1,000 Predictions: What’s New and What’s Old in a Retrospective Analysis of the Sea Ice Outlook, 2008–2020

Abstract:

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. Then, each September, we learn how accurate those predictions were. More than 1,000 individual predictions by researchers using many different methods have accumulated over these 13 years. Earlier papers analyzed the ensemble skill of the first few hundred SIO contributions through 2013 (Stroeve et al. 2014) or 2015 (Hamilton and Stroeve 2016). Here we bring those analyses up to date with data through 2020. Many conclusions from the earlier papers have proven to be robust, but new insights emerge also — about why ice extent is so hard to predict, and about the prediction methods themselves. For example, median SIO predictions correlate not with same-year extent, which they aim to predict, but rather with extent in the previous year. Moreover, this lagged correlation has been stronger for dynamic model predictions than for statistical or heuristic methods (Figure 1). We conclude with a brief look at a very different realm: what the US public knows or believes about these Arctic changes.

References

Hamilton, L.C. & J. Stroeve. 2016. “400 predictions: The SEARCH Sea Ice Outlook 2008–2015.” Polar Geography 39(4):274–287. doi: 10.1080/1088937X.2016.1234518

Stroeve, J., L.C. Hamilton, C.M. Bitz & E. Blanchard-Wrigglesworth. 2014. “Predicting September sea ice: Ensemble skill of the SEARCH Sea Ice Outlook 2008–2013.” Geophysical Research Letters 41:2411–2418. doi: 10.1002/2014GL059388