Strain localization – manifesting as narrow shear bands in brittle rock masses under compressive stresses – is a critical yet contentious phenomenon in earthquake dynamics. Understanding the mechanisms driving this localization is essential, as it influences fault weakening and energy dissipation during seismic events. In this study, we investigate the spontaneous formation of highly localized shear zones, with thicknesses less than 1 cm, within fluid-saturated granular fault gouge using one-dimensional poromechanical numerical simulations. We utilize H-MEC (Hydro-Mechanical Earthquake Cycles), a novel two-phase flow numerical framework that couples solid deformation with pervasive fluid flow (Dal Zilio et al., 2022). This continuum-based model employs a staggered finite difference–marker-in-cell method, accounting for inertial wave-mediated dynamics and fluid flow in a poro-visco-elasto-plastic compressible medium. Global Picard iterations and adaptive time stepping enable accurate resolution of both long- and short-term processes, spanning timescales from years to milliseconds. Our simulations incorporate two frictional laws: the conventional rate-and-state-dependent friction and a newly developed rate-dependent friction. The rate-and-state model, while effective in various contexts, proves ill-posed in localization scenarios due to the absence of a diffusive term in the state variable, causing localization to collapse into a single grid cell regardless of resolution. Conversely, the rate-strengthening friction model with pore pressure diffusion governs localization through fluid pressure diffusion within the poroelastic medium. This approach eliminates the need for classical phenomenological parameters such as the evolutionary effect (b) and the characteristic slip distance (L), resulting in shear zones with finite thicknesses less than 1 cm for slip velocities on the order of meters per second. Additionally, under lower effective normal stress, the model predicts slow-slip events that localize over broader shear zones ranging from 4 to 6 meters. We further perform a linear stability analysis to delineate the poromechanical conditions that drive fluid-induced earthquakes. Our findings suggest that strain localization serves as a dynamic fault-weakening mechanism during seismic events, where the formation of shear bands reduces sliding stress and decreases frictional energy dissipation along the fault. This study provides a physically robust representation of strain localization, enhancing our understanding of the precursory processes leading to earthquakes and potentially informing early warning systems.
- Dal Zilio, L., Hegyi, B., Behr, W., & Gerya, T. (2022). Hydro-mechanical earthquake cycles in a poro-visco-elasto-plastic fluid-bearing fault structure. Tectonophysics, 838, 229516 (https://doi.org/10.1016/j.tecto.2022.229516).
How to cite: Dal Zilio, L. and Gerya, T.: Poromechanical modeling of strain localization during earthquake rupture, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14136, https://doi.org/10.5194/egusphere-egu25-14136, 2025.
The December 5, 2024, Mw 7.0 Off-shore Cape Mendocino earthquake, the largest in California since the 2019 Ridgecrest events, struck approximately 70 km southwest of Ferndale, near the Mendocino Triple Junction (MTJ) where the Pacific, North American, and Juan de Fuca/Gorda plates converge. The MTJ is California's historically most seismically active region, which has experienced multiple up to Mw 7 earthquakes over the past few decades. Understanding the dynamics of this earthquake is essential to better understand regional seismicity, structural and stress heterogeneity, and the resulting stress redistribution onto two adjacent high-hazard fault systems, the Cascadia subduction zone to the North and the San Andreas transform fault system to the South.
Rapidly characterizing large earthquakes is vital for effective disaster response and seismic and tsunami risk mitigation. Current assessments often rely on rapidly generated static or kinematic finite-fault models derived from geodetic data, teleseismic body waves, CMT solutions, scaling relationships, and regional waveforms (e.g., Hayes, 2017; Goldberg et al., 2022). Such models can also inform 3D dynamic rupture simulations, providing a physics-based perspective on earthquake behavior (e.g., Jia et al., 2023; Hayek et al., 2024).
In this presentation, we apply a new automated workflow to rapidly characterize the rupture dynamics of the recent Mw 7.0 Off-shore Cape Mendocino earthquake. We compare several reference finite-fault models, including those from the USGS, SLPINEAR/GeoAzur, and a new static geodetic inversion, to automatically constrain 3D dynamic rupture simulations. An ensemble of dynamic rupture models is explored, informed by the stress change of each finite-fault model, respectively. Preferred dynamic rupture models are automatically selected based on matching regional waveforms and moment rate release. This simple workflow can systematically assess the dynamic viability of kinematic slip models.
While the static geodetic inversion reveals a main ruptured asperity that broke a strongly coupled section of the fault and rupture ceasing at a previously identified creeping section of the fault, the associated dynamic rupture models cannot explain the complex rupture dynamics imprinting, e.g., on moment rate release. Instead, we find that a higher degree of smaller-scale initial stress complexity, such as resulting from the SLIPNEAR model, is required to explain observations. The resulting asymmetric rupture dynamics present challenges to rapid data-driven analyses and have significant implications for understanding future earthquakes in the Mendocino Triple Junction region.
How to cite: Gabriel, A.-A., Ulrich, T., and Magen, Y.: Automated 3D dynamic rupture simulations for rapid characterization of large earthquakes: Application to the December 5, Mw 7.0 Off-shore Cape Mendocino earthquake, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14836, https://doi.org/10.5194/egusphere-egu25-14836, 2025.
Standard characterization of weak earthquakes based on observed source spectra relies on corner frequency and radiated energy estimates only as general source characteristics. However, the routine analysis cannot provide physical details including the heterogeneity of the rupture process. Contrarily, those can be inferred by dynamic rupture modeling that combines elastodynamics and friction law; however, this approach is almost exclusively utilized for large events. Here, we develop a novel Bayesian dynamic inversion of apparent (station-specific) source spectra up to 25 Hz with removed path and site effects for slip-weakening friction parameters heterogeneous along a finite-extent planar fault. The approach is demonstrated through two real-world applications with distinct source radiation: a directive and a nondirective Mw~4 event in Central Italy. We find that, despite the event’s small size, the heterogeneity in the rupture propagation down to the smallest scales (~100 m) is necessary to accurately model the high-frequency radiation of the observed spectra. The inversion demonstrates the ability to resolve various mean source parameters with reasonable uncertainty and thus reliably characterize weak events. The study also pinpoints that the standard seismological estimates based on the omega-squared spectral model of the earthquake source can lead to inaccurate results when additional earthquake complexity, like directivity, is present. Finally, the dynamic inversion reveals fractal spatial characteristics of the governing dynamic parameters, which are essential for reproducing the observed high-frequency apparent source spectral decay. Such advanced studies promise to unravel so-far elusive small-scale characteristics of earthquake ruptures.
How to cite: Gallovič, F., Sgobba, S., and Valentová Krišková, Ľ.: Broadband earthquake source characterization by dynamic rupture inversion of apparent source spectra for two Mw~4 events, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3180, https://doi.org/10.5194/egusphere-egu25-3180, 2025.
Over the past two decades, advancements in seismological and geodetic observations have uncovered a diverse range of aseismic fault slip behaviors occurring at various depths, significantly contributing to the seismic cycle moment budget. One notable phenomenon is Episodic Slow Slip Events (SSEs), which occur in many subduction zones, at depths greater than the coupled seismogenic zone and shallower than the creeping zone. Lasting from days to weeks, SSEs are detected by GNSS stations through a reversal in station velocity, with amplitudes reaching several centimeters. The small deformation signal, combined with the rapid decrease in resolvable fault information with depth, suggests that data-driven models have limited constraints on the physics governing these events. Forward modeling of SSEs using rate-and-state friction laws offers valuable insights, but it is computationally intensive and constrained by the fast oscillating processes inherent to the system, limiting exploration of the controlling physics. In this study, we employ a two-dimensional subduction fault model with laboratory-constrained rate-and-state friction parameters to simulate SSEs in a Cascadia-like setting. We apply a model-order reduction technique to alleviate computational demands, facilitating detailed parametric studies of SSEs dynamics.
To model slow slip events, we use the open-source seismic cycle and aseismic slip (SEAS) simulation framework, Tandem (Uphoff et al., 2023). The Cascadia subduction zone is represented as a 1D planar fault that dips at an angle of 10°. We introduce SSEs into the system by creating a zone of low effective normal stress (σn) in the region where the fault transitions from slip weakening (up dip) to slip strengthening (down dip). The ratio of this low effective normal stress zone (W) with the critical nucleation size (h*), were found to control both the occurrence and rate of SSEs (e.g., Liu & Rice, 2007,
2009) and is given by
,
where a, b, and Dc are friction parameters.
To rigorously explore the parametric space controlling SSEs (W - width, σn- normal stress, a, b, Dc - friction parameters), we utilize a non-intrusive, data-driven Reduced Order Model (ROM). First, we transform the spatial fault distribution of simulated slip-time trajectories (time history data) into a latent space vector representation through spline interpolation across the slip-rate to state variable domain. This process compresses the simulated slip-time history by 90% allowing for efficient interpolation between latent state vectors. Next, we employ Proper Orthogonal Decomposition ROM using Radial Basis Functions to interpolate the latent state vectors over the parameter space. This two-step model order reduction approach significantly reduces the computational cost of obtaining slip-time trajectories compared to a traditional SEAS simulation, decreasing the run-time from thousands of CPU hours to just seconds.
This study emphasizes the potential of ROMs to enhance our understanding of earthquake physics, particularly the mechanisms behind SSEs. This advancement paves the way for improved models of seismic cycle dynamics and hazard assessments in subduction zones. By combining computational efficiency with physical insight, ROMs offer unique opportunities to explore the complex interplay of physical parameters that govern subduction seismogenic.
How to cite: Magen, Y., Gabriel, A.-A., and May, D. A.: Model-order reduction applied to rate-and-state friction earthquake cycle models uncovering the physics driving slow-slip events., EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15047, https://doi.org/10.5194/egusphere-egu25-15047, 2025.
Rupture speed plays a critical role in earthquake dynamics, seismic energy release, and ground shaking characteristics. While variations in rupture speed of earthquake fault slip from fast to slow are well-documented in nature and in the lab, the responsible mechanisms are not fully understood. Here we address the physical mechanisms for variations in rupture speed using an array of strain gage rosettes in direct shear experiments to estimate the rupture speed of stick-slip instabilities. The experiments were conducted with applied normal stresses spanning one order of magnitude, ranging from 2 to 20 MPa. High-speed records of shear strains at 13 equidistant locations along 15-cm-long granite faults were analyzed to understand the effects of normal stress on rupture dynamics. Our data follow the expectation that higher normal stress generally promotes faster rupture speeds, consistent with observations from natural fault systems.
Our analysis reveals the interplay between stress conditions, stored elastic energy, and fault behavior. The experiments provide insights into how changes in normal stress affect the propagation of frictional rupture along a simulated fault surface with a thin layer of moisturized quartz gouge (Min-U-Sil, 40). A concise relation between normal stress and rupture speed based on linear elastic fracture mechanics is derived to explain our observations.
Fracture energy scales linearly with normal stress, which tends to reduce rupture speed as normal stress increases. However, the greater difference between peak and residual strength at higher normal stresses allows for more energy to be released during fault slip. Thus, as normal stress increases, the energy release rate, which scales quadratically with normal stress, outpaces the linear increase in fracture energy, leading to higher rupture speeds.
Our results provide important information for seismic hazard assessment and the development of more accurate rupture models for earthquake forecasting. By clarifying the role of normal stress in modulating rupture speed, our work illuminates the complex interactions between stress conditions and earthquake rupture dynamics. Overall, our data underscore the significance of considering normal stress variations in seismological methods to improve earthquake estimations and hazard assessments.
How to cite: Ke, C.-Y., Chang, G., McLaskey, G., and Marone, C.: Earthquake Rupture Speed Dependence on Normal Stress in Laboratory Experiments, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8191, https://doi.org/10.5194/egusphere-egu25-8191, 2025.
Laboratory shear experiments provide valuable insights into the physical processes driving earthquakes. While the nucleation of lab earthquakes on bare rock surfaces has been extensively studied, the preparatory processes within fault gouge volumes remain poorly understood. To address this gap, we performed stick-slip experiments on a 5 × 5 cm² fault using granular quartz gouge under single and double direct shear configurations at normal stresses ranging from 40 to 50 MPa.
Our experiments reproduce the full spectrum of fault slip behaviors, from stable creep to slow and fast slip events. As strain accumulates and the internal structure of the gouge evolves, seismic cycles exhibit complex sequences, often with slow-slip foreshocks preceding rapid and energetic stress drops. We studied events with slip velocities spanning nearly two orders of magnitude (0.1 to 10 mm/s), highlighting the coexistence of slow and fast slip events on the same fault under identical boundary conditions. We used eddy-current sensors to track volumetric strain during the inter- pre- and co-seismic phases. Simultaneously, acoustic emissions (AEs) were recorded at 6.25 MHz using piezoelectric sensors embedded in the loading blocks. Event detection is performed with a custom-trained deep-learning model based on the PhaseNet model developed for seismic data. The temporal evolution of AEs, coupled with waveform similarity analysis, which serves as a proxy for the spatial progression of AEs, helps to constrain the preparation and nucleation processes of slip events characterized by different velocities. Despite the continuum between slow and fast slip modes revealed by mechanical and acoustic scaling, our results show that the acoustic behavior of slow and fast slip events in a well-developed gouge fault differs. Slow-slip events are characterized by longer durations and temporally distributed swarms of small AEs, while fast-slip events exhibit shorter durations and concentrated bursts of energetic AEs. Supported by seismological estimates of AE source parameters from calibrated piezoelectric sensors, we propose a micromechanical model in which the progressive failure of asperities, signaled by increasing AE rates, drives seismic slip to a critical nucleation point—reached only for fast-slip events and not for slow-slip events. Our results are framed within the rate-and-state framework, with nonlinear time-series analysis tools used to evaluate the predictability of laboratory seismic cycles.
Finally, to address the long lasting question of upscaling earthquake physical processes, we performed additional experiments on a larger 77 × 8 cm² fault at lower normal stresses with the same gouge material. High-resolution displacement and acoustic measurements in these tests provided detailed insights into the spatial evolution of slip, which can only be fully resolved with larger fault samples. This enables us to better constrain our previous results and investigate the impact of fault size on lab earthquake nucleation within the fault gouge volume, laying the foundation for upscaling laboratory observations to larger-scale experiments and, ultimately, to natural faults.
How to cite: Mastella, G., Pignalberi, F., Volpe, G., Marone, C., Corbi, F., Collettini, C., Giorgetti, C., and Marco, S.: Coexisting Slow-to-Fast Laboratory Earthquakes: Insights into Nucleation Processes in Fault Gouge Across Spatial Scales, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13100, https://doi.org/10.5194/egusphere-egu25-13100, 2025.
Inferring the spatio-temporal distribution of fault slip during earthquakes from seismological data is challenging due to the non-uniqueness, ill-posed nature, and high dimensionality of the inverse problem. Finite source inversion often relies on simplifying assumptions, and in the absence of ground truth observations it is difficult to assess the reliability of the resulting slip history. Therefore, evaluation of the inversion performance is typically limited to synthetic tests.
Laboratory earthquakes provide an alternative approach to address these challenges by providing "simulated real data" under controlled conditions with relatively well-constrained solutions. In this study, we are analyzing frictional ruptures obtained using a biaxial apparatus. The setup uses three independent vertical pistons to apply heterogeneous normal loads, and one horizontal piston to apply shear load along a 40 cm long fault interface. During each rupture, acceleration is recorded at 20 receivers positioned horizontally or vertically along the fault. These acceleration measurements are integrated twice to obtain displacements, which are then used to invert the slip history.
One of the critical aspects of slip inversion is the accurate definition of Green's function, which depends on numerous assumptions, including source geometry, medium properties and computational constraints, etc. To test the influence of Green’s functions, we first conduct a static inversion of the final slip using two distinct solutions: (1) the analytical Okada solution for a semi-infinite half-space and (2) finite-element simulations using COMSOL, which incorporate detailed information about our setup.
To perform the inversion, we use the Metropolis sampling algorithm, which provides a range of solutions, essential for assessing the uncertainty in our results and addressing the issue of non-uniqueness.
Our results demonstrate that while both solutions provide a good fit of the observed data, only the Green’s function obtained considering the geometry of the fault system and the loading conditions allows to obtain a solution that is consistent with the ground truth, measured individually during the experiments using laser sensors.
Using COMSOL-generated Green’s functions, our quasi-static inversion are able to reconstruct the evolution of the rupture, even for complex ruptures involving deceleration and subsequent acceleration, which are measured independently using high-speed photoelastic measurements. Our results demonstrate the robustness of our inversion procedure, and that accelerometers can be used to invert the evolution of fault slip along the fault, even during complex propagation sequences.
How to cite: Arzu, F., Twardzik, C., Fryer, B., Xie, Y., Ampuero, J.-P., and Passelegue, F.: Static and Quasi-Static Inversion of Fault Slip During Laboratory Earthquakes, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16507, https://doi.org/10.5194/egusphere-egu25-16507, 2025.
Earthquake swarms may be driven by slow processes within fault zones, such as fluid migration, or silent slip events. A prominent example is the 2024 Mw 7.5 Noto earthquake in Japan, which was preceded by a significant seismic swarm, likely triggered by upward fluid migration (Wang et al. 2024). These swarms provide insights into dynamic fault processes, and there is thus a need to detect smaller earthquakes for a clearer understanding of the temporal evolution of seismicity and the mechanisms driving fault activity (Shearer et al., 2022).
To address these challenges, we used a dense array of 300 seismic nodes deployed for three years at the Piñon Flat Observatory (South.Cal.) along the San Jacinto Fault, one of the most seismically active areas in California. Using advanced slant stacking techniques tuned to the fault geometry at crustal P- and S-wave velocities, we significantly enhanced our detection capabilities down to approximately magnitude -2. By detecting four times as many events as the standard USGS catalog, this allowed us to highlight three distinct swarm episodes that were not identified before. These episodes exhibit a characteristic progression: an initial activation phase, a steady state culminating in a peak, followed by a final decay.
We further investigate the magnitude distribution and spatial migration to propose a possible driving mechanism. This approach can be extended to other fault zones to unveil hidden fault activity.
How to cite: Higueret, Q., Brenguier, F., Aurélien, M., Sheng, Y., Vernon, F., Hollis, D., Aubert, C., and Ben-Zion, Y.: Detecting hidden seismic swarms using a 300 nodal long-term array, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11895, https://doi.org/10.5194/egusphere-egu25-11895, 2025.
Large continental strike-slip earthquakes usually present a complex rupture trace composed of several segments separated by discontinuities. Fault geometry may interact with stress concentration and rupture propagation and have an influence on the distribution of coseismic slip and the termination of ruptures. However, other factors can influence slip distribution and rupture propagation, such as the rheology of the fault and the bulk, which can be affected by crustal damage from earthquakes, and the initial stress state, which notably depends on slip history. Here we characterize these complex interactions in the case of the 2023 Mw 7.8 Pazarcık and Mw 7.5 Elbistan earthquakes that ruptured several segments of the East Anatolian Fault Zone (EAFZ) in south-central Turkey. We use a Bayesian framework to model coseismic slip in a layered elastic medium with geodetic data. In our model, most of the slip occurs above 15 km depth, with a shallow slip deficit. Shallow slip decreases at geometrical complexities supposedly due to off-fault deformation in these highly damaged areas. The termination of both ruptures also correspond to geometrical complexities. Aftershocks spread in wide fan-shaped damage zones around the southwestern tip of both ruptures, whereas they are more focused on the main fault or on subparallel planar structures to the northeast. We also build 2-year postseismic Sentinel-1 InSAR displacement time series and find that the segments with relatively strong shallow afterslip are located at the northeastern end of the rupture trace for both earthquakes. Preseismic InSAR time series computed by the FLATSIM service also show shallow creep on the Pütürge segment northeast of the Pazarcık rupture and more distributed deformation in the rest of the EAFZ. These observations suggest that deformation in the EAFZ is more localized to the northeast and more distributed to the southwest before, during and after the 2023 earthquakes. Our postseismic displacement time series also shows shallow creep on several secondary faults. Comparing these deformations to the stress changes caused by the 2023 earthquakes can give interesting rheological insights and help refine our understanding of the complex interactions between fault geometry, crustal damage and slip.
How to cite: Dérand, P., Jolivet, R., Raimbault, B., Dalaison, M., Hollingsworth, J., Klein, E., Fukushima, Y., Okur, Y., and Thomas, M.: Interactions between fault geometry, crustal damage and slip before, during and after the 2023 Kahramanmaraş earthquakes (Turkey) , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11400, https://doi.org/10.5194/egusphere-egu25-11400, 2025.
Reliable determination of the fault locking depths and slip velocities is of great importance for the assessment of maximum magnitude of a potential earthquake and in performing seismic hazard analyses, especially on large earthquake-prone fault zones. The North Anatolian Fault Zone (NAFZ), a nascent transform plate boundary between the Eurasian and Anatolian plates is characterized by high stress accumulation. The east-west trending earthquake sequence, starting with the M7.9 Erzincan earthquake in 1939 and ending with the M>7 İzmit and Düzce ruptures in 1999 in the NAFZ, revealed a westward migration of seismic energy release. The 1999 M7.2 Düzce earthquake occurred 3 months after the 1999 M7.4 İzmit earthquake, drew a particular attention to the Düzce region as it occurred in the east, reversing the westward migration movement, and resulted in an eastward supershear rupture. In this study, we aimed to analyze the interseismic locking depth and the surface creep parameters in Düzce region by using geodetic and geophysical data. For this purpose, we first examined the spatio-temporal variation of the surface deformation along the Düzce Fault segment of the NAFZ in order to better understand the interseismic loading parameters, and possible effect of the creep. Therefore, we implemented the InSAR Small Baseline Subset time series analysis technique using Sentinel-1 InSAR data for both ascending and descending orbits from 2017 to 2022 to estimate horizontal and vertical displacements and also to calculate the locking depth. Consistent with the previous studies, our study showed that the slip rate in Düzce segment of the NAFZ was ~25 mm/yr. Besides, we incorporate various geophysical properties (e.g. geo-electric resistivity, seismic velocity) through previously obtained by 2D and 3D modeling of magnetotelluric and seismological observations with InSAR-based surface deformation in and around the study area to have an insight into the impact of petrophysical rock properties on seismogenic zone characteristics.
This project is funded by the Bogazici University with the BAP Project No SUP-18161.
How to cite: Zoroğlu, Ç. S., Kaya Eken, T., Havazlı, E., Bletery, Q., and Özener, H.: Locking Depth and Interseismic Slip Rate Analysis of the North Anatolian Fault Zone: Insights from Geodetic and Geophysical Techniques in the Düzce Region, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-631, https://doi.org/10.5194/egusphere-egu25-631, 2025.
Morocco's High Atlas is an intracontinental orogenic belt located at the northern edge of the West African Craton (WAC). This major belt extends from what is now the Atlantic margin of Morocco to the Mediterranean coast of Tunisia, spanning the Sahara Atlas in Algeria. Within the context of the convergence of Nubian and Eurasian, GPS measurements across the High Atlas in Morocco indicate a very low surface deformation rate (<1mm/year). This study examines co-seismic deformation through InSAR modeling, Coulomb stress change analysis, and aftershock distribution. Additionally, it explores the crust-uppermost mantle's structure beneath the High Atlas and adjacent regions using ambient noise seismic tomography and P-wave coda autocorrelation to analyze the earthquake sequence within the context of regional geodynamics. Our InSAR modeling supports a NE-SW trending reverse faulting mechanism with a dip of ~69◦ towards the Northwest. The maximum slip is about 2m at an average depth of ~24km. The fault extends over 20km length from 10km to 40km depth. The reactivated fault stands beneath a flower structure and its surface expression coincides with Tizi n’Test Fault mapped fault. Analysis of Coulomb failure stress changes from the mainshock on nearby major faults showed a stress increase of ~5 bars above the rupture, aligning with areas of aftershock activity. Ambient seismic wave tomography and teleseismic P-wave coda autocorrelation from a network of 21 broadband stations allowed us to constrain the depths of the Moho and Lithosphere-Asthenosphere Boundary (LAB) beneath the Western High-Atlas. Our results showed that notable increase in Moho depth, with measurements ranging from 45 to 50 km in the epicentral area, therefore compensating the High Atlas topography. Conversely, in the surrounding low topography areas, the Moho depth remains relatively shallower, averaging around 35 km. Interestingly, we found that LAB is shallower in the region where we observed a thicker Moho, and vice versa. The 2023 High Atlas blind earthquake clearly ruptured at least 2/3 of a 45-50 km thick crust sitting above a 30-35 km thin lithospheric mantle.
How to cite: Ouammou, H., Aoudia, A., Javed, F., Thapa, H. R., Tahayt, A., and Wang, H.: The Morocco High Atlas 6.8 magnitude 2023 event: New insights from geodetic and seismological modeling, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18488, https://doi.org/10.5194/egusphere-egu25-18488, 2025.
The convergence of the Indian and Eurasian plates is accommodated through stick and slip on the detachment or Main Himalayan Thrust (MHT). The part of the MHT that lies under the Outer and Lesser Himalaya is seismogenic and slips episodically and accumulates strain during the interseismic period when it is locked, which is released during the earthquakes through sudden slip on the MHT. The MHT further north slips aseismically. Out of the three known seismic gaps (Kashmir gap, Central gap, Assam gap), we report results of GPS measurements from Jammu-Kashmir region. The 250 km long segment of the Kashmir Himalaya, known as the Kashmir Seismic Gap, has not experienced a major earthquake since 1555, although a possible event of magnitude Mw ~6 was recorded west of Srinagar in 1885. This region appears to be an anomaly, as the width of the Main Himalayan Thrust (MHT) is suggested to exceed 160–170 km compared to a width of approximately 100 km in other parts of the Himalaya. The seismic activity in the Kashmir region is diffused and does not indicate the location of the locking line. Additionally, the 3,500-meter contour, which encircles the Kashmir Valley and extends to the Pir Panjal and Zanskar Himalayas, does not guide its position either. We analysed GNSS data from 22 sites along with the published data from the region. To assess the convergence in the Jammu-Kashmir region, we examined two arc-normal profiles. Assuming the MHT extends to the Main Frontal Thrust (MFT) and that strain accumulation is uniform in the locked shallow MHT, we estimated the locking width in the Kashmir region (Profile-1) to be 169 ± 10 km, with a total convergence rate of 13.7 ± 1 mm/yr. Similarly, in the Jammu-Himachal region (Profile-2), we estimated the locking zone to have a width of 108 ± 10 km with a total convergence rate of 17.8 ± 1 mm/yr. The spatial variation of locking in the Kashmir and Jammu-Himachal regions shows high coupling. The Kashmir region's intermontane valley demonstrates lower coupling than surrounding high-coupling zones. While both areas show strain accumulation, the deformation and convergence in the Kashmir Himalaya are more distributed compared to the central Himalaya.
How to cite: Mohanty, A. and Gahalaut, V. K.: Seismic Hazard and seismogenesis in Kashmir Himalaya , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1035, https://doi.org/10.5194/egusphere-egu25-1035, 2025.
In early February 2018, a series of earthquakes occurred offshore Hualien. At least 16 M 4.5 earthquakes, including one M 6.1 event, were observed at depths between 3 and 15 km. On 4 February 2018, a moderate-magnitude earthquake (Mw ~6.0–6.2) struck offshore Hualien, followed two days later by a Mw 6.4 event approximately 5 km away. Although occurring in the near-source region of the 6 February Mw 6.4 earthquake and only two days before, this event has been overlooked. For the 4 February 2018 earthquake, we used GNSS and strain time series data to perform a finite source inversion. This analysis shows the rupture of a ~ 25 km × 15 km asperity located on the shallow section (5 to 8 km depth) of a subhorizontal fault plane with a dominant right-lateral strike-slip mechanism. We also performed the seismological analysis of the aftershocks that occurred between these two days. Aftershocks were clustered in the northeastern and southwestern directions of the 4 Feb event, with northeastern aftershocks situated within the maximum coseismic slip of 0.25–0.30 m, while southwestern aftershocks show a lower slip of 0.05 m. The total seismic moment of the earthquake is 2.4 * 1018 which corresponds to an earthquake of magnitude 6.18. The average static stress drop Δσ of the rupture is about 4 MPa, which is slightly larger than the characteristic earthquake stress drop of 3 Mpa. We also conducted a Coulomb stress analysis to analyze whether the southwestern aftershocks are linked to potential afterslip. The static Coulomb stress changes resolved onto the hypocenter of the 6 February M6.4 Hualien earthquake are approximately 0.1 MPa, which suggests that the latter may have been clock-advanced through static stress transfer.
How to cite: Rajkumar, R., Canitano, A., and Lin, H.-F.: The 4 February 2018 Mw 6.2 Hualien earthquake (Taiwan): geodetic finite-fault rupture model and its potential impact on the occurrence of the 6 February Mw 6.4 event, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11224, https://doi.org/10.5194/egusphere-egu25-11224, 2025.
A method is proposed for determining the most likely fault plane of an earthquake from the modeling of the directivity effect of the seismic source, which affects the half- and total- duration (rise time and pulse width) of the first P-wave at a fixed takeoff angle. The method is based on previously inferred relationships relating rise time and pulse width to source parameters and Q (Zollo and de Lorenzo, 2001; de Lorenzo et al., 2004; de Lorenzo and Zollo, 2006).
These studies employed a Sato-Hirasawa (1973) circular crack model with a constant rupture velocity, set to 90% of the S-wave velocity at the source. A set of nonlinear equations relate rise time and pulse width to the anelastic intrinsic factor (Q) and the source characteristics (strike, dip, fault radius). These equations allow determining which of the two solutions of the focal mechanism better reproduces the observed trend of rise time (and pulse width) vs. travel time.
Using a constant Q in the medium causes a problem of underfitting between theoretical and observed rise times (Filippucci et al., 2006). To overcome this problem, the novel technique accounts for the variability of Q along each source-to-receiver path.
In this new approach, after inferring an average Q for the study area, a line-search inversion is carried out to estimate the fault radius, for each of the two possible fault planes of each event. Differences between data and their theoretical estimates can be attributed to heterogeneity of Q along each path. When using data coming from different events, the retrieved Q differences can be averaged to reduce the misfit between data and their theoretical estimates. Data can therefore be corrected for the path effect. An iterative inversion procedure is stopped when the misfit between data and their theoretical estimates is less than the error of data.
The technique has been applied to a dataset of 21 earthquakes from the Amatrice-Visso-Norcia Sequence (2016) with ML ranging between 3 and 4, hypocentral depths between 8 km and 13 km and known focal mechanisms. For data selection, analysis, and plotting, custom software was developed in Matlab for this study.
How to cite: Perrucci, F., de Lorenzo, S., and Zollo, A.: A method for the determination of the fault plane from the analysis of data in the time domain, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12660, https://doi.org/10.5194/egusphere-egu25-12660, 2025.
Earthquake rupture directivity and source parameters provide key information to understand earthquake physics and constrain seismic hazard and risk, which is particularly important for faults near urban areas. We calculate rupture directivity and source parameters for earthquakes in the Marmara region, NW Türkiye, where a M > 7 earthquake is overdue. First, we analyze directivity patterns for 31 well-constrained 𝑀L > 3.5 earthquakes along the Main Marmara Fault, in close proximity to Istanbul. We calculate source mechanisms with a waveform modeling approach and analyze directivity from apparent source-time functions using empirical Green’s functions. Most of the strike-slip earthquakes to the west of the Princess Islands segment display a predominantly asymmetric rupture towards the east with a median directivity trending 85°, consistent with the fault strike. Consequently, earthquake ground shaking may be more pronounced towards Istanbul. This may hold potentially for a future large earthquake on the Main Marmara Fault. Second, we estimate source parameters for >1.500 earthquakes with ML [1.0, 5.7] over the last 15 years. Using a spectral fitting approach, we constrain the corner frequency, seismic moment and quality factor, and calculate the static stress drop. Statistically significant spatial variations of stress drops are observed along some segments, with locally lower values in partially creeping fault zones surrounding earthquake repeaters representing a proxy for aseismic slip. The recent occurrence of M > 5 earthquakes along the overdue Main Marmara Fault did not lead to significant stress drop variations, implying that those moderate events did not significantly modify the stress level in this region which is relevant given that a M > 7 event is pending. Combined, our results underline the importance of including rupture directivity effects and source parameters when estimating seismic hazard and risk near urban areas.
How to cite: Martínez Garzón, P., Chen, X., Kwiatek, G., Bindi, D., Ben-Zion, Y., Bohnhoff, M., and Cotton, F.: Eastward rupture directivity and source parameters variations in the Marmara region: implications for a future M > 7 earthquake, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7268, https://doi.org/10.5194/egusphere-egu25-7268, 2025.
Resolving the finite-source attributes of small to moderate earthquakes is essential for advancing our understanding of fault properties and earthquake physics. Time-domain approaches for source characterization are often unstable due to challenges in deconvolving path effects using co-located smaller earthquakes as empirical Green’s functions, as these methods are highly sensitive to seismic arrivals and noise levels. Frequency-domain analysis provides a more stable alternative by discarding phase information. However, the conventional approaches typically constrain only the rupture dimension and directivity using the corner frequency of apparent source spectra, assuming circular or linear rupture models, respectively. In this study, we present an advanced frequency-domain analysis by employing an elliptical rupture model and introducing a Bayesian inversion framework that fits the full apparent source spectra to estimate rupture length, width, propagation velocity, and directivity ratio. This approach fully leverages the observed spectra and widely deployed dense seismic arrays to constrain the rupture process beyond corner frequency analysis. Applied to several earthquakes with magnitudes ranging from 2 to 6, our method produces results consistent with seismic second moments and finite-fault inversion techniques. Unlike traditional methods, our framework estimates more accurate dynamic parameters such as stress drop, apparent stress, rupture energy, and critical slip distance without relying on circular rupture assumption. The method is easily automatable, enabling the development of an extensive earthquake catalog that includes dynamic and kinematic parameters for small to moderate events, thereby supporting statistical analyses of fault properties and deepening our understanding of earthquake source mechanisms.
Figure 1. The finite-source attributes inversion for the Mw 2.3 event in 2016 Oklahoma. (a) The 2-D posterior joint probability density resulting from the Bayesian inversion. (b) The parameters of the ellipse rupture model. a and b are the semi-major and semi-minor axes of the ellipse rupture model, respectively. defines the rupture directivity. represents the rupture velocity, controlling the rupture front propagation. denotes the directivity ratio, which controls the rupture's initial position. (c-d) The observed and predicted apparent source spectra, respectively. The estimated parameters of the rupture model are a = 85 ± 5 m, b = 46 ± 5 m, θ = 80 ± 1 °, vr = 2.19 ± 0.02 km/s, and e = 0.18 ± 0.01, which are consistent with previous studies (Fan et al., 2018).
How to cite: Liu, Z., Meng, H., Zhou, M., Ju, F., and Yu, C.: An Advanced Frequency Domain Methodology for Resolving Finite Source Attributes of Small to Moderate Earthquakes: Beyond Corner Frequency Analysis, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17061, https://doi.org/10.5194/egusphere-egu25-17061, 2025.
Stress release is a fundamental physical parameter of earthquakes. Yet the direct measurement of stress drop at depth is practically difficult; different estimations of stress drop often give different values, likely because of biases and uncertainties of the various methods and dataset. To better understand these biases and uncertainties, it is important to estimate stress drop with a wide range of approaches, which have different potential errors. So here we use and attempt to validate a different approach to determining stress drop, focusing on variability in the apparent source time functions observed at a range of stations.
Specifically, we estimate stress drops of earthquakes in the 2019 Ridgecrest, CA sequence. Over the past few years, researchers have used a variety of approaches to estimate stress drops of these earthquakes. We compare their estimates with stress drops obtained from our inter-station coherence approach. With these comparisons, we aim to assess which models of earthquake rupture are plausible.
The coherence method works by examining the differences in the apparent source time functions (ASTFs) observed at a range of stations. We note that signals coming from different locations within the rupture area have different arrival time at the observing stations. The arrival time differences are proportional to the largest distance between generated seismic waves, which is proportional to the rupture diameter D. Thus longer-period inter-station differences can arise when the rupture diameter is larger, and we identifythe periods where inter-station differences exist in order to identify the rupture diameter.
The 2019 Ridgecrest earthquake sequence occurred between 4 July 2019 to 17 July 2019. We aim to focus on 55 well-located earthquakes in the range magnitude of 2.01 to 4.52 at depths between 2 km to 10 km. For each earthquake signal, we perform signal pre-processing to the vertical component of seismogram by remove the effect of trend from the trace. We taper the signal to minimize the effect of discontinuities at the beginning and the end of time series and remove the instrument response of the seismogram. Then we obtain high-quality arrival times and cross-correlate signals from nearly co-located earthquakes to partly remove the path effect. After the path effect removal, we can examine differences in the source time functions among stations and infer the earthquake diameters.
Initial results for M3.81 and M3.99 earthquakes imply stress drops of 2 MPa and 19 MPa, respectively. These estimates are scattered around estimates obtained by other researchers, and we are currently working to obtain more stress drops to compare and to examine synthetic ruptures and understand why the estimates differ.
How to cite: Astra, I. M. K. A. and Hawthorne, J.: Estimating stress drops using inter-station phase coherence: comparison with the Ridgecrest stress drop validation study, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-20124, https://doi.org/10.5194/egusphere-egu25-20124, 2025.
Repeating earthquakes are an expression of repeated activation of the same fault patch (asperity), with full or significant rupture area overlap. They can be used to locally characterize slip behavior, dynamics, and fine structure of faults at depth which are inaccessible with other methods. Very recently, the enigmatic observation of so called anti-repeating earthquakes has resurfaced, which is the observation of repeater-like waveforms that bear opposite amplitude signs. Observations of such a phenomenon are rare, and the known examples are today limited and in parts disputable.
This is partly because identification and characterization of repeaters requires a sound analysis in order to discriminate between neighboring earthquakes and repeaters which truly activate identical fault-patches.
Here we used a recently compiled repeating earthquake catalog for North Chile using comparatively strict repeater identification criteria, such as the utilization of a time window that covers both p- and s-phase and a cc≥0.95 for a passband of 1-8Hz at at least two stations. It consists of 10,706 repeating earthquakes that are part of 3,179 sequences. We have explicitly searched for anti-repeating sequences, which are sequences, that include at least one event with flipped waveforms at all seismic stations. Among all the repeater sequences, we find only 4, which show such a behavior. Of those, 3 are doublets (pairs of two) and only one group of 34 events contains multiple anti-repeaters. We show detailed analysis of that group, applying highly precise relative relocation and clustering methods.
The existence of anti-repeating earthquakes raises the question on the in-situ nucleation and rupture conditions required to produce the phenomenon. While apparently being an absolute exception, better understanding their circumstances will potentially improve our understanding of earthquake source mechanisms and the subduction system.
How to cite: Folesky, J., Kummerow, J., and Hofman, L. J.: Anti-repeating earthquakes in North Chile: a very rare observation, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6837, https://doi.org/10.5194/egusphere-egu25-6837, 2025.
Earthquakes result from the transient frictional weakening of faults during co-seismic slip. Dry faults weaken due to the degradation of fault asperities by frictional heating (e.g. flash heating). In the presence of fluids, theoretical models predict faults to weaken by thermal pressurization of pore fluid. Despite theoretical predictions, not only numerical models seldom consider the Pressure-Temperature dependence of the fluid properties, but experimental data is also scarce on rock-fluid interactions during dynamic rupture under realistic stress conditions. This study seeks to elucidate how fluid thermodynamic properties influence the respective roles of thermal pressurization and flash heating in fault weakening.
Here, dynamic stick-slip events (SSEs) were experimentally produced under low and high pore fluid pressure conditions on samples of Westerly granite, previously heat treated to enhance their permeability. To investigate the mechanisms driving frictional weakening, fluid pressure was directly monitored on and off the fault during SSEs using in-situ pore fluid sensors. Acoustic emissions, both amplified and unamplified, provided microseismic counts, location, magnitude and rupture velocities of each SSE. The post-SSE temperature was assessed using Raman spectroscopy on a carbon layer deposited along the fault surface.
Preliminary experimental results highlight the transition from thermal pressurization (TP) to dilatant strengthening (DS) and off-fault damage depending on the stress regime. At low shear stress, TP was observed as a coseismic increase in pore fluid pressure for each SSE. On the contrary, in the later stages of our experiment, at higher shear stress, SSEs were preceded by a pre-seismic drop of on-fault pore fluid pressure, followed by a large coseismic one. Off-fault pore fluid pressure showed a slight increase throughout all SSEs. Strain responses in the sample bulk exhibit unique patterns: dynamic dilatancy followed by dynamic compression during early SSEs, and static dilatancy followed by dynamic compression during later SSEs. Rupture velocity inversions predominantly indicate supershear characteristics. Finally, during one of our experiments at Pc = 90 MPa and Pp = 45 MPa, the slow transition between TP and DS was accompanied by a long phase during which only slow stick-slip ruptures were observed. The mechanism underlying this inversion and the role of fluid pressure behaviors on fault weakening remains to be analyzed.
Eventually, key physical and seismic parameters derived from the experiments will inform numerical models, which will be compared against thermal pressurization theory—adjusted to account for fluid thermodynamic property dependencies—and extrapolated to crustal depths (~2–10 km) where natural earthquake nucleation typically occurs.
How to cite: Fan, C., Lin, G., Aubry, J., Deldicque, D., Bhat, H. S., and Schubnel, A.: From thermal pressurization to dilatant strengthening during stick-slip ruptures on saturated saw-cut thermally cracked westerly granite, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8303, https://doi.org/10.5194/egusphere-egu25-8303, 2025.
During an earthquake, the coseismic deformation of the ocean floor is transmitted through the water column. If the earthquake’s energy is sufficiently large, it can uplift the ocean surface, and the subsequent collapse due to gravity leads to the propagation of waves as a tsunami. This perturbation also creates current fields, as water is pushed away from the uplifted area, which carry information about the seafloor deformation, including its rate and distribution. By measuring the surface current fields, information about the earthquake’s underwater spatial and temporal part characteristics can be obtainable. Using data measured directly above the source, in conjunction with onshore measurements, may lead to better resolution of the inverted seismic source, especially near the shallower parts of the rupture, complementing traditional inversion methods, such as geodetic data based models.
As a first step, this work presents a method for inverting the sea surface current field induced by coseismic deformation, isolated from background currents such as tidal or wind-driven currents, to determine the distribution of deformation at the sea bottom, assuming a flat ocean floor and instantaneous deformation. We use a simple linear fluid model to relate the coseismic effects to surface ocean currents and test robust inversion methods, assessing the associated uncertainties, using synthetic data and, as a benchmark, the deformation distribution from the 8.8 2010 Mw Maule earthquake. This approach offers novel insights into the use of new datasets for retrieving seismic source information.
How to cite: Cifuentes-Lobos, R., Behrens, J., and Calisto, I.: Inversion of ocean surface currents to obtain coseismic seafloor deformation., EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12707, https://doi.org/10.5194/egusphere-egu25-12707, 2025.
Understanding how stresses accumulate and release along subduction zones, the regions hosting the world largest earthquakes, is essential for a better assessment of seismic hazard. Monitoring the surface deformation of the upper plate allows to infer processes taking place along the subduction interface during the different phases of the seismic cycle. Here we focus on the Mexican subduction zone, where large damaging earthquakes frequently occur. This subduction zone also hosts large and frequent slow slip events (SSEs) that are predominantly aseismic. Using geodetic techniques (InSAR and GNSS) along the 1000 km of the Mexican subduction zone allows different modes of slip mode to be investigated (interseismic, post-seismic, slow slip), but is challenging. Previous studies have shown the slow slip event and interseismic signal can be extracted from InSAR observation but have been limited to the 2016-2019 period, and were not completely homogeneous in term of processing. The large amount of data needed to cover the whole Mexican Subduction zone was one challenging issue. Here we analyze the extended Sentinel-1 InSAR dataset, processed through the French national FLATSIM facilities (The ForM@Ter LArge-Scale Multi-Temporal Sentinel-1 InterferoMetry Service) covering the 2016-2022 period. The post-processing of the FLATSIM data includes several steps. First, the noisy pixels (i.e., affected by low-coherence, unwrapping errors) or non-tectonic signals (strong subsidence) are masked using the quality indicators of the FLATSIM products. Then, we correct the InSAR time from co-seismic offsets using a parametric model, since several earthquakes occurred during the study period, and estimate for each pixel a linear trend corresponding to the interseismic deformation. We finally adjust the different tracks in a common reference frame, by correcting, for all the tracks, from ramp to adjust (1) the InSAR signal to the GNSS estimate for the same time period and (2) the coherency between the overlap zones of adjacent tracks. The result is a homogeneous map of interseismic deformation over the whole Mexican subduction zone for the period 2016-2022. Some areas are affected by post-seismic deformation and large slow slip events (in Guerrero and Oaxaca region), that should be taken into account when analysing coupling map of the subduction.
How to cite: Pathier, E., Radiguet, M., Maduel, A., Meridi, A., and Thea, R.: Quantifying interseismic deformation along the Mexican subduction zone from Sentinel-1 InSAR time series, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18120, https://doi.org/10.5194/egusphere-egu25-18120, 2025.
Advances in geodetic monitoring of faults have revealed slow deformation transients and complex interactions between slow aseismic events and rapid seismic events. The Haiyuan Fault (northeast China) is of particular interest to decipher fault slip behavior and the associated physical mechanisms, due to its dual slip modes, similar to that of the San Andreas Fault in California, with both locked segments prone to major earthquakes and segments where aseismic slip ("creep") is observed. Here we focus on its 35 km-long creeping segment in the junction area between the western termination of the 1920 Mw7.9 earthquake and the eastern termination of a seismic gap [1], revisiting the spatial distribution and temporal evolution of creep from the joint analysis of ERS, Envisat and Sentinel-1 data.
We primarily use Sentinel-1 displacement time series over a seven-year period (2015-2022) processed by the FormaTerre FLATSIM service [2], corresponding to two tracks along ascending orbits and three tracks along descending orbits covering the creeping section of the fault. We first analyze the linear and seasonal components of the displacement time series, then decompose the linear term into fault-parallel horizontal velocity and vertical velocity fields. The creep signature and spatial extent are clearly identified in the line of sight and horizontal velocity maps. The surface creep rate distribution shows along-strike variations with peaks reaching up to 5 mm/yr for the horizontal component. Subsidence at a rate of 6,5 mm/yr is also observed in the extensional relay zone at the eastern end of the creeping section. We then invert InSAR line of sight velocity maps for the slip distribution along the seismogenic zone using the CSI software [3], using GNSS data as additional constraint. Creep distribution is compared with those derived from ERS and Envisat data to discuss the potential evolution of creep over decades.
We also investigate such potential temporal variations in the creep rate as seen from previous ENVISAT observation. We analyze the temporal evolution of the cumulative relative creep for each track independently, considering both the raw displacement time series and the decomposed horizontal and vertical components. Applying Principal and Independent Component Analysis as exploratory tools, we separate the extensional relay zone from the western part of the creeping section. Preliminary results show specific spatial patterns associated with temporal evolutions of mixed horizontal creep and subsidence, consistent on all tracks. They highlight in particular the morphology and slip partitioning in the fault step-over area, as well as multiple periods with transient events. These transients include variations in the subsidence velocity in the basin marking the extensional relay zone, as well as in the horizontal displacement velocities in the creeping section and in the uplift velocity to the west of the basin. Investigations into the tectonic and hydrological origins of these transients are ongoing.
References
[1] Jolivet,R.et al,J.Geophys.Res.Solid Earth,2012,doi:10.1029/2011JB008732
[2] Thollard,etal,Remote Sens 2021,doi.org/10.3390/rs13183734
[3] Jolivet,R.et al,Geophysical Research Letters,2020,doi.org/10.1029/2019GL085377
How to cite: Mokhtari, F., Lasserre, C., Jolivet, R., Cavalié, O., Daout, S., Jianbao, S., Doin, M.-P., and Durand, P.: slip dynamics and morphology of a MAJOR CREEPING FAULT STEP-OVER at the eastern end of the Tianzhu seismic gap (Haiyuan fault, China), from INSAR, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16335, https://doi.org/10.5194/egusphere-egu25-16335, 2025.
Understanding the strain field induced by earthquakes and aseismic slip events is crucial for interpreting deformation data, constraining source parameters, and calculating the induced static stress, which may lead to further seismicity. The Fault Activation and Earthquake Rupture (FEAR) experiments at the Bedretto Underground Laboratory for Geosciences and Geoenergies (“BedrettoLab”) provide a unique opportunity to directly observe and study strain fields resulting from fluid-induced slow and fast fault slip. Understanding strain evolution before, during, and after seismic and aseismic slip events is critical for identifying patterns of deformation localization and rupture nucleation, which are key to advancing physical models of earthquake processes and hazards. In two preparatory experiments before the main fault activation experiments, we attempted to trigger an ML~0.0 quake with model-informed fluid injection protocols. During the first experiment, after 4.5 days of injection with 15 MPa injection pressure, and 18 hours at 20 MPa of injection, we triggered an event with Mw~ -0.4, after which we stopped the stimulation. The observed co-seismic static deformation confirmed our monitoring network’s sensitivity to detect strain changes induced by earthquakes of at least Mw ~-0.4, laying the groundwork for subsequent strain modeling analyses.
We modeled the displacement, strain, and stress fields resulting from the mainshock of the first experiment, treating it as a generic uniform dislocation source. To parameterize the source, we utilized a focal mechanism derived from P-wave first motion polarities and a radius estimate obtained through spectral fitting.
Our analysis focused on comparing strain data from a network of Fiber Bragg Grating (FBG) sensors located at 18 to 31 m from the quake, with analytical models simulating the static deformation induced in an elastic full space by the earthquake. Preliminary results show a strong correlation between the observed and modeled strain, validating the reliability of our simple shear dislocation model. By optimizing the model parameters, particularly the fault's rake, we improved the fit between the predicted and observed strain profiles, refining our estimates of the M-zero earthquake’s source characteristics.
Furthermore, our models enabled comparisons between the spatial patterns of seismicity and the stress fields induced by fault slip deformation—both seismic (Mw~–0.4 event, triggered in the M-zero experiment) and aseismic (triggered in the FEAR1 experiment). These comparisons contribute to understanding the mechanisms driving induced seismicity behaviors.
By modeling co-seismic strain and validating these models against novel observational data, this work lays the groundwork for future strain inversions, including those targeting the pre-seismic phase, to better constrain the physical mechanisms underlying fault slip behavior. This work highlights the potential of underground facilities like BedrettoLab to deliver detailed insights into fault slip behavior, setting the stage for further analyses using data from the recently completed FEAR1 experiment.
How to cite: Lambiase, A., Meier, M.-A., Spagnuolo, E., Nikkhoo, M., Rinaldi, A. P., Gischig, V., Selvadurai, P. A., Giardini, D., and Wiemer, S.: Induced fault slip events and their deformation fields: insights from FEAR experiments, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16493, https://doi.org/10.5194/egusphere-egu25-16493, 2025.
Modern earthquake source estimation studies increasingly employ nonlinear optimization strategies to determine kinematic rupture parameters, often integrating geodetic and seismic data. The Mw 5.9 Simav Earthquake, which occurred on May 19, 2011, in the western Anatolia region of Turkey, serves as a significant case study for comprehensive source characterization. Understanding the active normal faulting mechanisms and stress distributions in this region is crucial. The earthquake occurred along the Simav Fault Zone (SFZ), an active fault system approximately 15–20 km in length and 2–3 km in width, characterized by WNW-ESE trending listric normal faults. This structure forms part of the broader extensional tectonic regime and graben systems that dominate western Anatolia, shaped by slab rollback and extensional forces driven by the subduction of the African Plate beneath the Anatolian Block along the Hellenic and Cyprus arcs. Surface ruptures trending WNW-ESE observed during the Simav Earthquake confirm the active nature of the fault.
Integrating seismic and geodetic data allows for more accurate estimation of source parameters. This process includes several key steps: modeling fault geometry, calculating Green's Functions (typically within a layered elastic half-space), and estimating distributed final slip alongside other kinematic source parameters. Aftershocks of the 2011 Simav Earthquake, concentrated at depths of 10 to 22 km, provide critical insights into fault geometry and rupture dynamics. Additionally, Coulomb stress analysis highlights the essential role of stress transfer in this region.
In this study, Bayesian inference was employed to integrate data and model uncertainties, yielding posterior distributions of source parameters. For Bayesian analysis, the Bayesian Earthquake Analysis Tool (BEAT) was utilized. BEAT is a robust tool specifically designed for modeling complex earthquake sources by integrating seismic and geodetic data. The Bayesian approach accounts for measurement and estimation errors, thereby reducing model uncertainties. Informative priors were applied to narrow the parameter space, resulting in more efficient and reliable outcomes. In the case of the Simav Earthquake, this method facilitated the robust determination of source parameters within the context of a layered medium.
Innovative sampling algorithms further enhanced the analysis by efficiently exploring high-dimensional parameter spaces, leading to improved estimates of fault geometry and mechanisms compared to earlier studies. These advancements provide a more robust understanding of source model parameters and their uncertainties. Comprehensive investigations of the Simav Fault Zone, incorporating surface ruptures and deformation analyses derived from GNSS data, significantly contribute to our understanding of the region's stress regime and seismic risk assessment.
How to cite: Memikbese, Y. and Duran, P.: A Bayesian Source Characterization of the 19 May 2011 Mw 5.9 Simav Earthquake Using Joint Inversion of Seismic and Geodetic Data, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-838, https://doi.org/10.5194/egusphere-egu25-838, 2025.
We use coseismic GNSS data to estimate the source fault parameters and coseismic fault slip of the Mw 7.3 earthquake that occurred in Hualien City, Taiwan on 3 April 2024. We examined three coseismic fault scenarios, that is Model 1, Model 2, and Model 3, because of the elusive fault system in the coseismic region. In Model 1 and Model 2, the coseismic displacements were explained by a single source fault dipping to the east and west, respectively. In Model 3, we considered a combination of an east-dipping and a west-dipping fault. For Model 1-3, we explore the geometry parameters of source fault plane(s) and invert the coseismic fault slip through a Bayesian method. Although Model 1 and Model 2 produce reasonable fits to GNSS data, Model 3 better reproduces the first-order pattern of the GNSS data. In Model 3, the east-dipping fault accounts for most of the moment release of this earthquake with a thrust slip up to about 1.5 m. The slip on the west-dipping fault in Model 3 is smaller with a maximum slip of about 0.4 m. We calculate the Coulomb stress in Model 3 on major active crustal faults in this region. The Coulomb stress on the Longitudinal Valley Fault and Milun Fault increased by 21 and 5 bars, respectively, implying a higher seismic potential in the future.
How to cite: Sang, C., Hu, Y., Yang, S., Wang, K., and Cui, X.: Coseismic Slip of the 2024 Mw 7.3 Hualien Earthquake Constrained by GNSS Observations, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15814, https://doi.org/10.5194/egusphere-egu25-15814, 2025.
The rupture process of the 2024 Hualien MW 7.4 earthquake is determined using seismic and geodetic datasets. Joint inversion reveals a conjugate faulting scenario with rupture initiation on an east-southeast-dipping reverse fault near the 40-km deep fault intersection with a west-northwest-dipping reverse fault that began to rupture 8 s later. The slip zone on F1 extends ~60 km along strike and ~60 km along dip, while slip on F2 extends ~60 km along strike and ~40 km along dip, with overall rupture velocity of ~2.5 km/s. The heterogeneous slip distributions have a peak slip of ~2.4 m, and complementary aftershock patterns. The total seismic moment is 1.5×1020 N·m, most occurring within 45 s. Depth-dependent rise time patterns are observed, with longer values at shallower depths. This study contributes valuable insights into conjugate fault interactions during collision zone earthquakes, facilitating seismic risk assessment and complex fault dynamic simulations.
How to cite: Liu, C., Bai, R., Ding, C., Lay, T., and Xiong, X.: Conjugate Rupture of the 2 April 2024 MW 7.4 Hualien Earthquake Inferred from Seismic and Geodetic Observations, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19122, https://doi.org/10.5194/egusphere-egu25-19122, 2025.
On January 8, 2022, a significant Mw 6.6 earthquake struck Menyuan, Qinghai, resulting in substantial surface damage. To investigate the geological context behind the strong surface rupture generated by the Menyuan earthquake and its impact on inhibiting stress release in the eastern section of Tuolaishan fault, this study utilized the spectral element method to simulate the dynamic rupture process of the branching fault on actual terrain. The dynamic rupture simulation revealed that the rupture initiated bilaterally along an upward direction from the initial rupture point. Influenced by a high-speed P-wave anomaly located above the source area, the rupture displayed a non-continuous pattern. As the rupture progressed into the eastern section of Tuolaishan, there was a significant abrupt decrease in both sliprate and slip. Furthermore, the area with a sliprate of approximately 3.6 m/s near the Earth's surface could be considered a strong motion generation zone. The combined influence of these factors, along with their high-frequency radiation, likely played a pivotal role in causing the pronounced coseismic surface deformation during the Menyuan Mw 6.6 earthquake. The spatial distribution of strain, as calculated from the dynamic simulation results, revealed that the southwestern side of the eastern section of Tuolaishan and the northeastern side of the western section of Lenglongling experienced predominantly tensile stresses, with corresponding areas subjected to compression. This observation aligns with the focal mechanism solution and the geological context of the northeastern margin of the Qinghai-Tibet Plateau, where the principal compressive stress direction transitions from north-south to southwest-northeast. Furthermore, the dynamic rupture process in the eastern section of the Tuolaishan was strongly inhibited by the rupture of the branching fault. This resulted in incomplete stress release and a residual seismic magnitude of approximately Mw 5.1. Triggered by Coulomb stress from the Menyuan earthquake, the potential for further rupture in the future is a possibility.
How to cite: Zhang, W. and Xie, Z.: Dynamic rupture process of the 2022 Menyuan Mw6.6 earthquake, Qinghai, China, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-136, https://doi.org/10.5194/egusphere-egu25-136, 2025.
The Mw 5.9 Simav earthquake that occurred on May 19, 2011, in Western Anatolia serves as a significant case for understanding active tectonic processes in an extensional regime. This study investigates fault geometry, slip distribution, and stress transfer parameters by integrating seismic, GNSS, and Interferometric Synthetic Aperture Radar (InSAR) data. Through Bayesian inference and joint inversion methods, the study aims to better understand the co-seismic deformations caused by the earthquake.
Within the Simav Fault Zone, the co-seismic slip distribution and stress transfer between fault segments will be analyzed to calculate potential stress increases and correlate them with the spatial and depth distribution of aftershocks. The study will specifically examine the relationship between NW-SE-oriented stress accumulation and concentrated aftershock activity. Aftershocks will be analyzed using the double-difference relocation technique. Focal mechanism and relocation analyses indicate seismic activity extending to depths of up to 22 km. Furthermore, GNSS data from before and after the 2011 earthquake reveal an extension rate of 30-40 mm/year, shedding light on the evolutionary processes of the graben system.
Previous studies have provided significant insights into the Simav earthquake and have conducted comprehensive analyses regarding the extensional regime and active faulting processes in the region. Building upon these findings, this study aims to present preliminary results and offer new perspectives on fault parameters and stress changes.
How to cite: Duran, P.: Re-Evaluation of the 2011 Mw 5.9 Simav Earthquake: Bayesian Modeling of Co-Seismic Slip Using InSAR, GPS, and Seismic Data, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17275, https://doi.org/10.5194/egusphere-egu25-17275, 2025.
Subduction zones are regions where one tectonic plate slides beneath another. This phenomenon generates a variety of earthquakes: interplate, tsunami, intraplate, and deep earthquakes. The different types of subduction zone earthquakes exhibit variations in the frequency content of the seismic energy released (Venkataraman and Kanamori, 2004). For example, tsunami earthquakes (Kanamori, 1972) occur in the shallow portions of the subduction zone. These earthquakes are deficient in high-frequency energy; however, they release a significantly larger amount of slip compared to ordinary subduction zone earthquakes. Despite their seismological similarities, there is currently no universally accepted model that describes the structural or morphological conditions around these faults that are conducive to large tsunamis with minimal ground motion (Sallares and Ranero, 2019).
The ratio of seismic energy to seismic moment, or scaled energy, can be interpreted as the radiated energy per unit area and per unit slip on the fault plane (Izutani and Kanamori, 2001). Newman and Okal (1998) demonstrated that scaled energy, calculated from observed waveforms, is one of the most powerful discriminants for tsunami earthquakes. Tsunami earthquakes typically show scaled energy values ranging from 7x10-7 to 3x10-6 (Venkataraman and Kanamori, 2004).
We estimate the radiated seismic energy from teleseismic P-waves using the methodology proposed by Perez-Campos et al. (2003). We calculated the scaled energy of events near the subduction zones in Mexico, Central America, and South America. The highest scaled energy value corresponds to a normal earthquake (MW 6.6) off the coast of Chile at a depth of 15 km (April 9th, 2001). Many strike-slip earthquakes exhibit high energy, while thrust events generally have lower values.
In the case of shallow thrust events, we observe a weak relationship between scaled energy and depth. This relationship was reported by Bilek, Lay, and Ruff (2004) for Chile, Peru, and Mexico. Additionally, we explore the relationship between scaled energy and rigidity.
How to cite: Flores, K., Ito, Y., and Kaneko, Y.: Study of Scaled Energy in Latin America, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18023, https://doi.org/10.5194/egusphere-egu25-18023, 2025.
Posters virtual: Mon, 28 Apr, 14:00–15:45 | vPoster spot 1
EGU25-4961 | Posters virtual | VPS21
"Deep Learning Application for Seismic Fault Interpretation Based on 3D Convolutional Neural Networks and ESSAttn Attention Mechanism"Mon, 28 Apr, 14:00–15:45 (CEST) | vP1.6
This paper proposes a deep learning model based on 3D Convolutional Neural Networks (CNN) and a custom attention mechanism (ESSAttn) for seismic fault interpretation from 3D seismic data. The model combines the advantages of self-attention mechanisms and convolutional neural networks to enhance the ability to capture and represent features in three-dimensional seismic data. The core innovation of the model lies in the introduction of the ESSAttn layer, which applies a non-traditional normalization process to the input feature queries, keys, and values, thereby strengthening the relationships between features, especially in high-dimensional seismic data. Unlike traditional attention mechanisms, the ESSAttn layer normalizes feature vectors by squaring them and integrates features across depth, width, height, and channel dimensions, significantly improving the effectiveness of attention computation.
The model's role in seismic fault interpretation is reflected in several aspects. First, the 3D convolutional layers automatically extract spatial features from seismic data, accurately capturing the location and shape of faults. Second, the ESSAttn layer enhances critical region features and focuses attention on important areas such as fault zones, reducing the interference from background noise and significantly improving fault detection accuracy. Finally, by using a weighted binary cross-entropy loss function, the model can prioritize fault regions when handling imbalanced data, improving sensitivity to weak fault signals.
The network architecture consists of three main parts: encoding, attention enhancement, and decoding. Initially, two 3D convolutional layers and max-pooling layers are used for feature extraction and down-sampling, followed by the ESSAttn layer to enhance the extracted features. The decoding part restores spatial resolution through upsampling and convolution layers, ultimately outputting the fault prediction results. The model is trained using the Adam optimizer, with a learning rate set to 1e-4.
Experimental results show that the model performs well in seismic fault interpretation tasks, effectively extracting and enhancing fault-related features. It is particularly suitable for automatic fault identification and localization in complex geological environments. The model's automation of feature extraction and enhancement reduces manual intervention, increases analysis efficiency, and demonstrates strong adaptability to large-scale 3D seismic datasets. Furthermore, the model architecture was visualized and saved using visualization tools for easier analysis and presentation.
Keywords: 3D Convolutional Neural Networks, ESSAttn, Attention Mechanism, Fault Interpretation, Weighted Cross-Entropy, 3D Seismic Data, Deep Learning
How to cite: zhang, Y.: "Deep Learning Application for Seismic Fault Interpretation Based on 3D Convolutional Neural Networks and ESSAttn Attention Mechanism", EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4961, https://doi.org/10.5194/egusphere-egu25-4961, 2025.
EGU25-6545 | ECS | Posters virtual | VPS21
Earthquake Moment Tensor Inversion Using 3D Velocity Model in the HimalayasMon, 28 Apr, 14:00–15:45 (CEST) | vP1.16
The Himalayan region, shaped by the ongoing collision of the Indian and Eurasian tectonic plates, is one of Earth’s most seismically active and geologically complex areas. The Indian plate moves northeastward at a rate of approximately 5 cm per year, driving tectonic activity in this region. Understanding earthquake source mechanisms in this region is crucial for seismic hazard assessment and geodynamic studies. Moment tensor (MT) inversion, a widely used technique for analysing earthquake faulting mechanisms, matches synthetic waveforms to observed data by minimising the misfit. However, conventional 1D velocity models often fail to capture the region’s complex lateral heterogeneities, leading to inaccuracies in source characterisation. Synthetic waveforms, generated via Green’s functions using frequency waveform (FK) methods and 1D velocity models, are critical for MT solutions, with time shifts playing a pivotal role in achieving optimal waveform correlations.
This study employs a 3D velocity model to improve MT inversion for a Mw 3.5 earthquake on 9 January 2021 (30.76°N, 78.54°E). Green’s functions were generated using the spectral element method for six simulations. Each simulation resulted in three-component waveforms, with a total of 18 synthetics per station. Observed data from 24 broadband stations were analysed, and results were compared to those obtained using 1D models. Slight variations in strike, dip, and rake values underscore the limitations of 1D models in capturing Earth’s heterogeneities.
The study reveals that 3D velocity models significantly enhance MT solution accuracy, particularly in determining focal depths, faulting mechanisms, and seismic moment magnitudes. A probabilistic approach was also applied to quantify the uncertainty associated with MT estimates, providing confidence measures. Extending this approach, MT inversion was performed for another earthquake in the Uttarakhand Himalaya using the same 3D velocity model, further demonstrating the advantages of 3D wavefield simulations in seismically active regions.
Keywords: Himalayas, Moment Tensor, Green’s Function, Spectral element method.
How to cite: Maurya, S., Silwal, V., Mahanta, R., and Yadav, R.: Earthquake Moment Tensor Inversion Using 3D Velocity Model in the Himalayas, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6545, https://doi.org/10.5194/egusphere-egu25-6545, 2025.
