Arctic Sea Ice Loss Far More Rapid and Extensive than Projected Using Best-fit Models

Nature Communications Earth & Environment, July 15

Model-based projections of Arctic sea ice extent have traditionally lagged behind real-world observations, in part because a large portion of global climate change models do not adequately capture specific Arctic conditions; especially ocean-atmosphere-sea ice interactions and heat transport, as well as measurements of sea ice volume (not just extent, or area). This study chose a subset of the “CMIP6” models to be used with the next IPCC Assessment Report, AR6, that appear to have best captured actual observations of sea ice extent in the past few decades, and which also incorporate ocean interactions. They found that these better-performing models project complete summer sea ice loss (typically defined as under 1 million km2) once global mean temperatures pass around 1.7°C above pre-industrial. If today’s high emissions growth and low mitigation continue (the SSP5-RCP8.5 scenario), these global temperatures and ice-free conditions may regularly occur in September as early as 2035, according to the most sophisticated models. Simply combining all models, without regard to past model performance or attention to ocean heat transport, underestimated temperature and time point of total summer sea ice loss by nearly a degree, or almost three decades.

This lower temperature and time threshold for Arctic summer sea ice loss would have feedbacks on multiple Arctic dynamics with global feedbacks, such as additional emissions from permafrost thaw; disturbances in northern hemisphere weather systems from a warmer Arctic ocean and atmosphere; and more rapid melt and sea-level rise from Greenland. Findings underscore the need for earlier and more rapid emissions reductions to keep temperatures within Paris Agreement goals, which this study found would greatly slow and delay the point of total summer sea ice loss; and related global impacts, including loss and damage.