Mantle circulation simulations are now capable of a high level of precision and complexity that allows the creation of numerous "Earth-like" models. Likewise, advances in observational resources and methods have improved the quantity and quality of data documenting the Earth's interior. Combining these developments presents a unique opportunity to enhance our understanding of mantle dynamics and evolution over geological time scales. However, the crucial components of Earth’s mantle thermochemical convection remain debated (e.g. the existence and nature of primordial materials, core-mantle interactions, processes happening in the transition zone). Furthermore, constraining geodynamical simulations through observations poses multiple challenges, such as data availability and quality, or inverse modelling methodology.
This session will explore how observational data can inform the design of geodynamical simulations and assess their predictions, thereby advancing our knowledge of the physicochemical processes that govern Earth's mantle. We invite submissions from across the Earth sciences — including seismology, geochemistry, mineral physics, or geomagnetism — demonstrating how observations can constrain the building blocks of geodynamical simulations or assess their predictions. We welcome studies focusing on pure observational approaches, data-driven inversions, forward numerical modelling, or quantifying how Earth-like a model is. Our overarching goal is to isolate the parameters or processes that may drive models to behave like the Earth.