Publication Date:
2023-07-10
Description:
Machine learning (ML) is emerging as a tool used in an ever-growing number of branches of contemporary geophysical research. The study of Earth’s internal magnetic field and, more specifically, its generating process, the geodynamo has so far been mostly devoid of applying such tools.In this research, first attempts with one such inference method are presented using an image processing deep neural network trained both on synthetic and real core-mantle boundary (CMB) magnetic field maps.Results show that it is currently impossible to obtain reconstructed source models, which are able to account for both the spatial variations and the secular time variations (SV) of the field. Main features of the maps of actual field values could however be reproduced.
Language:
English
Type:
info:eu-repo/semantics/conferenceObject
Permalink