Research report: Trajectory Modelling

This project involved new model developments and validation techniques to improve sea ice modelling in view of oil spill trajectory forecasting. Three work packages were developed, focusing on model development; evaluation and validation; and uncertainties. Data sets of ocean and ice model results for inclusion in oil spill models were produced.

In this project, we have worked on what could become the foundations of a comprehensive oil spill trajectory prediction system. For short-term predictions (about a week), one could use ensemble forecast systems, such as the TOPAZ system to predict the movement of the ice. Here our work has shown that using the EB model would give substantially better results than current ice models. Such prediction models will never be able to accurately predict the ice motion on mid-term time scales, but using the mean flow and diffusion/dispersion properties from past model runs one can create a probabilistic forecast covering up to seasonal time scales. This would provide ice drift information for oil spill models on both short and long time scales. The results of this project show that such a system is feasible, even if some technical challenges remain.