Session ERE5.1

ERE5.1 | Induced/triggered seismicity in geo-energy applications: monitoring, modeling, mitigation, and forecasting

Induced/triggered seismicity in geo-energy applications: monitoring, modeling, mitigation, and forecasting

Convener: Alessandro VerdecchiaECSECS | Co-conveners: Hongyu Yu, Antonio Pio Rinaldi, Rebecca M. Harrington, Victor Vilarrasa

Orals

| Thu, 27 Apr, 16:15–17:55 (CEST)

Room 0.94/95

Posters on site

| Attendance Thu, 27 Apr, 14:00–15:45 (CEST)

Hall X5

Numerous cases of induced/triggered seismicity resulting either directly or indirectly from injection/extraction associated with anthropogenic activity related to geo-resources exploration have been reported in the last decades. Induced earthquakes felt by the general public can often negatively affect public perception of geo-energies and may lead to the cancellation of important projects. Furthermore, large earthquakes may jeopardize wellbore stability and damage surface infrastructure. Thus, monitoring and modeling processes leading to fault slip, either seismic or aseismic, are critical to developing effective and reliable forecasting methodologies during deep underground exploitation. The complex interaction between injected fluids, subsurface geology, stress interactions, and resulting fault slip requires an interdisciplinary approach to understand the triggering mechanisms, and may require taking coupled thermo-hydro-mechanical-chemical processes into account.
In this session, we invite contributions from research aimed at investigating the interaction of the above processes during exploitation of underground resources, including hydrocarbon extraction, wastewater disposal, geothermal energy exploitation, hydraulic fracturing, gas storage and production, mining, and reservoir impoundment for hydro-energy. We particularly encourage novel contributions based on laboratory and underground near-fault experiments, numerical modeling, the spatio-temporal relationship between seismic properties, injection/extraction parameters, and/or geology, and fieldwork. Contributions covering both theoretical and experimental aspects of induced and triggered seismicity at multiple spatial and temporal scales are welcome.

Orals: Thu, 27 Apr | Room 0.94/95

Chairpersons: Alessandro Verdecchia, Hongyu Yu, Rebecca M. Harrington

Forecasting and managing induced seismicity

16:15–16:25

EGU23-1341

solicited

On-site presentation

Uncovering the Processes that Control Induced Earthquake Sequences

Elizabeth Cochran, Kayla Kroll, Morgan Page, Zachary Ross, and Justin Rubinstein

Earthquake sequences induced by fluid disposal into the subsurface show strong variability in their magnitude distributions and clustering behavior. We attempt to untangle the processes that control the occurrence and evolution of disposal-induced earthquake sequences by integrating detailed seismicity observations, pore pressure modeling, and 3D physics-based earthquake simulations. Observations of earthquake sequences in Oklahoma and Kansas include some sequences that have near-Poissonian distribution of interevent times and robust foreshock sequences, while other sequences display more typical mainshock-aftershock clustering behavior. Pore-pressure modeling shows that these behaviors correlate with the amplitude of the pore-pressure changes. Close to disposal wells where pore pressure changes are high, seismicity is controlled by both diffusion and earthquake stress interactions. Farther from the wells where pore pressure changes are lower, seismicity appears driven primarily by stress interactions. We further explore the stress and fault conditions that may allow for these diverse behaviors using an earthquake simulator, RSQSim. We find that the maximum magnitude of the triggered events depends strongly on the pre-existing stress on the fault. The roughness, distribution, and background stress state of pre-existing primary and secondary faults also likely influence the sequence behavior and is an area of ongoing work.

How to cite: Cochran, E., Kroll, K., Page, M., Ross, Z., and Rubinstein, J.: Uncovering the Processes that Control Induced Earthquake Sequences, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-1341, https://doi.org/10.5194/egusphere-egu23-1341, 2023.

16:25–16:35

EGU23-11960

ECS

On-site presentation

Can earthquakes nucleate on nominally stable velocity-strengthening faults?

Luca Dal Zilio, Paul A. Selvadurai, Jean-Paul Ampuero, Elisa Tinti, Massimo Cocco, Frédéric Cappa, Stefan Wiemer, Domenico Giardini, and the Bedretto Team

Tectonic faults are often assumed to slip either slow due to stable, velocity-strengthening frictional behavior, or fast as a result of velocity-weakening friction leading to dynamic (seismic) rupture. As a consequence, velocity-strengthening faults may be regarded as intrinsically stable as they do not spontaneously nucleate seismic events. However, recent laboratory and in-situ experiments of fluid injection challenged such assumptions. Here we present a fully coupled hydro-mechanical fault model in which stable, rate-strengthening frictional behavior is combined with dynamic weakening due to rapid poroelastic effects, allowing unstable (seismic) slip to occur on nominally stable faults. In our numerical experiments, fluid injection reduces the effective normal stress and frictional resistance, thus bringing the fault to failure. The onset of fault failure is controlled by competing mechanisms of shear-induced dilation and shear-enhanced compaction, which cause fault weakening and the propagation of a slow-slip transient from the fluid injection point. When a critical size of the slow slip patch is reached, dynamic rupture eventually nucleates at the slow-slip event front and propagates beyond the fluid pressure perturbed region. Further numerical experiments indicate that, when the fault is critically stressed, the growth of the aseismic patch – prior to dynamic rupture – occurs in a few seconds, whereas at lower stress levels, the aseismic slip phase propagates slowly over hundreds of seconds. These results predict that poroelastic compaction and fluid pressurization can cause the transition from aseismic slow-slip to fast seismic slip and the propagation of dynamic rupture on velocity-strengthening faults. In particular, they demonstrate that compaction-induced fluid pressurization can overcome the initial phase of shear-induced dilatancy, thus allowing the propagation of dynamic rupture in the form of pulse-like pore-pressure waves. The implication that earthquake rupture may nucleate on rate-strengthening faults, presently considered to be nominally stable, requires a re-evaluation of seismic hazard in many areas, particularly in the case of fluid injection in enhanced geothermal systems and CO2 storage.

How to cite: Dal Zilio, L., Selvadurai, P. A., Ampuero, J.-P., Tinti, E., Cocco, M., Cappa, F., Wiemer, S., Giardini, D., and Team, T. B.: Can earthquakes nucleate on nominally stable velocity-strengthening faults?, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-11960, https://doi.org/10.5194/egusphere-egu23-11960, 2023.

16:35–16:45

EGU23-15861

ECS

On-site presentation

Extraction of fluids to mitigate the seismic risk associated with post-injection aseismic slip

Alexis Sáez and Brice Lecampion

Subsurface fluid injections are commonly accompanied by seismicity which can sometimes result in earthquakes of relatively large magnitude that pose a serious hazard for the geo-energy industry. Current efforts to manage the seismic risk associated with fluid injections work generally under the tacit assumption that mitigation measures will become shortly effective in preventing the occurrence of earthquakes of larger magnitude than some pre-defined threshold. A common operational measure is shutting in the wells indefinitely. Unfortunately, seismicity after shut-in is common and, even more, it is not rare that the largest events of injection-induced seismic sequences occur during the post-injection stage. Understanding the physical mechanisms underpinning post-injection seismicity is thus of first importance for the successful development of geo-energy projects. Moreover, gaining knowledge in this matter may ultimately help to design physics-based strategies to mitigate the seismic risk associated with fluid injection operations. From a pure hydro-mechanical perspective, there are two well-known triggering mechanisms for post-injection-induced seismicity, namely, the diffusion of pore-fluid pressure (Parotidis et al., 2004) and poroelastic stressing (Segall and Lu, 2015). Recently, Sáez and Lecampion (2023) have investigated a third mechanism where injection-induced aseismic slip in pre-existing fractures and faults may keep propagating after shut-in and continue stressing even larger and more distant regions from the injector during time scales that could span even months for fluid injections of only few days, if the reactivated fracture/fault is critically stressed. This result has motivated us to develop a physics-based strategy to mitigate the seismic risk associated with post-injection aseismic slip. The idea is to extract fluids as an operational measure instead of just stopping the injection. It is shown that fluid withdrawal has not only the effect of reducing the further increase of pore pressure during the post-injection stage, but also the effect of decreasing both the spatial extent and exposure time of the surrounding rock mass to quasi-static changes of stresses due to aseismic slip. The main parameter controlling the reduction of the spatial extent and exposure time is the ratio between the extraction rate and injection rate. We make use of realistic field configurations to provide examples that show the significant reduction of the exposure time that can be achieved by extracting fluids. We also discuss some field evidence that support the plausibility of this extraction strategy.

How to cite: Sáez, A. and Lecampion, B.: Extraction of fluids to mitigate the seismic risk associated with post-injection aseismic slip, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-15861, https://doi.org/10.5194/egusphere-egu23-15861, 2023.

16:45–16:55

EGU23-16584

On-site presentation

Pressure diffusion controls maximum induced earthquake magnitudes

Cornelius Langenbruch, Mohammad J. A. Moein, and Serge A. Shapiro

16:55–17:05

EGU23-9584

ECS

On-site presentation

Research on the Mechanism of Induced Seismicity Utilizing InSAR-Retrieved Land Subsidence, Source Model, and Geological Constraints

Wojciech Witkowski, Magdalena Łucka, Henriette Sudhaus, Anna Barańska, Ryszard Hejmanowski, and Artur Guzy

Mining-induced sub-surface rock-failure processes cause seismicity and permanent surface deformation that may result in infrastructural damages and endanger residents in the affected areas. In recent years, this issue has received increasing attention due to the suddenness of the occurrence, the lack of effective predictive capabilities and the adverse consequences. Importantly, seismic events have caused significant land subsidence and several fatalities in Polish underground mines. To better investigate such mining-induced sub-surface rock failure processes, we analyze rock deformation based on surface displacement data and also put local geology in context with seismological analysis results.

In Poland's Legnica-Glogow Copper District, between 2016 and 2020, 13 mining-induced earthquakes in the moment magnitude range from 3.2 to 3.8 have been recorded by the IS-EPOS platform. Based on co-seismic Sentinel-1 satellite radar images of ESA’s Copernicus program, and radar interferometry (InSAR) we measure the surface displacements caused by nine of these earthquakes. The maximum line-of-sight surface displacements are observed for the Mw3.6 mining-induced earthquake on October 17, 2016 with 144 mm and 124 mm away from the satellite in ascending and descending radar interferograms, respectively, and an estimated maximum land subsidence reaching -142 mm. We estimate the location and volume component associated with these surface effects by using an isotropic volume point-source model in a Bayesian inference. The Geodetic Bayesian Inversion Software GBIS has been applied for these calculations. Notably, only some of the earthquakes we analyse resulted in land subsidence. Although no significant relationship between earthquake magnitude and land subsidence is apparent, we find that the greater the thickness of loose Quaternary strata, the greater the land subsidence. The focal points of the deformation at depth of the investigated earthquakes, which we compare to seismic hypocenters, are located at depths of 437 m to 669 m, according to our results. Importantly, these focal points are located above the mining exploitation fields, within layers of rigid Triassic rock. Our results also show a mismatch between the spatial location of the earthquake epicentres and the maximum land subsidence. These differences are up to 440 m and are greater for the earthquake epicentres determined using seismic data than for the source point model. This discrepancy suggests that the epicentre of a mining-induced earthquake is not necessarily associated with the region of maximum rock layer compaction.

This study adds to our understanding of the impact of mining-induced seismicity on the occurrence of land subsidence and the seismic mechanism in areas of active mining paving the way towards more sustainable mineral extraction.

How to cite: Witkowski, W., Łucka, M., Sudhaus, H., Barańska, A., Hejmanowski, R., and Guzy, A.: Research on the Mechanism of Induced Seismicity Utilizing InSAR-Retrieved Land Subsidence, Source Model, and Geological Constraints, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-9584, https://doi.org/10.5194/egusphere-egu23-9584, 2023.

Geothermal energy

17:05–17:15

EGU23-5768

On-site presentation

Unravelling kinematic source parameters of induced earthquakes at St Gallen geothermal field, Switzerland

Vincenzo Convertito and Raffaella De Matteis

On 14 July 2013, a deep geothermal drilling project close to the city of St Gallen, in Switzerland was initiated with the aim of producing electricity and heating. The target of the project was a fractured carbonate aquifer at a depth of 4 km b.s.l.. The project started with a stimulation phase on 14 July 2013 that produced only microseismicity. From 14 July 2013 through 19 July 2013 acid stimulations involving about 290 m³ of fluids broke the seal to a gas reservoir and caused a gas kick. Afterwards, well control operations were done by injecting ~700 m³ of water and heavier liquids, which probably induced the largest event in the sequence with M_L 3.5. Well-control operations ended on 24 July 2013. In September 2013 fishing operations (that is removing lost or stuck objects from the wellbore) were done together with a cleaning of the well. A total of 346 earthquakes were recorded from July 2013 to October 2013 at the seismic network managed by the Swiss Seismological Service (SED). The events occurred at depths up to 4.2 km with M_L between -1.2 and 3.5.

We infer seismic source parameters by using both a single-event approach and an empirical Green’s function (EGF) approach. This latter method helps to limit the spectral source parameters trade-off and to infer source parameters for the smaller earthquakes (M_L≤2.0). In fact, performing the ratio between the spectra of co-located events recorded at the same stations allows to remove propagation and site effects. For the larger events (M_L>2.0), we use an iterative single-event approach where Q is fixed to avoid the trade-off with f_c. Specifically, the Q_P and Q_S values are obtained from the slope of the displacement spectra - at frequencies lower than the expected corner frequency - of the smaller events. In order to apply the EGF approach, for each of the smaller earthquakes, we select among the larger events those for which the cross-correlation in the time domain is higher than a given threshold. Next, for the selected couples we compute the spectral ratio and infer the seismic moments' ratio and the corner frequency ratios by using a grid-searching technique. We find self-similar parameters scaling down to Mo=10¹⁰Nm and an average static stress drop of 0.1 MPa. The mean value of seismic efficiency, estimated from the average apparent stress to static stress drop ratios is 0.05 suggesting an overshoot dynamic weakening mechanism.

This work has been supported by PRIN-2017 MATISSE project (No. 20177EPPN2), funded by Italian Ministry of Education and Research.

How to cite: Convertito, V. and De Matteis, R.: Unravelling kinematic source parameters of induced earthquakes at St Gallen geothermal field, Switzerland, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-5768, https://doi.org/10.5194/egusphere-egu23-5768, 2023.

17:15–17:25

EGU23-3100

ECS

On-site presentation

Triggering Mechanisms of Post-Injection Induced Seismicity Using The Enhanced Geothermal System of Basel (Switzerland)

Auregan Boyet, Silvia De Simone, Shemin Ge, and Victor Vilarrasa

Induced seismicity is a limiting factor in the widespread deployment of Enhanced Geothermal Systems (EGS). Induced seismicity occurs not only during injection, but also after the stop of injection, with the largest magnitude earthquakes usually occurring after the stop of injection at different scales of time (hours to months). The post-injection large magnitude seismicity is counterintuitive and is still not well understood. Multiple mechanisms have been identified as triggering induced seismicity in EGS systems. Pore pressure increase due to fluid injection is the most commonly accepted triggering mechanism to explain induced seismicity. Yet, coupled poromechanical effects and static stress transfer are also important mechanisms in the process of fault reactivation. We study the combination of these different mechanisms by simulating coupled hydromechanical processes of the Deep Heat Mining Project at Basel, Switzerland (2006), a well-known case of short-term post-injection large-magnitude seismicity. We apply the material characteristics and stress conditions of the site to a simplified 2-dimensional fault network that is based on the monitored seismic events. In our model, pore pressure has a dominant effect on the triggering of the seismicity in the vicinity of the injection well, but its effect decreases away from the well as static stress transfer and poroelastic stressing gain importance. Poroelastic stress can have a stabilizing effect when its direction is opposite to the fault slipping orientation; but an abrupt shut-in, and consequently a quick poroelastic relaxation, reactivates the fault soon after the shut-in. Later post-injection induced seismicity is partially due to the post-injection diffusion of the pore pressure, but is mainly induced by constant static stress redistribution from reactivations of the different faults of the domain. Understanding the processes inducing seismicity, and especially the post-injection large-magnitude seismicity, should enable developing better strategies of shut-in to mitigate the risks of post-injection large-magnitude seismic events.

How to cite: Boyet, A., De Simone, S., Ge, S., and Vilarrasa, V.: Triggering Mechanisms of Post-Injection Induced Seismicity Using The Enhanced Geothermal System of Basel (Switzerland), EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-3100, https://doi.org/10.5194/egusphere-egu23-3100, 2023.

17:25–17:35

EGU23-13227

ECS

On-site presentation

Near-real-time microseismic monitoring with machine-learning and waveform back-projection at the Utah FORGE geothermal site

Peidong Shi, Federica Lanza, Francesco Grigoli, and Stefan Wiemer

Deep geothermal energy exploitation necessitates establishing effective fluid circulation paths for heat transfer and managing induced earthquake risk. By detecting and characterizing induced microseismic events, we can provide insights into the fracture network growth and the induced earthquake risk during hydraulic stimulation and geothermal production in enhanced geothermal systems (EGS). During hydraulic stimulation, monitoring has to be performed in near-real-time to provide timely information for assessing potential earthquake risk and for adjusting the stimulation plan. In addition, high-precision microseismic event location is vital for evaluating the connectivity of the stimulated reservoir and designing the trajectory of the production wells. However, achieving real-time monitoring and high-resolution location in a single monitoring workflow is challenging due to the low signal-to-noise ratio and short inter-event time of microseismic events.

To address these challenges in microseismic monitoring, we build a near-real-time monitoring workflow that integrates machine-learning (ML) techniques for efficient event detection and waveform back-projection methods for high-precision event location. The proposed workflow is designed to utilize various pre-trained ML models to deal with the scarcity issue of training datasets in new EGS sites. We apply the proposed workflow to the microseismic dataset collected at the Utah FORGE geothermal site in a playback mode. Because most pre-trained ML models are trained on local earthquake datasets having larger event magnitudes and lower data sampling rates, we implement and evaluate various strategies, such as re-scaling, re-sampling, and filtering, to enhance the performance of pre-trained models on the microseismic dataset. We compare the obtained ML catalog with a reference catalog built from a conventional workflow consisting of automatic phase picking and manual refinement. Due to the application of ML and waveform back-projection techniques, our workflow can nicely separate microseismic events with very short inter-event times (in terms of a second) and cope with events with significant magnitude/amplitude differences, leading to more reliable event detections. Detailed comparisons show that the accuracy of ML phase identification is comparable to and sometimes even superior to manual picking (with a difference in milliseconds), which contributes to precise event locations.

How to cite: Shi, P., Lanza, F., Grigoli, F., and Wiemer, S.: Near-real-time microseismic monitoring with machine-learning and waveform back-projection at the Utah FORGE geothermal site, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-13227, https://doi.org/10.5194/egusphere-egu23-13227, 2023.

Underground research laboratory experiments

17:35–17:45

EGU23-13390

ECS

Highlight

On-site presentation

First results on induced seismicity and its source parameters during hydraulic stimulations in the Bedretto Underground Laboratory, Switzerland

Martina Rosskopf, Virginie Durand, Linus Villiger, Anne Obermann, and Domenico Giardini and the Bedretto Team

The Bedretto Underground Laboratory for Geoenergies and Geosciences (BedrettoLab) is a unique research environment in a tunnel in the Swiss Alps with an overburden of one kilometer. For a first experimental campaign in the context of enhanced geothermal reservoir creation (enhanced geothermal system, EGS), six 150-300 m long monitoring boreholes were drilled in the BedrettoLab and equipped with a multi-component geophysical monitoring system. The seismic sensors are composed of acoustic emission sensors, accelerometers and geophones to cover a magnitude range from nano to micro seismicity. An additional 400 m long stimulation borehole is equipped with a multi-packer system dividing it into 15 separate intervals of 8 to 55 m length.

In the first phase, we performed characterization stimulations in eight intervals using comparable small amounts of injected fluid (350-14000 l) dependent on the length of the interval and their initial transmissivity. In the second phase, we are restimulating the intervals with higher injection volumes. The objective of these experiments is to improve our understanding of the hydro-seismo-mechanical response of the surrounding rock mass to hydraulic injections.

For each interval, the seismic activity varies in number and magnitude and in its spatial and temporal evolution although comparable injection volumes were used. We interpreted the spatio-temporal evolution of event clouds together with the preconditions of the intervals, such as fracture network and transmissivity. Those preconditions can have a big impact on how the rock mass behaves when pressurized during hydraulic stimulations. Depending on the interval, event clusters are distinguished, for which first moment tensors are estimated. The moment tensors show different mechanisms not only for different intervals but also for different clusters. Finally, we compare the seismicity to other observables in the volume like stress and temperature measurements performed in monitoring boreholes to recognize fluid pathways and the response of the rock mass.

How to cite: Rosskopf, M., Durand, V., Villiger, L., Obermann, A., and Giardini, D. and the Bedretto Team: First results on induced seismicity and its source parameters during hydraulic stimulations in the Bedretto Underground Laboratory, Switzerland, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-13390, https://doi.org/10.5194/egusphere-egu23-13390, 2023.

17:45–17:55

EGU23-14231

On-site presentation

Coda Q analysis of active UT measurements to monitor the dm-scale fracture network in response to the STIMTEC hydraulic stimulations, Reiche Zeche URL (Germany)

Aglaja Blanke, Carolin M. Boese, Georg Dresen, Marco Bohnhoff, and Grzegorz Kwiatek

The occurrence of induced microseismicity is commonly used to characterize a stimulated geo-reservoir, e.g. in terms of monitoring the damage and stress evolution in the rock mass. Studies in Underground Research Laboratories (URLs) have the advantage that a close by Acoustic Emission monitoring system (earthquake source – receiver distances < 50 meters) allows to investigate the detailed evolution of induced seismicity at the decimeter scale in response to injection operations. Controlled hydraulic stimulation experiments often benefit from additional active seismic measurements conducted near the source regions of the induced seismic events to better characterize the rock mass. The so-called active Ultrasonic Transmission (UT) measurements produce stronger signals covering a broad frequency range from the centimeter to decimeter scale which may be used to investigate space-time varying fracture network development and attenuation properties. The STIMTEC project in the Reiche Zeche URL in Freiberg (Germany) provides more than 300 active UT measurements that were performed before and after hydraulic stimulations in two boreholes in the targeted rock volume, an anisotropic metamorphic gneiss. To investigate spatio-temporal changes of frequency-dependent scattering attenuation and thus to monitor variations in the local fracture network, we analyzed S-coda waves of 88 spatially representative UT measurements covering a signal frequency content of 1 – 60 kHz. We grouped neighboring UT measurements to estimate frequency-dependent mean-Qc (coda quality factor) values for in total eight UT groups, covering different borehole depth intervals and time periods. Stable Qc estimates for center frequencies between 3 – 21 kHz of octave-width frequency bands were obtained. Our estimates show a characteristic frequency-dependence as observed at the field scale in geological reservoirs. Temporal variations of Qc values are strongly connected to hydraulic stimulation, and these variations are more significant than those resolved from estimated velocity changes. We find indication for healing processes of injection-induced small-scale fractures during a two-months post-stimulation phase indicated by attenuation changes at high frequencies not resolvable at the earth’s surface. The coda analysis further reveals spatial differences of attenuation characteristics supporting previous assumptions based on borehole televiewer logs and mapped structures of an existing fault with larger damage zone that crosses the stimulated rock volume. We conclude that the coda analysis of active UT measurements complements established imaging methods used during experiments in URLs. In particular, coda analysis is a powerful tool for the detection of damage zones and for monitoring changes in local fracture networks with immediate application for imaging geo-reservoirs considered for exploitation or underground storage of gases and liquids.

How to cite: Blanke, A., Boese, C. M., Dresen, G., Bohnhoff, M., and Kwiatek, G.: Coda Q analysis of active UT measurements to monitor the dm-scale fracture network in response to the STIMTEC hydraulic stimulations, Reiche Zeche URL (Germany), EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-14231, https://doi.org/10.5194/egusphere-egu23-14231, 2023.

Posters on site: Thu, 27 Apr, 14:00–15:45 | Hall X5

Chairpersons: Alessandro Verdecchia, Hongyu Yu, Rebecca M. Harrington

X5.394

EGU23-2798

A probabilistic assessment of induced seismicity and fluid flow in the Carboniferous Limestone of northwest England: Implications for geothermal energy

David Healy and Cathy Hollis

Faults slip in response to changes in stress or fluid pressure, and these changes can be natural or anthropogenic. Estimating the likelihood of fault slip for a given change in loading is critical for safe geological storage and energy extraction in faulted rocks, as well as effective communication of risks to policy makers and the public. The energy transition and decarbonization are urgent and essential tasks: we will only be successful if we manage to balance public perceptions of risk with the technical challenges inherent to the exploitation of faulted rock. To accomplish both, we need to do a better job of quantifying the uncertainties in our mechanical and geometrical data.

Measures of fault stability include slip (T_s) and dilation (T_d) tendency, and fracture susceptibility (S_f, the change in fluid pressure to push effective stress to failure). The input values for any of these measures are always uncertain, and they are uncertain to varying degrees. For example, while the vertical stress can be well constrained from wireline density log data, the maximum horizontal stress is generally much harder to quantify from any source.

Probabilistic models of fault stability for the Mississippian (Lower Carboniferous) Limestone underlying much of northern England are presented. This is an undrilled target for geothermal energy. Fault maps are derived from published geological maps and recently reprocessed seismic reflection data. Stress and pressure constraints are derived from legacy onshore hydrocarbon wells and wireline logs. Coefficients of friction and cohesive strength remain poorly constrained, not only in terms of their magnitude, but critically in the shapes of their statistical distributions. In addition, the applicability of simplified indices of fault stability (T_s, T_d, S_f) to complex natural fault zones is questionable, and our predictions could be improved through weighting by information derived from long-term seismological records.

This contribution raises two key points: 1) Laboratory data rarely include repeat measurements from quasi-identical samples of the same rock, and therefore the statistical distributions of friction or cohesion are poorly known. This is important because skewed distributions – and the direction of that skewness (high or low) – can significantly affect predictions of fault stability; 2) What is the best way to weight the calculated probability of slip in complex natural fault zones to account for geometrical weakening? Specifically, is total fault length more or less important than fault smoothness (roughness)?

How to cite: Healy, D. and Hollis, C.: A probabilistic assessment of induced seismicity and fluid flow in the Carboniferous Limestone of northwest England: Implications for geothermal energy, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-2798, https://doi.org/10.5194/egusphere-egu23-2798, 2023.

X5.395

EGU23-4577

CO2 fracturing simulation and reservoir simulation case study in the Liard Basin of Canada

Taehun Lee and Taewoong Ahn

X5.396

EGU23-4598

ECS

Long-term Fluid Injection Can Expedite Fault Development: Riedel Shear Structures Illuminated by Induced Earthquakes in Alberta, Canada

Hongyu Yu, Honn Kao, Bei Wang, and Ryan Visser

Riedel shear structures (RSS) are often observed in the embryonic stage of strike-slip fault development. Typical RSS often involve a set of conjugated fault segments that interact with each other prior to the growth of a through-going fault, that is, the principal deformation zone (PDZ), sub-structures R, R’ and P. Ideally, sub-structures R and P should be symmetric to the PDZ while structures R and R′ are conjugated to each other. It is a critical concept linking the geomechanical behavior of individual earthquakes to structural geology at both local and regional scales. The depiction of RSS can help understand the geomechanical behavior of individual earthquakes. In the meantime, the influence of long-term fluid injections on the developing process of RSS, as manifested by the common occurrences of injection-induced earthquakes, has been rarely addressed.

Here we document a clear case in western Canada where the development of a local RSS system is expedited by 25 years of wastewater injection. RSS are manifested by an earthquake sequence consisting of 187 events (M_L ranging 1.3–3.9) between 2018/01/01 and 2021/07/15 in an area without any previous seismic history.

Focal mechanisms of these events exhibit various faulting types with the majority (87%) being compatible with the background stress regime. The orientation of derived nodal planes and six fault segments depicted from the refined earthquake distribution collectively define the overall geometrical characteristics of RSS, which consists of four primary strike-slip structures striking 19º (R’), 79º (R), 94º (PDZ) and 109º (P), respectively. Mohr-Coulomb failure analysis further suggests a cumulative stress perturbation of up to 10.0 MPa, probably bringing the local structures very close to the critically stressed state.

Overall, our observations suggest that long-term fluid injection can expedite the development of local fault systems. Consequently, it is probably necessary to consider the full dimension of the local/regional RSS rather than the size of individual events in the assessment of the overall injection-induced seismic hazard.

How to cite: Yu, H., Kao, H., Wang, B., and Visser, R.: Long-term Fluid Injection Can Expedite Fault Development: Riedel Shear Structures Illuminated by Induced Earthquakes in Alberta, Canada, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-4598, https://doi.org/10.5194/egusphere-egu23-4598, 2023.

X5.397

EGU23-7498

Deriving the triggering potential in the Lower Rhine Embayment using background seismicity near Weisweiler, Germany

Marco Pascal Roth, Rebecca M Harrington, Brigitte Knapmeyer-Endrun, Claudia Finger, and Marco Dietl

The Earth’s crust is permeated with faults and fractures due to its long tectonic history. Faults and fractures not only act as zones of weakness but can increase permeability and act as conduits to circulate fluids in the underground, making them ideally suited for geothermal energy production. However, human-induced changes to in-situ pressure fields have a documented history of leading to fault (re-)activation in the form of earthquakes (seismic slip) or aseismic slip that does not generate measurable ground motion.

This study aims to quantify the current background earthquake activity, the spatiotemporal relationship to the seismotectonic setting, and the remote earthquake-earthquake triggering propensity in the Lower Rhine Embayment (LRE) in western North-Rhine Westphalia. The study area is targeted for extensive exploration activities focused on geothermal energy production. While regional mean slip rates do not exceed 0.1 mm/yr, paleo-seismic studies suggest that the normal faulting system has hosted a series of ~14 earthquakes of M_w > 5.0 since the 14th century, including the 1992 M_w 5.3 Roermond earthquake. Therefore, estimating background seismicity rates independent of anthropogenic stress perturbation is one important element in developing strategies to minimize the probability of felt earthquakes caused by industrial activity.

We evaluate waveform data from a temporary deployment of 48 seismic stations operating between July 2021 and May 2022 in a radius of roughly 10 km around a future exploration well drilling site in Weisweiler. We implement a machine learning-based earthquake detection algorithm, SeisBench, to detect earthquakes and denoise the continuous seismic waveforms (DeepDenoiser), estimate P- and S-phase arrivals (PhaseNet and Generalised Phase Detection; GPD), and associate earthquake phases using a Bayesian Gaussian mixture model (GaMMA). We locate 81 local earthquakes to complement the 14 recorded in the Earthquake Observatory Bensberg (BNS) catalog for a total of 95 seismic events recorded between July 2021 and December 2021 with magnitudes ranging from 0 < M_L < 1.3. In addition, we use the continuous BNS catalog to evaluate the dynamic triggering potential in the LRE starting in 1990. We select 20 teleseismic mainshocks with M > 6 (1990 – 2015) and M > 7 (2016 – present), as well as the 1992 Roermond M_w 5.3 due to the high amplitude shaking it caused within the study area. Preliminary remote dynamic triggering results suggest that the passing surface waves of the July 2021 M 8.2 Chignik, Alaska earthquake may have triggered a seismic sequence of about 16 locatable earthquakes. The migrating aftershock sequence of the Roermond earthquake also suggests the mainshock caused limited dynamic triggering within the study area, which lies outside of the classical aftershock zone of ~2-3 fault lengths.

How to cite: Roth, M. P., Harrington, R. M., Knapmeyer-Endrun, B., Finger, C., and Dietl, M.: Deriving the triggering potential in the Lower Rhine Embayment using background seismicity near Weisweiler, Germany, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-7498, https://doi.org/10.5194/egusphere-egu23-7498, 2023.

X5.398

EGU23-11883

ECS

Modeling the waveforms of induced seismicity sequences with application to Utah FORGE

Laura Ermert, Federica Lanza, Federico Ciardo, Peidong Shi, Michael Afanasiev, and Stefan Wiemer

We present a modeling workbench for generating hours-long synthetic sequences of induced seismic events. As input, this tool takes observed or synthetic catalogs of induced seismicity including the location, seismic moment and source mechanism of the induced events. It outputs synthetic seismic waveform recordings with added noise.

The purpose of this tool is two-fold: First, it serves to generate synthetic test data that can be used to train detection and location algorithms for induced seismicity monitoring, and to test their efficacy. Monitoring induced seismicity is a key task for the risk management of enhanced geothermal systems, but the relative scarcity of manually labeled data hampers rigorous testing of newly developed algorithms. We intend to partially close the gap in labeled data with synthetics, which can furthermore be used to enhance training data sets for deep learning approaches.

Second, the tool serves as an extension to physics-based models of induced seismicity in geothermal reservoirs, such as hydro-mechanical models, providing an efficient way of turning their induced event clouds into seismic “recordings”.

The workflow is based on the spectral-element solver Salvus and on Python, and relies on source-receiver reciprocity to decrease computational cost for large sets of induced events. As first application site, we chose the Utah Frontier Observatory for Research in Geothermal Energy – FORGE. This first application comprises a realistic digital representation of the reservoir and its surroundings, including topography, geologic structure and crustal scattering. We will show exemplary synthetic induced seismicity sequences based on an observed catalog from the 2022 stimulation at the Utah FORGE, as well as on a hydro-mechanically modeled synthetic catalog. The output waveform sequences are to be publicly shared for benchmarking monitoring workflows for induced events, in an effort to contribute to the de-risking of enhanced geothermal systems.

How to cite: Ermert, L., Lanza, F., Ciardo, F., Shi, P., Afanasiev, M., and Wiemer, S.: Modeling the waveforms of induced seismicity sequences with application to Utah FORGE, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-11883, https://doi.org/10.5194/egusphere-egu23-11883, 2023.

X5.399

EGU23-8557

ECS

Quantification of pre-operational seismic hazard of deep geothermal systems exemplified in the greater Ruhr region (Germany)

Michal Kruszewski, Alessandro Verdecchia, Oliver Heidbach, Rebecca M. Harrington, and David Healy

In 2018, after more than 700 years, the last black coal mine ceased operation in the greater Ruhr region in western Germany. In addition to the repurposing of the unused subsurface infrastructure for heat storage projects, the utilization of deep geothermal resources, located below mining levels in the Devonian Massenkalk formations, is the most promising way of facilitating the green energy transition in this highly populated region. The operation of fluid injection or withdrawal during geothermal production alters subsurface stress and poses questions about the possible reactivation of faults in the greater Ruhr region, with strong implications for seismic hazard potential. Deep geothermal resources depend on permeable pathways (i.e., fault zones and fracture networks) in the subsurface, which are known to have certain intrinsic pre-operational seismic hazard. In this study, we evaluate the probability of reactivation (using the so-called slip tendency and fracture susceptibility indicators) and dilation tendency of major faults accounting for uncertainties of in situ stress conditions, fault geometries, and frictional properties. Furthermore, we investigate the spatio-temporal evolution of the slip tendency of a major fault in the region during geothermal operations utilizing coupled thermo-hydro-mechanical numerical models. This study benefits from a recently published in situ stress database of the greater Ruhr region derived from 429 hydrofracturing tests performed across several coal mines. The assessment of fault reactivation utilizing probabilistic and coupled numerical modelling approaches, as presented in this study, aims at de-risking the exploration of deep geothermal systems in the greater Ruhr region and can hopefully serve as an example for quantification of pre-operational seismic hazards in other regions. Based on preliminary results, we find that the NW-SE striking faults are much more prone to reactivation, in comparison to the NE-SW ones, which are oriented unfavorably under prevailing stress conditions. We also find that thermal stress contributes significantly to fault stability, especially during long-term geothermal production.

How to cite: Kruszewski, M., Verdecchia, A., Heidbach, O., Harrington, R. M., and Healy, D.: Quantification of pre-operational seismic hazard of deep geothermal systems exemplified in the greater Ruhr region (Germany), EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-8557, https://doi.org/10.5194/egusphere-egu23-8557, 2023.

X5.400

EGU23-12783

ECS

Performance comparison of newly developed hydro-mechanical (hybrid) models for real-time induced seismicity forecasting

Victor Clasen Repolles, Antonio Pio Rinaldi, Federico Ciardo, Luigi Passarelli, and Stefan Wiemer and the Bedretto Team

The Adaptive Traffic Light System (ATLS) is a seismic risk mitigation tool that can be used during geoenergy exploitation projects such as in the creation of Enhanced Geothermal Systems (EGS). In real-time applications, ATLS needs to include a data assimilation scheme in order to produce adaptive and time-dependent probabilistic seismicity forecasts by using the maximum available information at the moment of the assessment during such industrial operation. Numerical models capable to robustly forecast in real-time the temporal evolution of induced seismicity while properly accounting for uncertainties are the core elements of a functioning ATLS. In this respect, hydro-mechanical (HM) hybrid models are suitable models since they combine both statistical and physics-based assumptions to forecast induced seismicity. They include an accurate modeling of the physical processes involved in the generation of seismicity caused by human activity by keeping the associated computational cost low. In this work, we have developed two classes of simplified hybrid models that are based on 1D radial pore-fluid diffusion and fluid injection. The first class (HM0) accounts for linear pore-fluid diffusion from the fluid source, while the second class (HM1) accounts for the non-linear response of the medium due to pressure-dependent permeability variations upon fluid injection. Each class is sequentially coupled to two stochastic seismicity models that trigger seismicity if the respective space-time pressure solution reaches a critical value according to the Mohr-Coulomb criterion. However, both seismicity models differ in the way they simulate seismicity. One model uses an analytical approach based on probability density functions of model parameters to simulate seismicity (CAPS), while the other model simulates seismicity via stochastic seed approach (SEED). In a hindcast experiment, we test the models’ real-time forecasting capabilities within a data assimilation scheme using datasets from in-situ injection experiments that encompass different spatial scales. An essential step towards building a reliable ATLS is to properly weight each candidate model according to its respective contribution to the seismic hazard calculation. Therefore, an important part in our calculations is to evaluate and compare the forecasting performance of each model for each of the presented cases. Our findings show that models that take into account reliable distribution of uncertainties on a-priori selected model parameters as well as models accounting for a physically more accurate pressure solution in space and time generate better forecasting performances.

How to cite: Clasen Repolles, V., Rinaldi, A. P., Ciardo, F., Passarelli, L., and Wiemer, S. and the Bedretto Team: Performance comparison of newly developed hydro-mechanical (hybrid) models for real-time induced seismicity forecasting, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-12783, https://doi.org/10.5194/egusphere-egu23-12783, 2023.

X5.401

EGU23-15982

ECS

Detecting and mapping the induced seismicity of a planned EGS in Helsinki/Espoo, Finland

Tommi Vuorinen, Gergor Hillers, Kati Oinonen, Jennifer Hällsten, George Taylor, and Martin Gal

The company ST1 Oy planned to construct an Enhanced Geothermal System (EGS) with two boreholes drilled down to ca. 6 km depth beneath the Aalto University campus in Otaniemi, Espoo, on the border of Helsinki. The company performed two stimulations, in June–July 2018 and in May 2020, with a goal of opening up a water reservoir and achieving water circulation between the boreholes. The stimulation periods, which induced thousands of earthquakes, and their immediate surroundings were monitored by both permanent and temporary seismic networks with over 100 stations located within a few tens of kilometers of the site. Between and after the stimulations, the site was and is still being monitored by a relatively dense, consisting in total of ca. 20 stations, regional surface station and company installed borehole station networks.

We have developed a cross-correlation based event detector which uses the existing ISUH manually analysed catalogues primarily of the 2018 and 2020 stimulation period seismicity, complemented by automatically picked catalogues from IMS, to detect events from the collected continuous waveform database. The 4-step detector – templating, detecting, event filtering & relocating – can run on varying station configuration and is able to detect events down to ML -0.5 – -1.0 with the station-event geometry around the EGS. We present here the results of the detector run from the beginning of the dense stimulation monitoring in May 2018 to the end of 2022 providing a comprehensive anatomy of the EGS induced seismicity.

How to cite: Vuorinen, T., Hillers, G., Oinonen, K., Hällsten, J., Taylor, G., and Gal, M.: Detecting and mapping the induced seismicity of a planned EGS in Helsinki/Espoo, Finland, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-15982, https://doi.org/10.5194/egusphere-egu23-15982, 2023.

X5.402

EGU23-8515

ECS

The effect of steady-state flow on induced stresses along displaced faults

Pavan Cornelissen and Jan Dirk Jansen

Pore pressure changes due to fluid injection or withdrawal alter the rock stresses, which may potentially induce seismic events. Activities that have been associated with induced seismicity include geothermal energy production, subsurface gas storage, and natural gas production. Physics-based models are required to gain insight in the processes that lead to induced seismicity. When computing induced stresses with such models, a commonly made simplification is the assumption of uniform pressure changes across the entire reservoir. In reality, pressure gradients arise due to fluid production or injection. Here, we assess the effect of non-uniform pressure fields under steady-state flow conditions on induced stresses. We employ (semi-)analytical techniques to compute the corresponding pressure field and fault stresses. We are particularly interested in reservoirs with displaced faults (i.e., cases with nonzero fault offset), as shear stresses tend to concentrate at the reservoir corners along the faults. The stress profile along the fault becomes asymmetric under steady-state flow. The effect of fluid flow on fault stresses is larger in case of injection than in case of depletion. Injection with up-dip flow results in increased zones of fault slip near the bottom of the reservoir, while injection with down-dip flow results in increased slip near the top of the reservoir. The significance of steady-state flow on induced stresses can be estimated from the ratio of the average pressure change along the fault and the applied pressure gradient. The effect of steady-state flow is most relevant at the start of production or injection and diminishes with time. Thus, the effect of steady-state flow is only expected to be relevant when initially critically stressed faults are present. For non-critically stressed faults, the assumption of uniform depletion or injection is expected to be reasonable. An order-of-magnitude estimate of the effect of steady-state flow across displaced faults in the Groningen natural gas reservoir shows that the effect on fault stresses is probably negligible. A similar estimate of the effect in typical low-enthalpy geothermal doublets indicates that steady-state flow may possibly play a small role, in particular close to the injector. Nevertheless, site-specific assessments are necessary to quantify the effect in greater detail.

How to cite: Cornelissen, P. and Jansen, J. D.: The effect of steady-state flow on induced stresses along displaced faults, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-8515, https://doi.org/10.5194/egusphere-egu23-8515, 2023.

X5.403

EGU23-16577

ECS

GEOBEST2020+ baseline seismic monitoring workflow and network performance evaluation for deep geothermal projects in Switzerland

Verónica Antunes, Toni Kraft, Philippe Roth, Tania Toledo, and Stefan Wiemer

Deep geothermal is a clean and renewable source of energy with a high potential for heat and electricity production which can help Switzerland meet its energy and climate objectives. Worldwide, several geothermal projects have been successfully operated for decades. Unfortunately, some projects have also been suspended due to unexpected levels of induced seismicity. Thus, adequate risk management is essential to establish safe and economically viable geothermal projects.

In Switzerland, the subsurface is under the sovereignty of the cantonal authorities. Within the GEOBEST2020+ project, the Swiss Seismological Service (SED) supports the cantons in adequately handling the risk of induced seismicity associated with deep geothermal projects. Funded by the Federal Office of Energy (SFOE) in the scope of its SwissEnergy program, the GEOBEST2020+ program aims to provide operator-independent seismological consulting and baseline seismic monitoring services to the cantonal authorities.

In this framework, we deploy dedicated seismic networks in the vicinity of the monitored projects. These networks must be sensitive enough to follow the evolution of microseismicity and allow the operators to run traffic-light systems and take action before larger events occur. Before the station installation, we perform a careful site survey analysis, considering the background noise conditions and evaluating the signal-to-noise ratio at each site. To evaluate beforehand the detection sensibility of a seismic network, we estimate the Bayesian Magnitude of Completeness (BMC), optimized for Switzerland. We additionally estimate the theoretical location uncertainties inside the network considering the background noise level at the stations and the network geometry.

Here we show the comparison results between our methodology to the ground truths of several years of continuous monitoring data at Lavey-les-Bains, canton of Vaud. We use a combination of three types of detection methods: Machine Learning combined with migration methods (MALMI), coherence (Pyrocko/Lassie) and template matching (QuakeMatch), to produce a high-quality, manually revised seismic catalog. We compare our theoretical methods of network performance to the real data measured at the geothermal site.

How to cite: Antunes, V., Kraft, T., Roth, P., Toledo, T., and Wiemer, S.: GEOBEST2020+ baseline seismic monitoring workflow and network performance evaluation for deep geothermal projects in Switzerland, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-16577, https://doi.org/10.5194/egusphere-egu23-16577, 2023.

X5.404

EGU23-11422

An Attempt to Control Induced Seismicity and On-Fault Stress Pre-Conditioning in the Bedretto Underground Laboratory, Switzerland

Mohammadreza Jalali, Paul Selvadurai, Luca Dal Zilio, Virginie Durand, Anne Obermann, Men-Andrin Meier, and Florian Amann and the Bedretto Team

Hydraulic stimulation has been used as a standard practice for rock mass treatment to fulfill the intended engineering goal of the stimulation procedure in various applications such as oil & gas, mining, and enhanced geothermal systems (EGS). High-pressure fluid injection during the hydraulic stimulation perturbs the local stress field around the injection borehole which has a direct influence on the hydro-mechanical behavior of the rock mass and the measured induced seismicity. Hydraulic stimulation is usually conducted using a predefined injection protocol, a sequence of pressure and/or flow rate injection cycles. Extensive attempts have been designed to develop a proper injection protocol in laboratory and field scales to reach the stimulation goals as well as reduce the risk of induced seismicity.

In the context of the ERC “Fault Activation and Earthquake Rupture (FEAR)” project, injection protocols are developed to further our understanding of earthquake rupture processes such as nucleation and premonitory slip. A series of stress-preconditioning injection strategies were tested in the Bedretto Underground Laboratory for Geoenergies and Geosciences, Switzerland (Ma et al., 2022, Solid Earth), using an existing stimulation and monitoring array (Plenkers et al., under rev. with Sensors). The main objective of these tests was to develop atypical injection strategies that could be implemented in different geological conditions that target aspects of induced seismicity associated with high-pressure fluid injection. We performed two injections and monitored the rock mass response using six monitoring boreholes each equipped with multi-component seismic monitoring systems, pressure, temperature, and distributed strain sensing arrays. Two different injection protocols have been applied in two packed-off injection intervals of a stimulation borehole, allowing us to characterize the hydro-mechanical response of the rock mass around these intervals. Each test was associated with pre- and post-characterization tests including hydraulic and hydro-mechanical tests such as HTPF tests.

The preliminary results of these tests show a direct dependency of the location and magnitude of seismicity with injection protocols. Moreover, the effect of the injection protocols on the characteristics of the rock mass such as transmissivity and normal stress could be quantified. Preconditioning of the fault appears to require low permeability, preferably undrained, fault structures and suffers if any fluids pathway are mitigated by the long range, slow pressurization of this region. Complex fluid pathways made re-stimulation and precondition of the target structures difficult. The outcomes of these tests play a key role in the design of upcoming FEAR stimulation experiments to better understand the earthquake physics on a field scale.

References

Ma, X., Hertrich, M., Amann, F., Bröker, K., Gholizadeh Doonechaly, N., et al., 2022. Multi-disciplinary characterizations of the BedrettoLab–a new underground geoscience research facility. Solid Earth, 13(2), pp.301-322.

Plenkers K., Reinicke A., Obermann A., Gholizadeh Doonechaly N., Krietsch H., et al., under rev. with Sensors, Multi-disciplinary monitoring networks for mesoscale underground experiments: Advances in the Bedretto Reservoir Project.

How to cite: Jalali, M., Selvadurai, P., Dal Zilio, L., Durand, V., Obermann, A., Meier, M.-A., and Amann, F. and the Bedretto Team: An Attempt to Control Induced Seismicity and On-Fault Stress Pre-Conditioning in the Bedretto Underground Laboratory, Switzerland, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-11422, https://doi.org/10.5194/egusphere-egu23-11422, 2023.

X5.405

EGU23-13671

ECS

Maximum injection-induced earthquake magnitude in discrete fracture networks

(withdrawn)

Mohammad Javad Afshari Moein

X5.406

EGU23-15041

ECS

Highlight

Dynamic forecasting of the next largest earthquake during hydraulic fracturing stimulations

Iason Grigoratos, Grzegorz Kwiatek, and Stefan Wiemer

Most traffic-light systems (TLS) related hydraulic fracturing stimulations still adopt the maximum observed magnitude as the decisive metric to aid decision making by stakeholders. However, waiting for the red-light magnitude to be observed is not a proactive stance, especially given that jumps of up to two magnitude units are evidently common enough between events. Clearly there is a need to actively forecast rather than to passively record the size of the next largest earthquake (NLE). In this study, we demonstrate that we can do just that using an ensemble of 6 existing models from the literature designed with similar purposes in mind (Shapiro et al. 2013; McGarr 2014; Mendecki 2016; van der Elst et al. 2016; Galis et al. 2017; Cao et al. 2020). Following a logic-tree approach, these 6 models are calibrated and dynamically weighted in near real-time using as sole inputs the initial parts of the earthquake catalogs and the reported injection rates. The proposed forecasting tool is tested against 18 past stimulations from 9 different Enhanced Geothermal Systems around the world (Helsinki, Basel, Soultz, Cooper Basin, Basel, Pohang, FORGE, Paralana, Newberry). Overall, the results indicate a consistent (across sites and time) and accurate estimation of the next largest magnitude with a tight uncertainty range (1σ) of less than 0.5 magnitude units. Our proposed framework underestimated the next largest magnitude only in one occasion (out of the 18 stimulations), while reliably maintaining a tight safety margin of less than 1 magnitude unit. We recommend that the forecasted NLE replaces the largest observed magnitude as the default metric adopted by future TLS governing any type of fluid-injection operation.

How to cite: Grigoratos, I., Kwiatek, G., and Wiemer, S.: Dynamic forecasting of the next largest earthquake during hydraulic fracturing stimulations, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-15041, https://doi.org/10.5194/egusphere-egu23-15041, 2023.

X5.407

EGU23-13239

ECS

Repeating earthquakes induced during a decameter-scale hydraulic stimulation experiment at the Grimsel Test Site, Switzerland

Linus Villiger, Toni Kraft, Valentin Gischig, Hannes Krietsch, and Stefan Wiemer

We performed a similarity based hierarchical cluster analysis and carried out a master event relative relocation of seismicity induced during the hydraulic stimulation of a brittle and highly fractured shear zone. The shear zone was stimulated in an extend of about 10 m and is located in crystalline rock at the Grimsel Test Site, Switzerland. The objective of the experiment was to improve our understanding of stimulation processes associated with high-pressure fluid injections used for reservoir creation in enhanced geothermal systems. In addition to the seismic network exhibiting high spherical coverage at close distances (i.e., eight acoustic emission sensors ~10 – 15 m away from the injection interval), fluid pressures, rock mass deformations and fault dislocations were monitored around the injection location. 1/3 of the located seismic events could be assigned to 44 repeater families exhibiting high waveform similarity (i.e., >95%) when comparing waveform similarities sequentially in time. When comparing similarities to the chosen master event, similarities decay continuously within a repeater family which also reflects in a spatial migration of hypocenters. However, source areas of seismic events within repeater families mostly overlap assuming an appropriate stress drop. But the spatial extent of the observed migration of hypocenters compared to the measured mechanical deformation in the volume is larger and thus suggests that multiple asperities are responsible for seismic events within a repeater family. The repeater families themselves are distributed in clusters which are attributed to the stimulation of distinct fractures in the damage zone of the targeted shear zone. No repeating events were found in the seismicity accompanying a tensile fracture splaying off the stimulated pre-existing fractures. Statistically, events of repeater families are rather Gaussian than power-law distributed. The extended analysis of this multisensor array data presented here reveals a highly complex fracturing process on the 10 m scale during hydraulic stimulation (see also an earlier study of the same experiment focusing on source mechanism Villiger et al. (2021)).

References

Villiger, L., Gischig, V. S., Kwiatek, G., Krietsch, H., Doetsch, J., Jalali, M., Amann, F., Giardini, D., & Wiemer, S. (2021). Meter-scale stress heterogeneities and stress redistribution drive complex fracture slip and fracture growth during a hydraulic stimulation experiment. Geophysical Journal International. https://doi.org/10.1093/gji/ggab057 (Geophysical Journal International)

How to cite: Villiger, L., Kraft, T., Gischig, V., Krietsch, H., and Wiemer, S.: Repeating earthquakes induced during a decameter-scale hydraulic stimulation experiment at the Grimsel Test Site, Switzerland, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-13239, https://doi.org/10.5194/egusphere-egu23-13239, 2023.