Arctic Shipping: AI Maps Ice to Reduce Risks in Bering Strait

Researchers developed a system using satellite radar data to classify sea ice conditions in the Bering Strait, a key passage for Arctic shipping between the Republic of Korea and the Arctic Ocean. They used data from the Sentinel-1 satellite combined with a Support Vector Machine to identify open water, thin ice, and first-year ice with 97% accuracy.

The team then integrated this ice classification with the POLARIS risk standard, a widely used framework for assessing navigational hazards. This allowed them to map safe routes for different types of ships, from standard merchant vessels to icebreakers, considering ice thickness, water depth, and safety buffers.

The analysis showed that the feasibility of trans-Arctic shipping is highly dependent on a vessel’s ice class. While an icebreaker could navigate a 124 km path, a non-ice-class vessel and even an ice-class 1A merchant vessel faced blockages. This framework aims to improve the safety and sustainability of Arctic shipping as sea ice continues to decline.