The SuperMAG data processing technique

Abstract

In this paper I outline the data processing technique which is used in the SuperMAG initiative. SuperMAG is a worldwide collaboration of organizations and national agencies that currently operate more than 300 ground based magnetometers. SuperMAG provides easy access to validated ground magnetic field perturbations in the same coordinate system, identical time resolution and with a common baseline removal approach. The purpose of SuperMAG is to provide scientists, teachers, students and the general public easy access to measurements of the magnetic field at the surface of the Earth. Easy access to data, plots and derived products maximizes the utilization of this unique data set. It is outlined how SuperMAG processes the observations obtained by the individual data provider. Data are rotated into a local magnetic coordinate system by determining a time dependent declination angle. This angle displays a slow gradual change and a yearly periodic variation attributed to changes in the Earth main field and season temperature variations. The baseline is determined from the data itself in a three step process: (1) a daily baseline, (2) a yearly trend, and (3) a residual offset. This technique does not require so‐called quiet days and thus it avoids all the well‐known problems associated with their identification. The residual offset for the N‐ and Z‐components shows a distinct latitudinal dependence while the E‐component is independent of the latitude. This result is interpreted as being due to a weak ring current (likely asymmetric) which is present even during official quiet days. For the purpose of M‐I research using 1‐min data I find no difference between observatories and variometers. I finally argue that there is no correct baseline determination technique since we do not have a set of ground‐truth observations required to make an objective evaluation. Instead, the user must keep in mind the assumptions on which the baseline was determined and draw conclusions accordingly.