The geomagnetic axial dipole field assumption

The time-average geomagnetic field is commonly assumed to be a geocentric axial dipole field (GAD). This assumption is crucial to many subfields of paleomagnetism in which one needs to determine paleolatitude. In some cases an interval of only 104 years is sufficient to obtain the time-average field, while in others an interval of at least 105 years is required. Numerous spherical harmonic models for the time-average field for the past 5 Myr exhibit conflicting estimates for non-dipole terms. New criteria, most notably that the Gauss coefficients should exhibit the same time-average magnitudes for reverse and normal polarity states, are used to evaluate these models. We find that the GAD assumption is surprisingly good because of the character of temporal variations in the non-dipole field. In the time-average field, other than the GAD field, only a small axial quadrupole field seems to survive careful scrutiny. If the existence of such a time-average quadrupole field is confirmed, it probably reflects coupling of some form between the dipole and quadrupole families. We find no convincing evidence in the Gauss coefficients for longitudinal variation of the time-average field, but there is some evidence for such asymmetry in the paleosecular variation data.