Search

Megathrust geometric properties exhibit some of the strongest correlations with maximum earthquake magnitude in global surveys of large subduction zone earthquakes, but the mechanisms through which fault geometry influences subduction earthquake cycle dynamics remain unresolved. Here, we develop 39 models of sequences of earthquakes and aseismic slip (SEAS) on variably‐dipping planar and variably‐curved nonplanar megathrusts using the volumetric, high‐order accurate code tandem to account for fault curvature. We vary the dip, downdip curvature and width of the seismogenic zone to examine how slab geometry mechanically influences megathrust seismic cycles, including the size, variability, and interevent timing of earthquakes. Dip and curvature control characteristic slip styles primarily through their influence on seismogenic zone width: wider seismogenic zones allow shallowly‐dipping megathrusts to host larger earthquakes than steeply‐dipping ones. Under elevated pore pressure and less strongly velocity‐weakening friction, all modeled fault geometries host uniform periodic ruptures. In contrast, shallowly‐dipping and sharply‐curved megathrusts host multi‐period supercycles of slow‐to‐fast, small‐to‐large slip events under higher effective stresses and more strongly velocity‐weakening friction. We discuss how subduction zones' maximum earthquake magnitudes may be primarily controlled by the dip and dimensions of the seismogenic zone, while second‐order effects from structurally‐derived mechanical heterogeneity modulate the recurrence frequency and timing of these events. Our results suggest that enhanced co‐ and interseismic strength and stress variability along the megathrust, such as induced near areas of high or heterogeneous fault curvature, limits how frequently large ruptures occur and may explain curved faults' tendency to host more frequent, smaller earthquakes than flat faults. Plain Language Summary: Subduction zones, where one tectonic plate dives beneath another, generate the largest earthquakes worldwide. Our study investigates how the shape and tilt of these large offshore underground faults, termed "megathrusts," may determine the size of large earthquakes, how often they happen, and how similar or different subsequent events are. By creating computer simulations of earthquakes in subduction zones, we found that the angles and dimensions of the megathrust may set a limit on how big an earthquake can get. We also find that the presence of bends or curves along these faults can make earthquakes more unpredictable, sometimes leading to more variable series of smaller quakes before the biggest one hits. Our findings may help explain why some areas near subduction zones are prone to larger or more frequent earthquakes than others. Understanding these patterns can improve our ability to prepare for these natural disasters, potentially reducing their damaging effects on nearby communities and infrastructure. Key Points: Systematic earthquake cycle simulations reveal how subduction zone geometry controls megathrust earthquake size and timingDip and curvature affect characteristic slip style: periodic uniform ruptures versus supercycles of variably‐sized slow‐to‐fast eventsSeismogenic zone dip and dimensions limit maximum earthquake size; curvature‐linked stress and strength heterogeneity modulates recurrence [ABSTRACT FROM AUTHOR]

How to effectively mitigate atmospheric delay signals in InSAR observations has long been a pressing problem in the InSAR community. Here we propose a new method, DetrendInSAR, that addresses this issue by incorporating the spatiotemporal characteristics of displacements and atmospheric delays as a‐priori information. This enables simultaneously modeling and estimating surface displacements, atmospheric delays, and other unmodeled long‐wavelength errors (e.g., orbit errors) in InSAR time series. We use both simulated‐ and real‐data experiments to validate the performance of the proposed method, and results show a 20%–70% reduction of RMSE values for the DetrendInSAR‐obtained displacements compared with four commonly‐used atmospheric delay reduction methods. We also show that the DetrendInSAR method is capable of capturing the spatial and temporal details of sudden deformation jumps as small as 1 cm, due to a magnitude 5.4 earthquake, in InSAR time series. Based on the DetrendInSAR‐corrected ascending and descending Sentinel‐1 InSAR time series, we calculate the 2‐year postseismic east and vertical displacements of the 2021 Mw 7.4 Maduo earthquake, which better illuminate the postseismic deformation. Plain Language Summary: Satellite Interferometric Synthetic Aperture Radar (InSAR) is a powerful tool for providing detailed information about ground surface movements caused by geohazards (e.g., earthquakes, volcanoes, and landslides). However, InSAR observations also contain noise contributions (e.g., atmospheric delays) in addition to the information of interest, making it difficult to accurately depict the ground movements and characterize the underlying mechanisms of geohazards. To separate the noise components from the deformation component in InSAR observations, we propose the DetrendInSAR method that utilizes the spatiotemporal characteristics of each component in InSAR observations as additional constraints. Results from simulated‐ and real‐data experiments demonstrate that the proposed DetrendInSAR method effectively reduces the noise components and achieves 20%–70% better results than commonly‐used strategies. For the real‐data case study, we calculated the two‐year postseismic displacement time series of the 2021 magnitude 7.4 Maduo earthquake, revealing more clearer and longer‐term postseismic surface displacements of this earthquake than in previous studies. Key Points: DetrendInSAR is a new pixel‐based method that simultaneously estimates deformation and atmospheric delays in InSAR time seriesThe method is validated using both simulated data and two‐year postseismic deformation data of the 2021 magnitude 7.4 Maduo earthquakeResults corrected by DetrendInSAR show 20%–70% reduction of RMSE values compared with four commonly‐used atmospheric signal reduction methods [ABSTRACT FROM AUTHOR]

Ahi Tupua, the central section of the Taupō Volcanic Zone in the central North Island of New Zealand, encompasses Taupō and Ōkataina calderas and has been the most frequently active and productive silicic magma system worldwide during the Quaternary. The entire Taupō Volcanic Zone is underlain by the Hikurangi Subduction Zone where a Large Igneous Province, the Hikurangi Plateau, is being subducted, but Ahi Tupua exhibits much higher rates of magma output than either the northern or the southern sections. Trace element signatures of Ahi Tupua eruptive products suggest that both decompression melting and flux melting play a role in this rifted arc setting, with fluid flux signatures being more prevalent in regions of active caldera volcanism. Intermediate‐depth (50–300 km) earthquakes provide a means of studying faulting and fluid flow processes within and around the subducting slab beneath Ahi Tupua. Using data from the national seismic network and a 13‐station temporary network ("ECLIPSE"), we repick and relocate 397 intermediate‐depth earthquakes of magnitudes M2.5+ that occurred in a 29‐month interval, and compute focal mechanisms for a subset of 47 earthquakes. We observe some seismicity that may be within the mantle wedge but most of the relocated earthquakes occur within the crust and uppermost mantle of the subducted slab and exhibit a weak transition from predominantly normal‐faulting in the slab crust to predominantly reverse‐faulting in the slab mantle. No double seismic zone is observed beneath Ahi Tupua. These observations are consistent with dehydration of the slab and flow of slab‐derived fluids along existing faults into the mantle wedge, that drives flux melting and accounts for the distinctive geochemical signatures and voluminous output of Ahi Tupua calderas. Plain Language Summary: Ahi Tupua is a volcanic region in the North Island of New Zealand that is the most active silicic volcanic system on Earth. It includes the active caldera volcanoes Taupō and Ōkataina. This volcanic region lies above the Hikurangi Subduction Zone, and studies of past eruptions suggest that Ahi Tupua's high rate of activity is driven by fluids released deep in the subduction zone. In this study, we use seismic data from across the entire North Island to investigate earthquakes occurring at depths of 50–300 km within the Hikurangi Subduction Zone. We show that these earthquakes suggest that the subducting oceanic crust is releasing a substantial amount of water into the overlying mantle, causing a large amount of melting and, ultimately, the distinctive volcanism of Ahi Tupua. Key Points: We relocate 397 intermediate‐depth earthquakes beneath Ahi Tupua, central Taupō Volcanic Zone, and compute 47 focal mechanismsMost seismicity occurs within the slab crust and mantle, with subordinate mantle wedge and plate interface faultingHypocenters and focal mechanisms are consistent with slab‐derived fluid fluxing melting of the mantle wedge beneath Ahi Tupua [ABSTRACT FROM AUTHOR]

Plate motion obliquity along the dominantly transform Queen Charlotte plate boundary (QCPB) peaks offshore Haida Gwaii. To investigate the effects of obliquity on plate boundary deformation, we analyze continuous seismic waveforms from temporary and permanent stations from 1998 to 2020 to generate a catalog of ∼50,000 earthquakes across Haida Gwaii. We use an automated technique based on auto‐regressive phase detection and onset estimation to obtain the initial seismic catalog, integrate existing catalogs, invert for 3D velocity structure using data from the best constrained period, and relocate the entire catalog using the new 3D velocity model. We investigate the seismically active sections of the transcurrent Queen Charlotte fault (QCF), noting that little seismicity locates directly along its bathymetrically defined trace. Instead, seismicity illuminates a complex system of segmented structures with variable geometries along strike. Other clusters highlight active shallow faults within the highly deformed Queen Charlotte terrace. Few aftershocks appear on the thrust plane of the 2012 Mw 7.8 Haida Gwaii earthquake except near its inferred intersection with the QCF at 15–20 km depths, suggesting elevated residual stress at the juncture of slip‐partitioning. Deep crustal seismicity (up to ∼20 km depths) beneath central Haida Gwaii aligned parallel to the strike of the thrust plane may represent landward underthrusting of the Pacific plate. Our results suggest possible coseismic strike‐slip rupture on the QCF during the 2012 earthquake and add support to the thesis that highly oblique transform boundaries are viable settings for subduction initiation. Plain Language Summary: We investigated the complex tectonics offshore Haida Gwaii, western Canada, where the Pacific and North American plates slide past one another obliquely. By compiling and analyzing the most comprehensive earthquake catalog in the area, spanning 1998–2020, we present the most detailed report to date of the earthquake‐producing structures in the region, including previously unidentified and highly segmented faults. Clusters of seismicity illuminate (a) a highly deformed terrace of elevated seafloor west of Haida Gwaii, (b) a complex and segmented fault system adjacent to the previously‐mapped surface trace of the main Queen Charlotte strike‐slip fault, and (c) the inferred fault intersection at depth between the subvertical Queen Charlotte fault (which hosted the 1949 magnitude 8.1 earthquake) and the shallowly dipping Haida Gwaii thrust (which hosted the 2012 magnitude 7.8 earthquake). We also speculate that the 2012 earthquake may also have involved some motion on the Queen Charlotte fault. These results contribute to better constraints on regional tectonics and hazards, and provide insights into the mechanisms of subduction initiation. Key Points: Seismicity off Moresby Island is distributed along multiple segments slightly off of the Queen Charlotte fault traceAftershocks at intersection of Queen Charlotte Fault with the 2012 Mw 7.8 thrust plane reflect residual stress at slip partitioning juncturePreviously undocumented deep seismicity beneath Haida Gwaii is consistent with an underthrusting Pacific Plate [ABSTRACT FROM AUTHOR]

In multiple observed caldera‐forming eruptions, the rock overlying a draining magma reservoir dropped downward along ring faults in sequences of discrete collapse earthquakes. These sequences are analogous to tectonic earthquake cycles and provide opportunities to examine fault mechanics and collapse eruption dynamics over multiple events. Collapse earthquake cycles have been studied with zero‐dimensional slider‐block models, but these do not account for the complicated interplay between fluid and elastic dynamics or for factors such as the heterogeneous fault properties and non‐vertical ring fault geometries often inferred at volcanoes. We present two‐dimensional axisymmetric mafic piston‐like collapse earthquake cycle models that include rate‐and‐state friction, fully‐dynamic elasticity, and compressible viscous fluid magma flow. We demonstrate that collapse earthquake intervals and magnitudes are highly sensitive to inertial effects, evolving stress fields, fault geometry, and depth‐varying fault friction. Given the consistent earthquake cycles observed in most eruptions, this suggests that ring faults can quickly stabilize and often become nearly vertical at depth. We use the well‐monitored 2018 collapse sequence at Kı̄lauea as a case study. Our model can produce many features of Kı̄lauea seismic and geodetic observations, except for a significant amount of interseismic slip, which cannot be readily explained with simple rate‐and‐state friction parameterizations. Plain Language Summary: Many large volcanic eruptions involve caldera collapse, where the rock above a magma reservoir abruptly subsides as the reservoir drains. Collapse can cause large earthquakes and repressurize the magma reservoir, creating a repeating cycle that helps sustain eruptions over days‐months. We develop computational models to understand the rock and magma interactions involved in collapse earthquake cycles, using data from the 2018 eruption of Kı̄lauea. The models can explain many geophysical measurements and show that collapse eruptions are very sensitive to fault geometry and spatially variable rock properties. However, standard friction models do not readily explain large amounts of fault slip measured between collapse earthquakes, suggesting that other complexities should be considered. Key Points: We develop 2D axisymmetric fully‐dynamic caldera collapse earthquake cycle models with rate‐and‐state friction and compressible fluid flowCollapse earthquake sequences are highly sensitive to fault geometry as well as heterogeneous stresses and frictionOur models fit geophysical data except large interseismic slip distances suggesting processes beyond simple rate‐and‐state friction parameterizations [ABSTRACT FROM AUTHOR]

We present a new geodetic strain rate and rotation rate model for Greece that has been derived using a highly dense GPS velocity field. The spatial distribution and the resolved rates of the various velocity gradient tensor quantities provided updated constraints on the present‐day upper crustal deformation in the region and revealed new details not reported previously. The spatial distribution of the second invariant demonstrated that the overall magnitude of strain rates is highest across two well‐defined provinces. The first follows the North Anatolian Fault and its two branches within the north Aegean, crosses central Greece and through the Gulf of Corinth it terminates in western Greece, while the second encompasses the extensional province of western Turkey and the eastern Aegean Sea islands. Our estimates revealed that shearing affects some of the fault‐bounded grabens of central Greece that lie to the SW of the North Aegean Basin implying considerable oblique extension. We identified a narrow region of counterclockwise rotation whose location and kinematics have been induced by the net effect across the intersection of the clockwise rotating domains of western and central Greece. The Aegean microplate and the Anatolian plate are separated by a wide transition zone which accommodates the curved stretching of the entire plate system. In both edges of the Hellenic forearc the dominant mode of crustal strain is E‐W extension. We found that earthquakes of M ≥ 5.6 are spatially well‐correlated with high‐strain areas, indicating that strain rate mapping could be used to inform future probabilistic seismic hazard analyses. Plain Language Summary: Greece and its surroundings comprise a region of elevated seismic hazard because of the intense active deformation that is accumulated in earth's crust over time and which is expected to be released as earthquakes of various magnitudes. Given the importance of this relationship, we used a dense data set of GPS velocities to gain insight into the spatial distribution and patterns of the present‐day strain and rotation rates. We found that deformation is mostly concentrated across two zones. The first crosses the North Anatolian Fault, enters the north Aegean Sea, continues in central Greece and terminates in western Greece, while the second encompasses the eastern Aegean Sea islands and western continental Turkey. Across the first zone, the crust experiences various deformation modes due to the interaction of large lithospheric‐scale tectonic structures with smaller regions of localized faulting. Within the second zone, the crust is stretching in response to differences in direction and rate of movement between the Aegean microplate and the Anatolian plate. Comparison of strain rates with seismicity revealed that earthquakes with M ≥ 5.6 predominantly occur within high‐strain areas, giving implications that knowledge of the spatial distribution of tectonic deformation can augment seismic hazard analyses. Key Points: We present an improved geodetic strain and rotation rate model for Greece which reveals new details about active deformationThe strain rate field of Greece is a superposition of short‐ and long‐wavelength signals driven by local forces and large‐scale tectonicsEarthquakes over M 5.6 correlate well with high strain rates and consistently occur at areas where the second invariant exceeds 40 nstrain/yr [ABSTRACT FROM AUTHOR]

Large earthquakes can trigger smaller seismic events, even at significant distances. The process of earthquake triggering offers valuable insights into the evolution of local stress states, deepening our understanding of the mechanisms of earthquake nucleation. However, our ability to detect these triggered events is limited by the quality and spatial density of local seismometers, posing significant challenges if the triggered event is hidden in the signal of a nearby larger earthquake. Distributed acoustic sensing (DAS) has the potential to enhance the monitoring capability of triggered earthquakes through its high spatial sampling and large spatial coverage. Here, we report on an uncatalogued magnitude (M) 5.1 event in northeast Turkey, which was likely dynamically and instantaneously triggered by the 2023 M7.8 earthquake in southeast Turkey, located 400 km away. This event was initially discovered on ∼1,100 km of active DAS recordings that are part of an 1,850‐km linear array. Subsequent validation using local seismometers confirmed the event's precise time, location, and magnitude. Interestingly, this dynamically triggered event exhibited precursory signals preceding its P arrivals on the nearby seismometers. It can be interpreted as the signal from other nearby, uncatalogued, smaller triggered events. Our results highlight the potential of high‐spatial‐density DAS in enhancing the local‐scale detection and the detailed analysis of earthquake triggering. Plain Language Summary: Large earthquakes can trigger smaller ones even far away. This helps us understand more about how earthquakes start and develop. However, finding these smaller earthquakes can be difficult as sometimes they are hidden in the chaos of a bigger, nearby earthquake. There is a novel technology called Distributed Acoustic Sensing (DAS) that might help. DAS can listen for earthquakes over large areas and gives a more detailed picture of what's happening underground. In this study, we discovered a moderate‐size earthquake with a magnitude of 5.1 in Northeast Turkey that was triggered by a large earthquake with a magnitude of 7.8 in Southeast Turkey, using a DAS system that stretched over a thousand kilometers. We then checked this finding with conventional seismometers to be sure. Interestingly, this triggered earthquake showed precursory signals before its main shaking started, which could be signs of other smaller earthquakes happening nearby. Our findings deepen our understanding of how earthquakes interact with each other and offer insights into how earthquakes start. Our results also suggest that DAS could help us find and understand these triggered earthquakes, providing us with invaluable information for future seismology studies. Key Points: We find an uncatalogued M5.1 earthquake in NE Turkey dynamically and instantaneously triggered by the 2023 Mw 7.8 SE Turkey earthquakeWe uncover this event on a thousand‐kilometer linear distributed acoustic sensing array, then confirm its precise time, location, and magnitude with seismometersPrecursory signals of the M5.1 event are observed and are most likely resulting from other nearby events rather than its nucleation phase [ABSTRACT FROM AUTHOR]

Understanding the scaling laws of source parameters is a fundamental subject in seismology. This study conducts spectral ratio analysis to estimate the source parameters for 409 shallow crustal earthquakes with Mw 3.2–6.0 in Japan. Subsequently, the source parameter scaling relations are investigated for a wide magnitude range by combining the results of this and previous studies. The spectral ratio method applied in this study provides the finite source properties of the localized area with large slip. The single asperity model, a simple heterogeneous source model with a single localized area with large slip, is used to estimate the stress drop since it can be more suitable to characterize earthquake sources than the standard homogeneous circular source. This investigation calibrates the constant to link the corner frequency and the source radius by comparing the corner frequency and source area estimated by the spectral ratio analyses. This calibrated constant is used to re‐calculate the stress drop of small earthquakes from the corner frequency estimated in previous studies. The stress drop and apparent stress increase with increasing magnitude up to Mw ∼ 5 and become magnitude‐independent for larger earthquakes. The radiation efficiency ranges typically from 0.1 to 1.0 and is independent of magnitude. The slip dependences of the stress drop, apparent stress, radiation efficiency, and fracture energy observed in this study are explained by the ones predicted from the slip‐weakening model incorporating the thermal pressurization effect. Thermal pressurization is one of the possible mechanisms explaining the observed source parameter scaling relations. Plain Language Summary: This study investigates the relationship between earthquake sizes and parameters characterizing the earthquake sources (source parameter scaling relations) to understand the difference between small and large earthquakes. The source parameters are estimated for shallow crustal earthquakes in Japan. Then, the source parameter scaling relations are obtained over a broad magnitude range by combining the results of this and previous studies. This investigation is based on a simple heterogeneous source model rather than the homogeneous circular source model, which has been widely used. A familiar hypothesis is that the sources of small and large earthquakes are physically similar (earthquake similarity). The results suggest that the earthquake similarity holds for earthquakes with a magnitude larger than about 5.0 but does not for small earthquakes. This study demonstrates that the observed source parameter scaling relations can be explained by the source parameter scaling relations predicted from the stress relaxation model incorporating the effect of the pore pressure rise within the fault core by frictional heating during seismic faulting (thermal pressurization of pore fluids). Thermal pressurization of pore fluids can be one of the significant mechanisms explaining the differences in the source parameter scaling relations for small and large earthquakes. Key Points: The single asperity model, a simple heterogeneous source model, is used rather than the standard homogeneous circular source modelThe apparent stress and stress drop increase with magnitude up to Mw ∼ 5 and become magnitude‐independent for larger earthquakesThermal pressurization is one of the possible mechanisms explaining the observed source parameter scaling relations [ABSTRACT FROM AUTHOR]

The b‐value in earthquake magnitude‐frequency distribution quantifies the relative frequency of large versus small earthquakes. Monitoring its evolution could provide fundamental insights into temporal variations of stress on different fault patches. However, genuine b‐value changes are often difficult to distinguish from artificial ones induced by temporal variations of the detection threshold. A highly innovative and effective solution to this issue has recently been proposed by van der Elst (2021, https://doi.org/10.1029/2020jb021027) by means of the b‐positive estimator, which is based on analyzing only the positive differences in magnitude between successive earthquakes. Here, we demonstrate the robustness of the estimator, which remains largely unaffected by detection issues due to the properties of conditional probability. We illustrate that this robustness can be further improved by considering positive differences in magnitude, not only between successive earthquakes but also between different pairs of earthquakes. This generalized approach, defined as the "b‐more‐positive estimator," enhances efficiency by providing a precise estimate of the b‐value while including a larger number of earthquakes from an incomplete catalog. However, our analysis reveals that the accuracy of the b estimators diminishes when earthquakes below the completeness threshold are included in the catalog. This leads to the paradoxical observation that greater efficiency is achieved when the catalog is more incomplete. To address this, we introduce the "b‐more‐incomplete estimator," where the b‐more‐positive estimator is applied only after artificially filtering the instrumental catalog to make it more incomplete. Our findings show the superior efficiency of the b‐more‐incomplete method. Plain Language Summary: Earthquake magnitudes can vary widely, and the b‐value is a common metric used to measure the frequency of earthquakes with large versus small magnitudes. In addition, the b‐value could serve as an indicator of the stress state of different fault patches, making it a valuable tool in earthquake research. However, since small earthquakes are often obscured by previous larger ones, determining whether changes in the b‐value are genuine or simply caused by detection problems can be challenging. To address this issue, a new approach called the b‐positive estimator has been recently developed. The method only considers positive changes in magnitude between successive earthquakes. In this study, we confirm that the b‐positive estimator is a powerful and effective technique to estimate the b‐value and is largely unaffected by issues related to detecting earthquakes. We extend the method by considering positive differences in magnitude, encompassing not only successive earthquakes but also different pairs of earthquakes. In particular we show that the b‐positive estimator is more efficient when the catalog is more incomplete. This allows us to develop modifications to the b‐positive method providing a more efficient tool to monitor the b‐value in ongoing seismic sequences. Key Points: Conditional probability detecting consecutive earthquakes makes positive magnitude difference distribution weakly affected by incompletenessThe b‐positive estimator can be enhanced by including more earthquake pairs, not only consecutive onesThe b‐positive estimator can be enhanced by making the catalog more incomplete [ABSTRACT FROM AUTHOR]

Teleseismic back‐projection imaging has emerged as a powerful tool for understanding the rupture propagation of large earthquakes. However, its application often suffers from artifacts related to the receiver array geometry. We developed a teleseismic back‐projection technique that can accommodate data from multiple arrays. Combined processing of P and pP waveforms may further improve the resolution. The method is suitable for defining arrays ad‐hoc to achieve a good azimuthal distribution for most earthquakes. We present a catalog of short‐period rupture histories (0.5–2.0 Hz) for all earthquakes from 2010 to 2022 with MW ≥ 7.5 and depth less than 200 km (56 events). The method provides automatic estimates of rupture length, directivity, speed, and aspect ratio, a proxy for rupture complexity. We obtained short‐period rupture length scaling relations that are in good agreement with previously published relations based on estimates of total slip. Rupture speeds were consistently in the sub‐Rayleigh regime for thrust and normal earthquakes, whereas a tenth of strike‐slip events propagated at supershear speeds. Many rupture histories exhibited complex behaviors, for example, rupture on conjugate faults, bilateral propagation, and dynamic triggering by a P wave. For megathrust earthquakes, ruptures encircling asperities were frequently observed, with downdip, updip, and balanced patterns. Although there is a preference for short‐period emissions to emanate from central and downdip parts of the megathrust, emissions updip of the main asperity are more frequent than suggested by earlier results. Plain Language Summary: Back‐projection is an earthquake imaging method based on seismic waveforms recorded remotely at a group of seismometers (seismic array). Here, we develop a new approach to combine backprojections from multiple arrays and seismic waveforms and use it to derive a catalog of large earthquake rupture histories from 2010 to 2022, providing a map view of the high‐frequency radiation emitted along the fault. The method automatically estimates the earthquake rupture length, speed, directivity, and aspect ratio. Based on these estimates, we obtained scaling relations between the earthquake magnitude and rupture length that agree with classical relationships. We identified strike‐slip earthquakes propagating at supershear, that is, faster than the shear wave speed, the usual limit for self‐sustaining rupture propagation. We observed complex rupture behaviors, for example, multiple faults activated, bilateral ruptures, and triggering of the main phase of a rupture by a primary (P) wave from the earliest part of the rupture. For subduction earthquakes, high‐frequency emissions were often observed, forming a ring around the fault interface patches (asperities) where the main slip occurs. There was a preference for high‐frequency radiation to emanate from central and deeper parts of the subducting plate interface, but shallower emissions were more frequent than expected from previous literature. Key Points: We provide a complete catalog of high‐frequency rupture histories for M ≥ 7.5 events 2010–2022We develop a semi‐automatic method for estimating rupture length, speed, directivity, and aspect ratioBoth encircling ruptures and emissions updip of slip asperities common in megathrust earthquakes [ABSTRACT FROM AUTHOR]

Repeating earthquakes repeatedly rupture the same fault asperities, which are likely loaded to failure by surrounding aseismic slip. However, repeaters occur less often than would be expected if these earthquakes accommodate all of the long‐term slip on the asperities. Here, we assess a possible explanation for this slip discrepancy: partial ruptures. On asperities that are much larger than the nucleation radius, a fraction of the slip could be accommodated by smaller ruptures on the same asperities. We search for partial ruptures of repeating earthquakes in Parkfield using the Northern California earthquakes catalog. We find 3991 individual repeaters which have 4468 partial ruptures. The presence of partial ruptures suggests that the asperities of repeating earthquakes are much larger than the nucleation radius. However, we find that partial ruptures could accommodate only around 25% of the slip on repeating earthquake patches. A 25% increase in the slip budget can explain only a small portion of the long recurrence intervals of repeating earthquakes. Plain Language Summary: Repeating earthquakes happen on the same fault patch over and over again. They are thought to happen on locked patches surrounded by a slowly moving section of the fault. This slow‐moving fault loads the patch to failure. However, the observed slip on the repeating earthquake patches does not match the long‐term slip on the surrounding fault. This slip deficit means the time between earthquakes is longer than expected. We explore the possibility that some of the slip deficit is explained by slip happening in smaller earthquakes ("partial ruptures") in between the time of the larger magnitude repeating earthquakes. We search for partial ruptures in Parkfield, California using the Northern California earthquakes catalog, which contains many well‐located repeating earthquake sequences. We find that partial ruptures could accommodate up to 25% of the slip on repeating earthquake patches, but this is still not enough slip to explain why small repeating earthquakes occur about five times less often than one would expect. Key Points: We search for partial ruptures of repeating earthquakes in Parkfield, CaliforniaWe find partial ruptures, which suggests repeating earthquake asperities are many times larger than the nucleation radiusIncluding partial ruptures in the slip budget does not account for the repeaters' surprisingly long recurrence intervals [ABSTRACT FROM AUTHOR]

The magnitude‐frequency distribution (MFD) describes the relative proportion of earthquake magnitudes and provides vital information for seismic hazard assessment. The b‐value, derived from the MFD, is commonly used to estimate the probability that a future earthquake will exceed a specified magnitude threshold. Improved MFD and b‐value estimates are of great importance in the central and eastern United States where high volumes of fluid injection have contributed to a significant rise in seismicity over the last decade. In this study, we recalculate the magnitudes of 8,775 events for the 2011 Prague, Oklahoma sequence using a relative magnitude approach that depends only on waveform data to calculate magnitudes. We also compare the distribution of successive magnitude differences to the MFD and show that a combination of the magnitude difference distribution (MDFD) and relative magnitudes yields a reliable estimate of b‐value. Using the MDFD and relative magnitudes, we examine the temporal and spatial variations in the b‐value and show that b‐value ranges between ∼0.6 and 0.85 during the aftershock sequence for at least 5 months after the M 5.7 mainshock, though areas surrounding the northeast part of the sequence experience higher b‐values (0.7–0.85) than the southwestern part of the Meeker‐Prague fault where b‐value is the lowest (0.6–0.7). We also identify a cluster of off‐fault events with the highest b‐values in the catalog (0.85). These new estimates of MFD and b‐value will contribute to understanding of the relations between induced and tectonic earthquake sequences and promote discussion regarding the use of b‐value in induced seismic hazard estimation. Plain Language Summary: The magnitude‐frequency distribution and the b‐value are statistical parameters that describe how earthquake magnitudes are distributed in a catalog. These parameters are vital for hazard estimation of induced seismicity in the central and eastern United States, but they are not well understood in this region due to reasons including, but not limited to, catalog incompleteness, magnitude uncertainty between different magnitude types, and variable injection history. In this study, we reevaluate the previously cataloged magnitude estimates using a relative magnitude method to estimate the temporal and spatial variation in b‐value for the 2011 Prague, Oklahoma sequence. With the relative magnitude method, we reestimate magnitude for 81% of the events in the Prague sequence which are then utilized for b‐value estimation. We demonstrate the presence of consistently low b‐values during the aftershock sequence for at least ∼5 months following the mainshock and identify a trend of decreasing b‐value along the Meeker‐Prague fault as distance from the mainshock increases. These observations allow us to compare spatial and temporal b‐value variations of induced sequences to those observed in tectonic sequences. Key Points: The relative magnitude method allows for recalculation of earthquake magnitude without requiring location or geological dataRelative magnitudes combined with the b+ estimator gives a reliable estimate of b‐valueWe identify patterns of spatial and temporal variations in b‐value that are consistent with tectonic sequences [ABSTRACT FROM AUTHOR]

Elevated seismic noise for moderate‐size earthquakes recorded at teleseismic distances has limited our ability to see their complexity. We develop a machine‐learning‐based algorithm to separate noise and earthquake signals that overlap in frequency. The multi‐task encoder‐decoder model is built around a kernel pre‐trained on local (e.g., short distances) earthquake data (Yin et al., 2022, https://doi.org/10.1093/gji/ggac290) and is modified by continued learning with high‐quality teleseismic data. We denoise teleseismic P waves of deep Mw5.0+ earthquakes and use the clean P waves to estimate source characteristics with reduced uncertainties of these understudied earthquakes. We find a scaling of moment and duration to be M0 ≃ τ4, and a resulting strong scaling of stress drop and radiated energy with magnitude (Δσ≃M00.21 ${\Delta }\sigma \simeq {M}_{0}^{0.21}$ and ER≃M01.24 ${E}_{R}\simeq {M}_{0}^{1.24}$). The median radiation efficiency is 5%, a low value compared to crustal earthquakes. Overall, we show that deep earthquakes have weak rupture directivity and few subevents, suggesting a simple model of a circular crack with radial rupture propagation is appropriate. When accounting for their respective scaling with earthquake size, we find no systematic depth variations of duration, stress drop, or radiated energy within the 100–700 km depth range. Our study supports the findings of Poli and Prieto (2016, https://doi.org/10.1002/2016jb013521) with a doubled amount of earthquakes investigated and with earthquakes of lower magnitudes. Plain Language Summary: The vibration of the Earth's ground recorded at seismometers carries the seismic signatures of distant earthquakes superimposed to the Earth's natural or anthropogenic noise surrounding the seismic station. We use artificial intelligence technology to separate the weak signals of distant earthquakes from other sources of ground vibrations unrelated to the earthquakes. The separated signal provides new insights into earthquakes, especially those within the Earth's deep interior, most of which have not been investigated due to noise levels. In contrast with shallow earthquakes, deep earthquakes are less efficient at radiating energy, though they exhibit a higher rate of increase in both stress drop and radiated energy as they grow. This may suggest that deep earthquakes tend to be more confined fault surfaces. A dual mechanism between nucleation in the subduction‐zone core and propagation of larger events in the dry mantle explains our observations. Key Points: A neural network is used to double the number of earthquakes for source analyses compared to previous studies by improving the data qualityDenoising teleseismic waves improves the source signature in Mw5.0–6.0 events and reduces uncertainties over all magnitudesLarge deep earthquake ruptures are dissipative and compact relative to crustal earthquakes [ABSTRACT FROM AUTHOR]

Seismic activity in the Northern Canadian Cordillera is characterized by diffuse earthquakes that extend hundreds of km northwest from the Yakutat collision zone. We use 25 months of broadband seismic data from Mackenzie Mountain Earthscope Project (MMEP), USArray Transportable Array (TA), and permanent Canadian National Seismic Network stations to present a local earthquake catalog with high sensitivity to small regional events. Deep learning techniques are adopted for both seismic phase detection and association. Event relocations are performed to provide well constrained estimates of earthquake depth distributions. Clusters of seismicity spanning the upper crust are located in the central Richardson Mountains, along the Tintina fault, and in the northeast Selwyn Basin. These clusters suggest that the core of the Richardson Anticlinorium is tectonically active and that the Tintina fault is a locus for low levels of active deformation. We interpret seismicity in the northeast Selwyn Basin as primarily occurring in the hanging wall of the Plateau thrust fault and suggest that some combination of localized duplex structures and lithological strength contrasts both within the Selwyn Basin and between abutting Paleozoic shelf sequences may be responsible for seismicity in the Mackenzie Mountain foreland. Plain Language Summary: The Northern Canadian Cordillera is situated inboard of a transition from a right‐lateral tectonic regime in coastal southeast Alaska to a subduction setting beneath southern Alaska. It presents a unique case study to examine the length scales over which tectonic deformation is able to permeate into a continental interior. In this study we use data from a relatively dense regional network of seismometers in order to present the most spatially complete catalog of earthquakes in the Northern Canadian Cordillera to date. We use machine learning techniques for the detection and classification of small magnitude earthquakes and use our results to analyze regional tectonics. Active faults that host earthquakes are identified and we use the earthquake depth distributions to further distinguish between mechanisms for regional stress transfer. We find evidence that the Tintina fault is presently active along a confined area at a low rate. We also find that low magnitude earthquakes do not generally occur within the Mackenzie Mountain Fold‐Thrust belt but are prevalent within the foreland Selwyn Basin. We also detect a pattern of relatively deep (∼30 km) earthquakes in the northern Richardson Mountains. Key Points: Small magnitude earthquakes extending to 25 km depth are detected throughout the Selwyn Basin in the Mackenzie Mountain forelandThe Tintina fault hosts a range of low magnitude earthquakes in the uppermost 15 km of crust within a 400 km zone along strikeDeformation in the Richardson Mountains is localized onto a narrow band of faults centered on the Richardson Anticlinorium [ABSTRACT FROM AUTHOR]

Earthquake early warning (EEW) systems aim to forecast the shaking intensity rapidly after an earthquake occurs and send warnings to affected areas before the onset of strong shaking. The system relies on rapid and accurate estimation of earthquake source parameters. However, it is known that source estimation for large ruptures in real‐time is challenging, and it often leads to magnitude underestimation. In a previous study, we showed that machine learning, HR‐GNSS, and realistic rupture synthetics can be used to reliably predict earthquake magnitude. This model, called Machine‐Learning Assessed Rapid Geodetic Earthquake model (M‐LARGE), can rapidly forecast large earthquake magnitudes with an accuracy of 99%. Here, we expand M‐LARGE to predict centroid location and fault size, enabling the construction of the fault rupture extent for forecasting shaking intensity using existing ground motion models. We test our model in the Chilean Subduction Zone with thousands of simulated and five real large earthquakes. The result achieves an average warning time of 40.5 s for shaking intensity MMI4+, surpassing the 34 s obtained by a similar GNSS EEW model. Our approach addresses a critical gap in existing EEW systems for large earthquakes by demonstrating real‐time fault tracking feasibility without saturation issues. This capability leads to timely and accurate ground motion forecasts and can support other methods, enhancing the overall effectiveness of EEW systems. Additionally, the ability to predict source parameters for real Chilean earthquakes implies that synthetic data, governed by our understanding of earthquake scaling, is consistent with the actual rupture processes. Plain Language Summary: Earthquake Early Warning (EEW) systems can evaluate the shaking impact soon after an earthquake occurs and send alerts to areas where strong shaking has not yet arrived. However, predicting shaking intensity for large earthquakes face challenges such as magnitude underestimation and rupture complexity, leading to inaccurate prediction. In our previous study, we proposed a machine learning model, called M‐LARGE, which can predict the magnitude of large earthquakes accurately and rapidly by utilizing surface deformation patterns recorded by HR‐GNSS. Here, we expand the M‐LARGE model to also predict the source location and fault size. This capability allows us to construct a fault, enabling forecasts of strong shaking 35–40 s before the shaking onset, longer than other similar methods. Our approach has the potential to improve the accuracy of EEW and reduce the impact of seismic hazards. Because the model is fully trained with simulated data, the successful prediction of real Chilean earthquakes indicates that our assumptions in simulating large earthquakes are sufficiently consistent with the rupture dynamics of actual events. Key Points: We train a ML model that utilizes HR‐GNSS data to predict source parameters and shaking intensity for large (M7.0+) earthquakesThe model yields a long median warning time of 35–40 s for MMI 4+ when tested on synthetic and real HR‐GNSS recordsOur model fills a critical gap in current EEW systems for large events and can operate in parallel with other methods for better warnings [ABSTRACT FROM AUTHOR]

Tsunamis can have devastating consequences for coastal communities. Yet hazards from future tsunamis are difficult to quantify due to their rarity in the instrumental record. Statistical earthquake catalogs have previously been used to quantify tsunami hazards. For the first time, we use a physics‐based synthetic earthquake catalog to assess probabilistic tsunami hazard in a local region. We analyze a 30,000‐year catalog to capture a wide range of the source complexity evident in tsunamigenic earthquakes. We produce models of land‐surface displacements and subsequent tsunamis from 2,585 earthquakes with magnitude > MW 7.0. Modeled slip of the Hikurangi and Tonga‐Kermadec subduction thrusts generated maximum wave heights at the coast of up to 28 m and these earthquakes pose the greatest tsunami hazard along New Zealand's coastline. The results provide a "proof of concept" for using earthquake simulators for probabilistic tsunami hazard models. They further present a platform upon which next‐generation probabilistic tsunami inundation models may be constructed. Plain Language Summary: Tsunamis with devastating, wide‐reaching consequences are significant hazards to nearly all of the world's coastlines that needs to be analyzed and quantified. Previous tsunami hazard studies have relied on sparse data retrieved from geological and/or probabilistic earthquake models to examine this infrequent but potentially deadly hazard. Consequently, they lack crucial information about the broad range of tsunami hazards resulting from local tsunamis. Our study represents the first time that a physics‐based synthetic earthquake catalog has been used to assess probabilistic tsunami hazard. It overcomes many of the challenges of previous tsunami hazard assessments. We model the tsunamis generated from 2,585 earthquakes which have magnitudes greater than 7.0 from 30,000 years of a synthetic catalog of earthquakes nucleating on both the subduction zones and crustal faults through New Zealand. Tsunami wave heights at the coast of up to 28 m were recorded. Additional analysis indicates that modeled earthquakes on the Hikurangi and Tonga‐Kermadec subduction thrusts pose the greatest hazard along many parts of New Zealand's coastline. Key Points: First use of a physics‐based synthetic earthquake catalog in tsunami hazard assessmentOur assessment demonstrates that near‐field seismic tsunamis can generate maximum wave heights at the coast of up to 28 m around New ZealandOur results provide a pathway toward next‐generation probabilistic tsunami risk modeling [ABSTRACT FROM AUTHOR]

We introduce the Ensemble Earthquake Early Warning System (E3WS), a set of Machine Learning (ML) algorithms designed to detect, locate, and estimate the magnitude of an earthquake starting from 3 s of P‐waves recorded by a single station. The system is made of six Ensemble ML algorithms trained on attributes computed from ground acceleration time series in the temporal, spectral, and cepstral domains. The training set comprises data sets from Peru, Chile, Japan, and the STEAD global data set. E3WS consists of three sequential stages: detection, P‐phase picking, and source characterization. The latter involves magnitude, epicentral distance, depth, and back azimuth estimation. E3WS achieves an overall success rate in the discrimination between earthquakes and noise of 99.9%, with no false positive (noise mis‐classified as earthquakes) and very few false negatives (earthquakes mis‐classified as noise). All false negatives correspond to M ≤ 4.3 earthquakes, which are unlikely to cause any damage. For P‐phase picking, the Mean Absolute Error is 0.14 s, small enough for earthquake early warning purposes. For source characterization, the E3WS estimates are virtually unbiased, have better accuracy for magnitude estimation than existing single‐station algorithms, and slightly better accuracy for earthquake location. By updating estimates every second, the approach gives time‐dependent magnitude estimates that follow the earthquake source time function. E3WS gives faster estimates than present alert systems relying on multiple stations, providing additional valuable seconds for potential protective actions. Plain Language Summary: We designed E3WS, an Earthquake Early Warning System (EEWS) using the first 3 s of seismic waves recorded by a single station. E3WS is the first EEWS entirely made of Artificial Intelligence algorithms. Our results show an improvement in magnitude estimates compared to existing single‐station based EEWS, such as Urgent Earthquake Detection and Alarm System and ElarmS, and slightly better localization. In order to track the magnitude of M > 7.0 earthquakes, whose source duration is typically longer than 6 s, we test E3WS on large earthquakes in real‐time scenarios. E3WS provides time‐dependent magnitude estimates that follow the earthquake source time functions. E3WS is highly exportable. Thanks to its simplicity, it can be installed in 32‐bit Single Board Computers like a Raspberry Pi 4, and of course in more complex 64‐bit processors. E3WS is not only theoretical, it is already running in alpha test mode as the core of the EEWS of Peru in construction. We believe that E3WS will provide valuable seconds to populations living at constant risk, giving the user extra seconds that may give them enough time to "drop, cover, and hold on" or to perform mitigation actions like stopping traffic, stopping elevators or evacuating the ground floor of buildings. Key Points: Earthquake Early Warning System (EEWS) with faster and robust preliminary estimates of magnitude and location to generate a first alertFirst EEWS designed purely with Artificial IntelligenceEarthquake Early Warning with continuous updates that follow the earthquake source time function [ABSTRACT FROM AUTHOR]

Following laboratory experiments and friction theory, slow slip events and seismicity rate accelerations observed before mainshocks are sometimes interpreted as evidence of a nucleation phase. However, such precursory observations still remain scarce and are associated with different time and length scales, raising doubts about their actual preparatory nature. We study the 2017 Valparaiso Mw = 6.9 earthquake, which was preceded by aseismic slip accompanied by an intense seismicity, suspected to reflect its nucleation phase. We complement previous observations, which have focused only on precursory activity, with a continuous investigation of seismic and aseismic processes from the foreshock sequence to the post‐mainshock phase. By building a high‐resolution earthquake catalog and searching for anomalous seismicity rate increases compared to aftershock triggering models, we highlight an over‐productive seismicity starting within the foreshock sequence and persisting several days after the mainshock. Using repeating earthquakes and high‐rate GPS observations, we highlight a transient aseismic perturbation starting 1‐day before the first foreshock and continuing after the mainshock. The estimated slip rate over time is lightly impacted by large magnitude earthquakes and does not accelerate toward the mainshock. Therefore, the unusual seismic and aseismic activity observed during the 2017 Valparaiso sequence might be interpreted as the result of a slow slip event starting before the mainshock and continuing beyond it. Rather than pointing to a possible nucleation phase of the 2017 Valparaiso mainshock, the identified slow slip event acts as an aseismic loading of nearby faults, increasing the seismic activity, and thus the likelihood of a large rupture. Plain Language Summary: Both laboratory experiments and friction theory show that earthquakes do not begin abruptly but are preceded by an accelerating slip associated with a seismicity increase. On the field, however, such precursory observations still remain scarce and are associated with different characteristic time and length scales, raising doubts that they actually reflect the same nucleation phenomena. We study the 2017 Valparaiso M = 6.9 earthquake, which was preceded by both a slow slip and an intense seismicity suspected to reflect such nucleation phase. We complement previous studies, that have focused only on precursory activity, with a continuous investigation of seismic and slow slip before and after the mainshock. Using refined earthquake detection tools, we highlight a seismicity excess starting before and persisting several days after the mainshock. Using repeating earthquakes and high‐resolution GPS, we show that the slow slip does not accelerate toward the mainshock, but continues after it. Therefore, rather than pointing to a possible accelerating nucleation phase of the Valparaiso mainshock, we suggest that the slow slip drives an enhanced seismic activity that is not mainshock‐directed. Within such slow‐slip driven seismicity, the probability of triggering a large earthquake (subsequently considered as the mainshock) is increased. Key Points: We use a high resolution seismic catalog and GPS to investigate seismic and aseismic process before and after the Valparaiso mainshockAn unusually high seismicity and a slow slip event is continuously observed from the foreshock sequence up to days after the mainshockRather than a nucleation phase of the mainshock, the slow slip event acts as an aseismic loading of nearby faults during the entire sequence [ABSTRACT FROM AUTHOR]

Due to the lack of recognized creep on dip‐slip faults, the nature of fault creep and its role in the generation of large earthquakes is largely unexplored. The Chihshang fault in Taiwan serves as one of the best examples of a thrust creeping fault in the world, exhibiting a fast creep rate of 2–3 cm/year and the capability of producing magnitude six earthquakes. Land‐based geodetic measurements are less sensitive to fault slip with depth, especially for the deeper portion of the seaward dipping fault. Taking advantage of slip rates inferred from repeating earthquake sequences (RESs) at greater depths (up to 30 km), we present a method that embeds RES‐derived slip rates into the neighboring fault patch for geodetic data inversion. Using the geodetic and seismological data from 2007 to 2011, we achieve a higher resolution of inter‐seismic slip rate distribution below a depth of 10 km. The inferred low coupling ratio establishes an extensive creeping area that coincides with a location of abundant RES and earthquake swarms. The inferred high coupling ratio, on the other hand, delineates a locked area corresponding to the two co‐seismic slip zones of the 2003 Mw 6.8 Chengkung and 2022 Mw 6.7 Yuli earthquakes. The transition zone marking the gradual change from fully locked to shallow creeping areas appears to coincide with the post‐seismic slip area of the 2003 Chengkung event. Using the fault coupling model, the moment rate deficit measured in the two locked areas is capable of generating a Mw 6.8 event every 78 years to the south and a Mw 6.7 event every 28 years to the north. This moment rate deficit is decreased by 10% when the closely located Central Range fault is considered in a two‐fault model, leading to the slightly longer recurrence interval on the Chihshang fault. The two‐fault model potentially contributes to a better understanding of the regional inter‐seismic deformation, which is recommended to be adopted in the future studies of seismic hazard assessment. Plain Language Summary: Large earthquakes can occur on the creeping faults that slowly slip at the surface but accumulate energy at greater depth. How much different portions of the fault slip annually is the key question for seismic hazard assessment. However, for a seaward‐dipping thrust fault like the Chihshang fault in Taiwan, finding the answer is difficult because the loss of resolution at the greater depth (>10 km) due to a lack of offshore data. Taking advantage of the slip rates inferred from repeating earthquakes that mainly occurred at 10–30 km depth, we present a method that embeds repeating events data into the geodetic inversion for higher resolution of inter‐seismic slip model. We found that the areas slipped faster during the inter‐seismic period (higher inter‐seismic slip rate) coincides with the location of abundant repeating earthquakes and swarm events; The areas with lower inter‐seismic slip rate on the other hand, correspond to the zones that slipped during the 2003 Mw 6.8 Chengkung and 2022 Mw 6.7 Yuli earthquakes. By calculating how much and where the seismic energy was released through inter‐seismic slip rate, we are able to identify two separate zones on the Chihshang fault that is capable of generating a ∼Mw 6 event every 28 to the north and 78 years to the south of the fault, respectively. Key Points: We present a method that embeds repeaters‐derived slip rates into the neighboring fault patch for geodetic data inversionWe achieve a higher resolution of inter‐seismic slip rate distribution below a depth of 10 km, on a seaward‐dipping thrust faultThe inferred low coupling ratio establishes an extensive creeping area, coinciding with the location of active repeaters and swarm events [ABSTRACT FROM AUTHOR]

The recently identified Prompt Elasto‐Gravity Signals (PEGS), generated by large earthquakes, propagate at the speed of light and are sensitive to the earthquake magnitude and focal mechanism. These characteristics make PEGS potentially very advantageous for tsunami early warning, which relies on fast and accurate estimation of the magnitude of large offshore earthquakes. PEGS‐based early warning does not suffer from the problem of magnitude estimation saturation, that P‐wave based early warning algorithms have, and could be faster than Global Navigation Satellite Systems (GNSS)‐based systems while not requiring a priori assumptions on slip distribution. We use a deep learning model called PEGSNet to evaluate the possibility to estimate in real time the evolution of the magnitude of big earthquakes in the tsunamigenic zone of Chile. The model is a Convolutional Neural Network (CNN) trained on a database of synthetic PEGS—simulated for an exhaustive set of possible earthquakes distributed along the Chilean subduction zone—augmented with empirical noise. The approach is multi‐station and leverages the information recorded by the seismic network to estimate as fast as possible the magnitude and location of an ongoing earthquake. Our results indicate that PEGSNet could have estimated that the magnitude of the 2010 Mw 8.8 Maule earthquake was above 8.7, 90 s after origin time. Our offline simulations using real data and noise recordings further support the instantaneous tracking of the source time function of the earthquake and show that deploying seismic stations in optimal locations could improve the performance of the algorithm. Plain Language Summary: Tsunami early warning requires the fast and reliable estimation of an earthquake magnitude provided by Earthquake Early Warning (EEW) systems. EEW systems are currently limited by the propagation speed of P‐waves, which rely on as natural information carriers. Even more problematic, EEW systems based on the first seismic arrivals tend to saturate with earthquake magnitude, and can become unreliable for magnitudes above 8. The recent discovery of Prompt Elasto‐Gravity Signals (PEGS), which comprise gravitational changes generated by earthquakes, has raised hope to overcome these limitations because they travel at the speed of light, much faster than P‐waves. We use PEGS to re‐train the previously developed deep learning model PEGSNet to track the magnitude evolution of big earthquakes in the Chilean subduction zone, historically affected by tsunamis. Given the scarcity of big earthquakes, we simulate signals to train the model using synthetic sources and the seismic stations available in 2010 and 2021, augmented with empirical noise recorded by those stations. PEGSNet tracks the moment release 90 s after the origin time. The performance of PEGSNet is limited by the seismic network configuration, the number of stations and the level of noise in the data, but could be useful for tsunami warning. Key Points: A Prompt Elasto‐Gravity Signals‐based early warning algorithm for ChileOur algorithm could have estimated the magnitude of the Maule earthquake 90 s after origin timeDeployment of seismic stations at optimal locations could greatly enhance the algorithm performance [ABSTRACT FROM AUTHOR]

Stress accumulation and release reflected by acoustic emissions (AEs) during shearing of granular materials provide important information on failure mechanisms in seismic faults and landslides controlled by stick‐slip. Among many characteristics (amplitude, energy, counts, and frequency) of AE signals generated by stick‐slip, stress changes corresponding to various frequency AEs in different stages of the stick‐slip process are not clear, which limits our knowledge of the characteristics of precursory signals before stick‐slip failure. To better understand the physical mechanisms of granular stick‐slip, we monitored the mechanical and AE signals using high‐frequency (2 MHz) synchronous acquisition during constant‐speed shear of packs of uniform glass beads with different sizes at different normal stresses. The release rate of AE energy was found to accelerate with the dilatation of the sample volume, and the stress drop of stick‐slip was augmented with the increase of normal stress and particle size. Three characteristic events of single cycle stick‐slip were observed in this study: main slip, minor slip, and microslip. We analyzed the AE frequency spectra of these three event types. Both main slip and minor slip corresponded to stress drop and generated high‐frequency AEs (about several hundred kHz), while the AE frequencies generated by microslip were lower (about tens of kHz) and exhibited stress strengthening, which were not apparent in previous studies due to the low frequency of acquisition. We propose that the microslip is mainly due to sliding on grain contacts, while the main slip and minor slip resulted from breakage and reforming of force chains. Low‐frequency AEs from microslip may suggest a crucial precursor of seismic faults and landslides. Plain Language Summary: The occurrence of earthquakes of similar magnitude (quasi‐periodic earthquakes) in a fault zone every few decades is thought to be controlled by stick‐slip of fault gouges. The mechanical evolution of fault gouges prior to earthquake occurrence is extremely complex and difficult to predict. However, many studies have observed the presence of significant microslip prior to stick‐slip failure of seismic faults, which may provide precursor information for an impending earthquake. Laboratory study of stick‐slip nucleation and failure mechanisms in granular materials provides important understanding of the physical mechanisms of quasi‐periodic earthquakes. Our experimental results of high‐frequency acoustic emission (AE) and ring‐shear simultaneous acquisition showed that stick‐slip failure (corresponding to the occurrence of an earthquake) leads to a decrease in the shear resistance of a granular system, generating high‐frequency acoustic signals. Increments of microslip increased the shear resistance of the granular system, and were recorded as low‐frequency acoustic signals before failure occurred. We suggest that the stick‐slip nucleation and failure were controlled by grain friction and collision, respectively. The different AE frequencies of the two may shed light on the precursor monitoring of quasi‐periodic earthquakes. Key Points: Stick‐slip failure was associated with high‐frequency (∼105 Hz) acoustic emissions (AEs) and micro‐slip during stick‐slip nucleation was associated with low‐frequency (∼104 Hz) AEsHigh‐frequency AEs may be caused by particle collisions after force‐chain breakageLow‐frequency AEs may be caused by particle friction during force‐chain adjustment and may suggest important precursors of laboratory earthquakes [ABSTRACT FROM AUTHOR]

The question "what arrests an earthquake rupture?" sits at the heart of any potential prediction of earthquake magnitude. Here, we use a one‐dimensional, thin‐elastic‐strip, minimal model, to illuminate the basic physical parameters that may control the arrest of large ruptures. The generic formulation of the model allows for wrapping various earthquake arrest scenarios into the variations of two dimensionless variables τ¯k ${\bar{\tau }}_{k}$ (initial dimensionless stress parameter on the fault) and G¯c ${\bar{G}}_{c}$ (dimensionless fracture energy), valid for both in‐plane and antiplane shear loading. Our continuum model is equivalent to the standard Burridge‐Knopoff model, with an added characteristic length scale, H, that corresponds to either the thickness of the damage zone for strike‐slip faults or to the thickness of the downward moving plate for subduction settings. We simulate the propagation and arrest of frictional ruptures and derive closed‐form expressions to predict rupture arrest under different conditions. Our generic model illuminates the different energy budget that mediates crack‐ and pulse‐like rupture propagation and arrest. It provides additional predictions such as generic stable pulse‐like rupture solutions, stress drop independence of the rupture size, the existence of back‐propagating fronts, and predicts that asymmetric slip profiles arise under certain pre‐stress conditions. These diverse features occur also in natural earthquakes, and the fact that they can all be predicted by a single minimal framework is encouraging and pave the way for future developments of this model. Plain Language Summary: Untangling the dynamics that governs the propagation and arrest of earthquakes is still challenging, mainly because of the few constraints available on the fault zone geometry, the constitutive properties of fault materials, as well as fault rheology during the rupture event. The present study aims at formulating a model containing a minimal number of parameters to describe the dynamics of large earthquakes. Despite its simplicity, this minimal model is able to reproduce several salient features of natural earthquakes that are still debated (e.g., various arrest scenarios, stable pulse‐like rupture, back‐propagating front, asymmetric slip profiles). We demonstrate how the proposed model can be used to simulate the propagation and arrest of large earthquakes, which are controlled by local variations of shear stress and material properties on the fault. With this simple and generic description, the proposed model could be readily extended to account for additional processes controlling the dynamics of large earthquakes. Key Points: A minimal model of frictional rupture describes large earthquake rupturesTwo parameters, the dimensionless pre‐stress and fracture energy, account for previously proposed mechanisms of earthquake arrestThe model illuminates the different energy balance that drives crack‐like and pulse‐like ruptures and control the asymmetry of slip profiles [ABSTRACT FROM AUTHOR]

Nepal is an actively deforming region due to its tectonic setting that hosts destructive earthquakes including the recent 2015 Mw 7.8 Gorkha earthquake. To better understand the physics of earthquakes and better assess the seismic hazard in the region, a highly resolved 3‐D structure of the crust is essential. This study presents a new 3‐D S‐wave velocity structure of the crust using ambient noise tomography (ANT). In relation to the 2015 earthquake, we show that the lateral variation of S‐wave velocity likely controls the rupture propagation and arrest. This study further constrains crustal discontinuities beneath Nepal including the Main Himalayan Thrust (MHT) using teleseismic P‐wave coda autocorrelation. The MHT geometry correlates well with sizeable low S‐wave velocity zones obtained from the ANT and separated by two ramps. The southernmost first ramp is where the 2015 Gorkha earthquake nucleated that we map up to the source model of the 1934 historical earthquake in Eastern Nepal. The northernmost second ramp is adjacent to the mid‐crustal low velocity zone beneath south Tibet. In between the two ramps, we interpret the horizontal low velocity zone as the interseismic creeping patch of the MHT with down‐dip extent of 50–70 km length and 90–100 km in correspondence with the Mw 7.8 2015 and 8.4 1934 source models, respectively. A systematic mapping of the low‐velocity zones in relation to the MHT integrated with GPS geodesy along with the respective slip from individual past earthquake ruptures can help decipher the paradox of maximum Himalayan earthquake magnitudes. Plain Language Summary: Nepal is located near the plate boundaries between the Indian and Eurasian plates, resulting in damaging earthquakes in the region. However, we do not know exactly what controls the earthquake processes beneath Himalaya Nepal due to its complex geological settings and lack of detailed seismic properties beneath this region. Here we present the detailed shear (S)‐wave velocity image up to 60 km depth beneath Himalaya Nepal that shows how fast/slow the S‐wave propagates, relating to the material properties in crust beneath Himalaya. We used data extracted from recordings of noise coming from, for example, ocean and atmosphere. We also utilized the conversion and reverberation of compressional waves that are sensitive to the sharp transition of material properties. The estimated S‐wave velocity near surface is consistent with material found on the surface of the crust. Here, we show that MHT can be described as a system of low‐velocity zones connected by two ramps. The first ramp is located near the hypocenter of 2015 Gorkha earthquake and second beneath South Tibet. The low velocity between two ramps is considered as interseismic creeping patch. A systematic mapping of such low‐velocity zone beneath all the Himalaya may allow a better understanding of the earthquake hazard. Key Points: Main Himalayan Thrust (MHT) as a system of sizeable low‐velocity zones connected by rampsInferred MHT interseismic patches with different length scales in correspondence of the 2015 and 1934 earthquakesLateral variations of S‐wave velocity across MHT likely control the rupture of the 2015 Gorkha earthquakes [ABSTRACT FROM AUTHOR]

On 3 August 2014, a destructive earthquake with a magnitude of 6.5 occurred in the Ludian region in Yunan, China, causing heavy casualties. However, the seismogenic structure of the Ludian earthquake is still unclear. To address this issue, we have developed a three‐dimensional joint inversion algorithm for magnetotelluric data and seismic body wave arrival times based on the cross gradient structural similarity constraint, which utilizes the complementary advantages of the two different datasets. We use arrival times of P‐ and S‐waves from 6,369 local earthquakes recorded by 30 seismic stations and magnetotelluric impedances from 127 MT stations to perform the joint inversion to construct a combined 3‐D P‐wave velocity (Vp), S‐wave velocity (Vs), and resistivity models and the locations of relocated aftershocks in the source region of the Ludian earthquake. Our models show significant low Vp/Vs ratios and low resistivity in the upper crust of the main shock zone, and the relocated aftershocks form an L‐shaped zone. Based on petrophysical studies and analysis of the cross‐plots of different physical properties, we infer the existence of cracks/fractures and fluids as well as high quartz contents in the source region. These fluids may play a key role in promoting the Ludian earthquake by reducing the effective normal stress of the fault zone, and the existence of high quartz contents further makes the crustal rocks in the source region more brittle. Our study provides a new evidence how the properties of the crust in the focal area have some control on the genesis of moderate to large earthquakes. Plain Language Summary: On 3 August 2014, a magnitude Ms 6.5 earthquake struck the Ludian County in Yunnan Province, China. Although the earthquake magnitude is moderate, it caused unusually severe damages. To further investigate the genesis mechanism for this earthquake, we have developed a 3‐D joint inversion system for two types of geophysical data: Magnetotellurics (MT) and body wave arrival times. We present this new method of using MT and seismic arrival time data to construct 3D subsurface models of the Ludian earthquake region based on both datasets. Our study shows low Vp/Vs ratios and low resistivity in the source area, and the aftershocks form an L‐shaped zone. On the basis of petrophysical experiments and cross‐plot analysis of different physical properties, we infer that there are cracks/fractures and fluids in the source area. In addition, the source region is also rich in quartz, which make the crustal rocks in the source region more brittle. These fluids may play an important role in promoting the occurrence of the Ludian earthquake. Our study shows that the structure properties have important control on the genesis of moderate to large earthquakes. Key Points: 3D joint inversion of magnetotelluric data and seismic body wave arrival times has been developed based on cross gradient constraintMore structurally consistent velocity and resistivity models are obtained for the source region of the 2014 M6.5 Ludian earthquakeThe earthquake source region is associated with low Vp/Vs and resistivity values, indicating the presence of fluids and high quartz contents [ABSTRACT FROM AUTHOR]

The Húsavík‐Flatey Fault Zone (HFFZ) is the largest strike‐slip fault in Iceland and poses a high seismic risk to coastal communities. To investigate physics‐based constraints on earthquake hazards, we construct three fault system models of varying geometric complexity and model 79 3‐D multi‐fault dynamic rupture scenarios in the HFFZ. By assuming a simple regional prestress and varying hypocenter locations, we analyze the rupture dynamics, fault interactions, and the associated ground motions up to 2.5 Hz. All models account for regional seismotectonics, topo‐bathymetry, 3‐D subsurface velocity, viscoelastic attenuation, and off‐fault plasticity, and we explore the effect of fault roughness. The rupture scenarios obey earthquake scaling relations and predict magnitudes comparable to those of historical events. We show how fault system geometry and segmentation, hypocenter location, and prestress can affect the potential for rupture cascading, leading to varying slip distributions across different portions of the fault system. Our earthquake scenarios yield spatially heterogeneous near‐field ground motions modulated by geometric complexities, topography, and rupture directivity, particularly in the near‐field. The average ground motion attenuation characteristics of dynamic rupture scenarios of comparable magnitudes and mean stress drop are independent of variations in source complexity, magnitude‐consistent and in good agreement with the latest regional empirical ground motion models. However, physics‐based ground motion variability changes considerably with fault‐distance and increases for unilateral compared to bilateral ruptures. Systematic variations in physics‐based near‐fault ground motions provide important insights into the mechanics and potential earthquake hazard of large strike‐slip fault systems, such as the HFFZ. Plain Language Summary: The Húsavík‐Flatey Fault Zone (HFFZ) is the largest strike‐slip fault in Iceland, located in the Tjörnes Fracture Zone in Northern Iceland where the largest earthquakes in Iceland have occurred. At present the seismogenic potential of HFFZ suggests that an earthquake of magnitude ∼7 is possible, which poses a high earthquake hazard in the region. In this study, we generate a set of plausible earthquake rupture scenarios on the HFFZ that account for multi‐physics, regional geology and topo‐bathymetry. We simulate the corresponding seismic ground motions by exploring various assumptions, for example, in terms of slipping fault geometry and hypocenter locations. Our simulated scenarios have comparable magnitudes with historic events. The physics‐based ground motion scaling conforms to new empirical ground motion models, but shows varying ground motion variability with distance. Our study provides an overview of multiple rupture scenarios on the HFFZ and suggests that an ensemble of physics‐based dynamic rupture scenarios can complement classical seismic hazard assessment methods to better characterize the hazard in tectonically and seismically complex regions, especially in data‐scarce regions. Key Points: Observational constrained dynamic rupture scenarios of the Húsavík‐Flatey Fault Zone match expected regional earthquake scaling and magnitudesSegmented fault geometry, hypocenter location, fault prestress and roughness affect rupture dynamics and ground motionsOur physics‐based scenarios show magnitude‐consistent average attenuation relationships and match empirical ground motion models [ABSTRACT FROM AUTHOR]

The Cascadia subduction zone (CSZ) is known to host M9 megathrust ruptures; however, no such event has occurred in historical observation. The distribution and characteristics of small‐ to moderate‐sized earthquakes can be used to determine the behavior of the megathrust fault but are notably absent offshore the CSZ due to the distance from onshore seismometers. We use automated subspace detection coupled with an onshore‐offshore seismic deployment to find small‐magnitude earthquakes in the offshore seismogenic zone and analyze their locations in the context of interseismic locking and seismogenic zone extent. We detected and located 5,282 earthquakes, 4,096 of which had been previously undetected. We find that the downdip extent of the seismogenic zone as defined by interplate seismicity agrees with the 20% locking contour of the Schmalzle et al. (2014, https://doi.org/10.1002/2013GC005172) geodetic model and extends deeper than predicted by previous thermal models. We cannot determine the updip extent of the seismogenic zone; this may be due to a lack of templates for detection in the updip source area, stress shadows updip of asperity loading, and/or strong locking to the trench. We present a map of possible asperities determined by the small earthquakes in this study. Our asperity locations and extents show some, but not complete, agreement with the asperities modeled from the 1700 M9 rupture and geodetic locking models, and good agreement with the paleo‐rupture extents determined from offshore turbidites and forearc basin‐based asperity estimates. This highlights the need of continued offshore observations over time, and to elucidate fine‐scale variation in locking. Plain Language Summary: Subduction zones host the world's largest earthquakes, with many subduction zones capable of M9 megathrust ruptures. The size and ground motion of these great earthquakes depend on how much of the fault slips seismically and the portions of the fault that promote stronger plate locking between earthquake ruptures. Great earthquakes are known to have occurred in the Cascadia subduction zone, the most recent of which was a M9 in 1700. Despite the geologic evidence for these past great earthquakes, there are no historical observations of large‐ to great‐earthquakes to directly assess where the fault slips seismically or is strongly locked. Additionally, few interplate earthquakes have been observed along this plate boundary. We use a data set acquired with temporary ocean‐bottom seismometers and coastal land‐based seismometers to search for interplate earthquakes that were missed using land seismic stations alone, due to their location primarily offshore. We use the locations of these newly detected earthquakes to map seismogenic extent on the plate boundary and areas of strong plate locking, which will produce higher energy release during an earthquake rupture. Key Points: We develop an earthquake catalog using an amphibious data set and subspace detection targeting Cascadia subduction zone interplate eventsLocations of interplate seismicity detected in this study are used to delineate downdip extent of the seismogenic zone and asperity distributionAsperities agree somewhat with models of geodetic locking and 1700 M9 asperities, and closely match paleo‐rupture extents and forearc basins [ABSTRACT FROM AUTHOR]

Rupture dynamics along a heterogeneous fault is studied in the framework of the recently developed damage‐breakage rheological model. The model utilizes a fault structure accommodating shear and wear in a heterogeneous weak fault‐core a few centimeters thick, separating two blocks with damage level gradually decreasing toward the host rock. We first demonstrate the similarity of the static stress field around a fault, represented by a three‐body tribosystem, to the one developed around a rough frictional interface. We show that both models predict heterogeneous stress field pattern. We then apply simulations of rupture on heterogeneously damaged faults. We show that increasing initial heterogeneity amplitudes is associated with smaller events with lower slip rates. The simulations further allow us to quantify the amount of accumulated damage correlative with wearing. During the propagating rupture, the strength of the fault‐core evolves, leading to higher wear generation along relatively strong zones or barriers. The total wear production in a given event is strongly dominated by the initial damage heterogeneity. Processing of synthetic seismograms shows excess energy radiation of high‐frequency seismic waves comparing to the expected radiation from planar faults. This radiation is enhanced with the increase in the variability of fault strength heterogeneity. Further calculations of the scaled energy show good fit with previous seismological and laboratory observations and demonstrate the impact of fault heterogeneity on radiated energy during dynamic rupture. Therefore, initial fault heterogeneity, manifested by fault‐core strength or by geometrical irregularity controls many aspects of earthquake rupture, including slip displacement and velocity. Plain Language Summary: Earthquakes are triggered when the stress build‐up on rocks exceeds the resistance to fracture of fault material. When faulting occurs, energy is generated in the fault zone and radiated through the rock body to the surface. The ground motions and their durations strongly depend on the earthquake rupture characteristics such as the amount of displacement, velocity, and directivity. Many studies modeled natural faults as planar frictional interfaces that failed and were displaced during earthquakes. Yet, geological observations and seismological measurements demonstrated that this is probably a naive picture, and fault zones are very heterogeneous, including complicated geometry and variation of strength along and across them. Here we examine the effect of such initial heterogeneities on earthquake and show that the heterogeneity of the rock strength in the fault core itself strongly affects earthquake dynamics. In particular, larger magnitude earthquakes are predicted for the more homogenous fault. We also find that higher fault heterogeneity would result in excess high‐frequency radiation. Considering these effects on ground motion around the fault, we show that seismological records may capture these differences in fault characteristics during earthquakes. Key Points: Initial damage level and its heterogeneity along the fault core control dynamics and seismic moment of simulated rupture eventsWear production is affected by the initial damage distribution. During slip, higher wear accumulation is calculated for barriers along faultsHigh variability of fault damage heterogeneity significantly enhances high‐frequency radiation [ABSTRACT FROM AUTHOR]

By providing unrivaled resolution in both time and space, volcano seismicity helps to chronicle and interpret eruptions. Standard earthquake detection methods are often insufficient as the eruption itself produces continuous seismic waves that obscure earthquake signals. We address this problem by developing an earthquake processing workflow specific to a high‐noise volcanic environment and applying it to the explosive 2008 Okmok Volcano eruption. This process includes applying single‐channel template matching combined with machine‐learning and fingerprint‐based techniques to expand the existing earthquake catalog of the eruption. We detected an order of magnitude more earthquakes, then located, relocated, determined locally calibrated magnitudes, and classified the events in the enhanced catalog. This new high‐resolution earthquake catalog increases the number of observations by about a factor of 10 and enables the detailed spatiotemporal seismic analysis during a large eruption. Plain Language Summary: Volcanic eruptions are some of the most challenging events to study seismically. The eruption itself continually excites seismic waves that hide the signatures of the accompanying earthquakes. The jetting of the plume, explosions, and impact of material on the ground all combine to create a blind spot in the seismological record. In this work, we develop a workflow that can overcome these limitations and reveal events that would otherwise remain hidden in the high‐noise, continuous seismic record. We utilize a combination of machine‐learning and related methods to increase the number of observed events by almost one order of magnitude. We determine the locations of the events using modern, high‐precision methods and use these locations with a customized, calibrated local magnitude scale for the region, which is necessary for volcanoes where seismic waves decay much faster than in the regions where standard magnitude scales were developed. We also automatically separate ordinary earthquakes from the more unusual long‐period ones and see an increase in detectability of both populations. The combined workflow now allows us to illuminate potentially the most interesting, and certainly most obscure, parts of the volcanic seismic cycle. Key Points: We develop a complete workflow suitable for co‐eruptive seismicity, which is inaccessible by normal seismic methodsWe combine different detection algorithms to increase the number of events of an explosive volcanic eruption by almost an order of magnitudeAutomated classification detects both long‐period and ordinary volcanic earthquakes [ABSTRACT FROM AUTHOR]

Volcanic inflation and deflation often precede eruptions and can lead to seismic velocity changes (dv/v $dv/v$) in the subsurface. Recently, interferometry on the coda of ambient noise‐cross‐correlation functions yielded encouraging results in detecting these changes at active volcanoes. Here, we analyze seismic data recorded at the Klyuchevskoy Volcanic Group in Kamchatka, Russia, between summer of 2015 and summer of 2016 to study signals related to volcanic activity. However, ubiquitous volcanic tremors introduce distortions in the noise wavefield that cause artifacts in the dv/v $dv/v$ estimates masking the impact of physical mechanisms. To avoid such instabilities, we propose a new technique called time‐segmented passive image interferometry. In this technique, we employ a hierarchical clustering algorithm to find periods in which the wavefield can be considered stationary. For these periods, we perform separate noise interferometry studies. To further increase the temporal resolution of our results, we use an AI‐driven approach to find stations with similar dv/v $dv/v$ responses and apply a spatial stack. The impacts of snow load and precipitation dominate the resulting dv/v $dv/v$ time series, as we demonstrate with the help of a simple model. In February 2016, we observe an abrupt velocity drop due to the M7.2 Zhupanov earthquake. Shortly after, we register a gradual velocity increase of about 0.3% at Bezymianny Volcano coinciding with surface deformation observed using remote sensing techniques. We suggest that the inflation of a shallow reservoir related to the beginning of Bezymianny's 2016/2017 eruptive cycle could have caused this local velocity increase and a decorrelation of the correlation function coda. Plain Language Summary: Before eruptions, volcanoes inflate due to the rising magma from below. Previous studies have found that these deformations can lead to small changes in the properties of the surrounding rock. We use passive image interferometry, a method that relies on the omnipresent background vibration of the Earth—mostly induced by the oceans, to measure these changes at the Klyuchevskoy Volcanic Group in Kamchatka, Russia. However, in Kamchatka, this background noise is masked and distorted by small earthquakes and tremors originating from the volcanoes themselves. We combine machine learning techniques with established monitoring methods to find times when these tremors remain similar. Afterward, we use data from these time periods in the conventional way to observe changes in the soil and the rock. Our results show that rain‐ and snowfall and the thickness of the snow cover exert the strongest influence on the properties of the rocks. Additionally, we found that a large magnitude 7.2 earthquake, which struck Kamchatka during our study, caused a slight weakening of the rocks due to microstructural damage. We register changes shortly before an eruption and suggest a connection to the beginning of an eruptive cycle in 2016. Key Points: Fluctuating noise conditions lead to distortions in noise interferometry studies, which we avoid with the help of machine learningThe seismic velocity on Kamchatka is affected by numerous mechanisms, amongst them environmental, tectonic, and volcanic eventsWe observe a velocity increase at Bezymianny during February 2016 and link it to the beginning of the eruptive cycle [ABSTRACT FROM AUTHOR]

Hjalli‐Ölfus is the westernmost segment of the east‐west transform South Iceland Seismic Zone (SISZ), which is the eastward extension of the ∼ENE‐trending transtensional Reykjanes Peninsula (RP). Historically, the area has shown an interactive behavior with the Hengill volcanic system to the north and the central SISZ to the east. We analyzed the state of stress and faulting mechanisms in Hjalli‐Ölfus between July 1991 and December 1999, in connection with the Hengill inflation episode (Feigl et al., 2000, https://doi.org/10.1029/2000JB900209) and the 13th November 1998 Mw 5.1 Hjalli‐Ölfus earthquake. We find that this region predominantly hosted oblique‐normal and left‐lateral strike‐slip events (4–10 km‐depth), with most nodal planes oriented along ∼ENE or ∼WSW directions (75° ± 15° or 255° ± 15°). We identify 5 stages of stress evolution from January 1991 to December 1999 over which Hjalli‐Ölfus experiences both spatial and temporal shifts in stress‐states. The Hengill inflation likely loaded both the fissure zone and western Ölfus, culminating in the Mw 5.4 (Hengill) and Mw 5.1 (Hjalli‐Ölfus) earthquakes. Following these events, the maximum compressive stress (SHmax) orientation near the location of the Mw 5.1 earthquake showed a ∼5°–7° counterclockwise swing, compared to SHmax before June 1998. The average SHmax (∼40° ± 1°) and minimum principal stress (σ3 ${\sigma }_{3}$ ∼ 130° ± 1°) are comparable to geological trends in the RP. We conclude that Hjalli‐Ölfus shows clockwise SHmax rotation upon loading, while a stress‐drop reverses the rotation. We also posit that the region, especially the western end, behaves like the RP during interseismic periods. Plain Language Summary: Iceland hosts a predominantly rifting plate boundary that is offset by two ∼east‐west trending, horizontally sliding (transform‐faulting) segments, one in the north and the other in the south. The southern segment, known as the South Iceland Seismic Zone (SISZ), is seismically productive and is flanked by diverging volcanic arms. The SISZ has hosted several moderate to large earthquakes on north‐south faults that cut across it. However, the westernmost end of the SISZ, also known as Hjalli‐Ölfus, differs from the rest of the SISZ as it seems to host earthquakes in an ∼east‐northeast direction, similar to its western, rifting neighbor, namely, the Reykjanes Peninsula. The activity in Hjalli‐Ölfus also seems to be responsive to volcanic/magmatic activity in the Hengill volcano to its north. This suggests the existence of multiple stress fields—volcanic, tectonic, or both—acting on the Hjalli‐Ölfus segment. Here, we study earthquakes in Hjalli‐Ölfus from January 1991 to December 1999, including a magnitude 5.1 earthquake in November 1998, to identify possible stress changes along the segment over time. Results indicate that magmatic deformation and seismic activity near the Hengill volcano directly influence the seismic productivity of Hjalli‐Ölfus, and that this seismicity is similar to that of the Reykjanes Peninsula. Key Points: Hjalli‐Ölfus predominantly hosts oblique‐normal and left‐lateral strike‐slip earthquakes on faults that orient in ∼ENE‐WSW directionsStress‐changes due to the Hengill inflation and seismicity are reflected as rotations in the maximum horizontal stress (SHmax) orientationsThe state of stress in Hjalli‐Ölfus mimics that of the Reykjanes Peninsula during interseismic periods [ABSTRACT FROM AUTHOR]

We use a dense seismic array of 1,108 vertical‐component geophones within a 600‐m footprint to detect thousands of small earthquakes near an active strand of the San Jacinto fault zone in southern California during a 26‐day period. We first correct site effects using multichannel cross‐correlations of the P‐waves of 256 cataloged earthquakes, and then perform beamforming analysis on the continuous waveforms in a slowness range from −0.4 to 0.4 s/km in both the east and north directions. At each time step, we identify the beam slowness with maximum amplitude and apply a picking algorithm to identify 13,408 events. These detections include over 55.6% of the events in the Quake Template Matching (QTM) catalog for all of southern California during the same time period and 70% of those within 100 km of the array. In addition, we detect over 10,000 new events, not in the QTM catalog. Many of these events can also be seen in records from nearby borehole seismic stations. Measured slownesses for the catalog and newly‐discovered events group into clusters that can be associated with QTM earthquake locations, but with slowness values considerably distorted from predictions based on a 1‐D velocity model, presumably owing to strong velocity heterogeneity near the San Jacinto Fault. Amplitudes of the detected events obey a Gutenberg‐Richter distribution with a b‐value close to one. Foreshocks are common among these detected events, increasing in rate before mainshocks following an inverse Omori's law. Plain Language Summary: We show that an array of over 1,000 seismometers located on a small patch of the San Jacinto Fault in southern California can detect many times more earthquakes than those in standard earthquake catalogs. Although accurate locations cannot be obtained for most of these newly identified events, their amplitudes indicate they obey similar scaling with size as larger magnitude earthquakes. Many of the larger events are preceded by foreshocks and the foreshock rate increases in the time leading up to the mainshocks. Our results suggest that future deployments of small‐aperture seismometer arrays could be used to complement existing seismic networks to probe earthquake activity in great detail. Key Points: We detect over 13,000 seismic events during 26 days in 2014 using a dense nodal array on the San Jacinto fault zone in southern CaliforniaMost of the detections are newly discovered earthquakes not in existing catalogsForeshocks are common among these earthquakes, increasing in the rate before their mainshocks following an inverse Omori's law [ABSTRACT FROM AUTHOR]

High‐velocity friction experiments on simulated fault gouges sheared at high normal stress and to low displacement are particularly relevant to induced seismicity, which is becoming an important topic in fault mechanics. Using a new, improved set‐up, which allows simulation of fault stress and fluid pressure (P $P$) conditions approaching in‐situ reservoir values, we performed ring‐shear experiments on simulated fault gouges prepared from the source‐, reservoir‐, and caprock‐formations of the Groningen gas field. Pre‐sheared gouges were subjected to a rotational slip pulse reaching ∼1.0 m/s peak velocity and 13–16 cm total displacement at effective normal stresses (σne ${\sigma }_{n}^{e}$) of 5–31 MPa and P $P$ up to 5 MPa, using water or dry nitrogen as pore fluid. All water‐saturated gouges show strong dynamic weakening within a few cm of slip, with the lowest dynamic friction (0.2–0.4) measured at the highest σne ${\sigma }_{n}^{e}$. By contrast, the weakening was subtle in experiments using nitrogen. Our analyses focus on the high‐P $P$ experiments, which are more realistic and show a distinct dependence of constitutive parameters (e.g., slip‐weakening rate) on σne ${\sigma }_{n}^{e}$, in the form of empirical linear, power‐law or exponential relations. The results provide much‐needed constraints for numerical modeling of induced rupture propagation in the Groningen field. Based on temperature‐ and P $P$‐measurements made in near‐direct contact with the active shear band, and using "post‐mortem" microstructures, we exclude previously‐proposed dynamic weakening mechanisms (e.g., flash heating or thermal pressurization) and suggest that water pressurization at heated asperity or grain contacts explains the weakening seen in our high‐P $P$ experiments. Plain Language Summary: In the Groningen gas field of the Netherlands, induced seismic events have occurred since the 1990s. The strongest event occurred in 2012 with a magnitude of MW 3.6. How faults slip during such small magnitude earthquakes (i.e., magnitude 3–4) is not clear, despite several recent studies. We conduct laboratory experiments to simulate the pulse‐like (i.e., short but rapid) fault slip behavior characteristic for small magnitude human‐induced seismicity, using a newly designed experimental set‐up enabling monitoring of the slip behavior under stress and displacement closer to conditions relevant to faults in Groningen. We find that the different rock types present in the Groningen field offer various levels of resistance to rupture of small earthquakes, which is crucial information for modeling earthquakes and understanding seismic hazard and risk. In addition, based on observations of fluid pressure, temperature and dilatation during the experiments, we propose a fault weakening mechanism occurring on the micrometer scale, that can explain our observations. Key Points: Wet faults in Groningen gas field lithologies weaken significantly in a short seismic slip pulse while dry ones do notWe observed fluid depressurization and dilation during seismic slip pulse, which is opposite to what is expected from thermal pressurizationSmall earthquake rupture experiences different resistance and slip weakening behaviors within different lithologies in the Groningen field [ABSTRACT FROM AUTHOR]

We use continuous and survey GNSS data to image the spatial and temporal evolution of afterslip during the 2 months following the Mw 8.3 2015 Illapel earthquake. Our approach solves for the incremental daily slip at the subduction interface using nonnegative least squares with spatial and temporal Laplacian regularization constraints. We find that afterslip developed at three specific areas at the megathrust, surrounding the coseismic rupture. In addition, well resolved afterslip also occurs within the coseismic rupture area that experienced ∼4 m of seismic slip. Our afterslip model shows striking correlations with the spatial distribution of aftershocks and repeating earthquakes. We capture the local afterslip triggered by a Mw 6.8 and two 6.9 aftershocks that ruptured downdip and north of the coseismic rupture, respectively. The latter ones were possibly triggered by the afterslip that developed north of the rupture. We also find a pulse of enhanced aseismic slip lasting a few days south of the rupture that spatially and temporally correlates with a seismicity burst. We finally find that areas of enhanced afterslip spatially correlates with areas having experienced regular seismic swarms observed during the years prior to the Illapel earthquake. This correlation supports the view of localized fluid high pore pressure areas behaving aseismically and surrounding a highly locked asperity, preventing the seismic rupture to propagate into them. Plain Language Summary: After a large earthquake, aseismic slip occurs in the vicinity of the seismic rupture. In this study, we focus on the aseismic slip that occurred after a large subduction earthquake of magnitude 8.3 in 2015 in central Chile. Using GPS time series, we obtain a spatiotemporal view of the aseismic slip during the 2 months following the earthquake. We find that aseismic slip occurred at three preferential areas located at the periphery of the rupture, as observed for many earthquakes. However, we also identify a patch of aseismic slip within the rupture area. We could image the daily afterslip rate and compare it to the daily seismicity and found striking correlations with the spatial distribution of aftershocks. In particular, we find that the largest aftershocks with magnitude 6.8–6.9 are tightly related to the evolution of afterslip. This result suggests that precisely monitoring the evolution of afterslip after a large earthquake might help to define preferential areas of large aftershocks. We also find a potential slow slip event south of the mainshock rupture which correlates with a local increase of seismicity. Finally, we find that the areas of enhanced afterslip had experienced seismic swarms decades before the earthquake, possibly reflecting areas of the fault with high fluid pore pressure. These results highlight new aspects about the dynamics of afterslip. Key Points: A full time‐dependent inversion provides a detailed view of afterslip during the 2.5 months following the Illapel earthquake in ChileAfterslip developed at three preferential areas, one of which overlaps a region of significant coseismic slip (2–4 m)Afterslip shows local and short‐term acceleration after large aftershocks and a slow slip event [ABSTRACT FROM AUTHOR]

We conduct a detailed study of the foreshock sequence preceding the 2010 Mw 6.7 Yushu, Qinghai earthquake in the Tibetan plateau by examining continuous waveforms recorded at a seismic station near the mainshock rupture zone. By using a deep learning phase picker—EQTransformer and a matched‐filter technique, we identify 120 foreshocks with magnitude ranging from −0.7 to 1.6, starting with a Mw 4.6 foreshock approximately 2 hr before the Mw 6.7 Yushu mainshock. Our analyses show that the foreshock sequence follows a typical Omori's law decay with a p‐value of 0.73 and the Gutenberg‐Richer frequency‐magnitude b‐value of 0.66. We do not find any evidence of accelerating events leading up to the Yushu mainshock. Hence, they could be considered as aftershocks of the Mw 4.6 earthquake. We further invert for the focal mechanisms and rupture directions for both the largest foreshock and the mainshock. The Mw 4.6 foreshock likely occurred on a NE‐SW trending fault conjugating to the NW‐SE trending fault of the mainshock. Coulomb stress analysis suggests the Mw 4.6 foreshock induces negative stress on the mainshock source area. These observations do not support either the pre‐slip or the cascade triggering model for foreshock generation. The occurrence of the foreshock, mainshock and large aftershocks appear to be modulated by the Earth's tidal forces, likely reflecting the role of high pore‐fluid pressures. Our observations, together with other recent studies, suggest that extensional step‐overs and conjugate faults along major strike‐slip faults play an important role in generating short‐term foreshock sequences. Plain Language Summary: We use both deep learning and matched‐filter methods to detect foreshocks that occur 2 hr before the magnitude 6.7 Yushu, Qinghai earthquake in Western China on 13 April 2010. The foreshock sequence begins with a magnitude 4.6 foreshock and the activity gradually decays with time. We do not observe an increase of magnitude or number of the events right before the mainshock. Therefore we interpret this as a typical aftershock sequence of the magnitude 4.6 earthquake. We also resolve the fault geometry and rupture direction. The results suggest the magnitude 4.6 earthquake ruptures along a fault that is nearly perpendicular to the fault broken during the magnitude 6.7 mainshock. Due to the fault geometry and relative location, the magnitude 4.6 foreshock might have inhibited the occurrence of the mainshock by reducing the stress on the fault. These observations do not support either the pre‐slip or cascade triggering models for foreshock generation. We calculate the theoretical tidal stress and strain in the study area and observe a correlation of earthquake activity and the tidal stress change. This correlation might provide evidence for fluid driven process. In conclusion our research shows complex geometry along major strike‐slip faults can influence foreshock behaviors. Key Points: We identified missing foreshocks of the 2010 Mw 6.7 Yushu, Qinghai earthquake by using deep learning and matched‐filter techniquesThe foreshock sequence can be considered as aftershocks of Mw 4.6 foreshock given no evidence of an acceleration leading to the mainshockThe Mw 4.6 foreshock likely occurred on an extensional step‐over which induced negative stress change at the Yushu mainshock hypocenter [ABSTRACT FROM AUTHOR]

We use 12 temporary and 9 permanent broadband seismometers that were operating for ∼400 days from October 2018 to November 2019 to generate the first published earthquake catalog and local magnitude function for southern New Caledonia (SNC). Local hypocenters mostly have depths <20 km, but east of the New Hebrides‐Vanuatu subduction zone they deepen to >100 km. Our local magnitude estimates ML for 107 earthquakes in the subduction zone are consistently 1.1 units smaller than Mw and mb over a range of Mw from 4.5 to 7.5, as determined by the United States Geological Survey. Our catalog has 460 earthquakes with Mw≥3.7 in the subduction zone and the largest event in SNC has ML 3.8. Seismicity rates in SNC are low, but ML> 5 earthquakes are 2.7 times more frequent than elsewhere in the northern Australian plate. The probability of an ML> 5 event in 50 years is 0.6 in SNC. The hazard of damaging seismic shaking in SNC is dominated by local moderate‐magnitude earthquakes, rather than large‐magnitude subduction events. The predicted peak ground acceleration for Nouméa at 10% probability of exceedance in 50 years is 0.08 g. Residual analysis of ground accelerations demonstrates that the hazard for Nouméa from subduction zone earthquakes is currently overestimated and that new regionally specific ground motion prediction equations (GMPEs) are needed. Our results highlight the inadequacy of current GMPEs in subduction zone footwall settings and the need for additional studies of this type of setting. Plain Language Summary: We used seismometers from a local network near Nouméa, which is the capital city of New Caledonia, to record earthquakes for 14 months. About 13% of the earthquakes were shallow crustal events (<20 km depth) that clustered near our network. The others clustered ∼250 km away in the New Hebrides‐Vanuatu subduction zone (NHV), where large earthquakes (recorded on global networks) regularly occur. Earthquakes beneath southern New Caledonia (SNC) are smaller and less frequent than in nearby Vanuatu, but they are more hazardous for local communities because they are at shallow depths and nearby. Our work demonstrates the necessity of long‐term monitoring of local earthquakes in SNC to assess hazard there. Our analysis shows that current models used to predict shaking from large earthquakes in the NHV over‐estimate hazard in New Caledonia, and this may have implications for other regions in similar tectonic settings. Key Points: We use 14 months of data from a local network to determine location and magnitude of 1,374 earthquakes near southern New CaledoniaGround motions in New Caledonia, which is in a subduction footwall, are smaller than expected from ground motion prediction equationsEarthquakes in New Caledonia are smaller and less frequent than in the nearby subduction zone, but induce greater local shaking hazard [ABSTRACT FROM AUTHOR]

Earthquake surface deformation provides key constraints on the geometry, kinematics, and displacements of fault rupture. However, deriving these characteristics from past earthquakes is complicated by insufficient knowledge of the pre‐event landscape and its post‐event modification. The 1987 Mw 6.5 Edgecumbe earthquake in the northern Taupō volcanic zone (TVZ) in New Zealand represents a moderate‐magnitude earthquake with distributed surface rupture that occurred before widespread high‐resolution topographic data were available. We use historical aerial photos to build pre‐ and post‐earthquake digital surface models (DSMs) using structure‐from‐motion techniques. We measured discrete and distributed deformation from differenced DSMs and compared the effectiveness of the technique to traditional field‐ and lidar‐based studies. We identified most fault traces recognized by 1987 field mapping, mapped newly identified traces, and made dense remote slip measurements with a vertical separation resolution of ∼0.3 m. Our maximum and average vertical separation measurements on the Edgecumbe fault trace (2.5 ± 0.3 and 1.2 m, respectively), are similar to field‐based values of 2.4 and 1.1 m, respectively. Importantly, this technique can discern between new and pre‐existing fault scarps better than field techniques or post‐earthquake lidar‐based measurements alone. Our surface displacement results are used to refine subsurface fault geometries and slip distributions at depth, which are further used to investigate potential magmatic‐tectonic stress interactions in the northern TVZ. Our results suggest the Edgecumbe fault dips more gently at depth than at the surface, hosted shallow slip in 1987, and may be advanced toward failure by interactions with nearby magma bodies. Plain Language Summary: Understanding earthquake behavior relies heavily on information about how past earthquakes changed the landscape. Detailed information about the surface topography before the earthquake is often limited, creating challenges for accurately measuring earthquake surface deformation. Aerial photos are widely available and can be used to create 3D digital surface models in places where other topographic information is lacking. We use historical aerial photos to make 3D datasets of the surface topography before and after the 1987 Edgecumbe earthquake in New Zealand. We calculated the difference between the two 3D datasets to identify and measure how the earthquake changed the landscape, and compared our results to previous measurements. We found that this method generally works as well as field methods for identifying and measuring fault movement, and has some advantages over other techniques. In particular, this approach can separate deformation from individual earthquakes, which had previously been a challenge. The results refine our understanding of the faults below the surface, provide insight into the earthquake's relationship to surrounding faults and volcanic systems, and allow us to better characterize seismic hazard both here and in other similar geologic settings. Key Points: Historical aerial photos can be used to generate pre‐ and post‐earthquake digital surface models for measuring displacementThis method performs at least as well as field surveys, and better captures single‐event displacements to constrain earthquake behaviorThe Edgecumbe earthquake is important for understanding low dip‐angle normal fault ruptures and possibly magma‐tectonic interactions [ABSTRACT FROM AUTHOR]

Strong earthquakes with larger magnitude and longer durations trigger many landslides, however, how magnitude and duration affect landslides is still unclear. Many factors could contribute to this, including additional shear stress provided by strong ground motion, or "seismogenic liquefaction"; herein, we hypothesize that the dynamic weakening of sliding zone gouge is important. We explored the influence of earthquake magnitude and duration on landslide triggering by simulating the seismic response of sliding zone gouge using a dynamic ring‐shear device and glass spheres. The experiments showed that vibration with larger amplitudes and longer durations more easily trigger deformation and even instability in dry granular materials. We used a dynamic triaxial‐bender system to find that the shear modulus of these materials decreased with the increase in duration and amplitude of cyclic loading. We suggest that this universal decrease in shear modulus is an important landslide‐trigger mechanism. Our results revealed how magnitude and duration of earthquakes affect co‐seismic landslides and why earthquakes with larger magnitude and long durations can trigger more co‐seismic landslides. Plain Language Summary: Strong earthquakes with larger magnitudes and longer durations trigger many landslides. Some mechanisms have been proposed to explain the triggering, for example, seismic waves provide additional driving force which increases the shear stress on the sliding plane. Earthquakes may cause an increase of pore pressure, which reduces the effective normal stress on the sliding plane. However, the mechanisms that apply to wet and dry conditions are not fully understood. We hypothesize that the dynamic weakening of sliding zone gouge caused by earthquakes is an important mechanism. We used a dynamic ring shear device and glass spheres to simulate the seismic response of sliding zone gouge, and our experiments show that larger amplitudes and longer durations more easily trigger deformation and even instability. Using a dynamic triaxial‐bender system, we found that the shear modulus decreased with the increase in duration and amplitude of vibration. Our results reveal how earthquakes with larger magnitudes and long durations trigger more co‐seismic landslides. Our experimental method may be helpful for monitoring and early warning of earthquake‐triggered landslides in the future. Key Points: Dynamic ring shear tests show the deformation characteristics of granular materials under vibrationLonger durations and larger amplitudes of vibration are more likely to trigger slip and shear modulus weakeningThe reduction and recovery of shear modulus is an important mechanism for triggering failure of earthquake‐triggered landslides [ABSTRACT FROM AUTHOR]

Accurate earthquake location and magnitude estimation play critical roles in seismology. Recent deep learning frameworks have produced encouraging results on various seismological tasks (e.g., earthquake detection, phase picking, seismic classification, and earthquake early warning). Many existing machine learning earthquake location methods utilize waveform information from a single station. However, multiple stations contain more complete information for earthquake source characterization. Inspired by recent successes in applying graph neural networks (GNNs) in graph‐structured data, we develop a Spatiotemporal Graph Neural Network (STGNN) for estimating earthquake locations and magnitudes. Our graph neural network leverages geographical and waveform information from multiple stations to construct graphs automatically and dynamically by adaptive message passing based on graphs' edges. Using a recent graph neural network and a fully convolutional neural network as baselines, we apply STGNN to earthquakes recorded by the Southern California Seismic Network from 2000 to 2019 and earthquakes collected in Oklahoma from 2014 to 2015. STGNN yields more accurate earthquake locations than those obtained by the baseline models and performs comparably in terms of depth and magnitude prediction, though the ability to predict depth and magnitude remains weak for all tested models. Our work demonstrates the potential of using GNNs and multiple stations for better automatic estimation of earthquake epicenters. Plain Language Summary: The location and magnitude of an earthquake can be best determined by relating the motion recorded at multiple stations in a network, and it would be beneficial to combine the waveforms from multiple seismic stations for source characterization. Machine learning‐based approaches have recently become prevalent in seismological tasks such as earthquake source characterization. Because of the irregular spatial distribution of seismic stations, graph convolutional neural networks (a deep learning architecture which handles graph‐structured data) have great potential in combining both spatial and temporal information from different seismic stations. In this work, waveforms recorded at multiple stations are passed through neural networks with connective links based on the similarity of waveform features and geographic locations. The model is tested on two data sets and compared to two published baselines (a graph convolutional neural network and a fully convolutional network). Compared with the baselines, STGNN achieves improved accuracy for epicenter estimation and comparable accuracy for depth and magnitude estimation. Key Points: We design a multistation‐based graph neural network for earthquake source characterizationBoth geographic distance and waveform feature similarity are considered in a graph convolutional neural network for feature combinationOur network is capable of automatically selecting and combining relevant seismic stations to characterize earthquake source parameters [ABSTRACT FROM AUTHOR]

Multiple physical mechanisms have been proposed to explain the cause of intermediate‐depth and deep earthquakes, but they are still under debate. Source parameters such as stress drop, have the potential to provide insight into their physical mechanisms. We develop a modified spectral decomposition method to analyze 1‐year seismic data from temporary land‐based and ocean bottom seismographs in a complex subduction zone. By applying this method to investigate 1,083 intermediate‐depth and deep earthquakes in the Tonga slab, we successfully resolve the source spectra and stress drops of 743 MW 2.6–6.0 earthquakes. Although the absolute stress drops are subject to the choices of source model parameters, the relative stress drops are more reliably resolved. The median stress drop of Tonga earthquakes does not change with respect to magnitude but decreases with depth by 2–3 times in two separate depth ranges of 70–250 and 400–600 km, corresponding to intermediate‐depth and deep earthquakes, respectively. The median stress drops show spatial variations, with two high‐stress‐drop (five times higher than the ambient value) regions, coinciding with strong local deformation where the Tonga slab bends or tears. In the Tonga double seismic zone at 120–300 km depths, the median stress drop appears smaller in the lower plane than in the upper plane, suggesting a slower rupture velocity or a higher fluid content in the lower‐plane region. Our results suggest that intermediate‐depth and deep earthquakes in the Tonga slab generally follow the earthquake self‐similar model and favor the fluid‐related embrittlement hypothesis for both groups of earthquakes. Plain Language Summary: Although earthquakes deeper than 70 km are observed in many subduction zones and sometimes cause damage to human societies, their nature is puzzling. This is because under high temperature and pressure conditions at great depths, rocks are expected to deform in a ductile way, like silly putty, rather than in a brittle way, like concrete. Multiple physical mechanisms have been proposed to explain these deep earthquakes, but they are still under debate. Stress drop, the change of shear stress after an earthquake, may provide insights into the physical mechanisms of these earthquakes. In this study, we estimate the stress drops of hundreds of deep earthquakes beneath the Tonga Trench where the Pacific plate subducts underneath the Australian plate. We find that stress drops of these earthquakes do not change with respect to earthquake magnitudes but show depth‐dependence in two separate depth ranges (70–250 and 400–600 km). We also find abnormally high stress drops in two regions where the subducted Tonga slab undergoes strong deformation. At 120–300 km depths, earthquakes occur along with two separate parallel layers. The lower‐plane earthquakes have smaller stress drops than the ones in the upper plane, possibly implying water carried to great depths by subduction. Key Points: Earthquake stress drops show no magnitude dependence and slight depth dependenceHigh stress‐drop zones coincide with dramatic changes in the Tonga slab geometryIn the Tonga double seismic zone at 120–300 km, the lower‐plane earthquakes have smaller stress drops than the upper‐plane earthquakes [ABSTRACT FROM AUTHOR]

A refined model of fault structure in the active New Madrid Seismic Zone (NMSZ) is presented based on relocated hypocenters and application of a statistical clustering method to determine fault planes. Over 200 earthquakes are recorded in the NMSZ every year, but the three‐dimensional (3‐D) fault structure is difficult to determine because the zone is covered by thick, Mississippi Embayment sediment. The distribution of earthquakes in the NMSZ indicates four major arms of seismicity, suggesting the presence of a northeast‐southwest trending strike‐slip fault system with a major northwest trending, contractional stepover fault. Relocation of 4,131 earthquakes using HypoDD resulted in major improvement in the depiction of fault structure in the NMSZ including three‐dimensional structural variations along the along the Reelfoot fault (RF) and a well‐defined intersection of the Axial and RFs. Optimal Anisotropic Dynamic Clustering analysis of the relocated hypocenters produced a fault model consisting of 12, well resolved planes. The RF is continuous along strike from the northern end to the Ridgely fault, located south of the intersection with the Axial fault (AF). The crosscutting Ridgely fault may serve as a barrier to rupture propagation along the entire fault. The strike‐slip arms of seismicity are well resolved and correspond to near vertical planes. Three planes are resolved in the southern part of the AF and are associated with the Osceola igneous complex. Plain Language Summary: A refined fault model is determined for the active New Madrid seismic zone using earthquakes that occurred from 2000 to 2019. The seismic zone is located in the central United States and poses a hazard to critical infrastructure and numerous population centers. The major faults are the Reelfoot thrust fault and the strike‐slip Axial fault (AF). The first step in constructing the improved fault model involved relocation of the earthquakes to decrease location error. The second step involved fitting planes to the relocated earthquakes under the assumption that the earthquakes cluster along fault segments. The resulting model provides a three‐dimensional representation of the fault structure. A major fault disrupts the southern portion of the Reelfoot fault, making it discontinuous. The northern part of the AF is very well defined. Earthquakes are more scattered along the southern part of the AF, but three fault planes are resolved that are located along the side of a major intrusion. The fault model can be used in studies involving possible rupture length and magnitude of large earthquakes. Key Points: New Madrid Seismic Zone hypocenters for the years 2000–2019 are relocated and have a median relative location error of 65 mPlanes are fit to the relocated hypocenters producing a detailed three‐dimensional model of fault structureThe Reelfoot fault is segmented by the Ridgely fault [ABSTRACT FROM AUTHOR]

Paleoseismic evidence from the late‐interseismic Alpine Fault suggests key section boundaries conditionally inhibit rupture. We utilize a year of data from a two‐part seismometer network (Dense Westland Arrays Researching Fault Segmentation) to characterize ∼7,500 earthquakes (−0.7 ≤ MLv ≤ 4.2) and ∼800 focal mechanisms, producing high‐resolution structural images of these boundaries to study effects of material and structural heterogeneities on mode‐switching rupture behavior. Lithologically‐controlled frictional behavior and crustal strength appear to influence lateral and vertical on‐fault seismicity distributions. Ultramafic hanging‐wall serpentinite and serpentinite‐related fault core minerals along the South Westland (SW) boundary, result in a locally shallow seismogenic cuttoff (∼8 km) and abundant on‐fault seismicity. Maximum horizontal compressive stress rotations (14° anti‐clockwise and 20° clockwise near the SW and North Westland (NW) boundaries, respectively, relative to the Central Section), coupled with spatially variable fault frictional properties, are more important than geometry alone in controlling Sections' relative frictional stability. Whereas the SW and Central Sections are well‐oriented for failure, the NW Section is severely misoriented compared with favorably oriented faults of the Marlborough Fault Zone, which possibly facilitate a preferred rupture route. Geometrically, a 40° dip change at the SW boundary may be accommodated either by a single through‐going fault plane ‐ a difficult geometry across which to obtain multi‐segment earthquakes when considering rupture dynamics ‐ or by a deeper vertical fault strand truncated by a shallower listric plane. Our new observations have implications for Alpine Fault rupture scenarios and highlight the need to consider a range of spatially heterogeneous, interdependent physical factors when evaluating controls on rupture segmentation. Plain Language Summary: New Zealand's Alpine Fault last ruptured in 1717 AD and the likelihood of it producing a major earthquake in the next 50 years is estimated to be 75%. The magnitude of the next earthquake will be strongly determined by the area of fault rupture. Two key localities along the fault, which correspond to the boundaries between sections of the fault with differing geometry, slip rate and frictional behavior, have been proposed to sometimes, but not always, act as barriers to fault rupture, limiting the sizes of the resulting earthquakes. However, it remains unclear what physical properties control this process. Using seismic data collected from dense seismometer networks, we construct detailed catalogs of small earthquakes occurring near these localities, at Martyr River and Inchbonnie, and precisely map the distribution of active seismicity near the fault. From these data we improve estimates of (a) the subsurface geometry of the fault and how it changes along length, (b) how seismicity is distributed with depth, to understand possible maximum depth of rupture in conjunction with observations of crustal temperatures and structure, and (c) the orientation of the stresses acting on the fault, to understand its stability to tectonic loading and its frictional behavior. Key Points: Dense microseismicity catalogs constrain fault structure, seismogenic depths and the stress field bookending the fault's Central SectionOn‐fault seismicity, a 40° dip change and a shallow, rheologically controlled cutoff depth are observed at the South Westland boundaryStress analyses indicate South and North Westland Sections are favorably and unfavorably oriented for frictional failure, respectively [ABSTRACT FROM AUTHOR]

We investigate induced seismicity associated with a hydraulic stimulation campaign performed in 2020 in the 5.8 km deep geothermal OTN‐2 well near Helsinki, Finland as part of the St1 Deep Heat project. A total of 2,875 m3 of fresh water was injected during 16 days at well‐head pressures <70 MPa and with flow rates between 400 and 1,000 L/min. The seismicity was monitored using a high‐resolution seismic network composed of 10 borehole geophones surrounding the project site and a borehole array of 10 geophones located in adjacent OTN‐3 well. A total of 6,121 induced earthquakes with local magnitudes MLHel>−1.9 ${M}_{\mathrm{L}}^{\mathrm{H}\mathrm{e}\mathrm{l}} > -1.9$ were recorded during and after the stimulation campaign. The analyzed statistical parameters include magnitude‐frequency b‐value, interevent time and interevent time ratio, as well as magnitude correlations. We find that the b‐value remained stationary for the entire injection period suggesting limited stress build‐up or limited fracture network coalescence in the reservoir. The seismicity during the stimulation neither shows signatures of magnitude correlations, nor temporal clustering or anticlustering beyond those arising from varying injection rates. The interevent time statistics are characterized by a Poissonian time‐varying distribution. The calculated parameters indicate no earthquake interaction. Focal mechanisms suggest that the injection activated a spatially distributed network of similarly oriented fractures. The seismicity displays stable behavior with no signatures pointing toward a runaway event. The cumulative seismic moment is proportional to the cumulative hydraulic energy and the maximum magnitude is controlled by injection rate. The performed study provides a base for implementation of time‐dependent probabilistic seismic hazard assessment for the project site. Plain Language Summary: We investigate anthropogenic seismicity associated with fluid injection into the 5.8 km deep geothermal OTN‐2 well near Helsinki, Finland, as a part of St1 Deep Heat Project. A total of 2,875 m3 of fresh water was injected during 16 days at well‐head pressures <70 MPa and with flow rates between 400 and 1,000 L/min. The seismicity was monitored using a seismic network composed of 20 borehole geophones located in Helsinki area and in the OTN‐3 well located close by the injection site. A total of 6,121 earthquakes indicating fractures of 1–30 m size were recorded during and after stimulation campaign. Using a handful of statistical properties derived from earthquake catalog we found no indication for earthquakes being triggered by other earthquakes. Instead, the earthquake activity rates, as well as the maximum earthquake size stayed proportional to the fluid injection rate. The spatio‐temporal behavior of seismicity and its properties suggest earthquakes occurred not on a single fault, but in a distributed network of similarly oriented fractures, limiting the possibility for occurrence of violent earthquakes. The performed study provides evidence that the induced seismicity due to injection performed within St1 Deep Heat project is stable and allow to constrain seismic hazard. Key Points: Induced seismicity associated with stimulation campaign in a 5.8 km deep geothermal OTN‐2 well passively responds to injection operationsSeismicity is a non‐stationary Poisson process with seismicity rate and maximum magnitude modulated by the hydraulic energy input rateSeismicity clusters in space and time in response to fluid injection but no interaction between earthquakes is observed [ABSTRACT FROM AUTHOR]

Viscoelastic processes in the upper mantle redistribute seismically generated stresses and modulate crustal deformation throughout the earthquake cycle. Geodetic observations of these motions at the surface of the crust‐mantle system offer the possibility of constraining the rheology of the upper mantle. Parsimonious representations of viscoelastically modulated deformation through the aseismic phase of the earthquake cycle should simultaneously explain geodetic observations of (a) rapid postseismic deformation, (b) late in the earthquake cycle near‐fault strain localization. To understand how rheological formulations affect kinematics, we compare predictions from time‐dependent forward models of deformation over the entire earthquake cycle for an idealized vertical strike‐slip fault in a homogeneous elastic crust underlain by a homogeneous viscoelastic upper‐mantle. We explore three different rheologies as inferred from laboratory experiments: (a) linear Maxwell, (b) linear Burgers, (c) power‐law. The linear Burgers and power‐law rheologies are consistent with fast and slow deformation phenomenology over the entire earthquake cycle, while the single‐layer linear Maxwell model is not. The kinematic similarity of linear Burgers and power‐law models suggests that geodetic observations alone may be insufficient to distinguish between them, but indicate that one may serve as an effective proxy for the other. However, the power‐law rheology model displays a postseismic response that is non‐linearly dependent on earthquake magnitude, which may offer a partial explanation for observations of limited postseismic deformation near some magnitude 6.5–7.0 earthquakes. We discuss the role of mechanical coupling between frictional slip and viscous creep in controlling the time‐dependence of regional stress transfer following large earthquakes and how this may affect the seismic hazard and risk to communities living close to fault networks. Plain Language Summary: The solid Earth is a viscoelastic material that displays both solid and fluid‐like behaviors depending on the observational time window and the applied stress. We develop numerical simulations of how the uppermost solid Earth responds to a sequence of periodic earthquakes and the earthquake cycle. Our simulations test a range of proposed viscoelastic materials. The predicted surface displacements from each model are compared with observational features extracted from geodetic datasets compiled over the past few decades. All existing viscoelastic material descriptions can satisfactorily explain observational features in the first few years following an earthquake; significant differences between the viscoelastic models emerge 10–100 years following a large earthquake. Identifying the most appropriate viscoelastic description requires the integration of geodetic data that constrains the velocity evolution from a sequence of earthquakes (as opposed to a single event) with observations from rock physics laboratory experiments. A unified description of visceolasticity in the uppermost solid earth has important implications for understanding stress evolution in fault networks, and improving models of seismic hazard. Key Points: We develop a kinematically consistent periodic earthquake cycle model with mechanical coupling between frictional slip and viscous creepSteady‐state power‐law and linear Burgers rheologies are consistent with earthquake cycle observations while linear Maxwell is notDistributed viscous creep alters the (stress) loading rate of a large volume around the fault and is important for multifault interactions [ABSTRACT FROM AUTHOR]

Lateral variation in seismic attenuation of the Jammu and Kashmir Himalaya has been studied from local‐earthquake coda waves using 588 single‐trace measurements and back‐projection tomography of coda quality factor (Qc(f)). Single‐trace measurements are performed for a range of lapse time and coda‐window length, and a final choice of 2ts and 90 s, respectively, have been made based on an unbiased distribution of Q0 as a function of epicentral distance. These choices ensure coda start time beyond transient arrivals and satisfies the assumption of coda‐wave generation through multiple forward‐scattering. From single‐traces we obtain Qc(f) = 231 ± 21f0.93±0.05, with a negative correlation between Q0 and η. Qc(f) tomography are presented as 2D‐anomaly maps computed from Qmean at every f, and as absolute Q0 and η maps. These reveal strong frequency‐dependent lateral variations in attenuation. At f = 3–5 Hz, the central Shiwalik Himalaya, Jammu‐Kishtwar Higher Himalaya, Zanskar Ranges, and eastern Kashmir Valley have ∼10% positive Qc anomaly; and high Q0 (270–300) and low η (0.8–0.85), indicating low absorption of seismic energy. The Kishtwar Window region, the reentrant of the Main Boundary and Main Central thrusts, and the western Kashmir Valley have ∼10% negative Qc anomaly, low‐to‐intermediate Q0 (210–250) and high η (0.9–1.0), revealing strong attenuation at structural heterogeneities. At f = 8–14 Hz, the lateral variation in Qc anomaly between the western Kashmir Himalaya (positive) and eastern Jammu‐Kishtwar Himalaya (negative), possibly arises from the crustal thickness variations and steep‐to‐shallow NE dipping Moho from east‐to‐west. Plain Language Summary: Jammu and Kashmir (J&K) Himalaya lies between two devastating earthquakes of Kangra (1905) and Kashmir (2005), and is known as a "seismic gap," where the last major earthquake occurred in 1555 (magnitude ∼8). This region has accumulated sufficient strain energy to drive a future earthquake of the same magnitude, which poses unprecedented risk to J&K and the densely populated Indo‐Gangetic plains. To assess the risk from such future earthquakes it is critical to quantify seismic hazard, which requires estimation of ground shaking from future earthquakes. This in turn depends on the propagation characteristics (attenuation) of earthquake waves through the Earth. In this study, we quantify the seismic attenuation in the J&K Himalaya using the later part of local earthquake waveforms (coda waves). Our study reveals strong frequency‐dependent seismic attenuation with spatial variations. At lower frequencies, the Shiwalik Himalaya, Jammu‐Kishtwar Higher Himalaya, and the eastern Kashmir Valley have the most efficient propagation of seismic energy through crystalline rocks. On the other hand the Kishtwar Window region, Lesser Himalaya, and the western Kashmir Valley have inefficient propagation of seismic energy due to structural heterogeneities, where seismic energy is strongly absorbed. At higher frequencies, the east‐to‐west variation of the Qc anomaly from negative‐to‐positive is a function of the subsurface structure and geometry of the boundaries. Key Points: A seismic coda‐wave attenuation model of Qc = 231 ± 21 f0.93±0.05 is obtained for the Jammu and Kashmir HimalayaPositive Qc anomalies are observed at low frequencies in the Shiwalik Himalaya, Jammu‐Kishtwar Higher Himalaya, and Zanskar RangesKashmir Valley is characterized by E–W variation of positive‐to‐negative Qc anomaly, correlates with laterally varying crustal structure [ABSTRACT FROM AUTHOR]

The 2021 Mw 6.1 Yangbi earthquake in southwest China is preceded by three major foreshocks: 05/18 Mw 4.3, 05/19 Mw 4.6, and 05/21 Mw 5.2. It provides a valuable chance to revisit two end‐member foreshock models: cascade‐up and pre‐slip models. We first determine the fault structure with relocated aftershocks and the focal mechanisms obtained from a multipoint‐source inversion. We find that the mainshock and two smaller foreshocks occur on an unmapped near‐vertical fault, and the largest foreshock occurs on a mapped stepover fault that dips to northeast. Second, for each major foreshock, we estimate and delineate the rupture area based on its aftershock location and a spectral ratio analysis. Based on the inferred rupture model, we finally calculate the evolution of Coulomb stress and examine potential interactions between each major event. Results show that the Yangbi sequence can be explained by the cascade triggering mechanism, while we also find evidence for aseismic slip that contributes to the triggering process: the first foreshock is preceded by a short‐term localized cluster, three repeater sequences are detected in the foreshock period, and the aftershock zones expand logarithmically with time. The Yangbi mainshock is probably triggered by multiple major foreshocks through both seismic and aseismic processes. This detailed seismological characterization of Yangbi sequence lends supports for an improved understanding on the foreshock mechanism: (a) the controlling mechanisms are not limited to cascade‐up and pre‐slip. Instead, multiple mechanisms can operate together; and (b) aseismic slip does not always provide more deterministic information on the mainshock. Plain Language Summary: The 2021 magnitude 6.1 Yangbi earthquake in southwest China is preceded by three magnitude 4–5 earthquakes, which are known as foreshocks. Whether the foreshock sequence can be used to forecast future occurrence of large earthquakes is of great scientific interest. Here, we analyze the interaction between these foreshocks with a high‐resolution seismic catalog and the modeling of stress transfer after each major foreshock. We find that the migration directions of foreshocks are not consistent, and that the stress transfer following each foreshock is large enough to explain the triggering of the Yangbi mainshock. This supports the cascade triggering model for foreshock generation, where the foreshocks are not precursors. However, we also find several lines of evidence for aseismic fault slip. These include: the first foreshock is preceded by a short‐term localized cluster, three repeater sequences are detected in the foreshock sequence, and the aftershock zones of three foreshocks expand logarithmically with time, suggesting that they are driven by afterslip. Our results suggest that multiple mechanisms (e.g., cascade‐triggering and aseismic slip) can act at the same time to drive foreshocks, and it is relatively challenging to use foreshocks to forecast future occurrence of large earthquakes. Key Points: The Yangbi mainshock and two smaller foreshocks occur on an unmapped near‐vertical fault, and the largest foreshock occurs on a mapped stepover fault that dips to NEThe rupture directivity and source parameters of major foreshocks are estimated by aftershock distribution and spectral ratio analysisThe Yangbi sequence can be explained as a cascade sequence, but aseismic processes can also be inferred, including pre‐slip cluster and logarithmic expansion of aftershocks [ABSTRACT FROM AUTHOR]

Dense arrays deployed near the hydraulic fracturing (HF) wells greatly enhance the understanding of injection‐induced seismicity. In this study, we revisit the continuous recordings that are acquired by 69 three‐component nodes at an HF site in Alberta, Canada, taking advantage of a machine learning‐based seismic detection and location workflow. The obtained new earthquake catalog contains 21,619 events with relative location errors of <1 m, which exceeds ∼20% of the number of earthquakes (18,040) reported previously (Igonin et al., 2021). This high‐resolution catalog reveals the distribution of earthquake sequences at much improved spatiotemporal resolution and illustrates several previously unmapped faults/fractures. Earthquake frequency‐magnitude distribution reveals that the average b value increases with depth from ∼1.1 above 3.5 km to ∼2.5 at greater depths. Further spatial analysis of seismic clusters indicates that the b value varies laterally (∼1–1.7) at shallow depths and is inversely related to the proximity to injection wells in conjunction with the change in structural types (i.e., reactivated fault and pre‐existing fracture). The seismic sequence on the north‐south oriented faults also shows a distinctive occurrence pattern and temporal affinity to the fracturing network reactivation between the two stages of HF operations. The change in faulting behavior could reflect a shift of dominating triggering mechanisms and physical processes from (a) the rapid diffusion of pore fluid pressure along pre‐existing fracture corridors to (b) the cascade migration of earthquake sequences in response to the cumulative Coulomb stress perturbation on the fluid‐lubricated, critically stressed faults. Plain Language Summary: The potential for fluid‐related industrial activities (e.g., hydraulic fracturing, HF) to induce felt earthquakes has long been recognized. In the past few years, a number of felt events (i.e., M ∼ 4) have occurred in western Canada, which are often linked to HF operations (e.g., Kao et al., 2018, https://doi.org/10.1190/tle37020117.1; Mahani et al., 2019, https://doi.org/10.1785/0220190040). To better understand processes that induce earthquakes, the Tony Creek dual Microseismic Experiment (ToC2ME) deployed a dense seismic array around an HF well in late 2016 (26 October to 30 November). In this study, we utilize a recently developed earthquake detection and location workflow to revisit the ToC2ME data set and build a high‐resolution earthquake catalog. We observe over 20,000 earthquakes with magnitudes down to −2 with high location accuracy at meter scale. The improved earthquake locations allow us to better delineate the fractures and faults activated during the HF. Our results further reveal the complexity of the triggering mechanism of HF‐induced earthquakes according to the statistical analysis of the frequency‐magnitude distribution of earthquakes. Our study suggests that multiple factors, including pore fluid pressure due to fluid injection and perturbation in the stress field caused by earthquakes, can jointly control the spatiotemporal variations of regional seismicity during multi‐stage HF operations. Key Points: We revisited the Tony Creek dual Microseismic Experiment data set and built a high‐resolution catalog with a new earthquake detection and location workflowThe 21,619 events revealed spatially varying b values associated with changes in structural type and pore‐fluid pressureMultiple triggering mechanisms may control earthquake occurrence during multi‐stage hydraulic fracturing operations [ABSTRACT FROM AUTHOR]

Absolute tectonic stress estimation is of fundamental importance to investigating regional earthquake hazards, while traditional algorithms only invert relative stress and its orientations. We use topography data and rupture models to invert the absolute tectonic stress in the source region of the 2016 Mw 7.0 Kumamoto earthquake. We adopt a basic assumption that the slip direction of each sub‐fault is in the shear direction of the total stress tensor, which is a combination of tectonic and topographic stresses. We use the Bayesian theorem to describe the probability density function of stress parameters and perform a grid search for the optimal values and uncertainties of stress parameters. Our synthetic test shows that the true parameters can be reliably estimated over a broad range of tectonic stress magnitudes. In the application to the Kumamoto earthquake, we find that the optimal tectonic stress is uniaxial tensional with a magnitude of 35 MPa and orientation of N6.5°W, which is at an angle of ∼47.5° to the fault strike. The average total shear stress on the fault plane is 16 MPa on the fault plane. With the inverted stress parameters, we evaluate the pore pressure based on the absolute stress field, the result of which is within a range of 0.3–0.45 of the lithostatic stress. Plain Language Summary: We estimate the absolute stress of the 2016 Kumamoto earthquake source region using rupture models and ground topography. We assume the overall slip direction is controlled by the tectonic stress loading, while the topographic stresses introduce spatial variation of slip directions. The tectonic stress field in the Kumamoto source region is estimated as uniaxial‐tensional stress of about 35 Mpa close the north‐south direction. This estimation is slightly higher than the previous estimations using other assumptions. Key Points: We develop an algorithm to invert for the absolute tectonic stress in the Kumamoto earthquake regionWe obtain a tectonic stress in uniaxial tension with magnitude 35 MPa and orientation N6.5°WWe estimate that the pore‐pressure above 5 km depth is about 0.3–0.45 times lithostatic stress, which is close to a drained system [ABSTRACT FROM AUTHOR]

A great earthquake struck the Semidi segment of the plate boundary along the Alaska Peninsula on 29 July 2021, re‐rupturing part of the 1938 rupture zone. The 2021 MW 8.2 Chignik earthquake occurred just northeast of the 22 July 2020 MW 7.8 Simeonof earthquake, with little slip overlap. Analysis of teleseismic P and SH waves, regional Global Navigation Satellite System (GNSS) displacements, and near‐field and far‐field tsunami observations provides a good resolution of the 2021 rupture process. During ∼60‐s long faulting, the slip was nonuniformly distributed along the megathrust over depths from 32 to 40 km, with up to ∼12.9‐m slip in an ∼170‐km‐long patch. The 40–45 km down‐dip limit of slip is well constrained by GNSS observations along the Alaska Peninsula. Tsunami observations preclude significant slip from extending to depths <25 km, confining all coseismic slip to beneath the shallow continental shelf. Most aftershocks locate seaward of the large‐slip zones, with a concentration of activity up‐dip of the deeper southwestern slip zone. Some localized aftershock patches locate beneath the continental slope. The surface‐wave magnitude MS of 8.1 for the 2021 earthquake is smaller than MS = 8.3–8.4 for the 1938 event. Seismic and tsunami data indicate that slip in 1938 was concentrated in the eastern region of its aftershock zone, extending beyond the Semidi Islands, where the 2021 event did not rupture. Plain Language Summary: The Pacific plate underthrusts the North American plate along the Alaska‐Aleutian subduction zone producing great earthquakes along most of the plate boundary. The region, that ruptured in 1938, hosted another great earthquake with magnitude 8.2 on 29 July 2021 offshore of Chignik, adjacent to a magnitude 7.8 event in 2020 beneath the Shumagin Islands. Using seismic wave recordings, Global Navigation Satellite System static displacements, and tsunami wave recordings at regional and distant locations, we determine the space‐time evolution of slip during the 2021 rupture process. The rupture extended northeastward ∼170 km, ending just short of the Semidi Islands. The slip is concentrated on the megathrust, from 32‐km to 40‐km deep. Shallow slip at depths <25 km is precluded by the tsunami observations, confining all rupture to beneath the shallow continental shelf. Most aftershocks are up‐dip of the large‐slip regions with a concentration in the shallow region of the southwestern rupture zone. Comparisons of observations and rupture models with the 1938 event indicate distinct slip distributions and seismic energy release, with the 1938 event being somewhat larger and extending further northeast than the 2021 rupture. This provides strong evidence that the 2021 event is not a simple repeat failure of the 1938 rupture. Key Points: The slip distribution of the 2021 Chignik MW 8.2 earthquake is constrained by joint analysis of seismic, geodetic, and tsunami recordingsDuring the ∼60‐s long rupture, patchy slip of up to 12.9 m occurred at depths of 32–40 km on the megathrust below the continental shelfThe 2021 slip pattern differs from the 1938 MW 8.2 rupture, with larger slip in the west and no slip in the easternmost 1938 rupture zone [ABSTRACT FROM AUTHOR]

Rupture directivity and its potential frequency dependence is an open issue within the seismological community, especially for small‐to‐moderate events. Here, we provide a statistical overview based on empirical evidence of seismological observations, thanks to the large amount of high‐quality seismic recordings (more than 30,000 waveforms) from Central Italy, which represents an excellent and almost unique natural laboratory of normal faulting earthquakes in the magnitude range between 3.4 and 6.5 within the time frame 2008–2018. In order to detect an anisotropic distribution of ground motion amplitudes due to the rupture directivity, we fit the smoothed Fourier Amplitude Spectra (FAS) cleared of source‐, site‐ and path‐effects. According to our criteria, about 36% of the analyzed events (162 out of 456) are directive and the distribution of rupture direction is aligned with the strikes of the major faults of the Central Apennines. We find that the directivity is a band‐limited phenomenon whose width may extend up to five times the corner frequency. The results of this research provide useful insights to parameterize directivity, to be explicitly implemented in future ground motion modeling and scenario predictions. Plain Language Summary: In seismology, directivity is one of the source phenomena that causes large spatial variability of earthquake ground motions and is related to the features of the rupture propagation along the fault. The importance of this effect is very well known for large‐magnitude events (M > 6), while it is still an open issue for small‐to‐moderate events. The aim of this paper is to recognize directive events and quantify their strength in Central Italy, using a large data set of earthquakes in the magnitude range 3.4–6.5 that occurred between 2008 and 2018. We find that about 36% of the analyzed events are directive, caused by rupture propagations oriented along the NW‐SE alignment of the Central Apennines fault systems. Furthermore, we find that directivity is a band‐limited phenomenon and that as the directivity gets stronger, the frequency band becomes wider. Our contribution provides a useful insight with the possibility to improve the parametrization of directivity within the empirical seismic hazard assessment. Key Points: Statistical procedure to detect directivity effects in small‐to‐moderate earthquakes analyzing the aleatory residuals of the ground motionThe directive events are the 36% of the data set and the distribution of rupture direction is aligned with the strikes of the faults systemThe observed directivity is a frequency‐dependent band‐limited phenomenon, which can occur up to five times the value of corner frequency [ABSTRACT FROM AUTHOR]

Rupture size is a fundamental earthquake source parameter that is challenging to infer independently from far‐field seismological observations. Here, we develop a novel observational constraint on source size based on the decay rate of wavefield coherence across a seismic array. For a given earthquake, waveform coherence decays with increasing interstation distance or, more precisely, with increasing projection difference defined as the difference between the takeoff vectors associated to the two stations projected along the rupture direction. We find that coherence generally falls off with projection difference faster for earthquakes of larger magnitudes. The magnitude dependence of the coherence decay rate can be explained by a finite source effect: larger source sizes cause larger differences of phase delays between waves arriving from different parts of the rupture at different stations, hence a stronger spatial variability of the wavefield, resulting in a breakdown of waveform coherence. Assuming a 1‐D Haskell's source model, the rupture size can be estimated from the coherence decay rate. We apply this method to USArray data of earthquakes in three subduction zones, the Sea of Okhotsk, South America and Japan. The source sizes inferred from the coherence decay patterns are consistent with scaling relations intermediate between width‐saturated L‐models and quasi equi‐dimensional rupture models. Our observation captures a unique pattern of array waveform coherence and demonstrates the potential of utilizing waveform coherence to study earthquake source parameters. Plain Language Summary: In this paper, we observe that for a given earthquake, the similarity of the waveform (coherence) recorded by stations in a far‐field array decays with increasing interstation distance. The waveform coherence generally falls off with interstation distance faster for earthquakes of larger magnitudes and therefore, larger rupture sizes. We develop a novel method using array waveform coherence to estimate the earthquake rupture size, and apply this method to USArray data of earthquakes in three geological regions: the Sea of Okhotsk, South America and Japan. The source sizes inferred from the coherence decay patterns are consistent with previous studies using different methods. According to our observation, there is a potential to infer earthquake rupture independently from far‐field observations by waveform coherence. Key Points: Waveform coherence decays as a function of interstation distance faster for large earthquakes (Mw > 7)The waveform coherence pattern is governed by earthquake source sizeRupture sizes inferred from waveform coherence satisfy empirical scaling relations [ABSTRACT FROM AUTHOR]