Surface core flows are inverted directly from geomagnetic data, or from their interpollation through Gauss
coefficients, using the radial induction equation at the core mantle boudary (CMB).
We provide here :
Earth's core surface (Aitoff projection)
Orthographic projection
Cross-section of secular variation along Φ
Time series of spherical harmonics coefficients
Comparison with observatories.
Spectra of averaged core quantities
Three core flow models are provided and can be downloaded:
- Gillet_gnm_2015
Quasi-geostrophic motions covering 1940-2010 [Gillet et al, 2015],
inverted from the COV-OBS magnetic field model.
It consists in an ensemble of 20 realizations, which have been obtained using a weak stochastic temporal constraint
and an iterative estimate of the subgrid induction at the core surface, under the frozen flux approximation.
- Barrois_gnm_2017e
Surface core flows over 1955-2020 obtained under a dynamo spatial norm with a stochastic Kalman filter
[Barrois et al, 2017].
These have been inverted from the COV-OBS.x1
magnetic field model. An augmented state approach allows to account for an unmodeled term that includes effects from both subgrid-scale induction and magnetic diffusion
[Barrois et al, 2017 + erratum].
Spatial cross-covariances for the core flow, the magnetic field and the unmodeled term are estimated through statistics derived from geodynamo simulations.
- Barrois_VO_2018e
Surface core flows over 1998-2018 obtained with the same algorithm as Barrois et al, 2017 (+ erratum),
but inverted directly from magnetic records above Earth's surface, namely ground-based series and virtual observatory series
[Barrois et al, 2018].
Magnetic data have been provided by C. Finlay and M. Hammer from DTU Copenhagen.
In addition, visualisation and download of two archeomagnetic models are provided:
-
COV-LAKE and COV-ARCH
The COV-ARCH and COV-LAKE ensemble of models describe Gauss coefficients for the geomagnetic field over the period spanning 1200 BC to 2000 AD. COV-ARCH is obtained by inverting archeomagnetic and lava data, while COV-LAKE also incorporates sediment series. They are constructed using as a temporal prior cross-covariances corresponding to order 2 Auto-Regressive processes [Hellio and Gillet, 2018].
Magnetic series are also compared with the CHAOS-6 field model
[Finlay et al, 2016].
This project is supported by the French National Space Science Agency (CNES).