Joint application of ground-based transient electromagnetics and airborne electromagnetics

Abstract

The focus of this thesis lies on the joint application of ground-based and airborne electromagnetic methods for the investigation of a glacial valley. For the first time two different airborne electromagnetic (AEM) surveying methods were employed to determine the resistivity structure of a single geological target: the frequency-domain helicopter-borne electromagnetic (HEM) system operated by the Federal Institute for Geosciences and Natural Resources (BGR), Germany, and the time-domain SkyTEM system developed at the University of Aarhus, Denmark. For verification of the airborne results, ground-based transient electromagnetics (TEM) and 2D resistivity surveying were also performed. The target survey area was the Cuxhaven valley in northern Germany, a significant local groundwater reservoir. The course of this buried valley was revealed by drillings, and the shape determined by reflection seismics along several transects across the valley. Electrical and electromagnetic methods were applied to investigate the structure of the valley fill, consisting of gravel, sand, silt and clay. Here, the extension and the thickness of clay layers are of particular interest. They have a low hydraulic permeability and often serve as protection for underlying aquifers against pollution from the surface. The standard tools for presenting AEM data are apparent-resistivity maps and resistivity-depth sections. Although the large and dense data sets are favorable for 3D interpretation, it is still not common to perform 3D inversion in AEM, as the effort in terms of computing time is too high. Therefore, 1D inversion models are still used to display 3D resistivity distributions by stitching together the 1D layered inversion models. Besides the 3D inversion the combination of different data sets in one inversion scheme is an ongoing research issue. One approach is the classical joint inversion, which results in one resistivity model at the shared survey sites, whereas each site is regarded as individual. In this thesis I follow a different approach: spatially constrained inversion (SCI). SCI is a technique where different data sets are combined in one inversion scheme and spatial constraints are applied to the resistivity structure revealed at adjacent survey sites. Thus, the method is particularly useful for large data sets as obtained in AEM. Using spatial constraints, information can be propagated horizontally to adjacent models. With this technique it is then possible to resolve layers which are locally poorly resolved. SCI was originally developed at University of Aarhus for SkyTEM data. In this thesis I adapt the technique for the use on HEM data and apply it to both, SkyTEM and HEM data of the Cuxhaven valley using a priori information from geology, drilling, and seismics. Systematic studies of the SCI parameters show that a) the inversion results applying SCI are less dependent on the starting model in comparison to single-site inversion, b) HEM data resolve the base of a conductive layer which can be identified as the Lauenburg clay, and c) SkyTEM data reveal the base of the Cuxhaven valley which is also confirmed by a seismic section and high-moment TEM measurements. The influence of the valley geometry on the 1D inversion results were systematically studied by 3D forward modeling of different slope geometries. 1D inversions of the synthetic data across the slope were performed with and without constraints between neighboring sites. The resulting 1D models are affected by the slope and simulate a valley base at shallower depth than in the 3D model. Using constraints again decreases the dependency on the starting model. This thesis demonstrates that 1.) by the joint application of ground-based and airborne electromagnetic methods, 2.) by the application of the SCI including a priori information and 3.) by explaining 2D effects of valley slopes using a 3D forward code, a better understanding of the structure of the Cuxhaven valley is obtained.