Sensitivity of the geomagnetic axial dipole to thermal core-mantle interactions

Since the work of William Gilbert in 1600 (ref. 1), it has been widely believed that the Earth's magnetic field, when suitably time-averaged, is that of a magnetic dipole positioned at the Earth's centre and aligned with the rotational axis. This 'geocentric axial dipole' (GAD) hypothesis has been the central model for the study of the Earth's magnetic field-it underpins almost all interpretations of palaeomagnetic data, whether for studies of palaeomagnetic secular variation, for plate tectonic reconstructions, or for studies of palaeoclimate(2). Although the GAD hypothesis appears to provide a good description of the Earth's magnetic field over at least the past 100 Myr (ref. 2), it is difficult to test the hypothesis for earlier periods, and there is some evidence that a more complicated model is required for the period before 250 Myr ago(3). Kent and Smethurst(3) suggested that this additional complexity might be because the inner core would have been smaller at that time. Here I use a numerical geodynamo model and find that reducing the size of the inner core does not significantly change the character of the magnetic field. I also consider an alternative process that could lead to the breakdown of the GAD hypothesis on this timescale, the evolution of heat-flux variations at the core-mantle boundary, induced by mantle convection. I find that a simple pattern of heat-flux variations at the core-mantle boundary, which is plausible for times before the Mesozoic era, results in a strong octupolar contribution to the field, consistent with previous findings(3).