Hemispheric Dichotomy of Mantle Dynamics Revealed by Machine Learning
Adam J. Stewart, Yanchong Li, Zebin Cao, Lijun Liu

TL;DR
This paper uses machine learning to better estimate Earth's mantle contribution to topography, revealing hemispheric differences in mantle dynamics and challenging previous assumptions about mantle forces.
Contribution
Introduces a novel machine learning approach to estimate mantle-related topography using surface and crustal data, reducing reliance on empirical models.
Findings
Residual topography amplitudes are smaller than previous estimates.
Hemispheric contrast in mantle dynamics is indicated by a prominent degree-one topographic component.
Mantle convection is more vigorous beneath the Atlantic than the Pacific.
Abstract
Past efforts to interrogate the mantle's contribution to Earth's topography suffer from inadequate in-situ measurements of true bathymetry and the involvement of an empirical plate model whose presumed lithospheric density profile interferes with the interpretation of other mantle forces. Here, we introduce a machine learning algorithm that estimates the oceanic residual topography based solely on surface and crustal attributes, providing a more objective proxy for the topographic contribution from the mantle. Ablation studies show that seafloor age is the most important factor, while other properties help further improve the fit to bathymetry. The resulting residual topography has notably smaller amplitudes than previous estimates, indicating that more surface features can be explained by crustal properties than previously thought. This exercise, designed to allow detection of…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsHigh-pressure geophysics and materials · Seismic Imaging and Inversion Techniques · earthquake and tectonic studies
