge along the slab interface at this latitude as interpreted earlier from the same data set. By comparing relocated fore- and aftershock seismicity of the Iquique earthquake with the resulting depth image, we can divide the continental wedge into two domains. First, a frontal unit beneath the lower slope with several eastward dipping back-rotated splay faults but no seismicity in the upper plate as well as along the plate interface. Secondly, a landward unit beneath the middle slope with differing reflectivity that shows significant seismicity in the upper plate as well as along the plate interface. Both units are separated by a large eastward dipping mega splay fault, the root zone of which shows diffuse seismicity, both in the upper plate and at the interface. The identification of a well-defined nearly aseismic frontal unit sheds new light on the interplate locking beneath the lower continental slope and its controls.

Active normal faults on the Mediterranean island of Crete form prominent limestone scarps together with basin and range topography. These faults mainly strike E‐ESE and N‐NNE in southern and northern Crete, respectively, with fault sets commonly intersecting and northerly trending faults being a factor of 3 more abundant. Lengths, displacements, and displacement rates have been analyzed for 84 active faults sampled over 2 ± 0.5 Ma (long‐term) and 16.5 ± 2 ka (short‐term) time‐intervals, with half showing no resolvable short‐term activity. Active faults record earthquake processes on timescales of thousands to million years and constrain sampling biases, which can lead to under and over estimates of fault parameters. The available data provide no evidence for fault propagation and support a model in which fault lengths were established early in the development of the fault system. Short‐term displacement rates (0.09–1.2 mm/year) are generally higher than long‐term rates (0.002–0.7 mm/year), with a factor of 4 disparity in the average recurrence intervals for the two time periods (∼2.5 Kyr vs. ∼11 Kyr). We attribute these differences to “clustering” of surface‐rupturing (e.g., >Mw6) earthquakes on individual faults over millennial timescales, and to preferential sampling of the most seismically active faults during the short‐term. Displacement rates are comparable when averaged for each time interval on the longest faults (>10 km), indicating that for these faults earthquake “clustering” spans time‐intervals of Mw6 on Crete are at least three times more frequent than historical earthquakes since ∼1920, possibly because multi‐fault surface‐rupturing earthquakes are double counted in the paleo‐record.

Summarization: The importance of splay-thrust faults in subduction seismogenesis is increasingly acknowledged; however, their elastic interaction with the plate interface remains unclear. Here, we use GPS velocities, constrained by millennial fault slip rates, to study elastic fault-interactions between the plate interface and its upper-plate splay-thrust faults from the southern Hellenic Subduction System (HSS). We find that, despite its largely aseismic character, the HSS plate interface zone is kinematically segmented, with slip rate deficits locally reaching ~85% and ~45% of the plate convergence rate on the western and eastern segments, respectively, and on structures different from those that ruptured historically. Although western Crete has been more active seismically during late Holocene, we find that the eastern HSS has higher seismic potential for large-magnitude (M > 6) earthquakes and its interface zone is closer to failure. Elastic fault interactions are responsible for both significant intersegment variability in strain accumulation and uniformity in earthquake rupture segmentation along the HSS over millennial timescales. Presented on: Geophysical Research Letters

Active normal faults on the eastern Mediterranean island of Crete form prominent limestone scarps together with basin and range topography. These faults mainly strike E-ESE and N-NNE in southern and northern Crete, respectively, with fault sets commonly intersecting and northern-trending faults a factor of three more abundant. Displacements, lengths and displacement rates have been analysed for 84 active faults sampled over 2±0.5 Ma (long term) and 16.5±2 ka (short term) time-intervals, with about half showing no resolvable short-term activity. Active faults record earthquake processes on timescales of thousands to millions of years and constrain sampling biases, which can lead to under and over estimates of the numbers, rupture lengths, recurrence intervals and single event displacements of paleoearthquakes. The available data indicate no fault propagation and, for the Quaternary, higher displacement rates on longer faults, supporting a model in which fault lengths and maximum earthquake magnitudes were established early in the development of the fault system. Short-term displacement rates (0.09-1.2 mm/yr) are generally higher than long-term rates (0.002-0.7 mm/yr), with a factor of four disparity in the average recurrence intervals for the two time periods (~2.5 kyr vs ~11 kyr). We attribute these differences to ‘clustering’ of >Mw 6 earthquakes on individual faults over millennial timescales, and to preferential sampling of the most seismically active faults in the post-~16.5 kyr time interval (i.e. faults with highest displacement rates or mid ‘cluster’). Displacement rates are comparable when averaged for each time interval on the longest faults (>10 km), interpreted to indicate that for these faults earthquake ‘clustering’ spans time-intervals of 16.5±2 kyr or less. Paleoarthquakes >Mw 6 on Crete are a factor of three more frequent than historical events for the last 100 yrs, which could be partly due to paleoevents rupturing multiple faults, either in the same event or in triggered slip events.

Subduction systems globally terminate, allowing plate-motion to be transferred from the oceanic megathrust onto continental and/or oceanic transform faults. The mechanism of this kinematic transition over earthquake timescales is, however, poorly understood due to a lack of relevant data. Here, we study the 2016 Mw 7.8 Kaikōura Earthquake in New Zealand, the first large instrumentally-recorded earthquake across a subduction-termination, to investigate this transfer mechanism in detail. We find that the Kaikōura Earthquake, unlike standard subduction earthquakes globally, involved a predominance (∼80%) of coseismic-slip on upper-plate faults and minor triggered-slip on the underlying oceanic subduction-thrust. In the months following the earthquake, the subduction-thrust accommodated most of the earthquake's afterslip down-dip of its co-seismic rupture zone. This top-down strain-release mechanism is in accord with local geological, geodetic and historical seismicity data which suggest that the bulk of plate-convergence (>75%) is accommodated in the upper-plate. We suggest, therefore, that this alternative strain-release mechanism, which is distinct from standard plate-boundary earthquakes, is characteristic of subduction-terminations and results in the majority of seismic/tsunami hazard being on steep near-surface faults.

Asperities are patches where the fault surfaces stick until they break in earthquakes. Locating asperities and understanding their causes in subduction zones is challenging because they are generally located offshore. We use seismicity, inter- and co-seismic slip, and the residual gravity field to map the asperity responsible for the 2014 M8.1 Iquique Chile earthquake. Until two weeks before the mainshock, seismicity occurred exclusively down-dip of the asperity, until a foreshock series broke first the upper plate and later its updip rim. Together the seismicity formed a ring around the future earthquake’s main slip patch. The asperity correlates both with high inter-seismic locking and a circular gravity low, suggesting that it is controlled by geologic structure. Most features of the spatiotemporal seismicity pattern can be explained by a mechanical model in which a single strong asperity is stressed by slab pull.

Increasing evidence suggests that large thrust-faults that splay from the plate-interface to extend within the upper-plate have a significant impact on subduction seismogenesis. The manner in which these two elements, the plate-interface itself and its splay-thrust faults, interact with one another during the earthquake cycle remains, however, poorly explored. Here, we use GPS velocities, constrained by millennial fault slip-rates, to quantify the accumulation (and partitioning) of strain on individual faults of the plate-interface zone and capture their possible interactions. We zoom into the southern Hellenic Subduction System (HSS), where the greatest (M8.3) earthquake and tsunami ever recorded in the Mediterranean was produced by slip on a splay-thrust fault. Our analysis shows that the HSS is kinematically segmented and strain is accumulated at spatially variable rates along individual structures of the plate-interface zone. We find that insterseismic locking reaches up to ~85% and ~45% on the western and eastern segments, respectively, and on structures different to those that ruptured historically. Although the western HSS has been more active recently (e.g. 365 BC), the eastern HSS carries currently higher potential for large-magnitude (M>6) earthquakes andits interface-zone appears to be closer to failure. Elastic fault-interactions are responsible for both significant inter-segment variability in strain-accumulation and millennial uniformity in earthquake rupture-segmentation between eastern and western HSS.

After large earthquakes at subduction zones, the plate interface continues moving due to mostly frictional afterslip processes. Below depths of 60 km, little frictional afterslip is to be expected on the plate interface due to low shear strength, lack of apparent geodetic interseismic locking, and low seismic moment release from aftershocks. However, inversion models that consider an elastic crust above a mantle with viscoelastic rheology result in a significant portion of afterslip at depths > 60 km. In this study, we present a forward 3D geomechanical-numerical model with power-law rheology that simulates dislocation creep processes for the crust and upper mantle in combination with an afterslip inversion. The linear rheology case is also considered for comparison. We estimate the cumulative viscoelastic relaxation and the afterslip distribution for the first six years following the 2010 Mw 8.8 Maule earthquake in Chile. The cumulative afterslip distribution is obtained from the inversion of the residual surface displacements between continuous GPS (cGPS) observations and predicted displacements from viscoelastic forward modelling. We investigate three simulations: two with the same dislocation creep parameters in the slab and upper mantle but different ones in the continental crust, and another with elastic properties in the crust and slab and a linear viscoelastic upper mantle. Our preferred simulation is the one with power-law rheology in the crust and upper mantle with a weak continental crust since the corresponding afterslip distribution shows the best overall fit to the cGPS displacements (cumulative and time series) as well as having a good correlation with aftershock activity. In this simulation, most of the viscoelastic relaxation occurs in the continental lower crust beneath the volcanic arc due to dislocation creep processes. The resulting afterslip pattern from the inversion is reduced at depths > 60 km, which correlates well with the spatial distribution of cumulative seismic moment release from aftershocks. We conclude that by allowing for non-linear stress relaxation in the continental lower crust due to dislocation creep processes, the resulting afterslip distribution is in better agreement with the physical constraints from the shear strength of the plate interface at depth, the predicted locking degree, and the aftershock activity.

The physical nature and the rheology of a subduction shear zone play an important role in the deformation and the degree of locking along its interface with the upper plate. Inspired from exhumed subduction shear zones that exhibit block-in-matrix characteristics (mélanges), we create synthetic models with different proportions of strong clasts within a weak matrix and compare them to natural mélange outcrops. Using 2D Finite Element visco-plastic numerical simulations and simple shear kinematic conditions, we determine the effective rheological parameters of such a two-phase medium, comprising blocks of basalt embedded within a wet quartzitic matrix. We treat our models and their structures as scale-independent and self-similar and upscale published field geometries to km-scale models, compatible with large-scale far-field observations. Exhumed subduction mélanges suggest that deformation is mainly taken up by dissolution-precipitation creep. However, such flow laws are neither well-established yet experimentally nor of ample use in numerical modelling studies. In order to make our results comparable to and usable by numerical studies, we assume dislocation creep as the governing flow law for both basalt and wet quartz and by using different pressures, temperatures and strain rates we provide effective rheological estimates for a natural subduction interface. Our results suggest that the block-in-matrix ratio affects deformation and strain localization, with the effective dislocation creep parameters varying between the values of the strong and the weak phase, in cases where deformation of both materials is purely viscous. As the contribution of brittle deformation of the strong blocks increases, however, the value of the stress exponent, n, can exceed that of the purely strong phase.

The release of elastic energy along an active fault is accommodated by a wide range of slip modes. It ranges from long-term slow slip events (SSEs) and creep to short-term tremors and earthquakes. They vary not only in their characteristic duration but also in their magnitude, spatial extend and slip velocities. As all slip modes are related to earthquakes, the understanding of the relationships between the different slip modes and the underlying mechanisms is crucial to assess earthquake hazards in various regions. The exact relation is unclear, as in some regions many slip modes occur simultaneously (e.g. Tohoku-Oki) and in others certain slip modes are completely absent (e.g. Cascadia).One of the driving factors in the generation of this large variety of slip modes is the interplay of fault heterogeneity and geometrical complexity of the fault system. Using a scaled physical model we test various settings in terms of fault heterogeneity and geometrical complexity. The experimental results are then validated and benchmarked using multi-scale numerical simulations. We describe the system using the rate-and-state frictional framework and introduce the on-fault heterogeneity with variable frictional properties. All properties are the same for analogue and simulation as far as they can be determined or realized experimentally (a-b, vload, Shmax, Shmin, etc...). As analogue material we use segmented, decimetre sized neoprene foam blocks in multiple configurations (e.g. biaxial shear at forces We are able to reliably generate frequent stick-slip events of variable size and recurrence intervals. The slip characteristics, such as slip distribution, are in good agreement with analytical solutions of fault slip in elastic media. In this contribution we will highlight the material properties, experimental results and used methodologies to monitor and process the experimental data. Additionally, we are going to give an outlook on the interaction behaviour of multiple faults in dependence of their geometric configuration and the generation of power-law type magnitude scaling relations.

The Pamir and Hindu Kush are located at the western tip of the India-Asia collision zone. Approximately a third of the northward motion of India’s western syntax is mostly accommodated by continental-scale underthrusting of the Indian plate beneath Asia. On its way northwards the arcuate, convex Pamir mountain range acts as a rigid indenter penetrating the weaker Eurasian plate, while lateral extrusion occurs to the west in the Tajik Depression.Intense present-day shallow seismicity indicates active deformation along the northern and north-western semi-arid margin of the Pamir, where over the last century several M>6 and three M>7 crustal earthquakes, including a recent M6.4 event in 2016, were recorded. Earthquakes are distributed in the proximity of three main fault systems: the Pamir thrust system to the north, and the Darvaz fault and Vakhsh thrust system to the north-west. The pronounced topographic expression of these lithospheric faults is associated to a deeply incised landscape, which was profoundly shaped by past widespread glaciations. The transient evolution of the landscape following deglaciation is observed in the dynamic river network, characterised by intense fluvial incision and changes in the fluvial connectivity of the drainage system.At depth, recent seismic tomography studies suggest delamination, stretching and tearing of the Asian slab beneath SW Pamir, and slab break-off underneath Hindu Kush. Slab break-off episodes are known to result in stress surges in the overlying lithosphere, potentially causing deformation and uplift.In this complex system characterised by an important interplay between tectonics, climate and surface processes, we use qualitative and quantitative analyses of the topography and of the drainage systems evolution, inclusive of numerical tools, in order to define what is –and has been- the role played by the main lithospheric active faults of this area. In addition, we aim at identifying how landscape and surface dynamics respond, temporally and spatially, to processes, such as slab tearing/break-off, occurring at depth.

It is increasingly apparent that the progression to eventual failure of large subduction earthquakes can be captured by continuous networks that record anomalous seismic and geodetic signals in the late interseismic period. Such precursory signals are generally understood to be related to a gradual decoupling of the mainshock area of the fault and can last from days to years. These natural observations are consistent with various numerical and laboratory models in which similar late-interseismic signals are generated.Here we analyse the continuous GNSS records of the final 5 years leading to the 2010 Mw 8.8 Maule, Chile and 2011 Mw 9.0 Tohoku-oki, Japan earthquakes. We implement the Greedy Automatic Signal Decomposition - a regression approach that builds upon existing tectonic trajectory models - to model the daily GNSS displacement time series as the sum of background seasonal oscillations, step functions, linear (1st order polynomial) motion, and a sparse number of multi-transient functions. The multi-transient functions are simply the sum of decay functions (e.g. exponential, logarithmic) that begin at the same time but have different characteristic decay constants. The inclusion of these versatile multi-transients allows the model to capture a variety of transient motion. We see that both subduction margins exhibit variability in their interseismic velocities. The most striking of these motions occur in the 5-7 months directly before both the Maule and Tohoku-oki earthquakes during which the sense of motion reverses in the trench-perpendicular component. These reversals manifest themselves as wobbles in the displacement time series with a peak-to-peak displacement between 4-8 mm and occur on a spatial scale in the order of thousands of kilometres. After investigating fluid loading and possible reference frame artifacts, we conclude that the wobbles are most likely of a tectonic origin.In the pre-Tohoku-oki case, for which we have a much denser surface coverage, kinematic models indicate an initial extension in the Philippine Sea Plate followed by a viscoelastic rebound. The spatial scale and approximate onset of this apparent extension are in agreement with the anomalous GRACE gravity signals reported in earlier work of Panet et al. (2018, Nature Geoscience). Furthermore, the speed that the trench-wards transient migrates along-strike of the subduction zone before Tohoku-oki indicates that deep slow-slip is also occurring. In the pre-Maule case, we see a similar reversal but lack the number of measurements to track any migration of the velocity front. Nevertheless, from inclusion of vertical displacement in analyses of both networks, we suspect that these late interseismic reversal signals are caused by a sudden enhanced slab pulling. Such an enhanced slab pull might be caused by sudden densification of metastable slab. Therefore, a main message of this work is that large asperities, while they might fail gradually local to the mainshock region, might also brought to failure by changes in the slab pull boundary conditions that can be several hundreds of km deep.