Inferring the paleo-longitude directly from the paleo-geomagnetic data

Abstract

It is thought that paleo-magnetism has the incapability in providing paleo-longitude. To obtain this important location parameter many other indirect methods have been developed based on different assumptions. Here we present a scanning method to derive the paleo-longitude from the usual paleo-magnetic measurements. This method takes into account the contributions to the Earth's magnetic potential from additional dipoles with their axes in the equatorial plane, which were omitted by the traditional paleo-magnetism. In this method, firstly we assume that $\theta_p$ and $\lambda_p$ are accurate (or determined well enough), and define a cost function; And secondly we minimize this function by systematically searching through all longitudes and latitudes in their domain; Finally when a local minima of this cost function reaches, the corresponding longitude is the paleo-longitude that we look for. Simultaneously the paleo-latitude is obtained. Synthetic experiments show that this method works very well when there are no errors in the geomagntic measurements (Components of magnetic field: $B_x, B_y, B_z$, or declination $D$ and inclination $I$). If there exist errors in geomagntic measurements, we recommend adding a Tikhonov regularization factor to the cost function for deriving reasonable paleo-longitude, and provide two examples. Error analysis shows that the main error sources for paleo-longitude are $B_y$ and/or $I$ in our method. In addition, such a cost function and its like could be used as a theoretical framework that can directly invert the paleo-longitude, paleo-latitude, and even the location of the paleo-geomagnetic poles simultaneously from the paleo-geomagnetic measurements through any appropriate inversion method.