Quantifying the Effect of Primary Field Modelling on TEMPEST Data — The Importance of Uncertainty

Abstract

The TEMPEST system is used widely for large-scale mapping. It is a fixed-wing time-domain system with the transmitter strung around the aircraft and the receiver coils towed in a bird behind and below the aircraft. The TEMPEST data are special in the sense that the data are presented as B-field, 100% duty cycle data. In this process the system self-response is removed, which means that one needs to compute and reinstate the primary field that was removed to accurately model the measured data.

In this paper we show that it is crucial to assign uncertainty to the reinstating of the primary field because it can be several orders of magnitude larger than the secondary field especially over resistive grounds and at late times. To quantify the effect of the uncertainty we have produced a number of inversions of a line in the Capricorn survey where we have added different levels of uncertainty when reinstating the primary field. The results have all been produced with the Aarhus Workbench which uses the AarhusInv algorithm and compared with results from the GA-LEI algorithm.

We show that reinstating the primary field into the forward calculations is necessary for accurate modelling of TEMPEST responses. Though, to achieve realistic and fitting inversion models (particularly over resistive ground when less signal is measured) it is crucial to allow for a small uncertainty on the primary field when this is reinstated to the forward response. This balances the inversion and allows for misfits in the range of the assumed data noise, which is not possible for the resistive areas without the assumed noise on the primary.