Bram Vaes , Douwe J.J. van Hinsbergen , Suzanna H.A. van de Lagemaat, Erik van der Wiel, Nalan Lom, Eldert L. Advokaat , Lydian M. Boschman, Leandro Cesar Gallo, Annika Greve, Carl Guilmette, Shihu Li, Peter Lippert, Leny Montheil, Abdul Qayyum, Cor Langereis

Apparent polar wander paths (APWPs) calculated from paleomagnetic data describe the motion of tectonic plates relative to the Earth’s rotation axis through geological time, providing a quantitative paleogeographic framework for studying the evolution of Earth’s interior, surface, and atmosphere. Previous APWPs were typically calculated from collections of paleomagnetic poles, with each pole computed from collections of paleomagnetic sites, and each site representing a spot reading of the paleomagnetic field. It was recently shown that the choice of how sites are distributed over poles strongly determines the confidence region around APWPs and possibly the APWP itself, and that the number of paleomagnetic data used to compute a single paleomagnetic pole varies widely and is essentially arbitrary. Here, we use a recently proposed method to overcome this problem and provide a new global APWP for the last 320 million years that is calculated from simulated site-level paleomagnetic data instead of from paleopoles, in which spatial and temporal uncertainties of the original datasets are incorporated. We provide an updated global paleomagnetic database scrutinized against quantitative, stringent quality criteria, and use an updated global plate motion model. The new global APWP follows the same trend as the most recent pole-based APWP but has smaller uncertainties. This demonstrates that the first-order geometry of the global APWP is robust and reproducible, indicating that paleomagnetism provides a reliable reference frame as basis for reconstructing, for instance, paleogeography and paleoclimate. Moreover, we find that previously identified peaks in APW rate disappear when calculating the APWP from site-level data and correcting for a temporal bias in the underlying data. Finally, we show that a higher-resolution global APWP frame may be determined for time intervals with high data density, but that this is not yet feasible for the entire 320-0 Ma time span. Future collection of large and well-dated paleomagnetic datasets from stable plate interiors are needed to improve the quality and resolution of the global APWP, which may contribute to solving detailed Earth scientific problems that rely on a paleomagnetic reference frame.