A helicopter transient electromagnetic (TEM) survey was flown over shallow coastal waters in Broken Bay, Australia, overlying several palaeovalleys and exposed reef sections. The infilled palaeovalleys contain unconsolidated sediments with variable thicknesses exceeding 100 m. Previous marine seismic reflection and vibrocore studies provide an estimate of sediment thickness and surficial sediment conductivity, respectively, which, combined with known bathymetry (sonar soundings) and measured seawater conductivity, can be used to generate a crude 1D geoelectric ground-truth model consisting of two layers (seawater/sediment) overlying a relatively resistive basement. The primary focus was to examine the accuracy of interpreted water depths obtained by 1D inversion of airborne TEM data, assuming a two-layer over resistive basement model, by comparing these depths with known water depths. The secondary focus was to compare the interpreted sediment thickness (i.e., the thickness of the second geoelectric layer, which combined with bathymetry, gives the bedrock depth) with thicknesses estimated from marine seismic data to test the potential of using airborne electromagnetic systems for remote sensing of the coarse features of the bedrock topography. Interpreted water depths obtained from TEM data resulted in absolute water depth accuracies of 1–2 m for depths between 10 and 30 m, and 0.3–0.5 m in water shallower than [Formula: see text]. More importantly, similar water depth accuracies were obtained using raw TEM data (with birdswing removal) and TEM data obtained by postprocessing using time-consuming empirical corrections based on the TEM half-space response over deep sea water. The interpreted sediment depths derived from TEM and marine seismic data showed good agreement, generally, for example, inversion of TEM data delineated a distinct palaeovalley that transects a beach, with a maximum depth of 60–70 m below the seafloor, in agreement with depths estimated from marine seismic data to within [Formula: see text].