Modelling iceberg tracks around Antarctica

The drift tracks of five large tabular icebergs drifting around Antarctica were simulated using a dynamic model similar to that used by Lichey and Hellmer (2001) with the exception that sea ice strength was not considered as a condition of iceberg trapping. Three of these iceberg tracks (737n165, 74n131 and BlOB) drifted in the Ross Sea for periods of time between 1995 and 2000. Iceberg BO9A drifted along the coast of East Antarctica along the Antarctic coastal current and entered the Weddell Sea. Iceberg BlOA drifted along the West Antarctic region initially moving west in the Antarctic coastal current then moved north near 130°W before entering the Antarctic Circumpolar current and moving east through Drake Passage. The input fields of wind, ocean and sea ice velocity, and sea ice concentration are used in the model to simulate the drift of these icebergs. The iceberg sizes range from 200 km\\(^2\\) to 2070 km\\(^2\\), with drift tracks lasting between 1 and 8 years. Only periods that indicated movement of the iceberg from one composite image to the next were simulated, providing a more focussed model track. Various model simulations are conducted using ocean velocity fields at either 103 m or 238 m depths and three variations of the iceberg trapping criteria dependent on sea ice concentration. Two of these variations are based on the sea ice concentration at the iceberg position (using 85% or 90% sea ice concentration as the trapping criteria) and the other excludes iceberg trapping altogether. The assumption of a trapped iceberg instantly moving at the velocity of the sea ice, regardless of previous iceberg velocity, returns poor results and at times moves the iceberg in the opposite direction. If the iceberg is released from the sea ice, all the forces in the model become active and large amplitude inertial oscillations are initiated in the drift track. An analysis of the force contribution to the iceberg velocity is conducted and it was found that the Coriolis and sea surface slope are close to opposing forces, keeping the iceberg within the ocean current and the air drag can influence the iceberg drift when sufficiently strong enough. An RMS analysis between the final position of the model iceberg and the position of the observed drift track for the same date are assessed for each model track. The model simulation ocean currents near the base of the iceberg (238 m) and no iceberg trapping attain the least rms error in position between the model and the observed iceberg track for all but two icebergs, B1OA and BO9A. For these icebergs, the model simulations providing the least error were 85% iceberg trapping criteria and 238 m ocean currents for both icebergs. This result shows that the sea ice movement is well represented in East and West Antarctica, and not so well in the Ross Sea, as model simulations performed better with no iceberg trapping. These results show that trapping an iceberg and instantly moving that iceberg with the sea ice floe doesn't fully necessarily capture the drift of large tabular icebergs. The iceberg model is strongly dependent on the data representing the ocean currents and sea ice movement to sufficiently model the drift of large tabular icebergs.