The area of the Arctic covered by sea ice, or frozen ocean water, is decreasing due to climate change. This reduced sea ice cover has increased commercial shipping and tourism in the Arctic. Navigation routes typically avoid sea-ice-covered regions, since they are slower and more expensive to traverse. As such, better predictions of where the sea ice edge is located will enable safer and more cost-effective navigation of the Arctic. In this talk, I present a probabilistic model for the sea ice edge contour. Unlike existing models, I directly represent the contour, treating it as a sequence of connected points. Using a Bayesian framework, I estimate a multivariate distribution for how points on the contour move jointly. Physical constraints such as land boundaries are also considered. I compare the performance of contour models estimated entirely from observations with hybrid models that also make use of bias-corrected output from physics-based climate models.