Skip to Main content Skip to Navigation

Assimilation de données et modélisation stochastique dans la réanalyse des données géomagnétiques satellitaires

Abstract : This thesis, entitled {sc Data Assimilation and Stochastic Modelling in Geomagnetic Satellite Data Reanalysis}, intends to retrieve information on the state of the Earth's core at the Core-Mantle-Boundary, by combining, first, spatial constraints coming from direct numerical simulations, and second, temporal information coming from stochastic equations.This purpose is achieved through inverse methods and a data assimilation augmented state algorithm.The proposed algorithm is designed to be flexible, textit{i.e.} able to integrate several types of data or constraints, and to be simple, textit{i.e.} with low computation time and easy to modify.This work fits in with the other studies on the geomagnetic data assimilation of the community, and with the opportunity to use the last satellite data from Swarm spacecraft (2014-....).We have worked in collaboration with Julien Aubert (IPGP), who has provided the spatial constraints from Coupled-Earth dynamo, and with Christopher C. Finlay (DTU) and Magnus Hammer (DTU), who have provided the satellites and ground observatories data.The major outcomes of this thesis are the design of a functional algorithm, validated through synthetic twin experiments (published), and applied, first, to the Gauss coefficients of a geomagnetic model, and second, to the measures of the CHAMP and Swarm missions.My algorithm is able to retrieve information, not only on the measured quantities, but also on the unobserved quantities like the core flows or the magnetic diffusion.This work has led to the production of a magnetic field and core flows model at the core surface which is not classically regularized.The geomagnetic field model shows results that are globally similar to the CHAOS-6 reference field model, and that are coherent with the other studies of the community.Thus, the maps of the magnetic field and the velocity field obtained, confirm that the dipole decay is principally driven by advection, and display the persistent presence of the Atlantic gyre associated with a Pacific hemisphere less energetic.The inverted magnetic diffusion is concentrated under Indonesia and Indian Ocean.Fundamentally, my thesis demonstrate the importance of taking into account the modelling errors in the geomagnetic data assimilation, which leads to strong biases and an underestimation of the textit{a posteriori} errors when those errors are neglected.Finally, the work presented in this manuscript is preliminary, and it paves the way toward an increased use of the satellite data, with in particular, the free release of my code in order to compared the results with the ones obtained by the community.
Document type :
Complete list of metadatas
Contributor : Abes Star :  Contact
Submitted on : Tuesday, April 10, 2018 - 11:50:12 AM
Last modification on : Wednesday, October 14, 2020 - 4:16:24 AM


Version validated by the jury (STAR)


  • HAL Id : tel-01762619, version 1



Olivier Barrois. Assimilation de données et modélisation stochastique dans la réanalyse des données géomagnétiques satellitaires. Sciences de la Terre. Université Grenoble Alpes, 2017. Français. ⟨NNT : 2017GREAU030⟩. ⟨tel-01762619⟩



Record views


Files downloads