Frequency-domain electromagnetic induction (EMI) is routinely used to detect the presence of seawater due to the inherent electrical conductivity of the seawater. This approach is used to infer sea-ice thickness (SIT). A time-domain EMI sensor is presented, which demonstrates the potential for correlating the spectroscopic properties of the received signal with the distance to the sea surface. This is a novel approach to SIT measurement, which differs from existing methods in that it uses measurements from 10 kHz to 93 kHz rather than a single frequency. The sensor was tested at a tidal pool containing seawater and measured to have a conductivity of 57.3 mS/cm. Measurements were performed at a range of heights between 0.2 m and 1.9 m above the sea surface and for inclinations from 0° to 45°. These measurements were correlated with Finite Element Modeling (FEM) simulations performed in COMSOL. The measured and simulated datasets are presented along with a proposed form of post-processing, which establishes a correlation between the distance to the sea surface and the measured EMI response. This forms a proxy measurement for the absolute distance from the EMI sensor to the sea surface. Because the air gap between the sensor and the seawater is indicative of the properties of sea ice, this study demonstrates a novel approach to non-destructive measurement of sea-ice thickness. The measurements show that this distance to the sea surface can be estimated to within approximately 10% of the known value from 0.2–1.5 m and 15% from 1.5 to 1.9 m.

One of the key challenges with developing pulsed induction (PI) electromagnetic induction (EMI) sensors for use in the Arctic is the inaccessibility of the environment, which makes in situ testing prohibitively expensive. To mitigate this, sensor development can be streamlined through the creation of a robust simulation strategy with which to optimize features such as coil turns and geometry. Building on work that previously presented a method for simulating an Arctic PI sensor via a time-domain finite element model (FEM), this paper presents a method for approximating a time-domain simulation with multiple frequency-domain simulations. A comparison between the fast Fourier transform (FFT) of a time-domain simulation and a collection of frequency-domain simulations is presented. These are validated against empirical data with a PI sensor over seawater, with an air gap used as a proxy for sea ice. Using the method described, a range of coils is simulated with dimensions from $0.5\times 0.5$ m up to $1.0\times 2.0$ m, demonstrating the ability of this approach to enable comparison of sensor performance over a wider parameter space. For a parametric sweep over 10 sensor-to-seawater lift-off distances, the improvement from the time-domain simulation (of a 402 $\mathsf{\mu }$s window) to the frequency-domain simulation (comprising 100 discrete frequencies) represents a reduction in simulation time from 38,013 min to 141 min.