Mid-August ice: what an extent figure does not tell you
Arctic sea ice extent stood at 5.33 million km² on 16 August, ninth lowest in the satellite record. That number is a poor guide to whether any particular route is passable.
Arctic-wide sea ice extent had fallen to 5.334 million km² by 16 August 2026, ranking ninth lowest in the record that begins in 1979. Extent has come down from 12.036 million km² on 19 May. The annual minimum is still roughly four weeks away.
Two details matter more than the headline ranking. First, the decline slowed markedly from late July into early August — regional slow-downs are common, an Arctic-wide pause less so. Second, modelled thickness suggests thicker ice extending further north from the Canadian Arctic than at the same point in 2025, and multi-year ice in the southern Beaufort sitting closer to the North American coast than last year. Total volume, at or near record low earlier in the year, is now appreciably above 2025 and effectively tied with 2024 for second lowest.
The NSIDC's Sea Ice Today analysis site resumed monthly science assessments on 19 August after a period without updates, which restores a reference point that route planners had been working without.
Operational perspective
Extent is a climate metric. It answers the question "how much ocean has at least 15 per cent ice cover" — and that threshold is the problem for anyone planning a passage. A cell counted as ice-covered may be 20 per cent cover with wide leads, or 95 per cent close pack. Both appear identically in an extent total, and they represent entirely different days on the bridge.
Volume and thickness get closer to the operational question, because they carry information about what the ice would do to a hull. But they are modelled fields, not observations, and their resolution is coarse against the scale at which a master decides whether to enter a lead. The southern Beaufort detail above is a good example: "multi-year ice closer to the coast than last year" is directionally useful for a route decision made in June and near-useless for a track decision made at the ice edge in September.
What that means in practice is a hierarchy. Extent and volume set expectations for the season and inform whether a voyage is worth planning at all. National ice service charts — concentration, stage of development, floe size — set the route. Operational products, including radar imagery where available, set the track. Substituting one level for another is the most common error in Arctic voyage preparation, and it usually shows up as a plan that looked comfortable in April and does not survive contact with the ice edge.
A late minimum with above-recent-average volume is a combination worth noting for anyone timing an eastbound passage into October: the window may close on thickness before it closes on extent.