Forecasting change of the magnetic field using core surface flows and ensemble Kalman filtering
Article first published online: 19 SEP 2009
Copyright 2009 by the American Geophysical Union.
Geophysical Research Letters
Volume 36, Issue 18, September 2009
How to Cite
2009), Forecasting change of the magnetic field using core surface flows and ensemble Kalman filtering, Geophys. Res. Lett., 36, L18303, doi:10.1029/2009GL039927., and (
- Issue published online: 19 SEP 2009
- Article first published online: 19 SEP 2009
- Manuscript Accepted: 19 AUG 2009
- Manuscript Revised: 17 AUG 2009
- Manuscript Received: 6 JUL 2009
- magnetic field;
 Accurate forecasting of the change of the Earth's internal magnetic field over short intervals of time (e.g., less than five years) has many applications for government, academic and commercial users. Forecasting can be achieved by making a number of reasonable assumptions about how the main field interacts with the flow in the liquid outer core. In particular, the magnetic field can be considered to be entrained in the large scale flow along the core-mantle boundary surface over short time periods, giving rise to measurable change at the Earth's surface. The observed change (or secular variation) at or above the surface of the Earth can thus be inverted to produce flow models; these can be used to propagate fluid parcels threaded by the field forwards in time to forecast the non-linear change of the magnetic field. In addition to prediction of field change by flow models, it would be advantageous to include observations of the field from satellite measurements or ground-based observatories. We therefore present a method using Ensemble Kalman Filtering (EnKF) to produce an optimal assimilation between magnetic field change as forecast from core flow models and direct observations of the field. We show, by assuming a steady flow and assimilating field observations annually, it is possible to produce a forecast over five years with less than 30nT root mean square difference from the ‘true’ field – within an assumed error budget. The EnKF method also allows sensitivity analysis of the field models to noise and uncertainty within the physical representation.