Item

Abstract

Paleomagnetism serves as the principal tool for the quantitative reconstruction of Earth’s paleogeography, thereby providing key input for geodynamics, paleoclimatology and paleobiology. Paleomagnetic data have been used for >60 years to quantify the apparent motion of the time-averaged paleomagnetic pole relative to a fixed continent or tectonic plate through time. By assuming that the time-averaged geomagnetic pole coincides with the Earth’s spin axis, this polar motion can be translated into the motion of a continent or plate relative to the geographic pole. However, obtaining a reliable estimate of the time-averaged pole position requires the averaging of the short-term, secular variation of Earth’s magnetic field. Here, I show that coeval paleomagnetic poles derived from datasets underrepresenting paleosecular variation or based on a smaller number of paleomagnetic sites are more dispersed and often biased compared to poles that adequately sample secular variation. This reveals that the degree to which paleosecular variation is averaged is an important contributor to uncertainties and bias in estimates of the paleomagnetic pole position. Moreover, the most widely used method to correct sediment-derived paleomagnetic poles for the notorious problem of inclination shallowing relies on both a model of paleosecular variation as well as on the assumption that secular variation is adequately represented by the dataset. In this study, I highlight the importance of a thorough understanding of secular variation throughout geological time for the use of paleomagnetism as a quantitative tool for the reconstruction of paleogeography.