Low signal-to-noise FDEM in-phase data: Practical potential for magnetic susceptibility modelling

In this paper, we consider the use of land-based frequency-domain electromagnetics (FDEM) for magnetic susceptibility modelling. FDEM data comprises both out-of-phase and in-phase components, which can be related to the electrical conductivity and magnetic susceptibility of the subsurface. Though applying the FDEM method to obtain information on the subsurface conductivity is well established in various domains (e.g. through the low induction number approximation of subsurface apparent conductivity), the potential for susceptibility mapping is often overlooked. Especially given a subsurface with a low magnetite and maghemite content (e.g. most sedimentary environments), it is generally assumed that susceptibility is negligible. Nonetheless, the heterogeneity of the near surface and the impact of anthropogenic disturbances on the soil can cause sufficient variation in susceptibility for it to be detectable in a repeatable way. Unfortunately, it can be challenging to study the potential for susceptibility mapping due to systematic errors, an often poor low signal-to-noise ratio, and the intricacy of correlating in-phase responses with subsurface susceptibility and conductivity. Alongside use of an accurate forward model - accounting for out-of-phase/in-phase coupling - any attempt at relating the in-phase response with subsurface susceptibility requires overcoming instrument-specific limitations that burden the real-world application of FDEM susceptibility mapping. Firstly, the often erratic and drift-sensitive nature of in-phase responses calls for relative data levelling. In addition, a correction for absolute levelling offsets may be equally necessary: ancillary (subsurface) susceptibility data can be used to assess the importance of absolute in-phase calibration though hereby accurate in-situ data is required. To allow assessing the (importance of) in-phase calibration alongside the potential of FDEM data for susceptibility modelling, we consider an experimental test case whereby the in-phase responses of a multi-receiver FDEM instrument are calibrated through downhole susceptibility data. Our results show that, while it is possible to derive approximate susceptibility profiles from FDEM data, robust quantitative analysis hinges on appropriate calibration of the responses.