We present a quantum computing concept for 1-D elastic wave propagation in heterogeneous media with two components: a theoretical formulation and an implementation on a real quantum computer. The method rests on a finite-difference approximation, followed by a transformation of the discrete elastic wave equation to the Schrödinger equation, which can be simulated directly on a gate-based quantum computer. An implementation on an error-free quantum simulator verifies our approach and forms the basis of numerical experiments with small problems on an actual quantum computer. As the presented approach promises exponential speedup compared to classical numerical wave propagation methods, it has the potential to significantly push the limits of global full-waveform inversion, particularly maximum feasible frequencies, on future quantum computers.

How to cite: Schade, M., Bösch, C., Hapla, V., and Fichtner, A.: A quantum computing concept for 1-D elastic wave simulation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10632, https://doi.org/10.5194/egusphere-egu24-10632, 2024.

Tectonic deformation plays a crucial role in shaping the Earth's surface, with strain localization resulting in the formation of shear zones in depth and faults on the surface. These structures accommodate a significant portion of the displacement between tectonic plates. While long-term deformation can be approximated as continuous visco-plastic processes, earthquakes involve cycles of stress loading and unloading that trigger rapid and catastrophic elasto-plastic deformation. Earthquake dynamic rupture models offer valuable insights into studying and comprehending earthquakes. However, these models heavily rely on initial conditions that are often challenging to obtain solely from observations. Particularly, a mechanically self-consistent prestress state loading a fault prior a seismic event and 3D fault geometry, especially in depth, are commonly poorly constrained. Nonetheless, the prestress state and the fault geometry significantly impact earthquakes initiate, propagate, and arrest and the associate radiation of seismic waves and ground shaking.

To address the lack of information on stress and fault geometry, one promising approach is to use long-term geodynamic numerical simulations. In this study, we employ pTatin3D, a visco-plastic finite element software, to simulate the evolution of strike-slip deformation in 3D over geological time scales. To ensure a physically consistent long-term model, the fault geometry is not prescribed but solved for based on the lithospheric rheology and tectonic plate velocities. However, the geodynamics model describes faults as continuous volumetric fields of finite deformation and strain-rate, rendering them 3D objects, while earthquake dynamic rupture models typically represent faults as 2D interfaces.

In this study, we outline a new and versatile method to link 3D geodynamic simulations to rupture dynamics earthquake and seismic wave propagation modelling. We first extract 3D volumetric shear zones from the geodynamic model and automatically convert them into surface representations. Next, we generate meshes including these as faults for dynamic rupture models. Finally, we showcase 3D dynamic rupture models utilizing the stress states and faults self-consistently as derived from the long-term geodynamic model as initial conditions.

How to cite: Jourdon, A., Hayek, N., May, D., and Gabriel, A.-A.: One-way linking of 3D long-term geodynamic models and short-term earthquake dynamic rupture models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19294, https://doi.org/10.5194/egusphere-egu24-19294, 2024.

Deep-focus earthquakes occur at depths of 300-700 km below the surface where brittle failure is unlikely due to the large lithostatic pressure. Instead, they require a ductile localization mechanism that can significantly reduce rock strength and create highly localized shear zones. The feedback loop of shear heating, temperature-dependent viscosity and localization is called thermal runaway and has been linked to deep-focus earthquakes.

We present one- and two-dimensional (1D and 2D) numerical, thermomechanical models that investigate the occurrence, nucleation and temporal evolution of thermal runaway in a simple shear setting. The models are characterized by a visco-elastic rheology where viscous creep is accommodated with a composite rheology of diffusion and dislocation creep as well as low-temperature plasticity. We utilize the pseudo-transient iterative method in combination with a viscosity regularization and adaptive time stepping to solve this nonlinear system of equations and avoid resolution dependencies.

Varying eight input parameters, we observe two distinct types of behavior. After elastic loading, models either release stress over hundreds to thousands of years, accompanied by low slip velocities and moderate temperature increase, or they release stress within seconds to minutes while slip velocity and temperature increase drastically – Thermal runaway occurs. With nondimensional scaling analysis, we unite the eight different input parameters into two nondimensional numbers that allow inferring the behavior. The ratio t_r/t_d describes the competition between heat generation by viscous dissipation and heat loss due to thermal diffusion whereas the ratio U_el/U_th compares the elastic and thermal energy density before stress relaxation.

2D experiments show that thermal runaway allows highly localized ductile ruptures to nucleate at small heterogeneities and propagate like brittle fractures. The ruptures accelerate during propagation and reach the highest velocities when two tips link up. Rupture trajectories are usually parallel to the direction of background deformation but bend in the vicinity of other ruptures to allow for a link up. The results demonstrate that thermal runaway can create highly localized, propagating shear zones that reach slip velocities in line with slow earthquakes at upper mantle and transition zone conditions.

How to cite: Spang, A., Thielmann, M., and Kiss, D.: Thermal runaway as driving mechanism of deep earthquakes – Constraints from numerical modeling, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10403, https://doi.org/10.5194/egusphere-egu24-10403, 2024.

Through the progression of seismic activity, natural fault zones undergo a complex evolution characterized by the accumulation of damage and the formation of gouge within the fault core across multiple scales. Even though this is believed to be among the key factors affecting the evolution of fault seismicity over time, a deep understanding of the mechanisms at play is still missing. In this study, we explore the role of gouge production in the self-organization process of loaded rough faults, focusing on the evolving dynamics of earthquake nucleation, recurrence and moment partitioning during the seismic cycle. We model the stress and sliding dependence of gouge evolution by linearly coupling Archard's wear law with rate-and-state friction through the critical slip distance ( Dc ). Including this new formulation in 2D quasi-dynamic, elastic simulations of rough faults, we can reproduce the effects of spatially and temporally heterogeneous gouge evolution. Following the build-up of gouge over many cycles, we observe a progressive transition from cascade-driven to creep-dominated nucleation processes, marked by an increase of precursory slow slip and foreshock activity. A clear shift in the moment partitioning from faster to slower slip rates becomes evident as heterogeneity grows larger, followed by a reduction of the total cumulative moment released. Finally, the recurrence interval is observed to grow initially, then drop abruptly and become more unpredictable as the amplitude of Dc continues to rise. Incorporating a new formulation of gouge production in earthquake cycles simulations, this work sheds light on the role of gouge accumulation in the maturation process of natural faults, offering critical insights for seismic risk assessment and mitigation.

How to cite: Castellano, M., Milanese, E., Cattania, C., and Kammer, D.: The influence of gouge formation on seismicity and fault slip behavior. , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6202, https://doi.org/10.5194/egusphere-egu24-6202, 2024.

Earthquakes are associated with the propagation of a dynamic rupture, which radiates elastic energy through seismic waves. The generation of seismic radiation is related to dynamic weakening of shear stress and stress drop. In modeling dynamic ruptures, shear stress evolution is commonly imposed through a constitutive law, such as the widely adopted slip weakening laws. According to these constitutive laws, shear stress evolves as a function of slip in each point of the rupturing fault, prescribing strength excess, stress drop and dynamic weakening.

Here, we compare two well-known slip weakening laws: namely, the classic Ida’s (1972) and the Ohnaka’s (1996) slip weakening laws. The former prescribes that fault stress increases from the initial stress to the peak stress with zero slip and then linearly decreases from the peak value to a residual value over a slip-distance Dc (dynamic weakening). The latter assumes that the initial stress hardening phase occurs over a non-negligible slip-distance Da and that shear stress decrease from the peak value is not linear. The Ohnaka’s law was validated with numerous laboratory experiments. The evolution of shear stress with slip allows the estimate of the breakdown work Wb, i.e. the excess of work above a minimum stress level with slip from 0 to Dc.

We collected data from high-velocity friction experiments to quantify yield, peak and residual stresses, Da and Dc distances for bare-rock samples of Carrara Marble and Gabbro deformed under various experimental conditions (room humidity, vacuum, pressurized fluids) and normal stress (from 5 to 40 MPa). The ratio Da/Dc is much lower for Carrara marble (0.015) than for Gabbro (0.12). We implemented the Ohnaka’s constitutive law in a 2D finite difference code for spontaneous dynamic ruptures characterized by a fault in a homogeneous elastic material.  We perform simulations using the two different slip weakening laws. We kept constant Dc, and we compared the results of the simulations in terms of rupture style, rupture velocity, breakdown work, and cohesive zone size. As expected both laws yield crack-like ruptures. Moreover, Ohnaka’s law in comparison to the linear slip weakening law produces:

• rupture velocity ~2 % higher;
• breakdown work (Wb) up to 60 % lower. Moreover, dividing the breakdown work into the energy dissipated between the yield stress and the peak stress over the slip-distance Da (Wb_a), we notice that Wb_a can reach up to the 30% of the total Wb in case of Gabbro (Da/Dc = 0.12).
• a cohesive zone size (defined as the portion of the fault in which the slip velocity is higher than zero and the stress is higher than its residual value) up to 50% larger.

Therefore, Ohnaka’s law generates more energetic ruptures (i.e. faster rupture velocity and peak slip-rate) despite having a larger cohesive zone due to the lower breakdown energy dissipated during rupture propagation. We discuss our results in terms of the difference between breakdown highlighting the implications on dynamic rupture propagation and earthquake energy budget. We emphasize that common interpretations of energy dissipated during rupture propagation are model-dependent.

How to cite: Cornelio, C., Murphy, S., Spagnuolo, E., Nielsen, S., and Cocco, M.: Differences in breakdown work and fracture energy in slip weakening constitutive laws, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16751, https://doi.org/10.5194/egusphere-egu24-16751, 2024.

Natural faults exhibit complex geometry. In this study, we model cycles of earthquake ruptures on non-planar faults governed by a friction formulation that combines rate and state friction for low slip velocity and enhanced weakening friction in the form of flash heating for high slip velocity, both consistent with rock friction experiments. The numerical method allows non-matching meshes across the fault, continuously updates the fault geometry, and employs variable time steps with quasi-static and fully dynamic time integration schemes during slow and fast deformation stages, respectively. To prevent the development of large stresses on the fault, the model also accounts for fault wear and inelastic off-fault deformation. We investigate the effect of macro-scale roughness on the fault slip behavior and rupture dynamics in terms of event magnitude, stress drop, and rupture style and speed. We analyze the relationship between the fault geometry, stresses from the preceding earthquakes, and rupture characteristics.

The simulation results show a significant increase in event variability with roughness levels, with both small partial ruptures and ruptures significantly more intense than those on planar faults. The planar faults host a sequence of earthquakes that rupture the entire fault, exhibiting similar magnitudes and stress drops. The substantial reduction in friction enables the ruptures to propagate under a low background shear-to-normal stress ratio as self-healing slip pulses, with a sub-Rayleigh rupture speed. Faults with low roughness levels generally show a similar pattern. Prior to some events, the stress ratio along the fault slightly increases, leading to ruptures with secondary slip pulses and larger magnitudes. As roughness increases, stresses become more heterogeneous, resulting in a more complex sequence of ruptures, some of which arrest at restraining bends with a low stress ratio. However, stress accumulation and slip deficit during these partial ruptures result in high stress ratios on the unruptured fault segments. These are eventually released by large events of crack-like ruptures with supershear propagation speed and stress drops and slip significantly larger than a typical event on a planar fault. Therefore, while fault roughness can cause rupture arrest, consistent with previous studies, it can also substantially increase earthquake magnitudes. This factor should be accounted for in earthquake hazard assessments.

How to cite: Tal, Y.: Rupture Dynamics and Characteristics During Earthquake Cycles on Nonplanar Faults with Strongly Rate-Weakening Friction, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11379, https://doi.org/10.5194/egusphere-egu24-11379, 2024.

As part of the Southern Alps Long Skinny Array (SALSA) project, ~35+ seismometers have been deployed with 10–12 km spacing along a 450 km-long   section of the Alpine Fault. SALSA is focused on determining the ground motions likely to be produced by a future Alpine Fault earthquake. This project is addressing three principal objectives: (1) Determine the Alpine Fault’s subsurface geometry, present-day slip rates, and spatial variations in how tectonic stresses are currently accumulating on the fault, (2) Estimate the ground shaking that would be recorded at seismometers throughout central and southern New Zealand by localised slip at different points on the Alpine Fault, focusing on the synthesis of long-period Green's functions  representing accurate path effects between sources distributed along the fault and population centres throughout the South Island, and (3) Calculate the ground shaking hazard from geologically informed earthquake rupture scenarios. In this presentation we will address the influence of inferred Alpine Fault source models derived from empirical data as well as current knowledge of the fault geological and geophysical parameters on regional seismic hazard.

How to cite: Holden, C., Townend, J., Chamberlain, C., Warren-Smith, E., Juarez-Garfias, C., Pita-Sllim, O., Van Wijk, K., and Denolle, M.: Inferred source models for Alpine Fault Earthquake Scenarios and influence on seismic hazard., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8160, https://doi.org/10.5194/egusphere-egu24-8160, 2024.

Earthquakes that nucleate at depths shallower than a few km are very rare but pose high near-fault hazard despite moderate magnitudes. Some very shallow earthquakes have been associated with surface mass removal processes, both natural (e.g. glacier melting) and anthropogenic. A notable recent case is the November 11 2019 Mw 4.9 Le Teil, France earthquake. It called strong public attention because of its very shallow depth (slip shallower than 2 km), very strong ground motion (>1 g) affecting the near-fault population, and proximity to nuclear power plants. It has been proposed that this earthquake could have been triggered by mass removal from a large cement quarry located close to the epicenter. Indeed, the estimated Coulomb stress change induced by the quarry activity on the fault is of several 100 kPa. Here, we further evaluate the mechanical viability of the quarry-triggering hypothesis through 3D earthquake cycle simulations.

We consider a dipping fault governed by rate-and-state friction, with velocity-weakening steady-state behavior, and a realistic mass removal history constrained by analyses of aerial optical images across ~180 years of quarry activity. To account for uncertainties about the recurrence time of natural earthquakes on the fault and the timing of the previous natural event, we consider mass-removal loads starting at different times relative to the simulated natural earthquake cycle. Our simulations show that realistic mass removal rates can advance the failure time by thousands of years. Simulations with a constant mass-removal rate but same cumulative removed mass at 180 years produce a similar triggering timing. This indicates that the induced clock advance mostly depends on the cumulative load, rather than on its rate. The dependence on loading rate manifests through the following mechanism: the model with constant rate can trigger slow slip events instead of regular earthquakes, which postpones the next regular earthquake by a long time, whereas the model with realistic loading history (and higher load rates) always triggers regular earthquakes. The quarry's proximity to the fault and the frictional heterogeneity on the fault also play important roles. For example, clock advance is higher if the quarry location is closer to the edge of the velocity-weakening zone and lower in the middle. Also, the model with the classical rate-and-state model shows negligible impact if the quarry is at the top of a steady-state behavior zone and far away from the velocity weakening zone.

While these models confirm the possibility that mass removal can trigger shallow earthquakes on velocity-weakening faults, we will also report on additional simulations that examine whether such triggering can occur on a fault with velocity-strengthening behavior at shallow depth or it requires a more sophisticated fault rheology, such as friction with a transition from velocity-strengthening to velocity-weakening at increasing slip rate. These modeling efforts will be further constrained by ongoing laboratory experiments on representative materials of the fault that ruptured in the Le Teil earthquake.

How to cite: Sopaci, E. and Ampuero, J. P.: Triggering of very shallow earthquakes by surface mass removal processes - case study of the 2019 Mw4.9 Le Teil, France earthquake , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9827, https://doi.org/10.5194/egusphere-egu24-9827, 2024.

How to cite: Chao, S.-H., Huang, J.-Y., Hsu, C.-C., Lin, C.-M., Sung, C.-H., and Kuo, C.-H.: Taiwan Non-Ergodic Ground Motion Prediction Equations for Spectral Accelerations and Instantaneous Powers, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18840, https://doi.org/10.5194/egusphere-egu24-18840, 2024.

The Statewide Southern California Earthquake Center (SCEC) has designed and implemented CyberShake (CS), a high-performance computing (HPC) workflow to undertake Physics-Based Probabilistic Seismic Hazard Analysis (PB-PSHA) in California (CA). Here, we have ported CS from CA to the South Iceland Seismic Zone (SISZ) and the Reykjanes Peninsula Oblique Rift (RPOR), which experience sinistral transform motion and pose a very high earthquake risk to about 2/3 of the Iceland population. We consider a realization of the 3D SISZ-RPOR fault system, where fault areas are estimated from event magnitude through a scaling law (Mai & Beroza, 2017),  that fits maximum fault extents observed from slip inversions and surface mappings. The magnitude variability across the modeling region (~63.8°- 64.1°N, ~20°-23°W) is Mw 5-7. In this work, we employ CS to model ~2100 kinematic earthquake ruptures and quantify the resulting ground motion (GM) in terms of Pseudo Spectral Acceleration (PSA) intensity measures. An important computational milestone is the software development of an open-source in-house workflow manager at the Marenostrum Supercomputer that replaces the one used in CA by SCEC based on Pegasus and HTCondor. This new workflow manager handles input data (fault-plane geometries, rupture magnitudes, surface stations for GM recording and hazard studies), orchestrates the execution of CS components, and stores results (particle velocity seismograms and hazard curves). Among these components, the Graves-Pitarka (GP) kinematic rupture generator is used to produce finite-fault source descriptions characterized by a few large asperities. The other important component is the open-source fourth-order finite-difference staggered-grid AWP-ODC earthquake simulation code that allows for reciprocity and efficiently simulates rupture and seismic wave propagation in 3D heterogeneous Earth models. CS uses an adjoint computational procedure in which simulations of wave propagation are performed using a polarized delta source to compute the Strain Green Tensors (SGTs) at each fault point. The convolution of SGTs with GP ruptures yields particle-velocity seismograms at each station. SGT time histories are memory demanding, but the adjoint calculations are completely independent and therefore embarrassing parallel, making CS a highly efficient earthquake simulation tool. In this study, SGTs are constructed using a source frequency range of 0-1.0 Hz, generating ground motion synthetics resolved up to 0.5 Hz. CS rotation-invariant PSA values (3 and 5 sec periods) computed from our study show a good agreement with updated Bayesian ground motion prediction equations (Kowsari et al, 2022). This study is a first step towards a PB-PSHA in the SISZ-RPOR region and to routinely apply Cybershake outside of California.

REFERENCES:

Mai, M., & Beroza, G. Source scaling properties from finite-fault-rupture models. Bulletin of the Seismological Society of America, 90(3), 604-615, 2000.

Kowsari, M., Sonnemann, T., Halldorsson, B., Hrafnkelsson, B., Snæbjörnsson, J. &  Jonsson, S. Bayesian inference of empirical ground motion models to pseudo-spectral accelerations of South Iceland Seismic Zone earthquakes based on informative priors. Soil Dynamics and Earthquake Engineering, 132, 106075, 2020.

How to cite: Rojas, O., Bayat, F., Monterrubio-Velasco, M., Abril, C., Callaghan, S., Rodríguez, J. E., Kowsari, M., Halldórsson, B., Olsen, K., Gabriel, A.-A., and de la Puente, J.: CyberShake simulation of strike-slip earthquakes on the Southwest Iceland transform zone, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13834, https://doi.org/10.5194/egusphere-egu24-13834, 2024.