Scientific program

• Key Topics
• Contributed Sessions
• Methods
• Areas of Applications

Program

The conference welcomes contributions to all areas of mathematical geosciences and geoinformatics. Contributions can either be submitted to

• contributed focus sessions,
  organized around current hot topics by conveners from these subject areas,
• or without a preselected session to general sessions
  organized for clusters of topically related contributions by the sciencific committee.

Contributed Sessions

Click on the titles marked to see a description.

1. Advances in Image Analysis and Machine Learning Algorithms for Quantitative Petrography: From Mineral Separates to Core Logging

   L.Paul Bedard, Arnaud L. Back, Julien Maitre, Kevin Bouchard

   Mineral identification and segmentation are challenging yet crucial tasks in most geological studies. Conventional petrography, which depends on written descriptions, lacks numerical precision and is susceptible to inaccuracies and discrepancies. However, with the growing availability of imaging systems and recent advances in computer image analysis, the potential of quantitative petrography has increased significantly. Recent deep learning algorithms have made substantial progress in the area of mineral identification in thin sections, crushed and unconsolidated sediments. They now boast accuracy rates approaching 90%. There have been significant advancements in the segmentation of minerals in thin sections and drilling core images. However, developing a general model applicable to diverse geological contexts, including igneous, sedimentary, metamorphic, hydrothermally altered rocks, and mineral deposits, remains a challenge. Finally, quantitative descriptions of mineral shapes, which are essential for interpreting geological processes, have also seen substantial progress. This session invites talks on mineral identification and segmentation, quantitative petrography, and core logging processes. We encourage participants to present their work on these subjects, highlighting any innovative methods or cutting-edge technologies that address current challenges.
2. AI-driven Mineral Prospectivity Modeling

   Emmanuel John Carranza, Renguang Zuo

   Mineral prospectivity modeling as a computer-based approach to delineate target areas for exploration of certain mineral deposits in a mineral system has evolved from being knowledge driven to artificial intelligence (AI)-driven. The applications of AI in mineral exploration are ever increasing nowadays to address the complexity of relationships among datasets and with known deposit occurrences. The session welcomes submissions for presentations of: (1) novel AI algorithms and applications for recognition and integration of geo-anomalies to support mineral exploration, in 2D or 3D; and (2) novel AI algorithms and applications for analysis and synthesis of a variety of geoscience datasets to model mineral prospectivity and associated uncertainty, in 2D or 3D.
3. Computational methods for quantifying hydrocarbon resource potential and exploration risk

   Jingdu Yu, Weihong Liu, Qian Zhang

   Advanced algorithms in hydrocarbon exploration and resource assessment, as well as risk computation, provide essential insights and strategies for exploration activities. This topic encompasses not only oil and gas but also resources such as uranium in sediments and helium in gas reservoirs. Developing methods for risk and potential analysis to achieve a higher degree of quantification is at the forefront of both theoretical and practical advancements in geology and exploration decision-making. We welcome the submission of methods and abstracts related to data fusion, geological information extraction, resource potential computation, risk quantification models for forecasting, seismic and well logging attributes from data processing, and mathematical data applications. Submissions should specifically link to hydrocarbon and associated resource assessment technologies. This topic holds significant potential and appeal for academic journals focusing on technological advancements and their practical applications in hydrocarbon exploration and exploitation.
4. Big Data Mining & Artificial Intelligence in Solid Earth Science

   Yongzhang Zhou, Marshall (Xiaogang) Ma, Hui Yang, Craig A. Knoblock

   Big data and machine learning have brought earth resource and environmental study into an artificial-intelligence research stage. Big data mining, machine learning and artificial intelligence algorithms and models have been applied to study multi-scale and multi-type ore deposit exploration and environmental observation. This session is devoted to highlight recent progress in the research and applications of big data and machine learning in the fields of earth science. We welcomes all big data and AI-driven ideas and research to address the development, especially current and future challenges in big data mining and machine learning in geoscience. It will cover the latest data science and machine learning advancements in combining such multi-disciplinary data to enhance sustainable and efficient decision-making in ore deposit exploration and environmental observations. Studies using new data-driven approaches for geochemical data analysis, geological modeling, geophysical inversion, mineral exploration, decision-making under geological uncertainty, are encouraged.
5. Theories and Practices in AI for Geosciences

   Marshall (Xiaogang) Ma, Tao Wen, Anirudh Prabhu

   Artificial intelligence (AI) is rapidly reshaping how geoscientists analyze, model, and understand complex Earth systems. This session explores both the theoretical foundations and practical applications of AI in and across many fields in geosciences. We invite contributions that examine new methodological advances, such as knowledge graphs, deep learning, generative models, and foundation models, and their ability to capture spatial, temporal, and multiscale patterns in geoscientific data. Equally welcome are studies that highlight real-world implementations, including data integration, anomaly detection, simulation acceleration, uncertainty quantification, and decision-support tools for environmental and resource challenges. A central focus of the session is bridging theory and practice. For theories, we welcome thoughts and insights on how mathematical principles guide model development, how domain knowledge can be embedded into AI systems, and how AI methods can be validated against physical constraints and geoscientific understanding. For practices, we encourage case studies from fields such as geology, hydrology, geophysics, geochemistry, geomorphology, planetary science, and more. Overall, this session aims to foster a dialogue on rigorous, interpretable, and impactful AI approaches and envision a roadmap that can guide the future applications of AI in geoscience research.
6. Geoinformatics for Mineral Exploration: Integrated Spatial Intelligence and Target Generation

   Mana Rahimi, Vesa Nykänen

   Geoinformatics brings together spatial data, digital mapping, and quantitative modelling to support mineral exploration from regional analysis to prospect-scale target definition. Today’s exploration teams work with multi-source datasets (geological, geophysical, geochemical, and Earth observation) across 2D, 3D, and time, and need integrated spatial intelligence to turn these complex data landscapes into actionable targets within realistic time and budget constraints. This session provides a forum for innovative concepts, methods, and operational workflows in applied geoinformatics for mineral exploration. We welcome contributions demonstrating how integrated spatial approaches—combining mapping, modelling, databases, visualization, and decision support—improve exploration targeting, support mineral systems thinking, and bridge the gap between research tools and exploration practice. Both methodological advances and significant case studies from industry, government, and academia are encouraged. Indicative topics include (but are not limited to): • Geoinformatics workflows for mineral exploration and target generation from multi-source data • Innovative approaches to spatial data processing and integration of geological, geophysical, geochemical, and remote sensing/Earth observation data • Mineral systems–based modelling applied to spatial prospectivity mapping for concept-to-target exploration workflows • Exploratory Data Analysis (EDA) to support data pre-preprocessing in mineral prospectivity mapping and target generation • Design and use of exploration geodatabases, data infrastructures, and FAIR geodata practices • Advanced methodologies for processing and analysing multi-/hyperspectral and multi-temporal remote sensing data in mineral exploration and target generation • GIS- and remote sensing–based methods for 2D/3D/4D geo-mapping and alteration/mineral system characterization • Data-, knowledge-, and hybrid-driven spatial models for exploration targeting and decision-making in mineral exploration • Multi-criteria and multi-scale spatial decision-support approaches in exploration • Interactive geo-visualization and dashboard-based tools for mineral exploration data and models. The emphasis is on advanced, practical geoinformatics solutions that increase the success of modern mineral exploration and help solve complex spatial problems in targeting and decision-making
7. Sustainable advances and applications in the modelling and forecasting of complex spatial and/or temporal patterns

   Xiao Xia Lang, Guillaume Pirot

   This session aims to highlight and showcase emerging sustainable practices and advancements in ML for spatial and/or temporal pattern prediction within the geosciences. We particularly welcome contributions that address the trade-offs between model complexity, computational efficiency, and environmental impact. On one hand, in recent years, the application of machine learning (ML) within the geosciences has expanded rapidly. As datasets grow larger and models become more complex, the computational resources required have increased substantially. The high performances of CPUs and GPUs have risen model computations, but also elevated energy consumption, contributing to a larger carbon footprint and, consequently, to climate change. Addressing the trade-off between scientific innovation and environmental sustainability has become essential. On the other hand, predictive uncertainty analysis in geoscience applications such as groundwater flow and transport is often relying on the simulation of non-linear physical processes. These processes are usually parameterized by spatial and/or temporal property fields whose structural patterns strongly influence predictive outcomes. Given the limited amount of field data available to estimate property fields of interest, we rely on expert knowledge and our ability to consider different scenarios to represent them, estimate their structure and their patterns. In order to support scenarios exploration and uncertainty quantification, we need parsimonious modelling and forecasting techniques. We invite contributions presenting novel developments, methodological advances, and practical applications across a range of techniques, including but not limited to geostatistics, multiple-point statistics, and deep learning for spatial and/or temporal pattern predictions. Submissions that critically evaluate the sustainability and performance trade-offs of ML approaches in real world geoscientific contexts are especially encouraged.
8. Intelligent Mineral Exploration: Data-Driven and AI-Empowered Decision-Making Under Uncertainty

   Behnam Sadeghi, Eric Grunsky, David Zhen Yin, Wenlei Wang, Glen Nwaila

   This session explores how artificial intelligence and data-driven reasoning can transform mineral exploration into a smarter, uncertainty-aware, and more sustainable process. In the age of energy transition, discovering and managing critical minerals—such as copper, lithium, nickel, cobalt, and rare earth elements—demands not only advanced data analytics but also intelligent frameworks that learn from uncertainty rather than avoid it. Participants are invited to share cutting-edge research and creative “AI brainstorming” ideas that push the limits of how data, models, and expert judgment converge to guide exploration decisions. The emphasis is on how probabilistic and machine-learning approaches—spanning geochemical, geological, geophysical, and remote-sensing domains—can be integrated into cohesive, adaptive systems that improve mineral prospectivity mapping, optimize survey design, and support transparent, risk-aware exploration workflows. Key themes include: • AI and machine learning in uncertainty-driven exploration decision-making; • Data fusion across geological, geochemical, geophysical, and environmental datasets; • Advanced geostatistical and multifractal models for spatial uncertainty quantification; • Cognitive AI and human-in-the-loop frameworks for decision optimization; • Sustainable exploration strategies balancing efficiency, accuracy, and environmental responsibility. This session aims to create an interactive space for AI-driven brainstorming combined with geomathematics and geostatistics, bridging geoscience intuition and computational intelligence to redefine how decisions are made in the face of uncertainty.
9. Adaptive and Nonlinear Data Analysis Techniques for Characterizing the Complex Earth Systems

   Enamundram Chandrasekhar, Sang-Mook Lee

   The digital manifestations of the Earth's natural complex systems, such as climate changes, various types of geohazards, ocean and space weather dynamics etc, depicted either in the form of timeseries or its associated maps are all fundamentally non-stationary and nonlinear in their behaviour. Traditional data analysis techniques, which are stationary and linear, often fail to adequately capture the inherent nonlinear, multiscale and nonstationary processes to characterize such complex systems. The primary goal of this session is to foster cross-disciplinary knowledge exchange on cutting-edge adaptive and nonlinear data analysis techniques to demonstrate their relevance towards better understanding of these complex Earth systems. We invite contributions focused on applications of various advanced adaptive data analysis techniques, including, but not limited to Empirical Mode Decomposition (EMD) and its variants, Hilbert-Huang transform, wavelet analysis, fractal and multifractal analysis and Recurrence Quantification Analysis (RQA), on such multidimensional geophysical data, that specifically address the dynamics of their complex phenomena. Contributions bridging these nonlinear mathematical techniques with latest data-driven methods, such as machine learning and deep learning algorithms, designed for nonlinear data analysis are also welcome.
10. Mathematical Modelling and Machine Learning Applications in Early Warnings and Mitigation of Natural Hazards

    Enamundram Chandrasekhar, T Hemalatha, Roberto Carniel, E Hari Chandana

    In an era of increasing exposure to geohazards, namely, earthquakes, landslides, volcanic eruptions, and tsunamis, the integration of mathematical modelling and time series analysis tools with machine learning (ML) and deep learning (DL) offers transformative potential for early detection, quantitative risk assessment, and mitigation of geo-hazardous events. Recent advances in computational power, sensor networks, and multiparametric data availability (e.g., seismic, geodetic, geophysical, geological, remote sensing and in-situ monitoring) have significantly facilitated the development of real-time predictive models, hybrid geophysics-informed AI systems, and robust early warning frameworks. Furthermore, progresses in probabilistic and stochastic modelling, uncertainty quantification, spatio-temporal geostatistics and explainable AI (XAI) are becoming essential for ensuring enhanced reliability and confidence in predictions. These tools can effectively increase disaster preparedness through faster response times and more accurate forecasts. This session aims to provide a suitable platform to foster interdisciplinary collaboration among geoscientists, statisticians, applied mathematicians, and ML/DL researchers. The session will also emphasize multi-hazard interactions, real-time data assimilation from IoT sensor networks, satellite constellations, and crowd-sourced observations, as well as the challenges of scalability and transferability across geologically and socio-environmentally diverse regions. We invite contributions that present innovative research and applications at the intersection of mathematics, data science, and natural hazards. Sustainability through energy-efficient and ethically responsible AI deployments in hazard-prone communities will also be a key consideration. We particularly welcome contributions discussing current trends, challenges, and opportunities in building the advanced early warning and mitigation systems driven by understanding of the physical processes and data-driven intelligence.
11. Advances in Quantitative Modelling of Earth Dynamics and its Manifestations

    Alik Ismail-Zadeh, Alexandre Fournier , Oleg E. Melnik, Pavel Novák

    Over recent decades, quantitative geodynamics has advanced substantially, deepening our understanding of the physical and chemical processes governing the Earth's interior. These developments have been driven by progress in geodesy, geophysics, and the integration of modern computational and mathematical techniques. An important emerging direction is the application of data assimilation methods, which enable the reconstruction of past geodynamic states and the optimal estimation of present dynamic and magnetic states by combining observational data with numerical models. This session focuses on recent advances in quantitative modelling in geodynamics and the development of innovative approaches to studying Earth’s dynamic behaviour. Particular attention is given to the coupling between deep Earth processes and their surface expressions, including seismicity, volcanic activity, and related numerical simulations of earthquakes, eruptions, and lava flow dynamics. We invite contributions addressing quantitative aspects of geodynamic research, the development of new modelling and computational methods (including applications of artificial intelligence and machine learning), and the application of these methodologies to a wide range of geodynamic problems — from core and mantle dynamics to lithospheric and shallow Earth processes.
12. Advances in Analysis, Mathematical and Numerical Modelling in Atmosphere, Hydrology, and Sea-Ice Dynamics

    Elisa Arnone, Elisa Mantelli, Leslie M. Smith

    The increasing volatility of climate variables is driving significant changes in the dynamics of the atmosphere, sea ice and the hydrological cycle, affecting both extreme events and hydrological regimes. In this context, numerical modeling and the integration of multiple models represent powerful tools, either to support decision-making or to serve as virtual laboratories for assessing the impacts of changing scenarios. This session invites contributions that leverage recent advances in numerical modeling, including multi-model approaches, to analyze the dynamics of the atmosphere, hydrology, and sea ice under changing conditions, or in response to the need for decision-support tools.
13. Geometallurgy, geostatistics and stochastic decision making along the minerals value chain

    K. Gerald van den Boogaart, Julian Ortiz, Jörg Benndorf, Raimon Tolosana Delgado

    The sessions welcomes contributions from all aspects of mining-relevant mathematical methods, with the aim to contribute to integrating the minerals value chain. Contributions can cover parts or the whole of the mining cycle from orebody modelling to mine closure and consider all scales from microstructure to long term mineral province, from characterisation to decision making. Important areas are: quantification and modelling of ore and waste rock texture; mining geostatistics of grade, hardness and any other geometallurgical variables, with two-point or with high-order methods; modelling of beneficiation processes from mining to smelting; structural modelling and geomodelling for ore body modelling (particularly if tracking uncertainty); or stochastic optimisation, model update and decision making along the value chain (mine planning and scheduling, adaptive processing, real time mining, etc). These are non-exclusive examples: any contribution to characterising, understanding, forecasting and optimising processes and materials along the minerals value chain is welcome.
14. Geological Process Modeling

    Daniel Tetzlaff, Ana M. Sosa Paez

    This session encompasses research and applications of quantitative models that simulate physical processes for generating geologic scenarios and petrophysical properties in the context of their evolution through time. The subject includes not only sedimentary and stratigraphic processes, but also structural, diagenetic, geochemical, and biological processes. Statistical and artificial intelligence models may be included as long as they relate directly to physical phenomena that control the formation and evolution of geologic systems. Examples of applications include erosion, transport and deposition of sediment, carbonate growth, weathering, diagenesis, rock deformation, fracturing, faulting and mineralization.

Methods

1. Two Point Geostatistics
2. Multi Point Geostatistics
3. Model-Based Geostatistics
4. Point Processes
5. Time Series Analysis
6. Space-Time Processes
7. Fractal and Multi-Fractal Modelling
8. Image Analysis
9. Stochastic Geometry and Stereology
10. Other Spatial and Space-Time Statistical Methods
11. Spectral and Hyperspectral Data
12. Compositional Data Analysis
13. Spherical Mathematics, Probablity and Statistics
14. Multivariate Statistics
15. Bayesian Statistics
16. Machine Learning
17. Optimisation and Operations Research
18. Other Non-Space, Non-Time Statistical Methods
19. Numerical modelling and numerical simulation
20. Data Assimilation and data Integration
21. Inverse Problems
22. 3D geodata and modeling of 3D objects
23. Geographic and Geoscience Information Systems
24. space/space-time geo-databases
25. e-Geosciences and geoscience information exchange standards
26. other computer sciences methods
27. other mathematical methods
28. other methods

Areas of Application

1. Geomorphology and Quaternary geology
2. Tectonics, Stratigraphy and Sedimentology
3. Stress, Strain, and Fractures
4. Petrology, Mineralogy and Crystallography
5. Natural hazard and geological risk assessment
6. Oil and Hydrocarbon Energy Resources
7. Mineral Resources, Mining and Geometallurgy
8. Soil mechanics and geotechnics
9. Water: Sea, Surface and Subsurface
10. Climatology, Meteorology, Cryology, Atmosphere
11. Palaeonthology, Paleoecology, Biogeology
12. Soil Science: pedology and edafology
13. Geochemistry
14. Isotope Geochemistry
15. Chemical Analytics in the geosciences
16. Geophysics and Geodesy
17. Remote Sensing
18. Geography
19. Other topics in the geosciences
20. Other topics outside the geosciences

Key Topics