Introduction
Ground investigations prior to large infrastructure projects are predominantly based on geotechnical drillings. Ground-based geophysical investigations are rarely used because they are often tedious and expensive. On the other hand, the investigation sites are often too small for time efficient airborne geophysical investigations. Drone based geophysical measurements could fill this gap, providing fast and efficient scanning of construction sites, allowing for low flight altitudes and high spatial sampling. Furthermore, the required depth of investigation for geotechnical applications is often limited to a few tens of meters, favouring small low power systems.
In this paper we describe the results from geophysical ground investigations conducted with a prototype for a new small airborne time-domain electromagnetic (A-TEM) system. This development is the first step towards a small TEM system that could potentially be carried by a drone. The system was field-tested at a road construction site called Åsen in Trøndelag, Central Norway in fall 2021. The purpose of the field test was to investigate whether such a significantly smaller system can have a sufficiently high dipole moment and sensitivity to conduct meaningful ground investigations. At the same site, A-TEM data with the much bigger SkyTEM304 (Sørensen 2004) was acquired in 2019, which is used for comparison and benchmarking.
Conclusion
Results presented in this paper have shown that a significantly smaller and lighter airborne TEM system than currently available commercial TEM systems, can produce useful geophysical data. The data can be interpreted and provide inside into subsurface resistivity structures and geology. In the case of geotechnical applications, where the focus is on the shallow subsurface, such a smaller system has in fact some significant advantages. The system operates with a much smaller magnetic dipole moment and state of the art transmitter and receiver electronics combined with digital data recording provide high quality early time data. Lower flight altitudes and higher spatial sampling along the flight lines further improve the sensitivity to the shallow subsurface.
Resistivity models recovered by TEM data inversion show strong similarities between the small SkyTEM21HR system and the much biggerSkyTEM304 system in the parts of the subsurface where both systems have good sensitivity. However, the resistivity models recovered from the SkyTEM21HR data provide a more detailed definition of the shallow resistivity structures, even though the depth of investigation is somewhat reduced.
Higher spatial and temporal data sampling requires more automation in the data processing and interpretation. With the results presented in this study we show that automated data processing and machine learning based data interpretation algorithms can be used to efficiently process and interpret this type of high resolution AEM data. We have successfully performed an automated bedrock topography interpretation using an artificial neuronal network. We also conducted an automated volume material classification based on random forest classifiers. Generally, the interpretation results for high resolution SkyTEM21HR data show strong similarities to the interpretation results for the SkyTEM304 data. However, there are some significant differences, especially in the shallow subsurface that can be linked to the higher resolution of theSkyTEM21HR data.
Acknowledgements
We would like to thank Eureka/Eurostars for the financial support of this research and development project (nr.E114169). We would also like to thank the collaboration partners within this project, the Norwegian Geotechnical Institute (NGI) and the Drone Centre at the University of Southern Denmark (SDU) for fruitful discussions and valuable contributions to this project. We further want to acknowledge Nye Veier, the project owner of the construction site presented here.
