Evaluating the importance of footloose-type failure in ice island deterioration modeling

Abstract

The drift and deterioration of large and tabular icebergs, also known as “ice islands” in the Arctic, are modeled for both operational (e.g., offshore risk mitigation) and research (e.g., oceanographic impact of melt water input) purposes. In this paper, we build a theoretical argument to show that the lateral deterioration of ice islands is controlled by the rate of sidewall notch growth at the waterline, with this growth leading to the development of underwater rams and buoyancy-induced calving via the ‘footloose’ mechanism. This dominance of footloose-type lateral deterioration allows for the majority of ice island deterioration to be simulated with only three oceanic variables: wave height, wave period, and sea-surface temperature. Information regarding the size and lineage of ice islands tracked in the Canadian Ice Island Drift, Deterioration and Detection (CI2D3) Database provides opportunity to assess our theoretical work, as the database serves as a validation dataset for simulations of ice island length and area change. When simulating the length reduction over time of ice islands tracked in the CI2D3 Database, the footloose model reduced the mean error over 80 d to +277 m, compared to −1545 and −1403 m with no-melt and thermal-melt models, respectively. We also demonstrate a new approach to simulating the areal deterioration of ice islands resulting from discrete footloose calving events. The approach utilizes two parameters: the length-to-width ratio of the ice island (r) and the width of a footloose calving event relative to the ice island's length (K). With r = 1.6 and K= 0.8, the mean error in modeled area was close to zero after 20 d of simulation. A comparison of stresses associated with footloose events from a 1D-beam model and those simulated with 3D finite-element modeling showed that the 1D and 3D simulations produce broadly similar results. This supports our approaches and parameter assignments for simulating ice island length and area reduction from footloose calving. These approaches can now be incorporated into ice island deterioration models. The benefit of this incorporation will be greatest for those interested in research of longer-term impacts of ice island deterioration on ocean properties given the greater improvements to model error over periods of time that are longer than those that usually concern offshore ice management operations.