Simulation of the recurrence probability of ice islands in the Arctic Ocean

Abstract

Ice islands, the most extreme ice features in the Arctic Ocean, are hazards to offshore structures. To determine the probability of ice island trajectories in coastal areas of the Arctic Ocean, a random simulation model has been established. The model consists of the random ice island generation, the ice island dynamic model, and the Monte Carlo model for random geostrophic wind generation. Based on statistics of observation data, the generation location was assumed as uniformly distributed along the northern side of Ellesmere Islands. A 4 year interval of generation event is considered in the simulation. The number of new ice islands calved from ice shelves in one calving event is automatically produced by deducting each ice island area from the random area of ice shelf calved in one time. As a driving force source, geostrophic wind field was calculated from monthly-averaged pressure charts. As an significant force, the pack ice force was formulated by theoretical analysis combined with an existing empirical formula for each case. The results of probabilities of simulated ice island trajectories show that there are two zones of highest recurrence of ice islands, one near the Canadian Beaufort Sea, and another near the Chukchi Sea. There is a broad area of 1 to 10 year recurrence interval in the central ocean, and a high probability zone near the north end of Greenland. The simulation also yielded the frequencies of ice island ejection, the lifetime of ice islands, and the number of live ice islands in the Arctic Ocean. Two basic drift patterns of ice islands have been displayed by the simulation: short drift patterns in which ice islands move directly out of the ocean after generation, and the large-scale circulation pattern in which ice islands circulate around the Beaufort Sea from one to four times.