3D fault architecture controls the dynamism of earthquake swarms

Fault material heterogeneity controls deep interplate earthquakes

Earthquakes may seem random, but are often concentrated in some localized areas. Thus, they are likely controlled by fault materials and stress heterogeneity, which are little understood. Here, we provide high-resolution observations of fault material and stress heterogeneity in the Japan subduction zone through an integration of material and source imaging with numerical simulations. Our results present evidence for localized, anisotropic structures with a near-zero Poisson's ratio in the medium surrounding 1 to 2 kilometer-sized earthquake clusters, suggesting that the fault medium is damaged, foliated, and enriched with fluid. Such localized structures may cause stress perturbations on faults that in turn favor the frequent occurrence of deep interplate earthquakes at depths of 60 to 70 kilometers. Therefore, identifying the distribution and properties of fault material heterogeneity is important for more informed assessment of earthquake hazards.

High-resolution investigations of fault architecture in space and time

Heterogeneous fault architecture affects crustal seismotectonics and fluid migration. When studying it, we commonly rely on static conceptual models that generally overlook the absolute time dimension of fault (re)activation. Heterogenous faults, however, represent the end-result of protracted, cumulative and intricate deformation histories. This may lead to inaccurate reconstructions of tectonic histories and flowed models of fault hydro-mechanical behavior. We adopt here a multitechnique approach building upon the examination of now juxtaposed but not coeval brittle structural facies (BSF), which offer multiscalar insights in the spatio-temporal-thermal fault evolution. Our approach is applied to the Carboneras Fault, unveiling a ~ 25 Myr-long polyphase structural and thermal evolution. This led to a complex fault architecture, where BSFs exhibit a stark heterogeneity in fault rock and permeability, ultimately generating very different space- and time-dependent fault hydro-mechanical behaviors. Therefore, fault architectures shall be seen as dynamic features from which to extrapolate time-integrated comprehensive fault models accounting for the entire deformed rock volume and fault life span. We demonstrate that high-resolution studies of fault architectures are required to elucidate modes of fault growth and evolution, decipher long-lived, polyphase tectonic and thermal histories, and understand the influence of heterogenous fault architecture on hydraulic compartmentalization and earthquake rupture dynamics.

Parallel dynamics of slow slips and fluid-induced seismic swarms

Earthquake swarms may be driven by fluids, through hydraulic injections or natural fluid circulation, but also by slow and aseismic slip transients. Understanding the driving factors for these prolific sequences and how they can potentially develop into larger ruptures remains a challenge. A notable and almost ubiquitous feature of swarms is their hypocenters migration, which occurrence is closely related to the processes driving the observed seismicity, in a similar way as seismicity accompanies slow-slip events at subduction zones. Here, we analyze global data on migrating sequences, and identify scaling laws for migration velocity, moment and duration measured on natural and injection-induced swarms, foreshock sequences, and slow slip events. We highlight two different behaviors among these sequences: one linked to slow slips, with elevated migration velocities and moments, and the other related to fluid-induced processes, featuring lower velocities and moments. These results provide metrics for distinguishing between the drivers of earthquake swarms, fluid or slow-slip related, and prompt a reevaluation of scaling laws of fault slip transients, especially for swarms.

Self-evolving artificial intelligence framework to better decipher short-term large earthquakes

Large earthquakes (EQs) occur at surprising loci and timing, and their descriptions remain a long-standing enigma. Finding answers by traditional approaches or recently emerging machine learning (ML)-driven approaches is formidably difficult due to data scarcity, interwoven multiple physics, and absent first principles. This paper develops a novel artificial intelligence (AI) framework that can transform raw observational EQ data into ML-friendly new features via basic physics and mathematics and that can self-evolve in a direction to better reproduce short-term large EQs. An advanced reinforcement learning (RL) architecture is placed at the highest level to achieve self-evolution. It incorporates transparent ML models to reproduce magnitude and spatial location of large EQs ([Formula: see text] 6.5) weeks before of the failure. Verifications with 40-year EQs in the western U.S. and comparisons against a popular EQ forecasting method are promising. This work will add a new dimension of AI technologies to large EQ research. The developed AI framework will help establish a new database of all EQs in terms of ML-friendly new features and continue to self-evolve in a direction of better reproducing large EQs.

Slow rupture in a fluid-rich fault zone initiated the 2024 Mw 7.5 Noto earthquake

The 2024 moment magnitude 7.5 Noto Peninsula (Japan) earthquake caused devastation to communities and was generated by a complex rupture process. Using space geodetic and seismic observations, we have shown that the event deformed the peninsula with a peak uplift reaching 5 meters at the west coast. Shallow slip exceeded 10 meters on an offshore fault. Peak stress drop was greater than 10 megapascals. This devastating event began with a slow rupture propagation lasting 15 to 20 seconds near its hypocenter, where seismic swarms had surged since 2020 because of lower-crust fluid supply. The slow start was accompanied by intense high-frequency seismic radiation. These observations suggest a distinct coseismic slip mode reflecting high heterogeneity in fault properties within a fluid-rich fault zone.

Dual-initiation ruptures in the 2024 Noto earthquake encircling a fault asperity at a swarm edge

To reveal the connections between the 2024 moment magnitude (Mw) 7.5 Noto earthquake in Japan and the seismicity swarms that preceded it, we investigated its rupture process through near-source waveform analysis and source imaging techniques, combining seismic and geodetic datasets. We found notable complexity in the initial rupture stages. A strong fault asperity, which remained unbroken in preceding seismic swarms, slowed down the rupture. Then, a second rupture initiated at the opposite edge of the asperity, and the asperity succumbed to double-pincer rupture fronts. The failure of this high-stress drop asperity drove the earthquake into a large-scale event. Our observations help unravel the crucial role of fault asperities in controlling swarm migration and rupture propagation and underscore the need for detailed seismological and interdisciplinary studies to assess seismic risk in swarm-prone regions.

Interplay of tectonic and dynamic processes shaping multilayer extensional system in southern-central Apennines

High-hazard seismic zones can remain silent over centuries with meager seismicity rates challenging our understanding of seismic processes. We focus on the comprehensive analysis of cascading episodes of swarms and seismic sequences following the 2009 L'Aquila mainshock (MW 6.3) in the southern-central Apennine that previously experienced ~ M7 earthquakes. We enhance the seismic catalog, unmasking low-magnitude seismicity down to completeness magnitude ML ~ 0, and we unveil that the microseismicity might be secondarily triggered by the L'Aquila mainshock, influencing the frictional properties in the nearby fault zones or opening fault valves generating the intense seismic activity detected from 2009 to 2013. The diffusivity, observed in the most seismic episodes, and the high Vp/Vs values (> 1.88) indicate fluid circulation promoting multilayered extensional seismicity within 11-15 km and 16-23 km depth ranges. Mapping the 3D distribution of seismicity alongside geological data reveals an evident tectonic influence, unveiling unknown geometric aspects and providing the first evidence of a NNE-dipping deformation zone bounding at depths of 11-15 km the overlying fault system. Deeper seismicity suggests a mantellic CO2 ascending shape. These findings enrich the literature on tectonic seismic swarms in extensional domains, providing essential constraints on fluid involvement in the seismic processes and contributing to forthcoming discussions on the seismotectonic setting in high-seismic-risk areas of the Apennines of Italy.

A laboratory study on fault slip caused by fluid injection directly versus indirectly into a fault: implications for induced seismicity in EGSs

Enhanced geothermal systems (EGSs) developed by hydraulic stimulation are promising for exploiting petrothermal heat by improving fluid pathways in low-permeable geothermal reservoir rocks. However, fluid injection into the subsurface can potentially cause large seismic events by reactivating pre-existing faults, which is a significant barrier to EGSs. The management of injection-induced seismicity is, therefore, essential for the success of EGSs. During the hydraulic stimulation of an EGS, fluid can be injected into a fault zone or into the rock matrix containing pre-existing faults adjacent to the injection well. The differences in hydromechanical responses between fluid injection into and adjacent to a fault have not been investigated in detail. Here, we performed triaxial fluid injection experiments involving injecting fluid directly and indirectly into a fault in granite rock samples to analyse the distinct hydromechanical responses and estimate the injection-induced seismicity in both cases. Our results suggest that in addition to directly injecting fluid into a critically stressed fault, injecting into nearly intact granite adjacent to the fault could also cause injection-induced seismic hazards owing to the high fluid pressure required to create new fractures in the granite matrix. It is, therefore, important to carefully identify pre-existing faults within tight reservoirs to avoid injecting fluid adjacent to them. Additionally, once prior unknown faults are delineated during hydraulic stimulation, appropriate shut-in strategies should be implemented immediately to mitigate seismic risks. This article is part of the theme issue 'Induced seismicity in coupled subsurface systems'.

Sharpen data-driven prediction rules of individual large earthquakes with aid of Fourier and Gauss

Predicting individual large earthquakes (EQs)' locations, magnitudes, and timing remains unreachable. The author's prior study shows that individual large EQs have unique signatures obtained from multi-layered data transformations. Via spatio-temporal convolutions, decades-long EQ catalog data are transformed into pseudo-physics quantities (e.g., energy, power, vorticity, and Laplacian), which turn into surface-like information via Gauss curvatures. Using these new features, a rule-learning machine learning approach unravels promising prediction rules. This paper suggests further data transformation via Fourier transformation (FT). Results show that FT-based new feature can help sharpen the prediction rules. Feasibility tests of large EQs ([Formula: see text] 6.5) over the past 40 years in the western U.S. show promise, shedding light on data-driven prediction of individual large EQs. The handshake among ML methods, Fourier, and Gauss may help answer the long-standing enigma of seismogenesis.

Physical state of water controls friction of gabbro-built faults

Earthquakes often occur along faults in the presence of hot, pressurized water. Here we exploit a new experimental device to study friction in gabbro faults with water in vapor, liquid and supercritical states (water temperature and pressure up to 400 °C and 30 MPa, respectively). The experimental faults are sheared over slip velocities from 1 μm/s to 100 mm/s and slip distances up to 3 m (seismic deformation conditions). Here, we show with water in the vapor state, fault friction decreases with increasing slip distance and velocity. However, when water is in the liquid or supercritical state, friction decreases with slip distance, regardless of slip velocity. We propose that the formation of weak minerals, the chemical bonding properties of water and (elasto)hydrodynamic lubrication may explain the weakening behavior of the experimental faults. In nature, the transition of water from liquid or supercritical to vapor state can cause an abrupt increase in fault friction that can stop or delay the nucleation phase of an earthquake.

Major southern San Andreas earthquakes modulated by lake-filling events

Hydrologic loads can stimulate seismicity in the Earth's crust1. However, evidence for the triggering of large earthquakes remains elusive. The southern San Andreas Fault (SSAF) in Southern California lies next to the Salton Sea2, a remnant of ancient Lake Cahuilla that periodically filled and desiccated over the past millennium3-5. Here we use new geologic and palaeoseismic data to demonstrate that the past six major earthquakes on the SSAF probably occurred during highstands of Lake Cahuilla5,6. To investigate possible causal relationships, we computed time-dependent Coulomb stress changes7,8 due to variations in the lake level. Using a fully coupled model of a poroelastic crust9-11 overlying a viscoelastic mantle12,13, we find that hydrologic loads increased Coulomb stress on the SSAF by several hundred kilopascals and fault-stressing rates by more than a factor of 2, which is probably sufficient for earthquake triggering7,8. The destabilizing effects of lake inundation are enhanced by a nonvertical fault dip14-17, the presence of a fault damage zone18,19 and lateral pore-pressure diffusion20,21. Our model may be applicable to other regions in which hydrologic loading, either natural8,22 or anthropogenic1,23, was associated with substantial seismicity.

The magmatic web beneath Hawai'i

The deep magmatic architecture of the Hawaiian volcanic system is central to understanding the transport of magma from the upper mantle to the individual volcanoes. We leverage advances in earthquake monitoring with deep learning algorithms to image the structures underlying a major mantle earthquake swarm of nearly 200,000 events that rapidly accelerated after the 2018 Kīlauea caldera collapse. At depths of 36 to 43 kilometers, we resolve a 15-kilometers-long collection of near-horizontal sheeted structures that we identify as a sill complex. These sills connect to the lower depths of Kīlauea's plumbing by a 25-kilometers-long belt of seismicity. Additionally, a column of seismicity links the sill complex to a shallow décollement near Mauna Loa. These findings implicate the mantle sill complex as a nexus for magma transport beneath Hawai'i and furthermore indicate widespread magmatic connectivity in the volcanic system.

Gauss curvature-based unique signatures of individual large earthquakes and its implications for customized data-driven prediction

Statistical descriptions of earthquakes offer important probabilistic information, and newly emerging technologies of high-precision observations and machine learning collectively advance our knowledge regarding complex earthquake behaviors. Still, there remains a formidable knowledge gap for predicting individual large earthquakes’ locations and magnitudes. Here, this study shows that the individual large earthquakes may have unique signatures that can be represented by new high-dimensional features—Gauss curvature-based coordinates. Particularly, the observed earthquake catalog data are transformed into a number of pseudo physics quantities (i.e., energy, power, vorticity, and Laplacian) which turn into smooth surface-like information via spatio-temporal convolution, giving rise to the new high-dimensional coordinates. Validations with 40-year earthquakes in the West U.S. region show that the new coordinates appear to hold uniqueness for individual large earthquakes ($$M_w \ge 7.0$$ M w ≥ 7.0 ), and the pseudo physics quantities help identify a customized data-driven prediction model. A Bayesian evolutionary algorithm in conjunction with flexible bases can identify a data-driven model, demonstrating its promising reproduction of individual large earthquake’s location and magnitude. Results imply that an individual large earthquake can be distinguished and remembered while its best-so-far model can be customized by machine learning. This study paves a new way to data-driven automated evolution of individual earthquake prediction.

Permeability-controlled migration of induced seismicity to deeper depths near Venus in North Texas

Migration of seismic events to deeper depths along basement faults over time has been observed in the wastewater injection sites, which can be correlated spatially and temporally to the propagation or retardation of pressure fronts and corresponding poroelastic response to given operation history. The seismicity rate model has been suggested as a physical indicator for the potential of earthquake nucleation along faults by quantifying poroelastic response to multiple well operations. Our field-scale model indicates that migrating patterns of 2015–2018 seismicity observed near Venus, TX are likely attributed to spatio-temporal evolution of Coulomb stressing rate constrained by the fault permeability. Even after reducing injection volumes since 2015, pore pressure continues to diffuse and steady transfer of elastic energy to the deep fault zone increases stressing rate consistently that can induce more frequent earthquakes at large distance scales. Sensitivity tests with variation in fault permeability show that (1) slow diffusion along a low-permeability fault limits earthquake nucleation near the injection interval or (2) rapid relaxation of pressure buildup within a high-permeability fault, caused by reducing injection volumes, may mitigate the seismic potential promptly.

Basal nucleation and the prevalence of ascending swarms in Long Valley caldera

Earthquake swarms are ubiquitous in volcanic systems, being manifestations of underlying nontectonic processes such as magma intrusions or volatile fluid transport. The Long Valley caldera, California, is one such setting where episodic earthquake swarms and persistent uplift suggest the presence of active magmatism. We quantify the long-term spatial and temporal characteristics of seismicity in the region using cluster analysis on a 25-year high-resolution earthquake catalog derived using leading-edge deep-learning algorithms. Our results show that earthquake swarms beneath the caldera exhibit enlarged families with statistically significant tendency for upward migration patterns. The ascending swarms tend to nucleate at the base of the seismogenic zone with a spatial footprint that is laterally constrained by the southern rim of the caldera. We suggest that these swarms are driven by the transport of volatile-rich fluids released from deep volcanic processes. The observations highlight the potential for extreme spatial segmentation of earthquake triggering processes in magmatic systems.