Improvements to In-situ Magnetometer Calibration Using Selective Signal Processing to Remove Local Stray Fields

High accuracy magnetic field measurements require that the magnetometer be calibrated in-situ to account for evolving instrument performances as the instrument ages. This can be done by minimizing the difference between the measured data and a reference magnetic field during geomagnetically quiet times. Proper characterization and removal of stray fields plays an important role in the accuracy of the minimization by reducing the variability of the parameters between calibrations. On missions with multiple magnetometers, using a 'Ness-style' gradiometer technique is sufficient for stray field characterization and removal provided the stray field can be approximated as a dipole. However, on missions such as Swarm-Echo, the most prominent stray field source early in the mission comes from the reaction wheels, which contain non-dipole terms, making the 'Ness-style' gradiometer insufficient for this. The stray field from the reaction wheels, coupled with limited data coverage in certain intervals, impacted the fit of the calibration, and resulted in the minimization returning non-physical variations in parameters between calibrations. To aid us in characterizing and removing this prominent noise source, we used Multi-Channel Singular Spectrum Analysis on the magnetometer data prior to August 2016 when the wheel rates were maximum, and data coverage was minimal. This signal processing technique is advantageous, as it is model free and does not require any assumptions about the signal prior to analyzing. We discuss the steps taken to implement Multi-Channel Singular Spectrum Analysis on the Swarm-Echo magnetometer data and discuss the impact it has made on the in-situ calibration results and measurements.