We are happy to present the IAMG program below. The exact order of the talks will be finalized at the beginning of July.
Preliminary Program Overwiew
Still subject to changes!Here you find a downloadable pdf version.
The conference will cover a broad range of general sessions on topics in geomathematics, geostatistics, geoinformatics and geomodeling:
- Compositional Data Analysis
- Computational Geophysics
- Computational Geodynamics
- Digital Outcrop models
- Earth System Modeling
- Environmental Geo-engineering
- Geotechnical Engineering
- Geothermal Enginering
- Hydrology and Hydrogeology
- Hydrocarbons and CO2 sequestration
- Hyperspectral methods
- Image Analysis
- Inverse problems and data assimilation
- Machine Learning
- Mineral Resources and Geometallurgy
- Modeling of Coupled processes
- Medical geology
- Modeling of Earth Resources and Energy Transition
- Natural risks and hazards
- Planetary Geosciences
- Remote sensing
- Urban Geology
Click on the titles marked to see a description.
- Beyond Gaussianity: what is the status of the GANs, MPS, Cumulants or Copula approaches?Julien Straubhaar, Thomas Mejer Hansen, Philippe RenardIn this session, we invite authors to show their latest research in the field of non multi-Gaussian random fields. This includes for example Generative Adversarial Neural networks, Multiple-Point Statistics, High-order Cumulants, and Copula approaches. All aspects from new theoretical developments to practical applications, comparison of methods, success or failure stories are welcome.
- Recent developments in machine learning techniques and quantum computing for geoscience applicationsTeeratorn Kadeethum, Daniel O’Malley, Hongkyu Yoon, Hamid M. NickRecent advancements in machine learning techniques and quantum computing have made their way into geoscience research. These approaches have been adopted and proposed to tackle long-standing challenges in geoscience or an enhancement of classical methods that have been used in this field. This mini-symposium invites presentations on advances within areas of machine learning and/or quantum computing in geoscience research. Topics include, but are not limited to, (1) machine learning algorithms and applications for model-reduction, optimization, inverse problems, uncertainty quantification, highly parameterized problems (e.g., parametrization of heterogeneous fields), and efficient dimensionality reduction of nonlinear operators and (2) quantum computing applications in geoscience research; for instance, seismic inversion with quantum annealing, quantum-computational hydrologic inverse analysis, or quantum optimization. The mini-symposium will bring together researchers working on fundamental and applied aspects of machine learning and quantum computing to provide a forum for discussion, interaction, and assessment of their presented techniques.
- Random patterns and shapes in spatio-temporal dataAila Sarkka, Radu StoicaThis session aims to bring together specialists in point processes, spatial statistics, and stochastic geometry, who have a common research question: characterisation and detection of patterns and structures in spatio-temporal data. The complexity of the available data today requires rigorous stochastic methodologies based on probabilistic modelling, simulation algorithms, and Bayesian inference, in order to obtain reliable and interpretable results. The purpose of this session is to present some new mathematical methods which have been developed for answering questions in environmental sciences, forestry, image analysis, stereology, astronomy and other fields, and which can be valuable additions in the methodological toolbox for the statistical analysis of geological data.
- Inverse problemsThomas Bodin, Kerry GallagherNumerous advances in the Earth sciences are driven by innovative inverse modelling techniques, and the ever-increasing power of computational resources. To promote further progress, this session offers a platform to discuss and learn about current and future approaches to solve (nonlinear) inverse problems in the geosciences. Examples of novel applications are also very welcome. Of particular interest are contributions focusing on: - Seismic tomography and full-waveform inversion at all scales, - Uncertainty analysis, - Bayesian inference, - Optimization methods, - Joint inversion of disparate datasets, - Multi-scale, multi-parameter inversion, - Strategies for exploiting massive data volumes, - Effective medium theory, downscaling, and inverse homogenization.
- Knowledge graphs in the cyberinfrastructure ecosystem of geosciencesXiaogang Ma, Chengbin WangVocabularies, schemas and ontologies have been increasingly created and applied in geosciences in the past decades, and have always been a topic of interest in geoinformatics. Inspired by the recent success of knowledge graphs in industry, the academia is beginning to use knowledge graph as an umbrella topic for works on vocabularies, schemas, and ontologies. In the geoscience community, we have seen many successful applications of knowledge graph in data curation and integration in the past decades, such as the global geologic map sharing enabled by OneGeology. Recently, there were also applications of knowledge graph together with machine learning algorithms to improve the quality of data analytics, such as those in hyperspectral remote sensing image processing, extension to mineral taxonomy, and petroleum exploration. Together, knowledge graphs have shown promising contribution to data science applications in geosciences, and more innovative developments are undergoing. This session welcomes submissions of all the topics mentioned above, to provide a venue for knowledge graph practitioners to share their results and experience, learn best practices, and discuss visions and potential collaborations for future work.
- Fractured geological media and fracture networks: flow, graphs, morphologyRachid Ababou, Israel Cañamón Valera"Fractured geological media and fracture networks: flow, graphs, morphology." This session is an opportunity to present various approaches for analyzing geometrical, topological, and hydraulic properties of 3D Discrete Planar Fracture Networks (DPFN), or their 2D counterpart (e.g., discrete flow networks represented by intersecting segments in the plane), or other discrete sets of Boolean objects (conductors, barriers, cavities, etc.). These generalized geological "networks" may be described deterministically or statistically (e.g. as random Boolean objects), and/or, based on graph theory concepts. Extensions to other mechanisms focusing on morphology and structure of discrete fracture networks are also welcome (e.g., two-phase flow, mechanical deformation, electrical conduction…). This topic is also related to upscaling issues in fractured media and networks : these are relevant to this session in connexion with morphology and structure. Overall, this session focuses on mathematical and algorithmic approaches for characterizing geological media and/or discrete fracture networks, in terms of topology, morphology, hydraulic/conduction properties, based on field observations as well as models.
- Computational Petrology and GeochemistryPierre Lanari, Marion Garçon, Paolo SossiNumerical modeling has become a critical aspect of modern research, especially in Earth Sciences as physical and chemical processes occurring in planetary interiors are not always directly observable from the surface. In petrology and geochemistry, a large variety of computational models have been developed to simulate and study these processes. To be realistic, computational experiments commonly rely on an algorithmic or mechanistic approach rather than deriving a mathematical analytical solution to the problem. In addition, the increase in the amount of geochemical data from state-of-the-art instruments (e.g. ICP-MS, microprobe, SIMS) has fostered the development of advanced software solutions for data reduction and interpretation. The ever-increasing size and availability of these datasets has, in turn, opened up new avenues for extracting statistical information on geological processes. The main goal of this session is to bring together geochemists, petrologists and data scientists who are either developing, using and/or applying numerical tools to understand geological processes. Topics of interest include (but are not limited to) geochemical and petrological modeling for fluids, melts and solids using major/trace elements or isotopes, thermodynamics and kinetics, thermo-mechanical simulations of petro/geochemical processes, provenance analysis, optimization and testing of databases. Model developers using machine learning, big data or minimization/inversion routines as well as those developing new techniques and tools for data visualization are particularly encouraged to submit an abstract.
- Analyzing compositional data in geosciencesJennifer McKinley, Karel Hron, Alessandra MenafoglioCompositional data analysis deals with vectors whose components show the relative importance of parts of a whole. This type of data appears in many geological applications commonly expressed as percentages, ppm, or the like. Since the log-ratio approach to compositional data analysis was introduced back in the 1980s, steady progress has been made in methodological approaches, visualization techniques, data modelling and applications. Submissions that cover these topics for analyzing compositional data in geosciences are welcomed, with particular emphasis on the problems related to earth resources and the environment.
- Filters and smoothers. Filters or smoothers?Jaime Gomez-Hernandez, Liangping LiAre you working with the Kalman filter, or the ensemble Kalman filter, or the ensemble smoother, or the extended Kalman filter, or the unscented Kalman filter, or the particle filter, or the diffuse Kalman filter, or the unscented particle filter, or the cubature Kalman filter, or the Gauss-Hermite quadrature filter, or the iterative ensemble Kalman filter, or the singular evolutive extended Kalman filter, or the iterative ensemble maximum likelihood filter, or the singular evolutive interpolated Kalman filter, or the error subspace transform Kalman filter, or the normal-score ensemble Kalman filter, or the restart ensemble Kalman filter, or the ensemble adjustment Kalman filter, or an square root Kalman filter, or the ensemble smoother with multiple data assimilation, or the iterative ensemble Kalman smoother? If so, this session is for you. Submit your abstract and explain your choice.
- Mining geostatistics, optimization and geometallurgyJörg Benndorf, Julian Ortiz Cabrera, Raimon Tolosana Delgado, K. Gerald van den BoogaartThe sessions aims to bring together all aspects of mining-relevant mathematics. The session integrates all mining related geomathematical methods: from microstructure characterization to an integrated spatiotemporal decision making for mining and processing including real time information updating. Important areas are: Potential Mapping, Microstructural Modelling and Observation, Geostatistics of Geometallurgical Variables, High Order Geostatistical Simulation, Structural Modelling with Uncertainty, Stochastic Mine Planning, Real Time Mining updating, and Predictive Process Optimization. Contributions from all fields of application or development of geomathematical methods for mining are welcome.
- Preserving realistic geology in statistical and mathematical geomodelsIngrid Aarnes, Jacob Skauvold, Carl JacquemynIn this session, we invite authors to share their research on how to represent realistic geology in statistical and mathematical geomodels. This could be for example stochastic facies models, process-based/process-mimicking models, implicit and surface-based representations or other statistical and mathematical solutions that ensure geological validity of models. The aim of the session is to bring together researchers working on translating geology into statistical and mathematical geomodels and learn from each other across scales, disciplines and applications. We welcome contributions from a wide range of approaches where the main aim is for numerical models to realistically represent geology as seen in outcrops and analogues.
- Spatial AssociationsYongze Song, Qiuming ChengThis session aims to invite researchers from multiple fields to present their latest methods and applications about the identification of spatial associations. In geosciences, spatial associations are generally explored based on spatial dependence, similarity, heterogeneity, singularity, and other structures of spatial data. In recent a few years, geospatial artificial intelligence provides more opportunities for more accurate and in-depth understanding of spatial associations. In this session, we encourage authors to submit papers in following areas: innovative methods of spatial associations; advanced applications of spatial associations; spatial statistics for spatial associations; and geospatial artificial intelligence for spatial associations.
- Modelling Land Degradation and RestorationJude Ndzifon Kimengsi, Chrétien Ngouanet, Reeves M. FokengLand degradation, a decline in land quality caused by human activities, has been a major global issue during the 20th Century and will remain high on the international agenda in the 21st Century, because of its adverse impact on agronomic productivity, the environment, and its effect on food security and the quality of life (Eswaran et al., 2001). GEF (2019) established that globally, about 25 percent of the total land area has been degraded. It is now established that land degradation and climate change are intrinsically linked affecting over 3 billion people and over 30% of Earth’s arable land. Land degradation and climate change will hinder the achievement of a plethora of Sustainable Development Goals. As a first step towards reversing this global environmental crisis and mitigating climate change, in 2011, countries got committed to the Bonn Challenge - a global effort to restore 150 million hectares of the world's degraded and deforested lands by 2020 and 350 million hectares by 2030. This session is aimed to bring to the scientific committee novel approaches and cross-edge research findings and new methodological insights of monitoring land degradation, landscape dynamics and restoration. We welcome papers from the following themes but not limited to:- • Remote sensing of land degradation at multiple scales • Modelling soil degradation with gridded data and geoinformatics • Novel directions in R/USLE modelling to improve its global applicability • Fine-scale mapping of land restoration opportunities • Land restoration and conservation typologies • Land restoration stakeholders and restoration challenges • Land degradation, restoration and people • The social and political ecology of land degradation and restoration
- Structural modelling: Parametrisation and Interpolation of Sub-Surface ArchitecturesGautier Laurent, Lachlan Grose, Simon LopezThis session is dedicated to new advances and remaining challenges in the field of geomodelling: the reconstruction of sub-surface architectures (e.g., stratigraphic layering, faults, folds, intrusions) from sparse and ambiguous spatial observations. This is a complex data assimilation problem, which is facing the hurdle of both aleatory and epistemic uncertainties. This problem has long been formalised as an interpolation process, with various mathematical implementations, but it generally remains dependent on expert inputs for adding geological controls or counterbalancing mathematical artefacts. This session welcomes presentations that illustrate this problem in original applications and/or propose original solutions, in particular - but not limited to - topics covering: (1) Limitations of numerical interpolation schemes and ways forward, (2) Formalisation of geological objects and concepts, (3) Knowledge integration within interpolation schemes, (4) Aleatory vs. epistemic uncertainties, (5) Original formalisms and modelling schemes.
- Machine learning-based mineral prospectivity mappingRenguang Zuo, John CarranzaMineral prospectivity mapping as a computer-based approach to delineate targeted areas for a specific type of mineral deposit has changed from being knowledge driven to data driven to today’s big data analytics. There are increasing applications of machine learning algorithms in mapping mineral prospectivity and identifying geochemical anomalies association with mineralization. The session welcomes the following submissions: (1) machine learning (such as logistic regression, random forests, support vector machines, and extreme learning machines), (2) deep learning (such as convolutional neural networks, deep autoencoder networks, recurrent neural networks, and generative adversarial networks, (3) simulation, and (4) case studies for identifying geochemical anomalies or mapping mineral prospectivity.
- Time series analysis in Geosciences: an homage to Professor Walther SchwarzacherJennifer McKinley, Eulogio Pardo-IgúzquizaTime series analysis is widely used in cyclostratigraphy, paleoclimatology, hydrology, geophysics, environmental geosciences, etc. The main objective of this session is the presentation of methodologies and cases studies dealing with the analysis of times series in Geosciences. New theoretical developments, new algorithms implementing known methodologies, new strategies for dealing with classical problems or classical methodologies dealing with new problems as well as case studies are appropriate for this special session. The session will be an homage to Professor Walther Schwarzacher a pioneer in mathematical geosciences in general and cyclostratigraphy in particular. Topics of special interest, among others, are: • Time series analysis in quantitative stratigraphy • Spectral analysis of uneven time series and categorical data • Uncertainty evaluation of the estimated power spectrum • New methodologies of spectral analysis • Time series analysis of compositional data • Soft-computing and learning machine methods in time series analysis • New developments in wavelets analysis of time series
- Reservoir/Petroleum GeostatisticsJuliana Leung, SANJAY SRINIVASANTraditionally, geostatistics has been applied in the petroleum industry for generating multiple geomodels conditioned to well, seismic and other reservoir-specific information. These models are subsequently used for flow modelling and recovery analysis. We are soliciting abstracts that showcase recent advancements in uncertainty quantification, characterization of complex reservoirs, integration of multifaceted data from diverse sources recognizing the scale and precision of such data, upscaling of flow/transport properties and coupling of artificial intelligence with traditional modelling workflows. Interesting case studies (including for oil and gas extraction, CO2 sequestration, hydrogen storage, geothermal reservoirs), as well as novel theoretical and computational developments, are welcome.
- Role of nonlinear signal analysis techniques in GeosciencesEnamundram Chandrasekhar, B S Daya SagarNonlinear signal analysis techniques, namely, the wavelet analysis, fractal and multifractal analysis, empirical mode decomposition (EMD) analysis, to name a few, have been increasingly finding their ways in a variety of applications in various branches of earth system sciences, such as geology, geophysics, geomagnetism, atmospheric sciences, ocean sciences, meteorological and climate change studies. The thorough mathematical formalism of these techniques and their ability to provide additional dimensions to unravel the hidden information in various kinds of nonlinear signals make these techniques very unique in their own right. This session provides a suitable platform to disseminate the knowledge to the geoscience community at large, about the applications of these actively progressing mathematical techniques. This session invites papers related to theory and applications of wavelets, fractals, multifractals and various data-adaptive techniques in different fields of geosciences.
- Spatiotemporal GeostatisticsDionissios Hristopulos, Sandra De IacoThe development of spatiotemporal models is at the forefront of current research in Geostatistics. Flexible and realistic models that can capture the complex behavior of dynamic processes are crucial for improved prediction of spatiotemporal processes and for reliable evaluation of the uncertainties involved. Significant modeling challenges involve topics such as the increasing size, heterogeneity, multivariate dependence, multiple correlation scales and non-Gaussian probability distribution of modern spatiotemporal data. This session seeks contributions that will advance spatiotemporal Geostatistics by proposing novel concepts and methodologies, computational algorithms, and innovative applications or studies that focus on spatiotemporal data analysis. A non-exhaustive list of topics includes the development of novel space-time covariance functions (e.g., non-separable models, models on the sphere and on manifolds, multivariate dependence, complex-valued models), covariance-free approaches (e.g., models based on stochastic partial differential equations and explicit precision operators), innovative simulation methods, computational advances and applications to interesting or challenging spatiotemporal datasets, as well as approaches for non-Gaussian space-time data and multiscale models. Contributions that combine Geostatistics with current developments in applied mathematics (e.g., uncertainty quantification and multifidelity frameworks), as well as in machine learning (e.g., hierarchical models, sparse and multiscale Gaussian process regression) are also welcome.
- Up-Scaling of Flow and Transport ModelsBenoit Noetinger, Marco DentzDespite the continuous improvement of the computing power, up-scaling remains a major step allowing to pass fluently from a detailed description of subsurface, to an effective flow model. Up-scaling is essential to address subsurface uncertainties and to effectively explore parameter spaces having large number of dimensions. Up-scaling can help to understand the large-scale behavior of the system at hand and to capture the most relevant parameters. That allows building a tractable model working in a lower dimensional space of parameters. This session welcomes contribution on recent advances in up-scaling research, from stochastic to homogenization theories, from heterogeneous to fractured media, from single to multiphase flows.
- Meshing and simulation of subsurface processesChristian Boehm, Tara LaForce, Jeanne PellerinNearly all fields in geoscience leverage numerical simulations to study subsurface processes, whether to infer unknown parameters, or to predict possible future evolution of a dynamical system. The widespread availability of massively parallel supercomputers enables researchers to create digital representations of the Earth with ever-increasing resolution and physical complexity. Representing subsurface structure in those numerical models is a key ingredient that can have a significant impact on the accuracy and efficiency of the simulations. Meshing the location and geometry of material interfaces, cavities, fault lines, infrastructure, etc., can be challenging, and the complexity of mesh generation algorithms often trade-off with the efficiency and accuracy of the numerical solvers. This session aims at discussing recent advances in (1) mesh generation algorithms and strategies to discretize complex Earth structure, (2) numerical methods to simulate subsurface processes and (3) case studies of applications in the geosciences. Examples include, but are not limited to, geodynamics, single and multi-phase geophysical flow and reservoir modeling, CO2 sequestration, nuclear waste disposal and seismic wave propagation.
- Uncertainty ModelingFlorian Wellmann, Clare BondUncertainties are prevalent in many geoscientific studies: from geostatistics and 3-D geological modelling to coupled process simulations. In this session, we invite contributions covering aspects of uncertainty estimation, modelling and visualisation in geosciences. A special focus will be on (1) methods to estimate uncertainties in data sets, expert elicitation and the analysis of bias (2) approaches to propagate uncertainties through modelling and simulation, including surrogate modelling, probabilistic programming and machine learning approaches, and (3) the quantification and visualisation of uncertainties in temporal and spatial domains. We encourage contributions both addressing theoretical and methodological challenges, as well as applications to geoscientific problems.
- Landscape evolution models: tectonics, relief, climateSebastien Carretier, Jean BraunLandscape evolution models simulate the dynamics of the continental surface by erosion and sedimentation on geological time scales (1ka-100Ma). The development of these models has accompanied the very significant progress in geomorphology since the 1980s. Despite the uncertainties in the erosion laws, these models provide a better understanding of the coupling between tectonic and climatic phenomena and surface processes. One of the major challenges of these models is using them in inversion procedures in order to reconstruct the evolution of topography, climate and tectonics, particularly from sedimentological data. This requires reducing their calculation time and defining the appropriate spatial and temporal scale for their parameterisation. Another challenge concerns the analysis and understanding of the internal dynamics of geomorphological systems at different scales, from that of a river to that of an alluvial plain, but accounting for all the couplings at different scales between hillslopes and rivers. The acceleration in the availability of very high resolution topographic data allows for an increasingly detailed analysis of the topography in 3D, both on the hillslopes of mountains and on alluvial plains. But these new scales of observation challenge models to develop new algorithms capable of accounting for processes at these fine scales while integrating them at the landscape scale over large time scales. Finally, there is a need to develop the modelling of submarine landform dynamics and the continent-ocean interface over long time scales, as well as wind erosion-sedimentation dynamics. This session welcomes contributions on the development of new algorithms to address these issues.