Co-estimating geomagnetic field and calibration parameters: modeling Earth's magnetic field with platform magnetometer data

TL;DR

This paper introduces a co-estimation method for geomagnetic field modeling that simultaneously calibrates platform magnetometers and estimates Earth's magnetic field, filling data gaps and improving model accuracy.

Contribution

It develops a CHAOS-based co-estimation scheme that integrates satellite, ground, and platform magnetometer data for comprehensive geomagnetic modeling.

Findings

Platform magnetometer data enhance secular acceleration estimates.

The co-estimation scheme reduces calibration bias and improves model reliability.

The method successfully models Earth's magnetic field from 2008 to 2018.

Abstract

Models of the geomagnetic field rely on magnetic data of high spatial and temporal resolution. The magnetic data from low-Earth orbit satellites of dedicated magnetic survey missions such as CHAMP and Swarm play a key role in the construction of such models. Unfortunately, there are no magnetic data from such satellites after the end of CHAMP in 2010 and before the launch of Swarm in late 2013. This limits our ability to recover signals on timescales of 3 years and less during this gap period. The magnetic data from platform magnetometers carried by satellites for navigational purposes may help address this data gap provided that they are carefully calibrated. Earlier studies have demonstrated that platform magnetometer data can be calibrated using a fixed reference field model. However, this approach can lead to biased calibration parameters. An alternative has been developed in the form of a co-estimation scheme which consists of simultaneously estimating both the calibration parameters and a model of the internal geomagnetic field. Here, we develop a scheme, based on the CHAOS framework, that involves the co-estimation of a geomagnetic field model along with calibration parameters of platform magnetometers. Using our implementation, we are able to derive a geomagnetic field model from 2008 to 2018 with satellite magnetic data from CHAMP, Swarm, secular variation data from ground observatories, and platform magnetometer data from CryoSat-2 and GRACE. Through experiments, we explore correlations between the estimates of the geomagnetic field and the calibration parameters, and suggest how these may be avoided. We find that platform magnetometer data provide additional information on the secular acceleration, especially in the Pacific during the gap period. This study adds to the evidence that it is beneficial to use platform magnetometer data in geomagnetic field modeling.