During the siting of hydraulic-fracturing (HF) wells within industrial activity areas, identifying potential seismogenic faults and effectively avoiding them is critical for mitigating induced seismicity risk. Meanwhile, characterizing the fine-scale structures of seismogenic faults provides the essential foundation for analyses of the mechanisms and rupture processes of induced earthquakes. However, multiple case studies have demonstrated that, even where seismic reflection data are available, it remains difficult to identify small-displacement seismogenic faults, particularly those dominated by strike-slip faults. Consistently, the four representative M5+ induced earthquakes in the Changning and Weiyuan shale gas blocks of the Sichuan Basin also exhibit difficulties in identifying the seismogenic faults from seismic reflection data. Moreover, the scales of faults that can be identified through seismic reflection data and related interpretation methods, and their corresponding seismogenic potential, remain to be systematically defined and quantitatively constrained. This study integrates spatiotemporal data from HF operations, seismicity data, and high-resolution 3D seismic reflection data, together with surface deformation measurements, to address the above questions.

The results show that potential seismogenic faults with moment magnitude (Mw) greater than approximately 3.3 that displace strong reflection horizons can be effectively identified using high-resolution 3D seismic reflection data. In addition, the associated structures of small-displacement strike-slip faults facilitate their recognition in seismic reflection profiles. A common feature of the seismogenic fault systems of the four representative earthquakes is that small-displacement subsidiary faults (including strike-slip faults) intersect the fracturing wells within the reservoir interval, forming downward migration pathways for fracturing fluids and thereby activating the underlying thrust or strike-slip seismogenic faults. More importantly, such small-displacement faults are widely developed within the fractured intervals of the Sichuan Basin shale gas fields, yet their identification remains challenging. As a result, numerous horizontal wells intersect these faults, constituting a key reason for the frequent occurrence of induced seismicity in these areas. The most effective approach to recognizing these faults is to trace multiple strong reflection horizons to construct structural maps. By applying multi-azimuth illumination and vertical stretching, fault traces can be visualized more clearly, in combination with various types of seismic reflection attribute volumes.

Beyond the Sichuan Basin, injection-induced earthquakes in most shale gas fields worldwide are also closely associated with small-displacement faults, particularly strike-slip faults. The failure to avoid such faults during the siting of HF wells is also likely a major reason for the frequent occurrence of induced seismicity in these areas. The small-displacement fault identification techniques presented in this study facilitate a more precise delineation of seismogenic fault system structure. More importantly, during well site selection, from the perspective of fault identification and avoidance based on 3D seismic reflection data, this study provides theoretical support and practical strategies for preventing induced earthquakes with a magnitude (Mw) greater than approximately 3.3. These findings also offer significant implications for the prevention of induced seismicity caused by fluid/gas injection in a broader range of applications.

How to cite: Ye, Y. and Lu, R.: Identification and Impacts of Small-Displacement Faults in Industry-induced earthquake: Insights from the Southern Sichuan Shale Gas Field, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-2497, https://doi.org/10.5194/egusphere-egu26-2497, 2026.

Hydraulic fracturing operations in the Southern Sichuan Basin have generated significant induced seismicity, raising important questions about the underlying rupture processes. We analyze stress drops of 3,369 induced earthquakes (ML > 0.5) using a non-parametric generalized inversion technique with rigorous reference-station corrections. Our analysis reveals two key characteristics of these induced events: first, they exhibit systematically low stress drops (median 0.07 MPa) that show positive scaling with seismic moment, challenging classical self-similarity assumptions; second, we observe pronounced spatial variations in stress release that correlate with depth and fault structure. Notably, fluid diffusion drives rapid activation of fault asperities, resulting in repeated high-stress-drop ruptures (0.3-6.0 MPa) within short timescales of days. This accelerated rupture cycle differs fundamentally from tectonic earthquake recurrence patterns. Our findings demonstrate that induced earthquake rupture dynamics are controlled by the interplay of heterogeneous fault strength and rapid fluid pressurization, providing critical insights for developing targeted hazard assessment strategies in energy-producing regions.

How to cite: Chen, X., Tao, J., Liu, D., and Chen, H.: Stress Drop Variability and Rapid Fault Activation in Hydraulic-Fracturing-Induced Earthquakes: Insights from the Southern Sichuan Basin, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-2228, https://doi.org/10.5194/egusphere-egu26-2228, 2026.

Mining districts in Chile are located within one of the most seismically active tectonic environments on Earth, which makes it challenging to distinguish between natural crustal seismicity and seismicity potentially influenced by mining. A key open question is whether large-scale mining operations produce a measurable and spatially coherent statistical signature in the surrounding shallow crust. In this study, we evaluate this hypothesis using the Gutenberg–Richter b-value as a quantitative proxy for local stress conditions and the degree of rock mass damage and fracturing. We focus on the seismic environment surrounding major mining districts in Chile, restricting the analysis to shallow crustal events with depths shallower than 10 km. This depth filter aims to isolate the seismic response of the upper crust that is most likely to be affected by mining-related stress perturbations, while reducing the contribution of deeper subduction-driven tectonic seismicity. To resolve spatial variations at kilometer scale, we implement a high-resolution concentric-ring analysis centered on each mining district, using 1 km radial bins extending outward from the extraction centers. To ensure statistical robustness and comparability across sites, the magnitude of completeness (Mc) is estimated dynamically using the maximum curvature method, yielding reference values close to ML ≈ 1.87 for the analyzed catalog. The Gutenberg–Richter b-value is then computed using the Aki–Utsu maximum-likelihood estimator, providing a rigorous and stable framework for inter-site comparisons under contrasting geomechanical and operational settings. The analysis reveals clear and systematic differences depending on the mining method. Underground mining environments show a pronounced increase in b-value (b > 1.5) within the first ~5 km, consistent with elevated rates of microseismicity and enhanced brittle damage associated with caving-related processes. In contrast, open-pit operations exhibit a comparatively stable b-value pattern with lower spatial dispersion. In both settings, b-values progressively converge toward the regional tectonic reference level (b ≈ 1.0) with increasing distance from the extraction centers, suggesting a characteristic radius of direct mining influence on the order of ~15–20 km. These preliminary results show that kilometer-scale mapping of the Gutenberg–Richter b-value provides a sensitive and interpretable metric to quantify the spatial footprint of mining-related seismic perturbations in the shallow crust. The observed b-value gradients offer a practical tool for regional-scale geomechanical monitoring, supporting the discrimination between background tectonic seismicity and elastic stress changes induced by excavation and/or large-volume rock mass caving in Chilean mining districts.

How to cite: Roquer, T. and Ravest, B.: Mining-Related b-Value Anomalies in the Upper Crust: A High-Resolution Ring Analysis Across Chilean Mining Districts , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20770, https://doi.org/10.5194/egusphere-egu26-20770, 2026.

Hydraulic fracturing often induces complex seismic sequences that migrate across stratigraphically distinct formations. However, the mechanisms governing delayed triggering and vertical interaction through lithological boundaries remain poorly understood. In this study, we report a novel "coseismic valve" mechanism observed in the Weiyuan shale gas field, southern Sichuan Basin, where the multi-stage evolution of seismicity was strictly governed by pre-existing 3D mechanical stratigraphy.

Using a dense local monitoring array, we constructed a high-resolution catalog by the deep-learning-based LOC-FLOW workflow. This catalog revealed a vertically partitioned fault system, where the deep and shallow seismicity clusters are distinctly separated by a ~400 m thick low-velocity ductile barrier.This barrier mechanically isolated a deep, critically stressed segment (characterized by a low b-value) from a shallower, compliant damage zone. Our analysis reveals a paradox in the role of ductile layers: initially, the barrier acted as a "pressure seal," preventing fluid leak-off and facilitating high differential stress accumulation in the underlying reservoir. This confinement culminated in the nucleation of an Mw 3.6 mainshock with an anomalously high stress drop.

Crucially, finite fault inversion and isochrone back-projection demonstrate that the mainshock rupture propagated upward, dynamically breaching the ductile barrier. This mechanical breach effectively functioned as a valve, establishing a vertical conduit for hydraulic connectivity. Following a distinct 6-day delay, a diffusive seismic swarm erupted in the previously quiescent shallow segment, driven by the upward surge of overpressured fluids through the newly created fracture network.

Our findings challenge the conventional view of ductile layers merely as passive aseismic buffers. We demonstrate that they can play a dual role: serving as stress-concentrating seals that prime the system for nucleation, and as structural valves that, once ruptured, enable cascading seismic hazards. This dynamic interaction highlights the necessity of integrating 3D structural frameworks into seismic risk assessment for geo-energy projects.

How to cite: Tao, J., Chen, X., Liu, D., Chen, H., Zhao, Y., Niu, F., and Zhang, L.:  The Dual Role of Ductile Barriers: From Stress-Concentrating Seals to Coseismic Valves in Induced Seismicity, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-2227, https://doi.org/10.5194/egusphere-egu26-2227, 2026.

Induced seismicity remains a major challenge for geothermal projects, with implications for public acceptance and operational risk management. Understanding how fluid injection interacts with fault structures to generate seismicity is therefore essential. The Húsmúli reinjection area in the Hengill geothermal field (SW Iceland) provides an ideal setting to investigate these processes due to its sustained induced seismicity and long operational history. Here, we present an improved seismicity catalog (2018–2021; COSEISMIQ project) and a waveform-based detection workflow that substantially increases catalog completeness and enhances spatiotemporal resolution.

We first improve the initial automatic catalog by re-picking events with unrealistic Vp/Vs ratios (Wadati analysis), high RMS location misfits, or unrealistic depths (e.g., airquakes). Phase picks are then refined using a cross-correlation (CC)-based repicking approach: events are clustered into waveform-similar families, traces are aligned and stacked to increase signal-to-noise, and consistent arrival times are obtained from a single-family reference pick. Missing picks are recovered by inspecting waveforms around the expected arrival time window and estimating phase onsets, accepting only traces with  CC ≥ 0.65 with respect to other family members.

3D spatial clustering of the refined catalog reveals NE–SW oriented seismic lineaments consistent with mapped faults and inferred fluid migration pathways. In contrast, nearby E-W structures show little to no seismicity, suggesting permeability barriers and reservoir compartmentalization. Repeating earthquakes occur along narrow fault segments, indicating repeated rupture of localized slip patches. To further enhance detection, we use QuakeMatch, a single-station template matching workflow using high-SNR events as templates at the station with the best waveform quality and data completeness. This expands the catalog from 3,647 to 12,899 events, lowering the magnitude of completeness and revealing numerous low-magnitude earthquakes previously missed by the automatic STA/LTA processing due to low signal-to-noise or waveform overlap. The resulting catalog shows swarm-like activity typical of fluid-driven seismicity and episodic bursts. A prominent sequence on 15 November 2020 (M_LX  = 4.08) is preceded by foreshocks and followed by multiple M_LX ≥ 3.0 aftershocks. Gutenberg–Richter analysis indicates a decrease in b-values prior to the mainshock, consistent with stress build-up and suggesting potential precursory behaviour relevant for operational monitoring.

How to cite: Kaur, S., Toledo, T., Kraft, T., and Simon, V.: Characterization of microseismicity at the Húsmúli reinjection area, Hengill Geothermal Field, Southwest Iceland​, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-8098, https://doi.org/10.5194/egusphere-egu26-8098, 2026.

Induced seismicity remains a significant challenge for the development of deep geothermal energy projects, with continued challenges at both the scientific and operational levels. 

Scientific level: Seismic monitoring at geothermal sites is commonly limited to periods of active operations such as hydraulic stimulation and testing, whereas datasets documenting the seismic response during shut-in and post-operational phases remain scarce. However, larger-magnitude earthquakes have been observed during shut-in phases, and in some cases months later, despite limited information on seismic activity during the active period. As a result, the processes governing delayed, larger-magnitude induced earthquakes remain poorly understood. 

Operational level: During active operations, the spatio-temporal evolution of induced seismicity provides one of the few direct indicators of subsurface processes. Real-time insight into whether seismicity evolves as expected or migrates toward potentially hazardous structures is essential for timely mitigation. Advanced Traffic Light Systems (ATLS) assess seismic hazard and risk based on observed seismic responses and rely on statistical and hydromechanical models to forecast the likelihood of induced events over the following hours to days. The reliability of these forecasts critically depends on the quality of the underlying earthquake catalog. Improved detection and location of small events and more robust magnitude estimates can substantially enhance hazard assessments and operational decision-making. 

To address these challenges, we introduce QuakeMatch (QM), a toolbox that leverages waveform similarity to improve seismic monitoring in both real-time and long-term applications. The workflow employs template matching based on events from a manually revised catalog, followed by refined magnitude estimation, event relocation of assembled events, and statistical analysis. 

We demonstrate the application of QM using the case studies from the Basel and Haute-Sorne deep geothermal projects. The Basel case is currently covered by earthquake catalogs with strongly varying location precision and completeness. A template-matched catalog by Herrmann et al. (2019), covering the period 2006–2019, does not include relocations and has not been updated since its publication. Here, QM is used to build a homogeneous long-term catalog of consistently high-precision earthquake locations that will improve our ability to assess the long-term response of this field over two decades up to the present day. For the Haute-Sorne case, we demonstrate the real-time application of QM, illustrating its potential to better inform advanced induced-seismicity-mitigation procedures (e.g., ATLS) with more reliable, consistent, and sensitive earthquake catalogs. Together, these examples illustrate the potential of combining long-term catalog enhancement with real-time monitoring to support safer and more informed geothermal operations. 

How to cite: Toledo, T., Simon, V., Kraft, T., and Diehl, T.: Enhancing Long-Term Seismic Analysis of Swiss Geothermal Projects through Waveform Similarity, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13011, https://doi.org/10.5194/egusphere-egu26-13011, 2026.

Injection-induced seismicity in Enhanced Geothermal Systems (EGS) can migrate hundreds of meters from the well and often persists after shut-in, raising operational and hazard concerns. Here we present a cross-site comparative analysis of three stimulation campaigns—Soultz-sous-Forêts (France, 1993), Basel (Switzerland, 2006), and Utah FORGE Stage 3 (USA, 2022)—to identify the dominant controls on intermediate-field seismic migration.

Using a unified dynamic time-windowing framework, we track seismic front evolution via three complementary distance metrics and evaluate their relationships with injection rate, wellhead pressure, cumulative injected volume, hydraulic energy, seismicity rate, and modeled pore pressure at the migration front. Across all sites, cumulative variables—particularly injected volume, hydraulic energy, and injection duration—show the strongest and most consistent correlations with seismic front expansion, whereas instantaneous parameters exhibit weaker or site-specific influence.

Post-injection behaviors distinguish three migration regimes: (i) a pressure-limited regime at Soultz, where the front halts immediately after shut-in; (ii) a diffusion-dominated regime at Basel, with continued post-shut-in propagation; and (iii) a stress-sensitive, limited-diffusion regime at Utah FORGE, characterized by rapid early migration followed by stagnation. Building on these contrasts, we introduce a six-indicator radar classification that quantitatively distinguishes the three regimes.

Our results show that cumulative hydraulic forcing provides transferable, physically interpretable predictors of intermediate-field migration and that distinct post-shut-in signatures reflect underlying connectivity and stress conditions. This comparative framework supports improved seismic hazard assessment and operational planning for geothermal reservoir stimulation.

How to cite: Wang, Z., Pankow, K., Rinaldi, A., Verdon, J., and Main, I.: Cumulative Controls on Intermediate-Field Seismic Migration: Comparative Evidence from Three Geothermal Stimulation Campaigns, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1317, https://doi.org/10.5194/egusphere-egu26-1317, 2026.

Between April and June 2013, the GRT-1 well at the Rittershoffen geothermal site in the Upper Rhine Valley (France) underwent three distinct stimulation phases: first a thermal stimulation, then a chemical stimulation, and finally a hydraulic stimulation. These fluid injections significantly enhanced the injectivity index of the well, rendering it suitable for economic exploitation. Throughout these operations, a local surface seismic network continuously monitored the site, recording thousands of unfelt seismic events.

This study builds upon and refines the findings of Lengliné et al. (2017), who focused solely on the hydraulic stimulation of GRT-1, and Maurer et al. (2020), whose interpretations were constrained by uncertain absolute locations of seismic events, particularly in terms of depth. By employing an improved template matching technique and a relative location method, we established a comprehensive seismic event catalog comprising over 3,000 events.

This reliable catalog enables precise tracking of the reservoir’s seismogenic response to the successive yet distinct stimulation types, with high spatial and temporal resolution. Consequently, it allows for an investigation into the potential seismic interplay between these stimulations. Our analysis examines the evolution of key characteristics, including event distribution and clustering, b-value, and seismic injection efficiency across the stimulation phases. The observed differences prompt critical questions regarding the reliability of using responses from prior stimulations to forecast seismogenic behavior during subsequent operations, even for the same site.

How to cite: Gaucher, E., Lengliné, O., and Schmittbuhl, J.: Variability of the Seismic Response of the Rittershoffen Geothermal Reservoir to the Series of GRT-1 Stimulations, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9185, https://doi.org/10.5194/egusphere-egu26-9185, 2026.

Injection of fluids during reservoir stimulation aims to enhance reservoir permeability but induces seismic activity that persists for several hours to several months after injection has ceased. Physical and hybrid models have been successfully applied to reproduce and forecast observed seismicity rates during and after injection. However, these models are typically site-dependent, raising the question of whether a general relationship between pressure and seismicity decay can be observed across different sites and operations.

In this study, we investigate the correlation between post-injection pore pressure decay and the decrease in seismicity rate using data from multiple EGS injection operations that share similar properties. First, the performance of several empirical statistical models is evaluated to describe the decrease in seismicity rate. Second, wellhead pressure decay is shown to be best described by a simple exponential model. Lastly, we introduce a time-to-fraction metric to compare the pressure and seismicity evolution after shut-in. We show that the times required to reach a given fraction of the initial rate for both pressure and seismicity are correlated, with pressure evolution being slower than seismicity rate evolution. No correlation is observed between seismicity decay and injection parameters such as injected volume, average injection pressure, or injection duration. These observations suggest that pore pressure has a limited influence on seismicity decay, which has strong implications for reservoir management.

How to cite: Minetto, R., Wang, Z., Lengliné, O., and Schmittbuhl, J.: Investigating the Correlation Between Post-Injection Trailing Seismicity and Wellhead Pressure Decay in Enhanced Geothermal Systems, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-14295, https://doi.org/10.5194/egusphere-egu26-14295, 2026.

Understanding how faults are activated and earthquakes are triggered is still a central challenge in seismology and seismic hazard assessment. Controlled hydraulic stimulation experiments offer a valuable opportunity to study these processes under well-constrained conditions and at spatial and temporal resolutions that are rarely achievable in natural settings. In this study, we present the results of three hydraulic stimulations experiments conducted at the Bedretto Underground Laboratory and monitored by a dense, high-sensitivity seismic network.

These experiments revealed a complex spatio-temporal evolution of induced seismicity, characterized by the activation of a multi-segment fault network. Two dominant seismic clusters were activated early on and show a clear spatial connection to the injection borehole, suggesting that pore pressure is the main driver of seismicity within these clusters. At later stages, a third cluster with a different orientation was activated, despite showing no obvious direct hydraulic connection to the injection interval. Seismicity within this cluster occurred with a temporal delay compared to the other two clusters. This suggests that the fault activation was likely driven by indirect processes such as aseismic deformation, stress transfer, and delayed fluid migration.

The observed fault network activation closely resembles patterns commonly reported in natural earthquake sequences. These findings suggest that the physical mechanisms controlling fault reactivation and earthquake triggering are largely independent of scale, linking controlled field experiments and natural earthquakes. Our results emphasize the importance of fault network geometry and stress interactions in understanding induced and natural seismicity.

How to cite: Rosskopf, M., Obermann, A., Rinaldi, A. P., Bröker, K., Villiger, L., and Giardini, D.: Fault Network Activation During Controlled Hydraulic Stimulation Experiments in the BedrettoLab, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12236, https://doi.org/10.5194/egusphere-egu26-12236, 2026.

A key challenge in induced seismicity is that fluid injection perturbs stress and pore pressure on faults with heterogeneous properties, leading to complex earthquake nucleation, migration and magnitude. The BedrettoLab (Bedretto Underground Laboratory for Geosciences and Geoenergies), located in the Swiss Alps, provides a unique natural testbed to study how these coupled hydro-mechanical processes interact with fault heterogeneity under controlled injection conditions, with direct access to well-characterized and densely instrumented fault zones. Previous characterization of the target MC fault zone at BedrettoLab show that layers of frictional velocity-strengthening (VS) fault gouge are embedded within velocity-weakening (VW) granitic bare rock, forming a strongly heterogeneous frictional architecture. However, how this frictional partitioning controls fault slip behavior and the magnitude of induced seismicity remains unclear. In this study, we use the newly developed 3D hydro-mechanical model HydroMech3D to explore the interplay between frictional heterogeneity and seismicity in fluid injection simulations, governed by rate- and state-dependent friction. We simulate injection scenarios using parameters and conditions derived from the ongoing FEAR (Fault Activation and Earthquake Rupture) experiments at Bedretto. Our simulations investigate how the spatial distribution of VS and VW patches control seismicity magnitude. By systematically changing the partition of VS and VW patches, we explore its influence on event size distributions and maximum magnitudes. Further simulations are conducted under various hydro-mechanical pre-conditioning conditions, by pre-determining the pressurized patch on the fault via the injection protocol prior to the main injection. These simulations allow us to understand whether fault pre-conditioning may influence the maximum magnitude of induced seismicity. Our results emphasize the critical role of frictional heterogeneity and injection strategy in fault dynamics, providing new insights into the hydro-mechanical behavior of complex fault zones during fluid injection and improving seismic risk assessment and mitigation strategies.

How to cite: Ye, J., Wang, Z., Ciardo, F., Rinaldi, A. P., Dal Zilio, L., and Giardini, D.: 3D Modeling of fluid-induced seismicity on fault with heterogeneous frictional asperities, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17827, https://doi.org/10.5194/egusphere-egu26-17827, 2026.

Fluid injection associated with geoenergy applications such as geothermal energy, CO2 sequestration, and hydraulic fracturing can alter fault stability through a combination of coupled hydro-mechanical processes. Laboratory experiments and underground observatories have provided valuable constraints on fault friction and near-fault pressure evolution, yet translating these observations to field-scale behavior requires physics-based numerical models that can resolve fault slip under realistic geometrical and mechanical conditions.

A major limitation of existing modeling approaches is the high computational cost of fully coupled three-dimensional simulations. As a result, many studies rely on one-dimensional fault representations or simplified elastic and hydraulic coupling. While such models have provided important insights into key physical mechanisms, they are not well suited to support the design, interpretation, and long-term forecasting of modern injection experiments equipped with dense monitoring systems. These experimental settings increasingly demand three-dimensional models capable of capturing realistic fault geometry, spatially variable frictional and hydraulic properties, and stress interactions beyond reduced-dimensional assumptions.

Here we present HydroMech3D, a physics-based numerical framework designed to efficiently simulate fluid-driven fault slip over earthquake-cycle timescales in three dimensions. The model employs a quasi-dynamic Boundary Element Method, discretizing only the fault surface embedded in elastic medium, thereby avoiding volumetric meshing. Fault slip is governed by rate-and-state friction and coupled to pore-pressure diffusion along the fault. Computational efficiency is achieved through a C++ implementation accelerated by hierarchical matrix from the Bigwham Library, enabling large-scale simulations with realistic fault geometry.

This framework allows systematic investigation of fault-scale heterogeneity, including asperities with contrasting frictional and hydraulic properties, and provides a platform to explore how three-dimensional fault structure influences aseismic slip, stress transfer, and earthquake nucleation during fluid injection. Benchmarking against established earthquake-cycle test cases validates the mechanical solver and establishes a baseline for ongoing fully coupled simulations. HydroMech3D offers a computationally efficient open-source tool to support experiment design, interpretation of near-fault observations, and assessment of induced seismicity in geoenergy applications.

How to cite: Wang, Z., Dal Zilio, L., Ciardo, F., and Rinaldi, A.: HydroMech3D: physics-based earthquake-cycle modeling of fluid-driven fault slip with realistic fault geometry, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-8716, https://doi.org/10.5194/egusphere-egu26-8716, 2026.

Accurate simulation of injection-induced seismicity requires to solve for strongly coupled hydro-mechanical physics describing processes acting over widely separated spatiotemporal scales, ranging from reservoir scale fluid diffusion to fault nucleation and rapid dynamic rupture. In this study, we present a monolithic hydro-mechanical dynamic framework based on the extended finite element method (XFEM) for modeling fluid-induced fault reactivation governed by rate-and-state friction. Faults are represented as embedded displacement discontinuities within a poroviscoelastic medium, enabling a consistent treatment of fault slip, unilateral contact constraints, stress-dependent permeability evolution, and fluid exchange between the fault and the surrounding porous matrix.

To overcome the computational cost associated with fully implicit time integration, we develop a hybrid implicit–explicit (IMEX) time-integration strategy. The implicit solver is employed during the quasi-static and nucleation phase, while an explicit scheme is activated only during the coseismic stage, once a prescribed slip-velocity threshold is exceeded. This adaptive solver switching allows accurate resolution of the dynamic rupture with substantial reduction of the computational effort. The approach is combined with adaptive time stepping to efficiently capture both slow interseismic evolution and fast seismic transients within a unified framework.

Numerical simulations of fluid injection into a faulted reservoir demonstrate that, despite unconditional stability, fully implicit schemes require minimum time steps comparable to the Courant–Friedrichs–Lewy limit to accurately resolve rupture nucleation and propagation. In contrast, the proposed IMEX formulation can reproduce fault slip evolution, stress redistribution, frictional weakening, seismic moment, and event magnitude with high fidelity, while reducing computational cost by approximately 60–77% relative to fully implicit simulations. Differences between the two approaches are primarily limited to peak slip velocities and rupture speeds, whereas rupture timing, accumulated slip and event-scale seismic metrics remain consistent.

The proposed XFEM-based IMEX framework provides a robust and computationally efficient tool for simulating injection-induced seismicity, offering a practical pathway toward reservoir scale simulations of coupled fault–fluid systems relevant to geo-energy applications and seismic hazard assessment.

How to cite: Sabah, M., Cacace, M., Hofmann, H., Blöcher, G., Reza Jalali, M., and R. Kivi, I.: A Hybrid Implicit–Explicit XFEM Framework for Fully Coupled Hydro-Mechanical Dynamic Simulation of Injection-Induced Seismicity, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17977, https://doi.org/10.5194/egusphere-egu26-17977, 2026.

Anthropogenic activity in the subsurface causes stress perturbations which can lead to the onset of seismicity. One of the notorious examples is the Groningen gas field in the northeast part of the Netherlands which is among the largest in Europe. Hydrocarbons have been produced there since 1963 until the field’s ultimate shutdown in October 2023. From December 1991 until January 1^st, 2026, total of 1561 events have been recorded in this area, with magnitude ranging from  to . The  events caused extensive damage to buildings and quite a societal unrest as well as scepticism towards subsurface operations in general. Considering, it is crucial to identify an envelope for safe utilization of the subsurface to be able to continue its usage for energy transition while limiting the risk of induced seismicity.

To be able to limit the risk of seismicity from subsurface operations, it is necessary to understand the non-stationary nature of induced seismicity, meaning the underlying physical causes of the observed spatial and temporal variations in event locations and frequency-magnitude distribution. This research is based on the hypothesis that the fault spatial distribution and geometry (dip angle, offset) in conjunction with operational parameters (pressure history, rates, injection temperatures) are the causal processes of the temporal and spatial variations in the Gutenberg-Richter parameters.

I will present the results from modelling production induced seismicity using the Groningen field as a study area. The results include synthetic earthquake catalogues obtained by modelling the event nucleation and magnitudes using a semi-analytical approach of slip weakening faults. For this model, fault geometry and pressure history serve as input. In order to obtain multiple catalogues spanning the full range of uncertainty, a Monte Carlo sensitivity analysis is conducted for different reservoir and fault properties. Subsequently, several statistical comparison tests of the simulated catalogue with the observed seismicity allows us to derive posterior estimates for our properties and provide crucial insight into how we are doing solving the puzzle of what is causing the observed spatiotemporal behaviour of induced earthquakes.

How to cite: Dvornik, L., Muntendam-Bos, A., Jansen, J. D., and Buijze, L.: Understanding the Non-Stationary Nature of Human-Induced Earthquakes and its Impact on Geothermal Energy Production, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17447, https://doi.org/10.5194/egusphere-egu26-17447, 2026.

The prediction of induced seismicity is a critical challenge for geological risk management and the safe operation of industrial facilities, such as geothermal projects. This study focuses on the Cooper Basin in Australia. We applied data science and machine learning techniques to analyze seismic time series, integrating two data sources: discrete seismological events (23,285 events) and continuous operational data sampled every 2 minutes (33,839 records).

The main objective was to develop machine learning models to predict, in future time windows of 10, 30, 60, and 90 minutes, two key variables: the number of seismic events or the maximum magnitude. The XGBoost and Random Forest algorithms were trained and compared. Model performance was evaluated using the R², RMSE, and MAE metrics, and their interpretability was analyzed using SHapley Additive exPlanations (SHAP).

The results demonstrate that both models generate predictions consistent with the observations, showing better predictive performance in the longer time windows (60 and 90 minutes). This approach provides a valuable framework for the monitoring and proactive risk assessment of geothermal operations.

How to cite: Garay Romero, L. R., Faenza, L., Garcia-Aristizabal, A., and Lombardi, A. M.: Application of Data Science and Machine Learning Techniques for the Prediction of Induced Seismicity, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12815, https://doi.org/10.5194/egusphere-egu26-12815, 2026.

How to cite: Sharia, T., Müller, B., and Rietbrock, A.: Stress perturbations and fault reactivation during cold fluid Injection - impact of hydraulic anisotropy , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-16611, https://doi.org/10.5194/egusphere-egu26-16611, 2026.

The increasing occurrence of injection-induced earthquakes has raised public concern and highlighted the importance of understanding subsurface processes to assess induced seismic hazards and risks. A feature of natural faults that has not received sufficient attention in induced seismicity modeling is their geometric roughness. We develop a simple physics-based model to investigate how fault roughness can control induced seismicity during fluid injection.

The first approach to modeling along-fault stresses prior to injection is to project the background stress tensor onto the rough fault. In this case, our models and theoretical analysis show that the apparent diffusivity of seismicity fronts can deviate significantly from the hydraulic diffusivity. Faults with realistic roughness generally display slow seismicity migration, producing apparent diffusivities far below the hydraulic values. Thus, seismicity fronts often lag behind the pressure front, especially at low background stresses and small roughness amplitudes. Only in the rare case of very rough faults stressed very close to failure, apparent diffusivity can exceed the hydraulic diffusivity, leading to seismicity fronts that outpace pressure fronts. 

The second approach to modeling along-fault stresses prior to injection is to simulate stress evolution after multiple tectonic rupture cycles. This ongoing work explores the resulting stress heterogeneity after multiple tectonic rupture cycles and examines whether seismicity migration follows the same trend as in the first approach, i.e., whether seismicity migration is generally slower than the pressure front on rough faults.

Apart from seismicity migration, the magnitude-frequency statistics are also analyzed. Along this single rough fault the frequency-magnitude distribution is bimodal. These results demonstrate how fault roughness and stress conditions control the induced seismicity through their influence on the criticality of the fault and stress transfer, and link long-term fault loading processes with short-term seismicity migration patterns in fluid injection scenarios.

How to cite: Lin, H.-F., Candela, T., and Ampuero, J.-P.: Injection-induced seismicity fronts and stress distribution on rough faults, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9772, https://doi.org/10.5194/egusphere-egu26-9772, 2026.

In 2017, an M_W 5.5 earthquake struck the Pohang region, representing the most damaging seismic event in South Korea, and has been linked in previous studies to hydraulic stimulation at the Pohang Enhanced Geothermal System (EGS) site. However, the relative roles of fluid injection, imposed stress state and fault-zone structure in nucleating this event remain a matter of debate, and the laboratory results presented here are intended to illuminate one mechanically plausible scenario rather than provide a unique causal explanation. Despite the scientific and societal importance of this earthquake, the frictional properties of rocks from the Pohang system are still poorly constrained. Here we experimentally characterize the frictional properties and slip behavior during fluid-induced reactivation of granodiorite wall rock powder and fault gouge recovered from the Pohang PX-2 borehole (~3.8 km depth). We first assessed the mineralogical assemblages of the two fault materials, which consist of mixtures of quartz, K-feldspar, plagioclase and phyllosilicates (mostly chlorite), with phyllosilicate contents varying between 15% and 23% for the wall rock and the fault gouge, respectively. We then measured friction, healing rate and the velocity dependence of friction for both materials under water-saturated conditions at normal stresses of 20–100 MPa using the BRAVA apparatus hosted at the National Institute of Geophysics and Volcanology (INGV). We performed frictional experiments in a double-direct-shear configuration, using a protocol consisting of a run-in at 10 µm/s, slide-hold-slide tests (SHS; hold times ranging between 3 and 3000 s), velocity-stepping tests (VS; velocities ranging between 0.3 and 300 μm/s), and fluid-injection tests (pore-pressure increases of 0.25 MPa every 5 min). Steady-state friction coefficients for both materials fall within the Byerlee range (μ ≈ 0.55–0.62). SHS tests reveal that both fault gouge and wall rock exhibit relatively high healing, with β in the range ≈ 0.0046–0.0092. Conversely, velocity-stepping tests reveal that, over the tested stress and velocity range, the wall rock has a slightly velocity-weakening to neutral behavior (a–b = −0.0007 to 0.0020), while fault gouge is predominantly velocity-neutral to strengthening (a–b = 0.0005 to 0.0028). Additional fluid-injection experiments indicate that, despite these slight differences in frictional properties, both the fault gouge and the wall rock can be reactivated under elevated pore pressure, with slip accelerating from creep to millimetre-per-second rates. Accompanying microstructural observations will examine whether differences in grain-size reduction, shear localization, or porosity evolution account for the similar reactivation behavior despite the slightly contrasting frictional properties. Overall, these measurements will help quantify how lithological heterogeneity, rate-and-state parameters, and pore-pressure evolution govern slip stability and the nucleation potential of injection-induced earthquakes in geothermal settings, with important implications for induced-seismicity hazard assessment in granitic EGS reservoirs.

How to cite: Woo, S., Volpe, G., Coppola, L., Collettini, C., and Son, M.: Frictional properties and fluid-induced reactivation of fault rocks from a granitic EGS reservoir, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12316, https://doi.org/10.5194/egusphere-egu26-12316, 2026.