Topics of interest include, but are not limited to:
- Advances in stress orientation and magnitude estimation
- New methodologies for 3D structural and geomechanical modelling, including deterministic, stochastic, and hybrid approaches
- Case studies highlighting the application of 3D structural modelling and stress state estimation
- Integration of geophysical and geological data in model-based inversion for improved subsurface characterization
- Advances in computational efficiency and uncertainty quantification in inversion techniques
- Innovative use of machine learning and AI in enhancing both geological models and inversion results
- Insights into the governing mechanics of seismotectonic processes
This session brings together geoscientists, modellers, and computational experts to discuss the latest advancements and challenges, offering insights into the future direction of characterizing the present subsurface stress state and 3D structural geological modelling.
The Swiss Geological Survey (SGS) is the competence centre for the subsurface and georesources of the Swiss Confederation. Between 2024 and 2030, the SGS is leading the Swiss Alps 3D (SA3D) project, which consists of eight modelling and research projects involving several universities. The aim is to develop a consistent, large-scale underground 3D geological model of the main contacts and structures of the Swiss Alps. This model will serve as a regional framework for future higher resolution 3D models, enabling a wide range of applications in infrastructure planning, groundwater studies, natural hazard assessment, education and research. Furthermore, the development of a large-scale, consistent model will promote the establishment of a collaborative scientific community in the field of Alpine geology and 3D geological modelling.
SA3D has been preceded by a four-year pilot study (2019 - 2023), which resulted in an explicit 3D geological model of the Aar Massif (Central Alps). The study highlights the importance and value of utilising 3D geological models when investigating complex geological systems, such as an orogen. In this contribution, we present the results of this pilot study to demonstrate the potential of the large-scale 3D geological models constructed in SA3D for a wide range of applications.
In fact, modelling 3D network of structures and lithostratigraphic contacts of mountain ranges provide strategic insights into the still largely unexplored subsurface of these regions. This is essential for a sustainable infrastructure development and regional assessment of primary resources. Furthermore, the characterization of large-scale 3D fault patterns is relevant for understanding the effects of tectonic preconditioning on the distribution of natural hazards and meteoric water penetration and upflow in orogens. This may have important implications for regional-scale hazards mitigation and for the exploitation of thermal anomalies in orogenic geothermal systems through realistic numerical simulations, as well as for the evaluation of the influence of meteoric water on the seismicity of regional faults.
How to cite: Musso Piantelli, F., Kurmann, E., Wehrens, P., Baland, P., and Müller, H.: Large-scale 3D structural geological models in alpine regions: impact and societal utilities., EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8213, https://doi.org/10.5194/egusphere-egu25-8213, 2025.
Geological 3D modelling in metamorphic belts remains a significant challenge in structural geology due to both mathematical and geological complexities. These challenges stem from the need for software capable of interpolating polydeformed surfaces explicitly or implicitly, while at the same time addressing the geological and topological meaning of these surfaces, i.e., the “geological legend” of the 3D model.
Traditional 3D geological modelling uses the boundary representation paradigm, where geological units are represented as hollow volumes bounded by discretized surfaces, typically stratigraphic boundaries or faults. Explicit interpolation methods generate these surfaces individually, possibly leading to inconsistencies. In contrast, implicit methods interpolate entire stratigraphic sequences in a single step, enabling faster workflows and ensuring mathematical consistency. Moreover, implicit methods produce a continuous (locally discontinuous at faults) volumetric “stratigraphic field” that assigns a scalar value representing a geological absolute or relative age, and boundaries are extracted a-posteriori (hence the name of the methods). Extensions of this approach, known as “GeoChron Model” or “time-aware geomodelling,” enable the assignment of ages to depositional, intrusive, or deformative events, linking the mathematical model to a well-defined sequence of geological events.
Here we propose a workflow that combines implicit and explicit modelling to facilitate conceptual interpretation, ensuring topologically and geologically consistent 3D model reconstruction in metamorphic belts. These regions pose particular challenges because time-aware geomodelling is often inapplicable due to the ill-defined or heterogeneous ages of tectonic boundaries, lithologies in tectono-metamorphic units, and deformation-related features like metamorphic foliations.
In our approach, 3D surfaces are analysed and labelled based on their topological relationships with surrounding geological objects in a preliminary conceptual modelling step, where both surface and volume perspectives are considered. Since boundary surfaces can have multiple roles depending on the geological context and might have been reactivated in polyphase deformation, it is essential to implement a systematic classification of volumes, that are distinguished as tectono-metamorphic, tectono-stratigraphic, or intrusive units (implying different boundary surfaces).
A critical strategy is the use of a time-aware legend wherever possible, such as for geological bodies with known absolute or relative ages. When age information is unavailable, as in very old basement complexes, or for coeval but spatially distinct units (e.g., ophiolite sequences emplaced at different crustal levels), a reasonable pseudo-stratigraphy is adopted (e.g. using relative structural levels instead of stratigraphic age).
Our combined workflow provides a structured and replicable methodology for addressing the unique challenges of 3D geological modelling in metamorphic belts. By systematically handling complex geological features, topological relationships, and polydeformed surfaces, it ensures more consistent and reliable geological models. This framework is expected to enhance interpretations in future studies and advance our understanding of metamorphic belts.
How to cite: Monti, R., Bistacchi, A., Hussain, W., Herwegh, M., and Musso Piantelli, F.: Building a geological legend for 3D geomodelling in metamorphic belts, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16567, https://doi.org/10.5194/egusphere-egu25-16567, 2025.
We present a transdimensional inversion to quantify stratigraphic and petrophysical uncertainties in 2D stratified subsurface models. The objective is to infer the number and position of geological units and their associated properties during inversion. The transdimensional framework relies on a reversible jump Markov chain Monte Carlo (RJMCMC) sampler, which provides self-adaptive capabilities for the parameterization to evolve with the data, and converge to parsimonious posterior solutions. These solutions balance model complexity with the information provided by diverse datasets, such as well logs, seismic surveys, and well tests, which can be integrated within a joint inversion framework. Nevertheless, parameterizations must be carefully defined, as ensuring a small number of parameters is required to maintain reasonable computational times. In this talk, we present an overview of the different geometrical and petrophysical parameterizations that can be used for this purpose. Starting from the classical 1D "layer-cake" model with piecewise constant properties, often employed in geophysics, we progressively introduce more complex parameterizations which better approach the complexity of subsurface layers. These include inclined layers, anticlines, synclines, faulted structures, and lateral variability. By moving towards increasingly realistic parameterizations, the methodology aims to improve the estimation of stratified properties while accounting for structural and stratigraphic variability. Synthetic and real-world applications with various data types will be briefly presented to demonstrate the ability of the sampler to recover coherent results when the parameterization and the noise model are appropriately defined. Overall, this approach provides a unified and adaptable framework for geomodeling, paving the way for improved subsurface characterization and uncertainty quantification in 3D.
How to cite: Herrero, J., Caumon, G., and Bodin, T.: From layer-cake models to complex subsurface structures: a flexible transdimensional inversion approach, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17387, https://doi.org/10.5194/egusphere-egu25-17387, 2025.
Neural fields (a.k.a. Spatial Neural Networks) are neural networks that take spatial coordinates as inputs and output target (interpolated) variable(s). They can learn arbitrarily complex functions and, because they are auto-differentiable, can be easily constrained by their spatial derivatives. In this contribution, we build on recent work to further explore applications of neural fields for geological modelling.
While scalar fields have been used to represent subsurface geology before, constraining these fields is a challenge. Geological models are under-constrained, requiring e.g. regularisation to derive geologically sensible results, making it difficult to learn high-frequency geometric details. Furthermore, unlike most applications of neural networks, neural fields have low dimensional inputs, which further limits their ability to learn high-frequency features during training.
We address these challenges by using random Fourier feature encoding, a technique inspired by computer vision which transforms spatial inputs into a higher-dimensional feature space by applying sine and cosine functions weighted by randomly initialized parameters. Loss functions based on the value and gradient of the output scalar field are then used to learn the geometry of subsurface geology. Significantly, we also impose a weak-harmonic constraint on the field by minimising the divergence of the scalar field’s gradient, which penalises the formation of closed scalar field isosurfaces (i.e., “bubbles”) which violate the layered topology of stratigraphic sequences.
We demonstrate our approach on several synthetic geological datasets, and show how the neural field approach can explore the possible solution space using different random initialisations, thereby helping quantify uncertainty. To conclude, we suggest that neural fields could provide a powerful tool for future geological modelling workflows, due to their flexibility and ability to constrain diverse aspects of geological models.
How to cite: Kamath, A., Thiele, S., and Gloaguen, R.: (Auto) Differentiating geology: Geological modelling with random Fourier features and neural fields, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9746, https://doi.org/10.5194/egusphere-egu25-9746, 2025.
Developing accurate 3D geological models of the subsurface is crucial, as they provide the foundations for multiple uses (e.g., resource exploration and exploitation, geohazard assessment, and environmental geoscience). The construction of these models is an intrinsically integrative task, which jointly takes into account all available data and information from multiple sources, i.e. structural geology, stratigraphy, petrophysics, geophysics. Despite the progress made in automating the integration, in particular with recent advances in artificial intelligence, human interpretation remains essential. Consequently, the performance and limitations of human geological interpretation need to be carefully assessed particularly when subsurface data are incomplete, sparse and imprecise. In this context, the French geological survey – BRGM – has set up a blind interpretation exercise that enables the geo-interpreters to test their ability to answer two main operational questions when jointly analyzing geological and multi-source geophysical datasets (seismic, gravimetric, electric/magneto-telluric): (q1) Is it possible to detect and characterize structural traps and potential migration pathways at several kilometers depth? (q2) Do the errors associated with each of the different datasets influence / affect / bias the geological interpretation? If so, how?
To this end, the following procedure was applied: (1) a simplified 3D geological model was constructed using a real exploration project dedicated to the characterization of helium reservoirs in a deep Permian sedimentary basin; (2) two cross-sections were extracted from the model with realistic petrophysical properties to constrain geophysical forward models, i.e. gravimetric, magneto-telluric, and seismic; (3) these geophysical "truths" were intentionally degraded to reflect measurement errors and realistic processing. During the 6-hour exercise, the degraded geophysical datasets along with geological data from one borehole and from the 1:1,000,000 scale geological map were provided to three teams of interpreters - each consisting of a geologist and a geophysicist, with the aim of interpreting the two cross-sections.
This communication summarizes the main lessons learned from this exercise by discussing the interaction between data resolution, quality and reliability, and cognitive biases. It points out the value of fostering recurrent exchanges with data producers during the geological interpretation process. Finally, we propose recommendations for improving the links between data-centric and human-centric inversion procedures.
How to cite: Rohmer, J., Allanic, C., Bitri, A., Dubois, F., Grataloup, S., Jacob, T., Stopin, A., Coueffe, R., Faure, A., Peyrefitte, A., Portal, A., Raingeard, A., Wawrzyniak, P., Chassagne, R., Coppo, N., Darnet, M., and Calcagno, P.: Uncertainties in joint analysis of geological and multi-source geophysical data: lessons from a blind interpretation exercise, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6893, https://doi.org/10.5194/egusphere-egu25-6893, 2025.
The Eastern Fram Strait is of high geo-mechanical interest for studying how regional forcing has influenced the continental margin’s hydrology, having an impact on slope stability and climate evolution. The region is subjected to tectonic stress fields induced by a cluster of mid-ocean ridges and transform faults as well as to glacial stresses associated with the evolution of the Svalbard-Barents Sea Ice Sheet. Studies over the last decade show that seafloor methane seepage is impacted by the spatiotemporal evolution of the aforementioned stress factors. However, in-situ stress measurements from the area have been lacking to constrain stress regime inferences from geophysical data and stress models.
During International Ocean Discovery Program (IODP) Expedition 403, borehole resistivity images were obtained in Holes U1618B within the second northernmost site and U1623D within the southernmost site in the expedition using Fomation-MicroScanner (FMS). The maximum and minimum principal horizontal stress orientations can be inferred from those borehole resistivity images that indicate failures of borehole walls subsequently caused after drilling. After processing and observation in combination with caliper logs, borehole breakouts and/or drilling induced tensile fractures were recognized in both boreholes. For U1618B (located on the Vestnesa ridge, an active seafloor seepage system), very scarce borehole compressive failures were recognized possibly due to weak horizontal compression implying normal stress regime. For U1623D (located offshore the Bellsund fjord), there were multiple borehole failures with large fluctuations of azimuth implying presence of local stress or small differential stress. Our results would provide the first actual data to discuss stress fields in the Eastern Fram Strait in combination with existing model-based studies contributing the advance of understanding geo-mechanics regarding seafloor fluid dynamics.
How to cite: Sakai, Y., Lin, W., Plaza-Faverola, A., G. Lucchi, R., St. John, K., and Ronge, T. and the IODP Exp.403 Science Party: Observation of borehole resistivity images from IODP Exp.403 implying stress fields exerted on the sedimentary succession in the Eastern Fram Strait, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14942, https://doi.org/10.5194/egusphere-egu25-14942, 2025.
Microseismicity continuously occurs within active seismogenic faults, where major earthquakes might be generated. These small events offer critical insights into the geometry and mechanical state of faults. To enhance the detection of low-magnitude events, often obscured by seismic noise, 200 seismic stations were deployed from September 2021 to August 2022 across the complex normal-faulting environment of the Southern Apennines, organized into 20 sub-kilometric arrays, as part of the DETECT experiment. Using this dense network, an enriched seismic catalog was generated by integrating machine learning and template matching techniques, which has allowed to identify ~3,600 earthquakes with magnitudes -1.5 < M < 2.8.
Here, we resolved focal mechanisms for 289 earthquakes in this catalog. Our analysis is based on the inversion, with the software FPFIT, of the P-wave onset polarities determined by leveraging a convolutional neural network, incorporating a tailored weighting scheme. Dense monitoring allows to increase the number of focal mechanisms by a factor of ~2 compared to six years microseismicity observed with ordinary seismic network. The retrieved fault parameters align with the orientation and normal kinematics of the primary fault segments associated with the 1980 M6.9 Irpinia earthquake, but they also reveal minor occurrences of inverse and oblique faulting. Fault plane solutions are used to constrain the orientation and relative magnitudes of the stress field components, iteratively discriminating between the principal and auxiliary nodal planes by introducing fault plane instability. Our analysis reveals a stress field characterized by a near-vertical maximum compressive stress (σ1) and quasi-horizontal intermediate (σ2) and least compressive (σ3) stress components. The azimuth of σ3 aligns with the anti-Apenninic direction of the extensional regional stress field, consistent with previous estimates derived from long-term microseismic observations. In the central sector, the stress field orientation supports the presence of a kinked structure identified through earthquake relocations. Moreover, the high number and spatial distribution of resolved fault planes enable the investigation of potential small-scale stress field variations. By inverting focal mechanisms within the Northern, Central, and Southern sectors of the Irpinia region, we retrieve individual stress tensors, which reveal spatially coherent stress orientations and relative magnitudes of stress components across the region. These findings demonstrate the feasibility of accurately resolving stress fields from short-term array monitoring, even in the absence of major earthquakes, highlighting the potential for detailed exploration of the stress field in tectonically complex regions.
How to cite: Scotto di Uccio, F., Martínez-Garzón, P., Meier, M.-A., Picozzi, M., and Festa, G.: Resolving focal mechanisms and stress field from microseismic events with short-term dense monitoring in the Southern Apennines, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10139, https://doi.org/10.5194/egusphere-egu25-10139, 2025.
The stress field is a key factor controlling the earthquake nucleation, rupture propagation, and arrest processes, which is essential for understanding the rupture process and evaluating the earthquake hazard. We utilize the focal mechanism rotation phenomenon that occurs after a large earthquake to determine the absolute stress field around faults by employing a non-linear inversion. We assume that there are numerous existing faults in the crust, and some of them with certain directions are allowed to rupture according to the Coulomb failing criterion. The co-seismic stress perturbance induced by the mainshock, calculated from co-seismic slip models, causes the focal mechanism rotation of aftershocks compared to earthquakes before the mainshock. Thus, we use the Bayesian method to invert the absolute stress tensor and friction coefficient before the mainshock for effectively explaining the focal mechanism rotation. Results of synthetic tests indicate that the true parameters can be tightly constrained by accurately fitting the P-axis distributions before and after the mainshock, especially when conjugate faults are absent, by incorporating the prior distribution of the P-axis and friction. Finally, we apply our inversion algorithm to the 2011 Tohoku earthquake. Based on the appearance of normal earthquakes at depths shallower than 10 km and comparisons of data fitting for different co-seismic slip models, we infer that the largest co-seismic slip occurred shallower than the hypocenter and extends upward to the seafloor. The optimal inversion results show an increase in deviatoric stress magnitude with depth, coupled with a rotation of the maximum compressional stress direction from horizontal to vertical. This suggests that deep creep-slip loading significantly influenced the stress field in the stick-slip zone. Moreover, we calculated the pore pressure from the isotropic stress magnitude, directly derived from our inversion algorithm. The average value in the source region of the mainshock is ~0.92. It's probably due to the existence of high-pressure fluid, the megathrust fault is relatively weak (~24MPa).
How to cite: Yun, N., Yue, H., Zhou, S., and Chen, L.: Absolute stress field inversion using focal mechanism rotation and co-seismic stress change: Application to 2011 M9 Tohoku, Japan, earthquake, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11324, https://doi.org/10.5194/egusphere-egu25-11324, 2025.
A robust prediction of the present-day stress state is of great importance for the safe usage of the subsurface, e.g., for borehole stability, mitigation of induced seismicity or the search and long-term safety of a high-level nuclear waste deposit. However, the state of knowledge concerning the stress state in Germany is limited as only unevenly distributed stress measurements are available. Two 3D geomechanical-numerical models created during the SpannEnD project (2018-2022) have improved this level of knowledge. Such geomechanical-numerical models - calibrated on available stress magnitudes - enable a continuum-mechanics based prediction of the present-day stress state. In the course of the follow-up project SpannEnD 2.0, a new, significantly improved model provides new insights into the stress state of Germany.
The new 3D geomechanical-numerical model combines information of 25 geological models and comprehensive additional data. The final geomechanical-numerical model comprises 52 geological units parametrized with individual mechanical properties (Young’s modulus and Poisson’s ratio) and densities. Linear elasticity is assumed and the finite element method (FEM) is used to solve the equilibrium of forces. Overall, the model contains ~10 million hexahedral elements providing a lateral resolution of 4 x 4 km2 and a vertical resolution of 45 m in the uppermost 5 km. A significantly enhanced stress magnitude database has been used for model calibration on magnitudes of the minimum (Shmin) and maximum horizontal stresses (SHmax). The model results show an overall good fit with these stress magnitudes indicated by a mean of the absolute stress differences of ~5 MPa for Shmin and SHmax. Furthermore, our results agree well with additional data sets not used for calibration, e.g., an absolute mean deviation of the orientation of SHmax with regard to World Stress Map data of ~10°.
How to cite: Ahlers, S., Henk, A., Reiter, K., Hergert, T., Röckel, L., Morawietz, S., Heidbach, O., Ziegler, M., and Müller, B.: SpannEnD 2.0 – Improved present-day stress prediction of Germany by a new 3D geomechanical-numerical model, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15638, https://doi.org/10.5194/egusphere-egu25-15638, 2025.
The mean stress state of the continental lithosphere is thought to align with the stresses generated by gravitational potential energy at mid-oceanic ridges. This equilibrium is manifested in an average continental crust of thickness ~40 km and elevations of a few hundred meters as suggested by Airy type compensation and geoid height data 1–3. Our recent data-driven thermomechanical model of the Alpine-Himalayan collision zone (AHCZ) suggested that this balance has a fundamental thermodynamical meaning in that such a state of continental lithosphere maintains a critical crustal thickness with optimal strength controlled by the radiogenic heat production 4. Such a state of critical crustal thickness is referred to as a “stable fixed-point attractor” 4, serving as a “reference tectonic state” 1 for the continental lithosphere. It facilitates comprehending intraplate continental deformation as a finite-amplitude perturbation, where thicker/thinner regions exhibit extension/compression at length scales devoid of flexural effects. We also demonstrated that the high amplitude orogen-type perturbations (e.g., Tibet, Alps) can evolve back to this reference tectonic state via damped oscillatory behaviour consistent with the Wilson Cycle timescale over a few hundred million years. In this study, we expand the data-driven thermomechanical models to a global scale to capture existing variability, particularly in the relatively less evolved orogen of Andes than the AHCZ. Observations of critical crustal thickness persist globally; however, the degree of weakening above the critical crustal thickness is less pronounced in the Andes than in Tibet.
References:
- Coblentz, D. D., Richardson, R. M. & Sandiford, M. On the gravitational potential of the Earth’s lithosphere. Tectonics 13, 929–945 (1994).
- Coblentz, D., van Wijk, J., Richardson, R. M. & Sandiford, M. The upper mantle geoid: Implications for continental structure and the intraplate stress field. in vol. i 197–214 (2015).
- Sandiford, M. Why are the continents just so…? J. Metamorph. Geol. 28, 569–577 (2010).
- Kumar, A., Cacace, M. & Scheck-Wenderoth, M. Thermodynamics of continental deformation. Sci. Rep. 13, 19920 (2023).
How to cite: Kumar, A., Cacace, M., and Scheck-Wenderoth, M.: Critical crustal thickness as a reference tectonic state: a global perspective, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3527, https://doi.org/10.5194/egusphere-egu25-3527, 2025.
Northern Bavaria in Southeast Germany is mainly covered by sedimentary rocks which form the Franconian Platform (mainly sandstones, mudstones and limestones). These continental to shallow marine sediments of the Permo-Mesozoic age overlay the Variscan basement and are regionally affected by Syn-Variscan compression (Freudenberger and Schwerd, 1996). Additional tectonic overprint including the Permo-Mesozoic basin extension, Cretaceous inversion and Cenozoic intraplate deformation resulted in a complex fault system (e.g., Wiest et al., in review). Structural tessellation – an amalgamation of similar fault systems to tectonically homogenous blocks, provides an effective tool for the development of the large-scale 3D model. Drone photogrammetry 3D reconstructions from regional limestone and sandstone quarries help to compensate the locally sparse drill core data and the lack of outcrops owing to large agricultural and forestry cultivation. The high-resolution drone photogrammetry models are used to transfer small to medium scale structural observations into the large-scale fault model. A combination of geological maps, all available drill core data, rare seismics and the regional drone photogrammetry provides enough data input to create a realistic tectono-stratigraphic model of Northern Bavaria. The finished 3D model of the Franconian Platform will be made publicly available through the Bavarian State Office for the Environment LfU (www.lfu.bayern.de).
References
Freudenberger, W., and Schwerd, K., 1996, Erläuterungen zur Geologischen Karte von Bayern 1:500 000, München, Bayerisches Geologisches Landesamt, 329 p.
Wiest, J.D., Köhler, S., Köhn, D., Stollhofen, H., Dengler, K., and Fazlikhani, H., A novel multi-scale approach to fault network analysis and visualization: test case Franconian Platform (SE Germany), in review.
How to cite: Lang, J., Koehn, D., and Prabhakaran, R.: The Franconian Platform in Northern Bavaria, Germany – A Drone supported, large Scale 3D Fault Model , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14135, https://doi.org/10.5194/egusphere-egu25-14135, 2025.
Three-dimensional (3D) geological modeling is a vital tool for visualizing subsurface geometries and understanding associated uncertainties, crucial for applications ranging from resource exploration to environmental management. Among the various modeling techniques, implicit methods have gained prominence due to their computational efficiency and ability to integrate diverse geological datasets. However, while methods such as dual kriging have successfully incorporated drift terms to enhance model accuracy, radial basis function (RBF) methods have traditionally not utilized this feature, limiting their adaptability in complex geological settings. This study addresses this gap by proposing an innovative approach that integrates geometrical external drift functions into the RBF framework. This enhancement allows the RBF models to converge to the geological expert’s conceptual geometries, significantly improving their ability to accurately model various geological structures such as planar strata, folded formations, and salt domes. The proposed methodology is demonstrated through two case studies on a synthetic fold model and real salt dome model, where its effectiveness is compared against traditional methods, showing notable improvements in both accuracy and computational efficiency. The findings suggest that incorporating external drift into RBF not only broadens the applicability of this method but also provides a more robust tool for subsurface modeling, particularly if the general subsurface geometrical configuration is understood.
How to cite: Yang, J., Carl, F., Achtziger-Zupančič, P., and Wellmann, F.: Three-Dimensional Modeling of Geological Bodies Using Radial Basis Function with External Drift Function, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11287, https://doi.org/10.5194/egusphere-egu25-11287, 2025.
3D modeling is a major asset for the understanding and quantitative characterization of subsurface geology. Geological survey organizations have produced 3D models for decades and are nowadays facing new kinds of demands for increasingly complex accurate representations of the subsurface. New challenges include "usual" difficulties such as integrating large sources of heterogeneous data, handling a wide range of possible model scales (from urban to national), but also new requirements on model uses. As an example, models should be easily (if not automatically) updatable and computable on various environments (not only in desktop software, but also in web / platform environments), models should also serve for multiple purposes and applications (which requires generating various kinds of representations of a single model), etc.
In this talk, we present the toolbox we have been developing at the French Geological Survey over the last years to progressively replace our two historical and homemade solutions for 3D geological modeling (GDM and GeoModeller software). The toolbox contains two parts. The first one is a set of python and C++ libraries that provide data structures and computational capabilities. These libraries can run on a wide range of software environments. The second one is a set of QGIS plugins. They provide access to 3D modelling capabilities to the geologists directly in the GIS environment without requiring them to have development skills.
Throughout the presentation, we illustrate how the different design choices we made helped us achieve our main objectives: extensibility of the toolbox capacities, reusability of the software components and performance. One of the key features to achieve these objectives is the design of modular and open software components. It enables models and workflows to be easily adapted to fit a wide range of production needs. The integration into the open-source ecosystem also provides numerous benefits, and we illustrate how we had to support the development of core QGIS functionalities to better manage 3D geological objects.
How to cite: Clausolles, N., Capar, L., Janvier, T., Lopez, S., and Quimerc'h, L.: An open-source toolbox for 3D geological modelling in QGIS, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18975, https://doi.org/10.5194/egusphere-egu25-18975, 2025.
Structural geological modelling methods currently depend on subjective stratigraphic interpretations, typically from geological maps and borehole logs. Implicit interpolation approaches can represent these interpreted geological units as scalar field values, to objectively derive a numerical representation of subsurface geometry, however sensitivity to the underlying geological interpretations (and biases or errors) remain.
In this contribution we present a neural-network based interpolation approach that removes the need for subjective value constraints. This network, or neural field, learns the relationship between input coordinates and scalar values, a flexible approach that has been recently demonstrated in the context of geological modelling. However, unlike previous approaches, we are able to constrain our model directly with objectively measured quantities (e.g., from geochemical assays, downhole petrophysical logs and/or hyperspectral core scan results). This is achieved by coupling the spatial neural field with a property neural field that learns to reconstruct measured quantities given a predicted scalar field value. Simultaneous training of these two neural fields encourages the spatial field to find a solution (subsurface geometry) that is most informative for predicting the measured properties. Constraints on the gradient (i.e. bedding orientation) and scalar value (i.e. stratigraphic unit) can also be included to further guide the neural fields, but are not required.
We demonstrate this weakly-supervised modelling approach on several synthetic datasets, and show how it could be applied to construct “self-updating” models that are iteratively updated as new geophysical, geochemical or hyperspectral data become available. These preliminary results indicate that unlabelled geological data can be used as powerful objective constraints for future geological modelling workflows, to ultimately derive accurate and unbiased representations of the subsurface.
How to cite: Thiele, S., Kamath, A., and Gloaguen, R.: Less is more: Weakly supervised interpolation using geological neural fields, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8735, https://doi.org/10.5194/egusphere-egu25-8735, 2025.
Quantification and comparison of 3D bodies is a scientific aim in many fields, such as medical diagnostics, computer graphics and biochemistry. We propose a methodology for the shape quantification in the context of natural subsurface structures: Dimensions, gradients and curvatures are determined on cross-sections along and across the horizontal main axis of salt structures. The acquired statistics of the dimensions are characteristic for the respective type of geological body, providing insight into the anisotropy of structures, the potential existence of overhangs and the geological processes that shaped the top of an evaluated structure. The statistics of the gradient and curvature carry information on the appearance of the outline and sphericity of the assessed structures. A total of 240 intrusive salt structures from the North German Basin have been analyzed. The statistical properties allow to cluster them into body types which correspond to regular geometrical end members that are linked to distinct formation processes.
How to cite: Carl, F., Achtziger-Zupančič, P., Yang, J., Colling Cassel, M., Kukla, P. A., and Wellmann, F.: What’s the top hat there? - A method for the quantification and comparison of subsurface bodies, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9032, https://doi.org/10.5194/egusphere-egu25-9032, 2025.
In the ongoing development of PZero within the Geosciences IR project led by the Italian Geological Survey (gecos-lab/PZero), the second phase of our research has been dedicated to enhancing seismic interpretation techniques and expanding data loading and slicing capabilities. Building on our earlier milestone of seamlessly integrating 2D and 3D seismic data, we introduced improved data handling alongside two advanced workflows for seismic horizon picking and structural interpretation.
First, we expanded the seismic data-loading functionality to support the straightforward import of SEG-Y files and other common formats. Users can now define arbitrary slicing orientations in the inline, crossline, and vertical (z) directions, managed by a newly implemented Grid Section Manager that specifies slice counts and orientations. This provides greater flexibility for tailored interpretation workflows and more robust seismic data analysis.
Second, we present a semi-automated A* edge tracking approach using Sobel filtering. By applying a Sobel filter to seismic slices, we enhanced the edges indicative of the horizon boundaries. The A* pathfinding algorithm tracks the horizon automatically once two points are selected on the filtered edges, considerably reducing manual picking while maintaining geological consistency across inlines, crosslines, or z-slices.
Third, an automatic interpretation method leveraged the Meta-Segment Anything Model (SAM2). A minimal user-provided guideline (such as a single polyline) on one slice is used by the SAM2 predictor to generate a horizon boundary mask, which is then propagated across neighboring slices in all directions. Once vectorized, these segmentation masks feed directly into PZero’s implicit or explicit 3D modeling framework, facilitating rapid updates and reproducibility across extensive seismic volumes.
Although the fully automatic SAM2 workflow significantly accelerates horizon picking, the semi-automated Sobel–A* approach remains indispensable in complex seismic settings, where automated segmentation can struggle to capture subtle geological details or correctly interpret noisy data. By allowing user interaction to guide the algorithm, the semi-automatic method ensures a higher fidelity and consistency of interpretive results in challenging areas.
Taken together, these integrated methods substantially enhance PZero’s capabilities for clastic sedimentary alluvial plain modeling. They enable more flexible data handling, efficient horizon picking, and reproducible workflows spanning both straightforward and intricate seismic environments.
How to cite: Hussain, W., Bistacchi, A., Benedetti, G., and Monti, R.: AI-Driven PZero Modeling: Enhanced Seismic Data Loading, Grid Section Management, and Automated Interpretation with Sobel, A* Pathfinding, and SAM2, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16273, https://doi.org/10.5194/egusphere-egu25-16273, 2025.
The need to describe the present-day crustal stress state has been recognized from a wide range of geo-disciplines. Furthermore, meeting the climate goals will require an energy transition and the associated phase-out of fossil, leading to increased and modified subsurface utilisation concepts as well as new demands on the integrity and long-term stability of subsurface operations. Thus, crustal stress data and geomechanical models for continuous predictions of the stress field in larger rock volumes will become more important. Stress data were already collected in the 1930s using surface relief methods, followed by flat jack and borehole relief methods in the 1950s, and hydraulic fracturing in the 1970s. Another important source of stress information was established in the 1980s using interpretations of borehole breakouts as stress indicator and later also drilling induced tensile fractures. Furthermore, due to the expansion of global seismological networks in the past decades, the number of earthquake focal mechanisms , primarily used as stress indicators for the deeper part of the Earth crust, has increased significantly. These developments resulted in the initiation of the World Stress Map (WSM) project (http://world-stress-map.org) in 1986.
The backbone of the WSM is a quality ranking scheme allowing the comparison of various stress indicators which sample the rock stress on a wide range of spatial scales. The latest WSM database was released in 2016. For the new WSM release 2025 we developed the new database infrastructure MaRS (Management and Repository of Stress) based on PostgreSQL. It has a web-based interface to insert new data and assess these data automatically with internal Python routines, streamlining data submission significantly. The new WSM release entails the following key changes:
- The WSM release 2025 has more than doubled the number of data records.
- Addition of high-quality data records from more than 3,000 boreholes including a study that uses a uniquely high-resolution dataset in Eastern Australia (see poster EGU25-5042 of Rajabi et al.).
- Integration of the global focal mechanism catalogue of the International Seismological Centre (ISC).
- Replacement of the 40 km depth limit using instead the global crustal model of Szwillus et al. (2019, JGR) to assign if data records from earthquake focal mechanisms are located in the crust or not.
- Updated WSM quality assessment scheme to make criteria programmable.
- Introduction of the new quality class X with three sub-classes for data records with missing information (Xmi), stress indicator that are rarely used (Xru), and stress indicator that are not established (Xne).
Quo Vadis WSM? The new database infrastructure MaRS allows us for a frequent release schedule of the WSM database to promptly provide the community with new data. MaRS was also developed to include to expand the WSM database in the next years with quality-ranked stress magnitude and pore pressure data. Adding this information is essential for model calibration widening the scope of WSM applications.
How to cite: Heidbach, O., Rajabi, M., Lammers, S., Morawietz, S., von Specht, S., Ziegler, M., Reiter, K., Di Giacomo, D., Harris, J., and Storchak, D.: The new World Stress Map database release 2025, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8603, https://doi.org/10.5194/egusphere-egu25-8603, 2025.
The majority of in-situ stress information in the World Stress Map database comes from earthquake focal mechanisms, and petroleum regions where oil and gas industry technologies enable the collection of contemporary crustal stress information using borehole logs and tests. As a result, there is a limited stress data in many other areas, particularly in regions with low seismicity due to their tectonic settings or limited hydrocarbon exploration and production. In recent years, borehole image logs have become a standard tool in the mining industry as well, used for geotechnical and structural analysis. These logs provide a pseudo-image of borehole walls, allowing the characterization of stress-related deformations, such as borehole breakouts and drilling-induced tensile fractures, to better understand the present-day stress state.
We investigated the orientation of present-day horizontal stresses (SHmax and Shmin) in various mine sites in Australia and Mozambique, inferred from the analysis of acoustic televiewer logs (ATVs) from over 1500 boreholes. This resulted in great understanding of in-situ stress orientation in regions with limited prior stress data. Unlike petroleum boreholes, where image log data is available for specific intervals (e.g., reservoirs), most open-pit mine boreholes are logged from near the surface, providing stress information from shallow depths and sometimes extending to 1.5 km. In addition, boreholes in mine industry have close spacing (sometimes less than 30 m apart) that provide a great opportunity to investigate the local variability of the stress state. It is e.g. possible to track rotations of the orientation of maximum horizontal stress SHmax near geological structures.
The SHmax orientations analysed at at the mine-site and basin scales in this study align closely with regional stress patterns, highlighting the role of large-scale tectonic forces as the primary drivers of crustal stress patterns. However, the high-resolution data used in this study — such as closely spaced boreholes (sometimes less than 30 meters) and SHmax orientation data spanning from near the surface to depths of 1.5 km — reveal small-scale SHmax rotations (ranging from 10° to 90° on a spatial distance of 1 to 100 meters) induced by stiffness contrasts, rock fabric, and geological structures. These small-scale SHmax rotations have significant implications for geotechnical and geomechanical applications across various disciplines.
How to cite: Rajabi, M., Heidbach, O., Ziegler, M., and Esterle, J.: High-resolution stress mapping using mine borehole data, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-5042, https://doi.org/10.5194/egusphere-egu25-5042, 2025.
A detailed understanding of the present-day stress state is key to understanding naturally occurring seismicity and safely and successfully conducting subsurface operations. For example, in the case of geothermal energy applications, the role of faults and fractures in both productivity and induced seismicity critically depends on the present-day stress state. In addition, knowledge of the subsurface stress state is also of significant importance to mitigate drilling risks. Here, in particular, the least principal stress controls the maximum allowable wellbore pressure before the drilled formation is unintentionally fractured. In its simplest form, the state of stress can be described by the magnitude of vertical stress and two horizontal stresses and their azimuthal orientations. Ideally, the state of stress includes the counteracting effect of pore fluid pressure (short: pore pressure) and is described as the effective stress tensor (effective stress is the difference between stress and pore pressure).
In this study, we investigate the magnitude of the least principal stress (minimum horizontal stress) in the North Alpine Foreland Basin in SE Germany using stress measurements such as Formation Integrity (FIT) and Leak-Off Tests (LOTs).
Whilst pore pressure magnitudes have been extensively studied and published in numerous publications in the North Alpine Foreland Basin in SE Germany, knowledge of the prevailing least principal stress is still quite limited, particularly in overpressured formations. So far, only subsets of the available FIT/LOT data, mainly concentrated around Munich, have been investigated. Recently, additional FIT/LOT data became available covering greater depths (up to 4,2 km) and overpressured formations. We investigate this new dataset in combination with data from previous studies to establish a minimum horizontal stress gradient model, which considers both pore pressure and rock type. To do so, we consider the ratio between the measured minimum horizontal and vertical effective stress using a previously established pore pressure magnitude model. The resulting effective stress ratio model is tested against the least principal stress measurements of deep geothermal wells in the study area's hydrostatically and overpressured regions, showing that considering both lithological and pore pressure variations is necessary to predict the least principal stress magnitudes. The established model can be used to improve the efficiency and safety of future drilling campaigns in the study area and can also serve as an input for mechanical subsurface modelling, e.g. for a better understanding of deformation or natural and induced seismicity.
How to cite: Breitsameter, J., Drews, M., Ziegler, M., Obermeier, P., and Shatyrbayeva, I.: Constraining lithologically differentiated minimum horizontal stress gradients in hydrostatically pressured and overpressured parts of the North Alpine Foreland Basin in SE Germany , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10263, https://doi.org/10.5194/egusphere-egu25-10263, 2025.
Understanding the elastic behavior of the oceanic lithosphere is crucial for interpreting plate dynamics and rheology. While various methods exist to estimate the lithosphere's ability to deform under load, the factors controlling this deformation across different tectonic settings remain poorly quantified.
We present a three-stage analysis to systematically evaluate controls on lithospheric flexure across the Western Pacific. First, we calculate a suite of metrics that characterize the elastic deformation properties of the lithosphere using gravity and bathymetry data. Second, we develop a random forest regression framework, a type of machine learning model, to reconstruct these observed deformation properties using a range of geophysical parameters, including gravity, bathymetry, sediment thickness, oceanic crustal age, heat flow, and hotspot proximity. By analyzing the feature importance within this model, we quantify the relative influence of each parameter on lithospheric deformation. Finally, we apply this framework to different tectonic settings (mid-ocean ridges, oceanic plateaus, abyssal plains, and seamount chains) to examine how the controlling factors vary by geological context.
This quantitative assessment, leveraging machine learning, advances our understanding of oceanic plate rheology and provides a framework for interpreting lithospheric behavior across different tectonic environments. The results have important implications for understanding plate dynamics and the evolution of the Pacific lithosphere.
How to cite: Sheng, H., Zhao, Y., Riel, B., and Lu, Z.: Thermal and Mechanical Controls on Pacific Plate Flexure under Seamount Loading, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3498, https://doi.org/10.5194/egusphere-egu25-3498, 2025.
Characterizing the crustal stress field is essential for understanding global processes such as earthquakes and plate tectonics, while also being critical for local applications, such as interim subsurface storage, and deep geological repositories for nuclear waste. A key challenge lies in understanding the interactions between the crustal stress field and pre-existing geological structures, especially with faults. Previous studies have aimed to understand the impact of faults on the stress field by making interpretations based on variation of stress magnitudes or rotation of the maximum horizontal stress (SHmax) orientation in larger regions. This approach cannot attribute the local perturbations in the stresses exclusively to the faults. Another common approach is the use of generic geomechanical-numerical models. Although instructive, generic models usually have limitations from a lack of site-specific calibration with in situ stress data.
The SHmax orientation is the only component of the reduced stress tensor that is systematically documented and accessible through databases such as the World Stress Map. The SHmax orientation reflects consistency on large scales, primarily driven by first-order tectonic forces and second-order buoyancy forces. However, significant SHmax rotations over shorter distances are often linked to third-order sources such as faults, and are challenging to model accurately due to computational complexity and the risk of numerical artifacts.
The hypothesis in this study is that the impact of local faults with a few tens of meters displacement on the in-situ stress state might be overstated. Here, we use 3-D geomechanical-numerical models that are calibrated against a unique and robust dataset of 50 stress magnitude data records. This dataset was acquired for evaluating the suitability of Zürich Nordost which is one of the three potential Swiss siting regions to build a deep geological repository for high-level nuclear waste. We vary the numerical resolutions and investigate the spatial scale at which faults influence the individual components of the far-field stress tensor and in particular the SHmax orientation. Finally, we compare models with and without faults.
Our results reveal that faults of this scale do not have a significant influence on the stress tensor orientation or principal stress magnitudes beyond a few 100s meters distance from the fault. Comparisons between the models reveal that the stress differences are not necessarily controlled by the mechanics of faults. The impact is rather due to lateral stiffness variations and density contrasts due to the offset between units that occurs at faults. Small lateral variations could be attributed to the mechanical behaviour of faults but these variations are generally less than the stress variations due to uncertainties in the rock property variability.
Our findings suggest that faults could be safely excluded from the modeling workflow for models focusing on large-scale stress predictions and not on stress changes close to the faults, such as those that characterize the geomechanics of potential deep geological repository regions. Removing faults from the modeling workflow reduces computational complexity and accelerates modeling process, without causing any significant differences in the model results at a distance of few 100s meters from the faults.
How to cite: Velagala, L. S. A. R., Heidbach, O., Ziegler, M., Reiter, K., and Henk, A.: The Role of Faults in Shaping Present-Day Stress Fields: Implications for 3D Subsurface Models., EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10251, https://doi.org/10.5194/egusphere-egu25-10251, 2025.
Due to the limited access to the underground, numerical models are essential in nearly all branches of geosciences to improve the general understanding or to estimate behaviour or properties in applied cases. Complex subsurface structures can be best represented by applying the finite element method (FEM) as it allows unstructured meshes during the discretization of the model geometry. The resulting model quality depends on the resolution of the mesh, the element type (shape), the element order (1st or 2nd), or special elements e.g. with reduced integrations points. However, always a balance between the effort of mesh generation, computing time, amount of model runs needed, and the justifiable expense needs to be found. As such factors can’t be tested for each project individually, we will test this with simplified and already existing, purely elastic geomechanical models. The derived conclusion can in turn be utilized to improve the numerical implementation of future studies.
To investigate the impact of a chosen mesh, 2-D models (mechanical in 3-D) are generated based on a cross section. Geologically, the models represent the crystalline basement, several slightly dipping thin Mesozoic sedimentary units, covered by Cenozoic deposits. The goal is, to represent the thin about 10 to 100 m thick Mesozoic units sufficiently well to reliably predict the present-day stress state. Varied within the target units are the mesh resolution, the element type (tetrahedra vs. hexahedra), the element order (1st and 2nd) and elements with reduced integration points provided by the used solver. All models are calibrated using in situ stress data from a borehole that is located at the model cross section which results in a best-fit model that minimizes the deviation between modelled and the in-situ stress calibration data by varying the displacement boundary condition of the model. Model results are always compared along the well trajectory using a reference model with a fine mesh resolution. The computational effort will be considered, too. Study results indicate that, flat (brick-like) hexahedrons provide better results than tetrahedrons, taking mesh resolution and computing effort into account. Above a certain level, the number of hexahedrons (fine vs. coarse resolution) in the vertical direction per layer exerts a discernible influence on the results in the proximity to material transitions only. Second order elements provide nearly the same results as first order elements, which means that the extra computational effort is not worth it. Differences due to the usage of special solver-provided elements are neglectable.
Additionally, we tested three site models based on a different model geometry, mechanical stratigraphy, and mesh resolution, by applying the same material properties whenever possible. Most of the observed differences are acceptable and mainly driven by the differences in geometry and the resolution of the mechanical stratigraphy. Deviation of model results is much bigger, when the original material properties (state of knowledge at the time) are applied.
How to cite: Reiter, K., Heidbach, O., Henke, A., Degen, D., and Achour, R.: Impact of resolution and finite element type in geomechanical-numerical modelling, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9583, https://doi.org/10.5194/egusphere-egu25-9583, 2025.
Geomechanical-numerical modelling aims to provide a comprehensive characterization of the stress tensor within rock volumes by leveraging localized stress magnitude data for model calibration. This calibration involves optimizing boundary conditions to achieve the closest alignment with in-situ stress measurements in boreholes that provide magnitudes of the minimum and maximum horizontal stress. However, the high cost of acquiring stress magnitude data frequently results in sparse and incomplete datasets which potentially prevents a meaningful calibration.
In this study, we use a comprehensive stress magnitude dataset of 50 stress magnitude data records acquired for the geomechanical characterization of the candidate siting region Zürich Nordost for a deep geological repository located in northern Switzerland. We demonstrate how the size of the calibration dataset influences the accuracy and uncertainty of stress magnitude predictions in geomechanical modelling of sedimentary formations. We introduce a novel statistical approach that incrementally increases the size of calibration data subsets. This approach evaluates how the amount of available data influences stress predictions across formations with varying rock stiffness. It achieves this by rapidly assessing the stress states associated with a large number of different combinations of stress magnitude data. The comparison of the resulting stress fields with increasing number of calibration point data allows to estimate the minimum number of calibration points that are required to achieve a stress prediction range that is as small as the range expected due to inherent uncertainties in the data. The results show that less than 20 data points are sufficient to achieve the same model precision and accuracy.
Furthermore, a detailed analysis of the dataset revealed a data outlier linked to a local stiffness anomaly. This outlier significantly impacts the stress predictions when calibration data are limited. However, as the calibration dataset size increased, the influence of the outlier diminishes. We also show that our statistical approach allows for the objective identification of clear outliers with respect to the model in the calibration dataset, which has an impact on the minimum number of data needed for the model calibration.
These findings underscore the significance of dataset size and composition in reducing uncertainties, thereby providing a framework for optimizing calibration strategies. This study offers valuable insights for subsurface projects, such as energy storage, CO2 sequestration, deep geological repositories, or geothermal energy, where precise stress predictions are critical.
How to cite: Laruelle, L., Ziegler, M., Reiter, K., Heidbach, O., Desroches, J., Giger, S., and Cotton, F.: Minimum amount of stress magnitude data for reliable geomechanical modelling, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6868, https://doi.org/10.5194/egusphere-egu25-6868, 2025.
Robust predictions of in-situ stress states are essential for the safety assessment and long-term stability of nuclear waste disposal sites. However, these predictions are inherently uncertain due to the variability in geological parameters and material properties as well as uncertainties of model calibration data. Thus, a large number of model simulations would be required for a complete investigation of the model uncertainties which is not feasible due to required high numerical resolution with several million discretization points. An alternative to classical full order solutions is to develop surrogate models that run much faster but perform with similar precision.
We propose to use a machine learning-aided methodology to set up and solve these surrogate models. Specifically, we use the non-intrusive reduced basis (NI-RB) method. The resulting surrogate models are 5-6 orders of magnitude faster compared to the initial full-order model which allows an extremely fast computation of many models with different parameters. The initially required full order geomechanical simulations are conducted using GOLEM, based on the MOOSE framework (a multiphysics simulation platform).
For our case study, we use benchmark models and a simplified model inspired by the potential siting area Nördlich Lägern for high-level nuclear waste in northern Switzerland. Preliminary results indicate that our surrogate model accurately replicates the findings of the full order solutions while significantly reducing computational costs. We primarily focus on global sensitivity analyses to identify the most critical parameters impacting the stress field. Our study explores seven scenarios for surrogate modeling, each focused on different model parameters. The first five scenario examine boundary conditions, rock properties (density, Poisson ratio, Young’s modulus), geometrical features and combinations of the three, using a benchmark model to demonstrate general implication for geomechanical studies. For these scenarios, we change between two to thirteen parameters. The sixth scenario uses the simplified study based on the Nördlich Lägern, adjusting 15 parameters (Young’s modulus of each lithological layer) illustrating the potential for future real-case applications.
We show an additional seventh scenario that integrates comprehensive fault considerations, including parameters such as geometry, geographical location, dip angle, and strike direction. These factors are vital in the context of subsurface engineering studies, as they significantly influence the stress fields and the overall stability of the geological formation. A thorough understanding of fault characteristics is paramount for assessing potential risks and ensuring long-term safety and structural integrity.
The results demonstrate that the surrogate models are much faster but keep a similar precision as the full order solution. This shows the potential of surrogate modeling for rapid uncertainty quantification in geomechanics, offering a useful tool for assessing nuclear waste disposal sites, but also different applications like, for example, geothermal exploration.
How to cite: Achour, R., Degen, D., Ziegler, M., Heidbach, O., Henk, A., Reiter, K., Cacace, M., and Wellmann, F.: Global Sensitivity Analysis to Improve Geomechanical Stress Characterizations Using Physics-Based Machine Learning Models, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10078, https://doi.org/10.5194/egusphere-egu25-10078, 2025.
Posters virtual: Tue, 29 Apr, 14:00–15:45 | vPoster spot 2
EGU25-9651 | ECS | Posters virtual | VPS28
The Fracture Induced Electromagnetic Radiation (FEMR) technique as a tool for stress mapping: A case study of the Ramon CraterTue, 29 Apr, 14:00–15:45 (CEST) | vP2.21
This study explores the application of Fracture-Induced Electromagnetic Radiation (FEMR) for stress analysis in the Ramon Crater, a tectonically “stable” region in southern Israel. FEMR, an innovative geophysical method, detects electromagnetic pulses emitted during micro-fracturing events to infer stress orientations. Unlike traditional seismic techniques, FEMR is sensitive to subtle stress changes, making it suitable for regions with limited seismicity. Field measurements were conducted at nine locations using the ANGEL-M device, capturing high-sensitivity electromagnetic signals to determine the stress azimuth. The results revealed a dominant mean stress azimuth of 308°, aligning closely with the acute bisector of two principal joint sets in the region, WNW-ESE and NNW-SSE. These orientations correspond to historical compressional stress from the Syrian Arc Stress (SAS) regime and more recent extensional stress from the Dead Sea Stress (DSS) field. The superimposition of these regimes has created a complex tectonic environment, evidenced by features such as joint sets, fault planes, and basaltic dikes. FEMR measurements correlate with these geological indicators, confirming the technique’s ability to detect regional stress directions and their evolution over time. In the past decade, the method of FEMR has progressively gained impetus as a viable, non-invasive, cost-effective, real-time geophysical tool for stress analysis in various parts of the world. Its range lies in delineating tectonically active zones, landslide-prone weak slip planes, highlighting stress accumulation in mines and tunnels, etc. This study highlights FEMR’s viability for stress field analysis, especially in stable tectonic zones. Its ability to capture micro-crack activity and subtle stress shifts offers a detailed understanding of how tectonic forces shape regional geodynamics. While FEMR enhances stress detection capabilities, careful calibration with geological models is essential to differentiate transient stress changes from long-term tectonic trends. This research advances FEMR’s application in geophysical studies, particularly for monitoring stress fields in regions influenced by ancient and ongoing tectonic forces.
How to cite: Das, S. and Frid, V.: The Fracture Induced Electromagnetic Radiation (FEMR) technique as a tool for stress mapping: A case study of the Ramon Crater, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9651, https://doi.org/10.5194/egusphere-egu25-9651, 2025.
