Melting of the submerged part of an iceberg, its keel, releases freshwater below the ocean surface, influencing water circulation, mixing, and local seawater conditions. Understanding how quickly the keel melts requires relating ice loss to seawater temperature and salinity. In her 2026 University of Manitoba master’s thesis, A Thermodynamic Rate of Ablation for Iceberg Keels, E. A. Marie combined laboratory experiments and published measurements to develop a model of ice ablation (loss of ice at the surface) as a function of these variables. She then applied it to a model iceberg using estimates of glacier density and trapped-air pressure, together with ocean conditions from a reanalysis product.
Within that broader work, finite element stress analysis of ice helped examine whether the experimental arrangement itself could influence measured melting. Marie used FEATool Multiphysics in MATLAB to perform finite element analysis (FEA) of a lead-core ice ball, comparing the calculated stress patterns with published melt shapes. The results supported her interpretation that mechanical loading could affect the observed pattern, helping assess the suitability of experiments used to investigate thermodynamic ablation.
In the reasearch, MATLAB ice stress analysis examined the effect of two opposing forces: gravity draws the lead core downward, while buoyancy pushes the surrounding ice upward. The resulting stress field includes compression where the bottom of the core meets the ice. FEATool was used to compute horizontal and vertical normal stresses, showing where the material was in tension or compression in each direction. These directional results gave Marie a way to compare internal loading with the locations where ice had disappeared. In Figure 2.10 of the thesis, reproduced above, the central circle represents the lead core and the surrounding colored region shows the stress distribution in the ice.
For comparison with the Vanier and Tien ice sphere melting experiment (1970), Marie used MATLAB’s Image Processing Toolbox to trace the original ice-ball outline from their published figure and scale it onto the stress plots. She also used ellipses to trace the reported remnant shape and the concavity that consistently developed at the bottom. Overlaying these shapes made it possible to inspect whether particular melt features coincided with particular stress regions. The outlines therefore represent observations from the earlier experiment, while the colored fields represent the FEATool calculation. Their combination provides a spatial comparison between the experimental geometry and the proposed mechanical explanation.
The comparison suggested a relationship between tensile stress and ice ablation, with the two stress components showing different spatial patterns. In the horizontal stress plot, the bottom concavity coincided with the highest tensile normal stresses. In the vertical stress plot, regions of greater ablation coincided with both high tensile and high compressive normal stresses, while the sides of the ball showed little or no ablation. Marie reported a similar overall melting pattern in her own experiments, but without the bottom concavity. Her ice balls were held underwater using a fine net rather than an internal lead core. She interpreted this difference, together with the stress maps, as evidence that the weighting method affected the result, and suggested that spatial differences in tensile stress were especially relevant.
By highlighting possible mechanical stress effects on ice melting, the analysis helped Marie assess the suitability of the experimental method. It offered a mechanical explanation for a feature that could complicate interpretation of temperature- and salinity-dependent melting measurements, and supported Marie’s preference for the net-based arrangement. The comparison remains qualitative: the thesis does not report measured stresses, a quantitative error assessment for the stress model, or a coupled FEATool calculation predicting the evolving melt boundary. Marie explicitly calls for further research on the relationship between stress and melting. The value of the analysis is in identifying and examining a plausible influence on the experiment, with the published ice shapes providing observational context for that interpretation.
Related FEATool structural-mechanics workflows include the Stress Analysis of a Thick Plate benchmark, which evaluates directional stresses under prescribed loading, and the Deformation of a Spanner tutorial, which applies structural loading to an imported geometry. Although they do not reproduce the ice-ball experiment, they illustrate the same general sequence of defining structural loading, solving the mechanical field, and inspecting stress or deformation results.
References
- Elise Athena Marie. A Thermodynamic Rate of Ablation for Iceberg Keels, MSc thesis, Department of Environment and Geography, University of Manitoba, 2026.