This thesis addresses the application of electromagnetic exploration techniques to reduce uncertainties in the subsurface, particularly in low-lying coastal and delta areas like the Netherlands that are increasingly vulnerable to flooding due to climate change. In response to climate change, countries must reinforce their water defenses, a challenging and costly effort that necessitates efficient resource allocation. The research focuses on magnetic-dipole electromagnetic induction (EMI) and direct-current electrical resistivity tomography (ERT) due to their sensitivity to the electrical resistivity of the subsurface, which correlates with other geotechnical properties. Both methods are easily deployable and can cover large areas relatively quickly. Chapter 2 delves into the theoretical aspects of EM data acquisition in dikes, demonstrating that EMI devices with far-offset receivers can capture large anomalies at a much faster rate than ERT. However, both methods perform poorly in detecting small, detrimental features such as thin layers, and are affected by groundwater salinity. Chapter 3 proposes a method to estimate the geometric variability of soil layers using geophysical tomograms. EMI and ERT tomograms are employed to estimate the orientations of soil layers, enabling an accurate estimation of geometric variability with reduced exploration effort. Chapter 4 highlights the value of high-resolution ERT in estimating the spatial variability of properties within homogeneous soil units. This method serves as an efficient alternative for estimating internal variability in geotechnical analyses of water defense structures and other geotechnical infrastructure. An essential contribution of this thesis is the proposal of quantitative and reproducible methods for characterizing subsurface heterogeneity in the context of water defenses. These insights help reduce uncertainties and optimize resource allocation for dike reinforcement. The integration of geophysical methods with other geotechnical site data enhances the understanding of the subsurface. Chapter 5 summarizes the main findings of this research regarding the geotechnical schematization of dikes.

The geometric variability of soil layers is a large source of uncertainty in the reliability assessment of dikes. Because direct samples of the subsurface soils are often insufficient to capture the complexity of the subsurface, geophysical methods provide a powerful source of complementary information. A combined approach to estimate the geometry of soil layers is presented. The approach combines local point data, i.e. data obtained from a CPT or a borehole log, and geophysical tomography in a universal cokriging framework. The approach uses the contact points between soil layers obtained from local point data and the orientations of the layers derived from geophysical tomography. To reduce subjectivity in the interpretation of tomographic images, an automated edge detection technique was used. The combined approach was applied to characterise two test sites where the presence of paleochannels locally change the geometry of soil layers. The results show that a combined approach enables the reduction of sampling efforts with an improved estimation of geometric variability.