Hemispheric Dichotomy of Mantle Dynamics Revealed by Machine Learning

Adam J. Stewart,Yanchong Li,Zebin Cao,Lijun Liu
2023-06-26
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 long-wavelength signals, uncovers a more prominent degree-one topographic component than the degree-two pattern, both with much smaller amplitudes than previously thought. We suggest that this result reflects a hemispheric contrast in mantle dynamics, especially the two large low shear-wave velocity provinces (LLSVPs). This inference is further strengthened by quantifying the respective topographic contributions from lithospheric isostasy and dynamic topography. We conclude that mantle convection beneath the Atlantic is more vigorous than that beneath the Pacific, likely due to their different thermal-chemical states.
Geophysics
What problem does this paper attempt to address?