Photo credit: Julie Noury-Soyer

Introducing the 2023 RING Meeting
By Guillaume Caumon, Pauline Collon and Paul Cupillard

Dear Colleagues, dear Friends,

Welcome to this new volume of the RING Meeting Proceedings! In this collection, you will find papers on new methodologies, applications, and results in integrative numerical geology. We hope you will find them useful to better map, understand, forecast, and manage uncertainties relative to the subsurface of the Earth.

The numerical representation of geological structures is often a very important step in geomodeling. In this area, ARIENTI et al. (2023) present a 3D structural model of the tectono-stratigraphic units of the Aosta Valley in the Alps, which highlights the value and the limitation of existing modeling techniques to effectively handle highly deformed formations in sparse data settings. To obtain acceptable pre-deformation geometries of such structures, OSORNO BOLÍVAR, CAUMON & CUPILLARD (2023) propose a numerical investigation for choosing the appropriate mechanical parameters for faults seen as thin viscous shear zones with the particle-in-cell FAIStokes code. As an alternative to this solver, LÉVY (2023) explains how partial semi-discrete optimal transport can be used to solve fluid dynamics problems with free surfaces, combining both theoretical and essential implementation aspects.

A very specific aspect in 3D modeling of geological structures concerns the management of surface intersections, for instance to create maps or sections from a 3D model, to account for faults and unconformities during model building and updating, and to create sealed geological models. This year, LEVY, LI & BORGESE (2023) present some very important advances to compute such intersections in an elegant and provably correct way using exact geometric representations. ANQUEZ, BOTELLA & SCHUH-SENLIS (2023) take a slightly different path to address this problem by temporarily allowing for incorrect mesh configurations, and then proposing iterative local mesh repair operations. Another important aspect in mesh design concerns the use of spatial adaptivity. On this, LEGENTIL et al. (2023) increase the level of detail in the neighborhood of wells for PEBI and hex-dominant unstructured meshes, and compare the simulation outcomes for different grid designs. RAGUENEL et al. (2023) generate hybrid meshes of a full-field model for CO2 storage simulation and obtain consistent results with both an open-source and a commercial simulator, clearing the path for coupled hydromechanical modeling on the same computational support. At a smaller scale, and precisely to help fill gap between core-scale observations and full-field models, TERTOIS (2023) proposes a surface-based bedset tetrahedral modeling approach using a smart extrusion technique and local mesh improvement methods.

Geophysical data are an essential source of information to understand and characterize the spatial structure of the subsurface. Most geomodeling workflows proceed linearly by interpretation, modeling and forecasting steps. In this context, reducing the gap between the way geophysicists and geologists approach the subsurface is essential to improve predictivity and better capture and reduce uncertainties. This calls for new methods to compute the geophysical response on complex geological media, and to confront candidate geological interpretations to geophysical data. Towards the first goal, KOREN, RAVVE & TERTOIS (2023) introduce a new numerical method for accurate ray tracing in geological media affected by large velocity contrasts. For full waveform modeling, RAPENNE et al. (2023) and SCUOTTO et al. (2023) present latest advances and validations on quadrangular meshing for spectral element simulation in homogenized two-dimensional heterogeneous models. In complement, RAPENNE, CUPILLARD & GOUACHE (2023) consider the homogenization of isotropic layered media up to order 1 with a focus on effective boundary conditions at the free surface. Towards the second goal, RUGGIERO, CUPILLARD & CAUMON (2023) present a numerical method to obtain uncertainties of elastic parameters in full waveform inversion using a low-rank Hessian approximation. Using potential field data, GIRAUD et al. (2023) propose a methodology to address geological interpretation questions concerning the existence of geological units and the inference of their geometry, and apply this approach to generate crustal-scale structural scenarios compatible with gravity anomalies in the occidental Pyrenean range.

The existence of geological objects is integrated rigorously in the inverse problem by TATY-MOUKATI et al. (2023), who address the stochastic fault interpretation of seismic sections using an advanced object-based stochastic simulation method (the Candy Model). In a similar spirit, MARCHAL et al. (2023) propose an elaborate spatial model to jointly generate the geological objects driving hydrothermal alterations and the geometry of the altered rocks. In a fracture characterization context, BONNEAU, CAUMON & STOICA (2023) go one step further to not only propose a dedicated spatial model allowing for specific fracture interaction terms, but also for inferring the corresponding parameters from analog data. Parameter inference for marked point processes is also addressed for astronomic applications by GILLOT, STOICA & GEMMERLE (2023) and for source identification in geochemical mixtures by REYPE et al. (2023). In the above approaches, the assessment of the model quality is essential for inference and validation purposes. This aspect is also considered for implicit interpolations of simple building blocks shapes (YANG et al., 2023) and of more regional models (CAUMON, 2023).

In sedimentary basins, the understanding and modeling of stratigraphic formations is very important, among others, for subsurface reservoir management tasks. For this, prior understanding of the depositional processes and forcing parameters is paramount. It is addressed this year by BAVILLE & ANDRIĆ-TOMAŠEVIĆ (2023), who study the impact of uplift and erodibility on the signature of coastal deposits. Another challenge in stratigraphic modeling is to find consistent ways to subdivide geological time depending on heterogeneity and model purpose. HERRERO et al. (2023) consider this problem by combining well logs and flow data, by making the number of units one of the unknowns of the inverse problem. In complement to changing the number of layers, numerical upscaling approaches can be essential when considering full-field models. For this, ZAKARI & CAUMON (2023) extend their flow-based permeability upscaling workflow for non-conforming grids to three dimensions.

Another way to summarize reservoir complexity for flow purposes is to use the concept of connectivity. For example, CHOLLET & VIARD (2023) show how medial axes, segmentation and various types of visual cues can be used to visualize connectivity and compartments. To relate various stochastic models to their flow behaviors in a quantitative way, SCARPA et al. (2023) perform an extensive analysis of flow simulation results in channelized deposits, which demonstrate not only the impact of carefully choosing the modeling strategy, but also the difficulty to relate static and dynamic reservoir summary measures. Karsts are another type of subsurface reservoirs in which connectivity and geometry are essential but challenging to control in numerical simulations. GOUY et al. (2023) propose a methodology and application integrating several types of data including connectivity, host rock type, fractures, and vadose zone information.

Beyond subsurface flow studies, geomodels have a significant potential to support geomechanical and coupled physical problems, as shown by MAGHAMI, GUGLIELMI & ULRICH (2023) on the analysis of the stress state of the fault zones in the Mont Terri Underground Laboratory. Geomodels are also essential for the selection and the study of nuclear waste storage (CARL et al., 2023).

In addition to the above domain-specific algorithms and approaches, machine learning and more generally artificial intelligence (AI) methods are showing great promises to help address subsurface challenges. This year, FRATANI et al. (2023) introduce a supervised technique to learn about multiple-point fault data association rules from 3D analog structural models. For facies modeling, GOLDITÉ et al. (2023) propose a simulation method based on generative adversarial networks, which uses a self-attention mechanism to integrate spatial correlation. At a much smaller scale, BERTAUD et al. (2023) describe a workflow based on convolutional neural networks and transfer learning to successfully detect microfossils on thin sections despite a relatively small annotated data base. As we all know thanks to the success of ChatGPT and other large language models, AI is also making significant progresses to represent knowledge in numerical form from natural language. BOUZIAT et al. (2023) provide an overview of these approaches and share their experience to effectively use these technologies for mining information in geoscientific reports and papers.

Finally, we are very pleased to include presentations on new software technologies. Indeed, numerical modeling is a rapidly evolving field, in which technologies can quickly change the landscape and the way to work. Part of the game nowadays to address a modeling challenge relies on choosing appropriate frameworks and combining existing components to work more effectively. Among these, CAVELIUS, MEYER & COUPEY (2023) present an open-source python package to facilitate the transition from developers to users, including automatic generation of user interface. To increase interoperability in geosciences, MCGAUGHEY et al. (2023) introduce an open format and python API based on HDF5 to represent, visualize and exchange a large variety of geological data and geomodeling objects. Web-based geomodel visualization is also addressed by the opensource microservice proposed by BOTELLA & CHAMPAGNOL (2023). This work illustrates the current trend towards browser-based applications in software development. To facilitate the development and execution of such applications both locally and on the cloud, MAERTEN, MAERTEN & REINISCH (2023) propose an open-source platform named YouWol.

To conclude this overview of this year’s meeting, we want to thank all the authors for the hard work they put into this volume’s contribution, and all the RING Consortium members for their lasting and essential support. We wish you a worthwhile and inspiring RING meeting!

References

Anquez P, Botella A & Schuh-Senlis M. (2023). One does not simply modify meshes: a robust framework for modeling and meshing. 2023 RING Meeting.

Arienti G, Bistacchi A, Caumon G, Dal Piaz G & Monopoli B. (2023). 3D geomodelling in polydeformed metamorphic mountain belts: an implicit geomodel-driven workflow applied to large scale modelling of the Northern Aosta Valley (North-Western Alps, Italy). 2023 RING Meeting.

Baville P & Andrić-Tomašević N. (2023). Can source area erodibility delay slab break-off signal in a depositional record? Insight from stratigraphic forward modeling. 2023 RING Meeting.

Bertaud L, Bouziat A, Ammar AB, Hamon Y & Ferraille M. (2023). Operational usability of Deep Learning methods for automated microfossil detection on thin sections: ongoing assessment and methodological insights. 2023 RING Meeting.

Bonneau F, Caumon G & Stoica RS. (2023). Fracture Network Characterization Using Stochastic Simulations of Marked Point Process And Bayesian Inference. 2023 RING Meeting.

Botella A & Champagnol J. (2023). Open-source geosciences visualization microservice: combining OpenGeode and VTK. 2023 RING Meeting.

Bouziat A, Nguyen M-T, Divies R, Khvoenkova N, Siccardi O, Dehghan K & Rumbach G. (2023). Knowledge mining in massive collections of geoscience documents with digital technologies: our journey in the TELLUS community. 2023 RING Meeting.

Carl F, de los Angeles de Lucio G, Yang J, Achtziger P, Kukla PA, Bense F & Wellmann F. (2023). Host rock analysis for the German nuclear waste disposal site-selection: review of subsurface geometries and input data for geological modelling. 2023 RING Meeting.

Caumon G. (2023). On some comparison metrics between 3D implicit structural models. 2023 RING Meeting.

Cavelius C, Meyer H & Coupey R. (2023). OneCode: an open-source Python library to collaborate on a unified code base. 2023 RING Meeting.

Chollet J & Viard T. (2023). Reservoir connectivity analysis and visualization. 2023 RING Meeting.

Fratani A, Caumon G, Stoica RS & Giraud J. (2023). Hypergraph-based fault observation association – theoretical framework and first results. 2023 RING Meeting.

Gillot N, Stoica RS & Gemmerle D. (2023). Study the galaxy distribution characterisation via Bayesian statistical learning of spatial marked point processes. 2023 RING Meeting.

Giraud J, Caumon G, Grose L, Ogarko V, Martin R, Cupillard P, Ford M, Arienti G & Aillères L. (2023). Level set inversion with geophysics-guided birth of rock units and equivalent model search: example in the Pyrenees. 2023 RING Meeting.

Goldité V, Bouziat A, Lecomte J-F & Faney T. (2023). Facies modelling in fluvial environments with Self-Attention Generative Adversarial Networks conditioned to well data. 2023 RING Meeting.

Gouy A, Collon P, Bailly-Comte V, Antoine C & Landrein P. (2023). How to control the morphology of simulated karst networks? 2023 RING Meeting.

Herrero J, Caumon G, Bodin T, Zakari M & Giraud J. (2023). Transdimensional inversion of flow data with a cascaded reversible jump algorithm on a layer-cake model. 2023 RING Meeting.

Koren Z, Ravve I & Tertois A-L. (2023). Kinematic and dynamic Eigenray method for 3D heterogeneous anisotropic media. 2023 RING Meeting.

Legentil C, Raguenel M, Lopez D, Li W-C, Borgese C & Darche G. (2023). Adaptation of cell resolution near wells in unstructured meshes for flow simulations. 2023 RING Meeting.

Lévy B. (2023). Partial optimal transport for a constant-volume Lagrangian mesh with free boundaries. 2023 RING Meeting.

Levy B, Li W-C & Borgese C. (2023). On mesh intersection robustness and efficiency. 2023 RING Meeting.

Maerten F, Maerten L & Reinisch G. (2023). Towards a web-based collaborative and open-source platform for fast prototyping numerical tools and applications for geoscientists. 2023 RING Meeting.

Maghami C, Guglielmi Y & Ulrich C. (2023). Geological modeling and stress analysis of a complex fault at MTUL. 2023 RING Meeting.

Marchal P, Caumon G, Collon P, Ledru P & Mercadier J. (2023). Joint stochastic modeling of alteration halos and geological structures using a multicomponent skeleton-based approach. 2023 RING Meeting.

McGaughey J, Brossoit J, Davis K, Fournier D & Hensgen S. (2023). GEOH5: A Framework for Geoscience Data and Model Portability. 2023 RING Meeting.

Osorno Bolívar JS, Caumon G & Cupillard P. (2023). Inverting for mechanical properties in the restoration of 2D geological structures with viscous behavior. 2023 RING Meeting.

Raguenel M, Li W-C, Borgese C, Mazuyer A & Darche G. (2023). Use of hybrid mesh for flow simulations: application on a North Sea CO2 storage study. 2023 RING Meeting.

Rapenne M, Cupillard P & Gouache C. (2023). First-order homogenization of the wave equation in isotropic non periodic stratified media. 2023 RING Meeting.

Rapenne M, Guillaume C, Paul C, Corentin G & Anquez P. (2023). Quadrangular adaptive mesh for elastic wave simulation in smooth anisotropic media. 2023 RING Meeting.

Reype C, Stoica RS, Gemmerle D, Richard A & Deaconu M. (2023). Hug model: parameter estimation via the ABC Shadow algorithm. 2023 RING Meeting.

Ruggiero G, cupillard P & Caumon G. (2023). Uncertainty estimation in elastic FWI images for structural interpretations. 2023 RING Meeting.

Scarpa E, Collon P, Panfilov I & Caumon G. (2023). Reproduction of channel stacking patterns in geomodeling: metrics and impact of the modeling strategy on reservoir flow behavior. 2023 RING Meeting.

Scuotto A, Rapenne M, Cupillard P & Caumon G. (2023). Introducing topography in a quadrangular adaptive meshing algorithm. 2023 RING Meeting.

Taty-Moukati F, Stoica RS, Bonneau F, Wu X & Caumon G. (2023). Stochastic seismic interpretation with a marked point process. 2023 RING Meeting.

Tertois A-L. (2023). Cross-Bedding, Bedforms, and Tetrahedral Meshes. 2023 RING Meeting.

Yang J, Colombo C, Carl F, de los Angeles de Lucio G, Achtziger P, Kukla PA & Wellmann F. (2023). Intelligent interpolation strategies in geological modelling for safe storage sites: comparison and uncertainty approach. 2023 RING Meeting.

Zakari M & Caumon G. (2023). Upscaling permeability using intersections of unstructured meshes and non-conforming structured grids. 2023 RING Meeting.