RESUMEN
The decline of Arctic sea ice extent is one of the most spectacular signatures of global warming, and studies converge to show that this decline has been accelerating over the last four decades, with a rate that is not reproduced by climate models. To improve these models, relying on comprehensive and accurate field data is essential. While sea ice extent and concentration are accurately monitored from microwave imagery, an accurate measure of its thickness is still lacking. Moreover, measuring observables related to the mechanical behavior of the ice (such as Young's modulus, Poisson's ratio, etc.) could provide better insights in the understanding of sea ice decline, by completing current knowledge so far acquired mostly from radar and sonar data. This paper aims at demonstrating on the laboratory scale that these can all be estimated simultaneously by measuring seismic waves guided in the ice layer. The experiment consisted of leaving a water tank in a cold room in order to grow an ice layer at its surface. While its thickness was increasing, ultrasonic guided waves were generated with a piezoelectric source, and measurements were subsequently inverted to infer the thickness and mechanical properties of the ice with very good accuracy.
RESUMEN
The propagation of elastic waves in polycrystals is revisited, with an emphasis on configurations relevant to the study of ice. Randomly oriented hexagonal single crystals are considered with specific, non-uniform, probability distributions for their major axis. Three typical textures or fabrics (i.e. preferred grain orientations) are studied in detail: one cluster fabric and two girdle fabrics, as found in ice recovered from deep ice cores. After computing the averaged elasticity tensor for the considered textures, wave propagation is studied using a wave equation with elastic constants c=ãcã+δc that are equal to an average plus deviations, presumed small, from that average. This allows for the use of the Voigt average in the wave equation, and velocities are obtained solving the appropriate Christoffel equation. The velocity for vertical propagation, as appropriate to interpret sonic logging measurements, is analysed in more details. Our formulae are shown to be accurate at the 0.5% level and they provide a rationale for previous empirical fits to wave propagation velocities with a quantitative agreement at the 0.07-0.7% level. We conclude that, within the formalism presented here, it is appropriate to use, with confidence, velocity measurements to characterize ice fabrics.