SUMMARY The geocentric axial dipole (GAD) hypothesis is key to many palaeomagnetic applications, for example plate-tectonic reconstructions; however, the validity of this hypothesis at high latitudes is not fully resolved. To address this, in this paper we determined the palaeomagnetic directional data of 156 lava units in Eyjafjarðardalur, Iceland, with the aim of determining the validity of the GAD hypothesis at high latitudes using time-averaged field (TAF) analysis. In addition to the palaeomagnetic directional data, we constructed an age model for the sequences using new 40Ar/39Ar dates, magnetostratigraphy and field data. The sequence age range is 2.6–8.0 Ma. We show that the mean virtual geomagnetic pole (VGP) for our data does not agree with the GAD theory at 95 per cent confidence, when only the standard tilt and tectonic corrections are made; however, when inclination-shallowing processes are accounted for, for example thermoremanence (TRM) anisotropy and refraction effects, the mean VGP can align with GAD at 95 per cent confidence. These inclination-shallowing processes are shown to reduce the inclination by up to 14° for some of the basaltic units. Applying the inclination-shallowing correction also reduces VGP dispersion to levels that agree with global model predictions. We propose that much of the scatter within the palaeomagnetic directional databases are due to inclination-shallowing process effects, which become more important as the natural remanent magnetization (NRM) intensity is high, for example >2 A m−1. We propose that inclination-shallowing processes can be identified and corrected for by examining the NRM intensity and dispersion.
Read full abstract