Last February I posted about a break up of sea ice that occurred in the Beaufort Sea, now another break up has happened over the last week. It demands a post, but one looking at the bigger picture, indeed were I to try to go over what has happened in the last week the result would read surprisingly similar to my post of February 2013.
I'm still not totally convinced this is unusual, but there is reason from the scientific literature to expect more sea ice break ups in the Beaufort Sea.
First it's worth looking at the ice state as seen in polar orbiting infra red satellite images, source Environment Cananda.
The Beaufort Sea is centre left of the image, fracturing is seen throughout the image, however Beaufort has been left heavily fractured, with fractures running along the coast of the Canadian Arctic Archipelago (towards the rose of compass points overlayed around the pole). As with last year this is because winds have set up a general clockwise motion of the pack placing ice under tension and giving rise to rapid parallel cracking.
Last year there was clear termination of cracks as they reached the thicker multi year ice, which could be identified from ASCAT, see first link of this post for details. That is less evident this year however it provided a clue that ice state in Beaufort (in 2013 virtually all first year ice following the 2012 record minimum) was playing a pivotal role in allowing the development of the fracturing.
Given the difficulty of ascertaining whether one, or even both of these events are unusual (I gave my opinion on that last year), perhaps a more fruitful question is this: Is there any reason to suspect that winter sea ice cracking might be becoming more intense than in the past?
In "Recent changes in the dynamic properties of declining Arctic sea ice: A model study" GRL 2012, Zhang et al use the PIOMAS model to examine dynamic physical properties of Arctic sea ice and how they have changed. The following figure has been amended to put years on the horizontal axis.
Above is an extract from Zhang et al's figure 1, it shows how modelled ice strength has fallen, in fact this fall is very similar (though not identical to) the decline in volume, which is a result of ice strength being determined by ice thickness and concentration, a finding based ultimately on empirical studies used to develop the forerunners of the PIOMAS model.
As both volume and mechanical strength have declined the correlation between wind and ice speed has risen (bottom panel of the above figure), this is to be expected because the wind has both less mass to carry, and is opposed by less internal ice strength. It is worth noting from the above graphic that since about 2005 the correlation between wind and ice speed has exceeded 0.85.
Zhang et al propose a wind/ice speed correlation of 0.85 as an (albeit arbitrary) benchmark above which ice can be considered in free motion. A correlation is a measure of how well two sets of data match, it ranges from -1 (a negative match e.g. as one goes up the other goes down), through 0 (no match between the two), to 1 (a positive match). So as the correlation between wind and ice speed approach 1 the ice is ever more responsive to the wind. Zhang et al's choice of 0.85 is an arbitrary threshold above which the ice can be considered as behaving like a floe of ice in open ocean would, being blown freely by the wind.
Zhang et al show that in the PIOMAS model the wind/ice speed correlation has increased post 2007, but that this varies throughout the year.
In late winter, when the ice is thickest, correlation drops. As the ice thins into the melt season the correlation rises as the ice loses mass and mechanical strength, as the refreeze starts much of the area of ice is now thin. But following the loss of volume (thickness) after 2007 the wind/ice speed correlation is always above that for the 1979 to 2006 average, and above the 0.85 threshold for far more of the time.
So as the ice in the PIOMAS model has thinned it has responded to the wind more efficiently, and this is for basic reasons, the mass of ice (inertia) and its mechanical strength have both decreased.
In "Arctic sea ice conditions in spring 2009–2013 prior to melt" Richter-Menge & Farrell, GRL 2012, looked at NASA IceBridge flight transect data over two regions, the Beaufort and Chukchi Seas and the central Arctic (off northern Greenland and the Canadian Arctic Archipelago). NASA IceBridge is a programme designed to bridge the gap between the ICESat system and Cryosat 2, also to provide validation data for Cryosat 2. It uses instrumentation in aircraft to monitor sea ice characteristics including, crucially, sea ice thickness.
I'll be comparing the Richter-Menge & Farrell results with PIOMAS in due course, but for now it is their figure 3 that interests me. The following figure has been amended to put the years legend in this pane of the source figure.
PIOMAS volume loss of 2010, here it is in NASA Icebridge data, not just the PIOMAS model.
So in Beaufort we have seen a massive shift to thinner ice after 2010, as shown by NASA IceBridge. Data from PIOMAS suggests that at least in the model ice response to wind forcing has increased, due to thinning of the pack. But the physics that give rise to this modelled behaviour concern a reduction of mechanical strength with thinning and loss of mass leading to reduced inertia, both being fundamental issues, so there is little reason to doubt this model finding.
So if indeed we are seeing earlier Beaufort Breakups in response to winds that would not have caused such events in the past, then there is some reason to expect an increase in such events, and I would not find it surprising.
EDIT 19/1/14, Wipneus has been working on 'Home brew AMSRE' data for some time, he has posted a close up animated gif showing the ice movement in Beaufort, see here.