Drone-towed controlled-source electromagnetic (CSEM) system for near-surface geophysical prospecting: on instrument noise, temperature drift, transmission frequency, and survey set-up

Drone-borne controlled-source electromagnetic (CSEM) systems combine the mobility of airborne systems with the high subsurface resolution in ground systems. As such, drone-borne systems are beneficial at sites with poor accessibility and in areas where high resolution is needed, e.g. for archaeological or subsurface pollution investigations. However, drone-borne CSEM systems are associated with challenges, which are not observed to the same degree in airborne or ground surveys. In this paper, we explore some of these challenges based on an example of a new drone-towed CSEM system. The system deploys a multi-frequency broadband electromagnetic sensor (GEM-2 uncrewed aerial vehicle, UAV), which is towed 6 m below a drone in a towing-bird configuration together with a NovAtel GNSS-IMU (global navigation satellite system-inertial measurement unit) unit, enabling centimetre-level position precision and orientation. The results of a number of controlled tests of the system are presented together with data from an initial survey at Falster (Denmark), including temperature drift, altitude vs. signal, survey mode signal dependency, and the effect of frequency choice on noise. The test results reveal the most critical issues for our system and issues that are likely encountered in similar drone-towed CSEM set-ups. We find that small altitude variations (± 0.5 m) along our flight paths drastically change the signal, and a local height vs. signal correlation is needed to correct near-surface drone-towed CSEM data. The highest measured impact was -46.2 ppmcm-1 for a transmission frequency of 91 kHz. We also observe a significant increase in the standard deviation of the noise level up to 500 % when going from one transmission frequency to five. We recommend not to use more than three transmission frequencies, and the lowest transmission frequencies should be as high as the application allows it. Finally, we find a strong temperature dependency (up to 32.2 ppm°C-1), which is not accounted for in the instrumentation.