International Union of Geodesy and Geophysics (IUGG)
Conference on Mathematical Geophysics (CMG2022)
Seoul, Republic of Korea, June 20-24, 2022
Overall theme: “Mathematics and Computer Science for Earth and Space Sciences”
Theme 1: Data Sciences, Machine Learning and Artificial Intelligence[Lead Convener: Sang-Mook Lee and Enamudram Chandrasekhar]
Big data and Artificial Intelligence, including machine learning and deep learning, are becoming essential elements of so-called Fourth Industrial Revolution, a new upheaval that will transform our society. Meanwhile, before such buildup, geoscientists were familiar with inferring new insights and knowledge from observational data through mathematical modeling and solving inverse problems. So an important question is what significant benefit can the new algorithms of Data Sciences provide where the traditional geophysical approaches fall short. In this session, we explore diverse applications of new Artificial Intelligence-based data scientific tools in diverse fields of geosciences ranging from atmospheric sciences to oceanography and geological sciences. Obviously Machine Learning and Deep Learning techniques have demonstrated their usefulness in cases involving image detection and real-time processing of large amounts of data. However, we look beyond such applications. Another interesting development of modern computing is that traditional boundary between applied and pure mathematics to some extent is being blurred. We also explore new ideas of thinking in modern mathematical computations and data processing.
Session 1.1 Nonlinear Signal Processing Techniques for Geophysics[Organizer: Enamundram Chandrasekhar]
Nonlinear signal analysis techniques using fractals, multifractals, wavelet-based and other data-adaptive approaches have been increasingly finding their way in a variety of applications in various branches of earth system sciences, such as geophysics, geomagnetism, atmospheric sciences, ocean science, meteorological studies, and climate change studies among others. While wavelet analysis helps to represent the nonlinear signals in a time-scale or space-scale plane for better understanding of their spatio-temporal behaviour, fractals and multifractal analyses facilitate to understand the intrinsic self-similarities and singularities present in them using scaling exponents as diagnostic parameters. Signals depicting fractal and/or multifractal behaviour are largely generated from nonlinear and dynamical systems. The thorough mathematical formalism and their ability to provide an additional dimension to unravel the hidden information in various kinds of signals make these techniques very unique, in the contemporary understanding of signals and systems. This session invites papers related to theory and applications of wavelets, fractals, multifractals and various data-adaptive techniques in different fields of geosciences.
Session 1.2 Nonlinear analysis and simulation of intermittent space and/or time fields[Organizer:Daniel Schertzer]
Intermittency is ubiquitous in nonlinear systems, particularly those in geophysics. In the past few decades, a key interdisciplinary topic of dynamical systems and turbulence has thus emerged, permitting Important progress to be achieved. However, much remains to be done — e.g., to clarify what are the similarities and differences between various approaches as well as between properties of various systems. Currently, there are approaches, both theoretical and empirical, based on (multi-) singular measures and on topological properties. Systems-wise, there are systems with only a few degrees of freedom, as well as systems with very large number of degrees of freedom.
This session therefore welcomes any communication aiming to foster new developments on intermittency ranging from data-driven research to mathematical physics.
Subject to discussion at the conference, we intend to prepare a special issue based on this session.
Convened by Daniel Scherzer, Ecole des Ponts ParisTech and Imperial College London, Daniel.Schertzer@enpc.fr and James Cho, Flatiron Institute, New York, email@example.com
Session 1.3: Inverse problems and machine learning[Organizer: Andrew Valentine]
In geophysics, we often find ourselves unable to directly observe quantities and phenomena of interest. Instead, we must infer their properties from indirect observations — for example, Earth structure is typically determined via its influence on the propagation of seismic waves. To achieve this, we rely on tools and ideas from inverse theory, and this session aims to showcase new developments in this field. Particular current challenges include inference in settings where the forward model is significantly non-linear; the rigorous quantification of uncertainties; and practical issues surrounding the use of massive datasets and expensive computational simulations. These are all areas where the burgeoning fields of data science and machine learning may have insights to offer, and we particularly welcome submissions that explore synergies and connections to these areas, in addition to more ‘traditional’ perspectives on geophysical inverse theory.
Theme 2: Geophysical Fluid Dynamics[Lead Conveners: William Dewar, Annick Pouquet and Ilya Zaliapin]
Geophysical fluid dynamics refers to the studies of flows in the atmosphere, hydrosphere, oceans and in the Earth’s interior and on its surface. The Gulf Stream, the atmospheric Jet Stream, El-Nino/Southern Oscillation, salt-water estuaries, landslides, and mud flows as well as magma flow inside of volcanic conduits, ductile halite flow in sedimentary basins, and thermal convection in the mantle are some examples of the fluid dynamics processes. This theme intends to cover mathematical and numerical aspects of geophysical fluid dynamics related to IUGG disciplines: atmospheric science (IAMAS), physical oceanography (IAPSO), hydrology (IAHS), cryospheric science (IACS), and volcanology (IAVCEI).
Session 2.1 Hydrologic connectivity in river channel and deltaic systems[Organizers: Efi Foufoula-Georgiou, Alejandro Tejedor, Ilya Zaliapin]
Understanding hydrologic connectivity – a water-mediated transport of matter, energy and biomass – is a fundamental problem with far-reaching science, engineering, and societal implications. This session focuses on complementary approaches to study hydrological, sediment, landscape, or biologic connectivity on different space and time scales in both diverging (deltaic) and converging (river channel networks) systems. We solicit conceptual models, statistical approaches, and mathematical theories that showcase and advance the connectivity science.
Session 2.2 Magma and lava flows[Organizer: Alik Ismail-Zadeh]
Volcanologists refer to magma, when they are talking about molten rock trapped underground. If these partially molten rocks make it to the surface during eruptions and spread on the surface from the volcanic edifice, it is called lava. The eruptions produce a variety of gravity currents depending on the chemical composition and temperature of the magmatic rocks, and the topography of the surface over which the lava flows. This session invites papers related to mathematical and numerical problems related to various aspects of magma and lava flows.
Theme 3: From the Core to the Space: Different Spheres with Common Mathematics[Lead convener: Roberto Carniel, Alexander Fournier, and Shin-Chan Han]
Mathematical theories, numerical analysis, and applications in geomagnetism (core dynamics) and aeronomy (evolution of the magnetic field), in physics and chemistry of the Earth interior as well as experts dealing with modeling core-mantle interaction, mantle & lithosphere interaction, lithosphere dynamics, and space research.
Theme 4: Mathematics for Natural Hazards Science[Lead convener: Alik Ismail-Zadeh and Salvatore Grimaldi]
This theme will focus on the essential methodological aspects of disaster science (that is, science related to all aspects of natural hazards and vulnerability) with emphasis on the interdisciplinary research in geosciences, mathematics, and computer science. Also, the changing climate and the related natural hazards and risks pose a multitude of pressing social and economic questions. Data assimilation, statistical approaches to analyses of extreme events, large-scale numerical modeling, dynamical system theory are some of the topics that will be discussed. The theme will be co-organized by the IUGG GeoRisk Commission, bringing theoretical experts in the field of geohazards, hydromet hazards, space hazards, climate hazards, and disaster risk research.
Session 4.1: Interactions between geophysical, biomedical and urban systems[Organizer:Daniel Schertzer]
The current debate on consistent adaptation and mitigation strategies with respect to environmental and epidemic threats underlines the key question of the multi-faceted interactions between geophysical, biomedical and urban phenomena. Uncertainties and knowledge gaps are numerous and appear to persist. For example, long-lasting difficulties in upscaling innovative climate solutions and determining optimal trade-offs between risks and resources are manifest.
These difficulties call for a development of disruptive concepts, methodologies and models to break the present scientific deadlocks, to better deal with nonlinearities, multi-component systems, stochastic synchronization, non-autonomous and delay aspects of the requisite models, tipping points and elements, and extreme variability over a wide range of scales in geophysical, biomedical and urban systems, as well as their interactions. This session proposes to bring together experts and interested parties from many relevant fields – as suggested by One Health or Global Health approaches – and help create a community active in this highly interdisciplinary area.
Convened by: Daniel Schertzer (Ecole des Ponts ParisTech and Imperial College London, Daniel.Schertzer@enpc.fr), Klaus Fraedrich(Max Planck Institute, Hamburg, firstname.lastname@example.org) and Michael Ghil (Ecole Normale Supérieure, Paris, and University of California at Los Angeles, email@example.com)
Theme 5: Geophysical Inversion: Theory, Algorithms, and Applications[Lead convener: Malcolm Sambridge and Alik Ismail-Zadeh]
The development of novel mathematical tools for inversion and inference remains central to progress in our knowledge of the Earth, where many geoscience observables only indirectly constrain properties of the system of interest. The scope of this theme will focus on ongoing research efforts directed toward i) the characterization of uncertainty and non-uniqueness; ii) the solution of weakly and highly nonlinear inverse problems in an effective manner; and (iii) managing the practical challenges associated with massive data sets and expensive forward models. This theme will highlight recent developments in geophysical inverse theory (broadly defined), focusing primarily on methodology and practical methods or solution with modern data sets and computation. We particularly encourage exploration of the emerging synergies and parallels between inversion algorithms and machine learning, as well as the potential for fundamentally new approaches to geophysical inference.
Session 5.1: Probabilistic approaches to inversion[Organizer: Kerry Gallagher and Jan Dettmer]
Probabilistic and statistical approaches to inverse modelling of different types of Geoscience data have become more popular in the last 15-20 years, partly due to advances in methodological approaches and algorithms, and also due to increased computing power. These approaches can facilitate model choice, quantitative assessment of multiple forward modelling scenarios and forecasting, and estimation of uncertainties in observations, model formulations and estimation of model parameters. In this session, we solicit submissions aimed at improving how we can identify and extract probabilistic information from data and models. These may include new algorithms; comparisons of probabilistic/statistical methods; innovation in use of high performance computing; as well as application or case in use of such inversion techniques.
Session 5.2: Advances in Geophysical imaging and applications[Organizer: Christian Boehm]
Nearly all fields in the Earth sciences combine numerical models and data measurements to infer unknown parameters or to image properties of a dynamical system. Recent advances in inverse theory and numerical methods, on the one hand, and the widespread availability of massively parallel supercomputers, on the other hand, provide unprecedented opportunities to improve our understanding of Earth’s interior and its mechanism. This session aims at discussing recent developments and applications in geophysical imaging across the scales. Contributions include, but are not limited to, the areas of earthquake engineering, passive and active-source seismic imaging, geodynamical modelling, magneto-fluid dynamics, etc. in conjunction with computational approaches such as scalable numerical solvers, optimization strategies, workflow management on HPC clusters, and big data.
Theme 6: Mathematics for Climate Science[Lead convener: Dick Peltier and Ute Herzfeld]
Our understanding of the future climate and its change comes in part from an analysis of climate models based on mathematical and physical ideas. Numerical climate models driven by a set of observations play essential role in the analysis. Understanding climate change on global, continental and smaller scales using theoretical tools, improving existing climate models, and reduction of uncertainties in the models are topics of the theme.
Theme 7: Mathematical Geophysics not listed above