Verification and validation of an in-ice oil spill trajectory model based on satellite-derived ice drift data

A future increase in hydrocarbon exploration and development activities driven by the probable existence of hydrocarbon reserves and an expected increase in shipping activities due to less severe ice conditions, pose a risk of potential oil spills in the offshore Arctic. Estimating oil spill trajectories is essential in quantifying risks and planning an effective spill response. An in-ice spill trajectory modelling, analysis and visualization tool suitable for spills in highly ice-infested waters has been previously developed at NRC. The source data is historical satellite-derived ice drift. The model has been enhanced by including time dependent land-fast ice extent to better estimate coastal spill trajectories. Two hypothetical in-ice spill scenarios in the Canadian Beaufort Sea were modelled based on 34 years of ice velocity data. In four months starting in November, a deep water spill in ice could travel over 700 km, while for a shallow water spill in ice, the travel distance could exceed 400 km. Depending on how fast an in-ice spill could be cleaned, both investigated deep water and shallow water spills could be an international issue, particularly the deep water spill scenario. Present model results were compared with an observed in-ice spill trajectory in the Barents Sea. Because of an underestimation of ice speeds in the input satellite-derived ice drift dataset, the present model underestimates the extent of the trajectory. However, the model estimated the trajectory of an observed buoy well. Present model results were also compared with an independent numerical study of oil spills in the Beaufort Sea. Coastward motions of an in-ice spill are found to be generally similar, however, along the coast, motions deviate after a certain time in the modelled period. Both models are based on data that are expected to be less accurate in the nearshore zone. We did not investigate what caused this deviation or whether the present model or the independent study is a better representation of reality.Peer reviewed: YesNRC publication: Ye