Economics Research Seminar

"From Melting Point to Tipping Point: Sea Ice, Shipping, and Emissions in the Arctic"

When HMS Terror and HMS Erebus sailed into the Northwest Passage in 1845, both carried hulls believed capable of withstanding Arctic ice, and both were lost. By the time their wrecks were found in 2014 and 2016, a commercial bulk carrier had already transited the passage. Retreating sea ice is opening the Arctic to commercial shipping, with consequences for emissions and fragile marine ecosystems. Yet how vessels respond to ice, and hence how much traffic future ice loss will bring, remains poorly understood. We construct a comprehensive dataset of all recorded vessel movements within the International Maritime Organization Polar Code area, covering 694,507 voyages across 18 Large Marine Ecosystems from 2012 to 2025. To test whether shipping responds to ice at a threshold, we develop a continuous-kink estimator for unknown thresholds in two-way fixed-effects Poisson panels, with a bootstrap sup-LR test. Vessel traffic responds non-smoothly: it is largely insensitive to ice below 46% concentration and falls sharply above it, a threshold higher than standard assumed navigability rules. Trip-level data show that the response operates through vessel entry rather than in-voyage behaviour. Unlike vessel counts, emissions exhibit no distinct threshold and rise proportionally less than traffic as ice retreats. Under CMIP6 scenarios, ice retreat could raise circumpolar traffic by 5-25% by 2100, concentrated in the Canadian High Arctic, among tankers and cargo vessels, and in shoulder and winter months.