Each Arctic summer since 2008, the Sea Ice Outlook (SIO) has invited researchers and the engaged public to contribute predictions regarding the September extent of Arctic sea ice. The public character of SIO, focused on a number whose true value soon becomes known, brings elements of constructive gamification and transparency to the science process. We analyze the performance of more than 400 predictions from SIO’s first eight years, testing for differences in ensemble skill across years, months and five types of method: heuristic, statistical, mixed, and ice-ocean or ice-ocean-atmosphere modeling. Results highlight a pattern of easy and difficult years, corresponding roughly to the distinction between climate and weather. Difficult years, in which most predictions are far from the observed extent, tend to have large positive or negative excursions from the overall downward trends. In contrast to these large interannual effects, ensemble improvement from June to July and August is modest. Among method types, predictions based on statistics and ice-ocean-atmosphere modeling perform better. Thinning ice that is sensitive to summer weather, complicating prediction, reflects our transitional era between a past Arctic cool enough to retain much thick, resistant multiyear ice; and a warmed future Arctic where little ice remains at summer’s end.



Publication Date


Journal Title

Polar Geography


Taylor & Francis

Digital Object Identifier (DOI)

Document Type



© 2016 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group


This is an article published by Taylor & Francis in Polar Geography in 2016, available online:

Included in

Sociology Commons