Loading
Geodynamics models are characterized by significant uncertainties. With rapidly expanding volumes of observational data and sustained improvements in both models and HPC, we now have an opportunity to use this data within the framework of inverse theory to reduce the uncertainties in geodynamics models, and to quantify uncertainties in the predictions of these models. This session will provide an opportunity for our community mostly composed of forward modelers to confront the challenges and potential rewards associated with an inverse view of the world.
Many geodynamics questions are driven by the need to understand how heat and material are transported into, out of, and through the Earth's interior. This session addresses the important transport mechanisms and material flux as constrained by observational data, exploring how geodynamics modeling addresses questions such as: what are the fluxes of volatiles in subduction zones? How is heat transferred across the core-mantle boundary and other interfaces in the Earth’s interior? How do melts form and migrate in the crust?
Geodynamics and seismological models have direct societal benefits: they increase our understanding of geologic processes; enhance human safety by improving predictions of risk, and provide a sense of wonder about the forces that shape our world. This sessions explores the predictive capability of modeling and what it reveals about the dynamics, structure and constituency of the Earth's interior; its relationship to observable deformation, and implications for volcanic and seismic hazard.
Speakers: Claire Currie, U. of Alberta, Eric Dunham, Stanford U.; Cynthia Ebinger, U. of Rochester
Geologists like to describe the Earth as a unique planet, but it is just one among many known worlds and its unique features need to be understood in this context. We are in a golden age of planetary discovery and exploration; the observations of other planetary bodies provide a new perspective on our own planet, while geodynamics models of the Earth provide insight into how other planets and moons may have evolved. The general principles which we apply to modeling the Earth from core to atmosphere apply throughout the solar system and to exoplanets. This session seeks to identify and discuss the common aspects and the challenges of developing geodynamics models of diverse planets and moons, and of the earliest history of Earth.
Speakers: Jon Aurnou, UCLA; Nick Featherstone, CU Boulder; Sarah Stewart, UC Davis
The outcomes and products of geodynamics and seismology modeling can play a distinctive role in disseminating scientific and technological understanding, improving understanding of the Earth as a dynamic system, and engaging a diverse workforce in the computational sciences. Conversely, engaging students in geodynamics modeling requires providing them with the mathematical, computational, and geophysical knowledge needed to become expert users and developers of geodynamics models. This session explores the approaches used to engage learners inside the classroom, in informal learning environments, and through public outreach, communications, and scientific visualization.
To further the discussion raised by the plenary session, an expert panel will provide diverse perspectives on UQ and inverse problems for geodynamics models, followed by open-ended discussion.
This panel explores a variety of questions that arise when software is released for use "in the wild" by the scientific community. Questions to be considered include: What are the obstacles to assembling a successful model? How can we make modeling results reproducible? Do we have the techniques (numerical methods, data integration, scientific visualization, software) required for geodynamics modeling? What computational resources (hardware, networking) are needed to sustain scientific advances using geodynamics modeling? How can I build on a published model I am interested in?
This panel explores issues in verification and validation in geodynamics for software, what the geodynamics community has learned from benchmarking efforts, and what is needed going forward. Key questions include: What considerations should go into designing and running performance and accuracy benchmarks? What are the key elements needed for verification and validation? How do scientists and software developers verify software? How do they validate software?