hide
Free keywords:
-
Abstract:
Modern geomagnetic field models are constructed from satellite and observatory data, while models on the millennial timescale are constructed from indirect records of paleomagnetic origin. An intermediate period, spanning the last four centuries, is covered by historical survey data and ship-log declination records. The magnetic field vector information is distributed over two uneven classes of data: Observations of the field intensity and inclination are sparse and unevenly distributed with clustering and a strong bias towards the northern hemisphere. Declination observations densely cover the globe and in particular the oceans back until the 16th century, and are available in high numbers. However, the unpaired declinations contain limited information about the magnetic field, as they are not sensitive to the field's intensity.
We employ recently developed statistical inference based on sequentialized Gaussian process regression in order to assess what information can be gained from integrating the large number of declination records into the modeling. Instead of uncovering a more detailed field, the availability of more records leads to better resolution of the global field structure. The availability of more records helps notably to constrain global field properties like the dipole moment. It also allows to resolve some detailed field structures more accurately. In principle, the method generalizes to any field represented by spherical harmonics.
