Refining Gondwana and Pangea palaeogeography: estimates of Phanerozoic non-dipole (octupole) fields

Reliable Phanerozoic paleopoles have been selected from the stable parts of the Gondwana continents and, upon appropriate reconstruction, have been combined in an apparent polar wander (APW) path, which can be compared with a previously compiled path for Laurussia. This comparison once again confirms that Pangea-A reconstructions for Late Palaeozoic and early Mesozoic times cannot be reconciled with the available paleomagnetic data, unless these data are corrected for latitudinal errors caused by non-dipole (octupole) field contributions or by inclination shallowing. Because the discrepancies persist even when only paleopoles from igneous rocks are used, inclination shallowing cannot be the sole cause of the problem. There is an apparent decrease in the percentage of octupole field contributions needed as a function of time; for Mesozoic and younger time, Gondwana–Laurussia comparisons require, on average, lower ratios of octupole/dipole fields than for Palaeozoic time. However, the Gondwana paleopoles for the Palaeozoic include a much greater proportion of results derived from sedimentary rocks than do those for the Mesozoic, so that this apparently diminishing octupole field contribution may be an artefact. We have also examined whether the clustering of coeval Gondwana poles improves with optimal G3 contributions, but found that while there are improvements, they are not systematic and not statistically significant. A combined APW path has been constructed for Pangea for times since the Mid-Carboniferous, which accounts for octupole fields, or equivalently, inclination shallowing. We argue that this ‘global’ path is an improvement over previous constructions as it represents a self-consistent plate tectonic model and does not violate widely accepted Pangea-A reconstructions.