Considerations for Latitudinal Time-Averaged-Field Palaeointensity Analysis of the Last Five Million Years

Central to palaeomagnetism and geophysics is the assumption that the time-averaged geomagnetic field is approximated by a geocentric-axial-dipole (GAD). In this paper, it is demonstrated through the use of a simple cap model that due to secular variation the time-averaged palaeointensity record will always have a smaller latitudinal dependency than a true GAD field. However, the simple cap model does not fully explain the behavior of the palaeointensity database (averaged over 0–5Ma) especially at high-latitudes. To investigate this dependency I use a Giant Gaussian Processes (GGP) model to estimate the contribution of permanent non-dipole features and determine their statistical significance. It was found that an axial quadrupole term between −5 and −10% of the GAD field combined with octupole term ∼ −15% of the GAD field, best explained palaeointensity latitudinal behavior. In particular, the octupole term with a sign opposite to that of the GAD, is required to describe the palaeointensity behavior at high latitudes, i.e., >60.