#	Pagina
attuale pagina	/open-h2020/projects/213916/results.html

Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - CoreSat (Dynamics of Earthâ€™s core from multi-satellite observations)

Teaser

Summary

Earthâ€™s magnetic field plays a fundamental role in our planetary habitat, controlling interactions between the Earth and the solar wind. Yet observations show a region of persistently weak field in the South Atlantic that has grown in size in recent decades. Pinning down the core dynamics responsible for this behaviour is essential if we are to understand the present geodynamo, and to forecast future magnetic field changes. Global magnetic observations from the Swarm satellite constellation mission, with three identical satellites now carrying out the most detailed ever survey of the geomagnetic field, today provide us with the means to probe the responsible core dynamics in exquisite detail and test hypotheses concerning the origin of field changes in the South Atlantic.

The objectives of the CoreSat project are to use multi-satellite magnetic field measurements to (a) reveal small scales and rapid time changes of the core-generated magnetic field (b) test whether rotation-dominated core convection can explain the recent time-dependence
of the South Atlantic Anomaly.

Work performed

Achieving the project objectives involves the three tasks listed below:

Task 1. Co-estimation of separate models for lithospheric and core magnetic fields, to reveal small scales of the core field

A PhD student, Mikkel Otzen has been hired for this task, and it is well underway. The PI working with 2 MSc students has developed a scheme for performing geostatistical simulation on a spherical surface that generates magnetic field realizations that honours specifiied histograms and semi-variograms of the radial magnetic field. This has been tested in the context of modelling the core and lithospheric magnetic fields and a manuscript describing this is in preparation. In parallel a Hamitonian Monte-Carlo inversion scheme for co-estimating core and lithospheric fields has been developed, validated in synthetic tests, and preliminary tests using Swarm satellite data have been carried out. These tests were successful, and it was concluded that including more detailed prior information on the lithospheric field structure would further enhance the separation. A PhD student, Mikkel Otzen, has been hired and is now undertaking this task. He has developed a lithospheric magnetic field simulator based on prior geological information regarding magnetic susceptibilities, crustal thicknesses, plate histories, and magnetic reversal chronologies. He has tested the impact of the assumed crustal thickness and is presently producing a geologically realistic ensemble of possible lithospheric magnetic fields.

Task 2. Characterisation of the quiet-time polar ionospheric current system, to reveal rapid time changes of the core field

A PhD student, Clemens Kloss, has been hired for this task and it is well underway. Field modelling software has been ported to python and made openly available to the scientific community at https://pypi.org/project/chaosmagpy/ . The ability to calibrate platform magnetometers has been included in the software, opening the way to utilize data from more satellites. Initial tests of the determination of horizontal ionospheric currents within the python field modelling framework have been carried out. In addition, project collaborator Karl- Magnus Laundal has developed a new scheme for better parameterizing nights side reconnection and substorms, an important ingredient in the polar ionospheric currents. A manuscript on this topic, with the PI as co-author is in preparation.

Task 3. Assimilation of high resolution satellite magnetic data into a quasi-geostrophic core convection model

A postdoc, Olivier Barrois, has been hired to work on this topic. In collaboration with new project partner Thomas Gastine (a CNRS researcher at IPGP, France) he has worked on the open-source quasi-geostrophic convection code PIZZA, modifying it for our purposes to include heat flux boundary conditions (with the option for these to be inhomogeneous) and more general background temperature profiles. A new hybrid code, including a quasi-geostrophic 2D velocity field along with a 3D representation of the magnetic and thermal fields, is currently under development. The thermal part has already been successfully implemented and tested and allows a thermal wind to be accounted for in a quasi-geostrophic framework. Work on integrating the 3D magnetic field with the 2D velocity field is progressing well and it is currently undergoing benchmark tests. This hybrid 2D-3D forward code is essential to the data assimilation experiments to be attempted in the second part of this task.

Final results

Task 1: Geostatistical simulation on a sphere has been implemented, we are not aware of any previous geostatistical simulation tools of this type. A publication and release of associated open-source software is planned. Hamiltonian Monte Carlo inversion has been used to probabilistically co-estimate core and lithospheric magnetic fields for the first time. It is planned to revisit this, once more informative prior models of the lithosphere are ready. Production and assessment of an ensemble of geologically plausible lithospheric field models is the next step. This is the most important remaining ingredient needed in order to carry out the planned probablistic separation of the core and lithospheric fields, which will allow smaller scales of the core field to be seen.

Task 2: The ability to construct field models from data collected by platform magnetometers, by co-estimating calibration parameters has been implemented. This was not previously possible and it opens the way to improving models of the quiet time ionosphere by providing data from multiple satellites at different local times. In addition, project partner Dr Karl Magnus Laundal has developed a novel parameterization of nightside reconnection. This was previously a weak point in statistical models of polar ionospheric currents. The next step is to first perform a simple co-estimation of the quiet-time ionospheric currents along with a core field model. PhD student Clemens Kloss is presently visiting Dr. Laundal in Bergen working on this topic. Once the co-estimation on the ionospheric field is implemented and validated an attempt will be made to estimate the core field to higher time resolution.

Task 3: We have developed the first quasi-geostrophic convection model in a spherical shell geometry to include inhomogeneous heat flux boundary conditions. Using this we have studied the behaviour of quasi-geostrophic convection at low Ekman number (down to 10^-9) and low Prandtl number (down to 10^-2) in a highly supercritical, turbulent regime that has been little explored. We reported a poster on this topic at the IUGG meeting and a manuscript on this topic is in preparation. The hybrid 2D-3D model presently in development will allow us to explore the impact of the magnetic field on such convection. Following this experiments will be carried out assimilating magnetic observations field from multiple satellite missions, within the virtual observatory framework and taking advantage of the models for the lithospheric and ionosphere developed in Tasks 1 and 2. These assimilation results will allow us to assess whether core convection can provide a detailed explanation for the present growth and deepening of the South Atlantic anomaly.