An iceberg drift model was developed for offshore oil and gas operations. Increasing forecasting accuracy is critical for making appropriate downtime decisions. Three sources of ocean current data were used to run test simulations of the model: estimated from observed iceberg drift and wind, measured from a drifting buoy, and a 3D ocean model. The average simulated iceberg position error at 24 hours was 12% smaller for the estimated current drift simulations compared with the 3D modelled current simulations. The 17-hour iceberg position error was 21% smaller for the buoy current drift simulation compared with the 3D modelled current simulation.


Iceberg drift prediction is an important component of ice management decision support for offshore oil and gas exploration and drilling operations. Improving the accuracy of iceberg drift trajectory forecasts can help to increase ice management efficiency in terms of allocation of ice management vessels, prioritization of tow targets when multiple icebergs are present, and ensuring operational shutdown decisions are appropriate. Inaccurate iceberg drift trajectory predictions can lead to unnecessary downtime and result in financial losses on the order of tens of millions of dollars. The Grand Banks offshore Newfoundland and Labrador, Canada, for example, have historically been the site of significant oil and gas exploration and drilling activities, with production, construction and exploration facilities threatened during iceberg incursion events occurring on a near-annual basis, typically during the months of March to June.

Previous models have typically operated on temporal integration of the momentum balance of the primary met-ocean forcings on iceberg drift: wind, current, and Coriolis. In most of the existing literature on iceberg drift modelling, iceberg drift simulations were run in a hindcast mode. Some of these models used wind and ocean current data from regional atmospheric and ocean prediction models, respectively, while others used site-specific measured wind and ocean current data. Some of the earliest iceberg drift modelling was performed offshore Labrador in the 1980s using locally measured winds and currents (Sodhi and El-Tahan, 1980; Smith and Banke, 1983; and Smith and Donaldson, 1987). Subsequent work used atmospheric and ocean circulation models to run iceberg drift simulations (Carrieres et al., 2001; Lichey and Hellmer, 2001; Kubat et al., 2005; and Eik, 2009).

This content is only available via PDF.
You can access this article if you purchase or spend a download.