When planning oil and gas exploration and production operations off the east coast of Canada, the potential for iceberg impacts must be considered. Environmental conditions in this region can be very harsh, and iceberg trajectories are notably unpredictable. When an iceberg has the potential to impact a Floating Production, Storage, and Offloading (FPSO) platform, ice management through towing will be attempted; and if this fails, the production system will be shut down, line flushed, the mooring and riser systems disconnected, and the platform moved off site.
If trajectory forecasting were highly accurate, only icebergs passing very close to the platform would require ice management and possible shutdown of the platform. Given natural variations in wind, currents, and waves, and challenges measuring and forecasting these parameters, there is considerable forecast uncertainty. This results in added expenses for extra ice management and unnecessary shutdowns.
Improvements in trajectory forecasting accuracy, characterization of forecast uncertainty, and methods to account for these uncertainties in operations would all be beneficial. This paper outlines an approach for simulating large numbers of iceberg trajectories in varied and realistic environmental conditions from hindcast met-ocean data in conjunction with a forecasting uncertainty model derived from forecast validation studies. A model, named BergCast, was developed so that proposed strategies for improving ice management operations can be evaluated, and the value of reducing forecasting uncertainty quantified.