The following brief communication was published on the The Cryosphere Discussions website two days ago: Does it matter exactly when the Arctic will become ice-free?
It's a good question, albeit a rhetorical one. The authors argue that a "robust definition of ice-free may reduce confusion in the community and amongst the public", and start by asking what the exact definition of ice-free is.
We consider four plausible definitions of the date of an ‘ice-free Arctic’. We apply the commonly-used threshold, for which northern hemispheric sea ice extent (defined as the total area of ocean with a sea ice fraction greater than 15%) is less than 1 million km2. The threshold of 1 million km2, rather than zero, is used because ice can be expected to remain for some time along the northern coast of Greenland, whilst for navigational purposes the central Arctic is ice-free. The ‘first ice-free year’ is then defined as:
A. The first year that at least one day is ‘ice free’.
B. The first year when the September mean is ’ice free’.
C. The first time the final year of a 5 year running mean of September monthly mean extents is ice-free.
D. The final year of 5 consecutive September monthly means which are ’ice-free’.
The question is an interesting one to ponder, and not just from a scientific perspective. Still, discussing the exact definition of ice-free may itself become a smoke screen that shrouds a more important issue, and thus cause even more confusion. Fortunately, the authors seem to be aware of this when they state in their paper's final paragraph: