Multi-frequency seismo-acoustic measurements of synthetic ice with fluid inclusions: implications for seismic and through-ice acoustic estimation of sea-ice thickness

Abstract. We conducted laboratory measurements of synthetic ice cores designed to mimic natural ice with variable fluid-inclusion content, primarily air inclusions and possible unfrozen water, using two frequency bands: P-wave velocity and attenuation at 4–10 kHz using an acoustic pulse tube, and P- and S-wave velocities and attenuation at ~300 Hz using a resonant column. Four synthetic ice cores of ~69 mm diameter were produced by directional freezing from above with water circulation using reverse osmosis water, providing control over inclusion concentrations. We classified the cores based on densities above and below 900 kg m⁻³. Two higher-density cores (>900 kg m⁻³) showed P-wave velocities of 3100 to 3500 m s⁻¹ at 4 to 10 kHz and ~3400 m s⁻¹ at ~300 Hz, with P-wave attenuation expressed as the inverse quality factor Qp⁻¹ generally below 0.1. Two lower-density cores (852 to 886 kg m⁻³) showed lower P-wave velocities of 2500 to 2800 m s⁻¹ at 4 to 10 kHz and approximately 3100 m s⁻¹ at ~300 Hz, with more variable attenuation and Qp⁻¹ values reaching up to ~0.2. S-wave velocities at ~300 Hz were approximately 1250 m s⁻¹ for the higher-density core and 1050 to 1150 m s⁻¹ for the lower-density core, with Qs⁻¹ ranging from 0.13 to 0.25 and from 0.23 to 0.31, respectively. These results indicate that lower bulk density, used here as a proxy for higher air inclusion concentration, is associated with reduced P and S-wave velocities and generally higher attenuation. The S-wave properties of the lower-density core were sensitive to applied strain during torsional vibration, with decreasing shear-wave velocity and increasing attenuation at higher applied strains. This suggests that lower-density ice cores has a more compliant and dissipative microstructure than higher-density ice cores. We adapted an effective medium sediment model originally developed for hydrate bearing sediment to ice, enabling joint interpretation of the pulse tube and resonant column data. The model captures the main trends in velocity and attenuation across both frequency bands and shows that permeability and gas bubble size strongly influence the elastic-wave properties. This dataset provides a reference for interpreting seismic and through-ice acoustic measurements of sea-ice thickness, in which part or all of the measured wavefield propagates within the ice. Our results suggest that air inclusions and possible unfrozen water can cause velocity variations of approximately 20–25 % in our measured frequency range of ~300 Hz and 4–10 kHz. If unaccounted for, these variations may lead to substantial errors in sea ice thickness estimates from geophysical velocity measurements derived from seismic or through-ice acoustic travel times. For example, for a 1 m thick ice layer, interpreting the same travel time using the measured end-member acoustic velocities would give apparent thicknesses of approximately 0.80–1.25 m, corresponding to potential errors of about 0.20–0.25 m.