Ice islands, massive tabular icebergs, are known to fracture as they drift. The footloose mechanism occurs when a large protuberance, known as a ram, develops along the submerged edge of the ice island and induces a buoyancy-driven bending stress. This study investigates the relationship between rams and footloose fracture using finite element models of ice islands with simulated underwater rams. Geospatial polygons of ice islands, derived from remote sensing imagery, were used to create three-dimensional shapes of ice islands at two thicknesses and with various ram sizes. Then, the location of maximum stress and fractures were predicted using finite element analysis (FEA) and the results were compared to remote sensing observations of the actual fractured pieces that calved from each of the 26 modelled ice islands. Accurate simulations of calving were achieved when a synthesized ram was placed along the ice island edge where the calving was observed. An empirical model was developed to predict the magnitude of stress from various ram sizes and shapes. The predictive ability of this empirical model suggests that ice island calving models can be improved and combined with drift forecasting models to help mitigate risks to offshore infrastructure and seafaring vessels.