Sea-ice drift speeds in the Arctic.

Hakkinen et al, 2008, use observations of sea-ice drift to examine possible trends in Arctic sea ice drift. Their aim is to examine whether sea-ice drift reveals anything about possible increase in storms in the Arctic, previous research has suggested that mid-lattitude storm systems are more frequently moving into the Arctic. As the authors note because sea-ice movement responds within hours to changes in wind this is a reasonable approach to the problem.

Hakkinen et al find an increasing trend in sea ice speed, the following graphs use the square of the sea ice velocity the data used are in a region of the central Arctic.

Figure 1 of Hakkinen et al, 2008, "Sea ice drift in the Arctic since the 1950s."

They note that:
"The envelope of summer ice drift increases
more steeply than that of the winter drift
due to increased open water, less resistance
from internal friction and probably due to
ice thickness reduction relative to previous
decades."

As noted in the paper this finding of an increase in sea-ice speed is dependent upon data quality, particularly in the early period of observations. However they also go on to examine the trends in speed for different speed-bands. They find that the faster speeds have increased, and the slower declined. They also find that the faster speeds account for less than 2% of the observations. This fits with the pattern I'd expect for storm driven increase, and I can't think of another process that would cause this observation. So the study's findings seem robust, even against concerns over early period data quality.

They go on to examine wind stress using NCAR/NCEP reanalysis data. Wind stress has increased across the Arctic Basin on an annual, summer, and winter basis for the period 1948 to 2006, the greatest increase being in summer and over the Trans Polar Drift. As the overwhelming driving force behind sea-ice motion is wind this explains the observations with regards sea-ice speeds.

The paper at the outset notes that there is some disagreement in the literature, with Serreze & Barrett in 2008 reanalysis study stated as finding no evidence of an increase in Arctic storms. However in my opinion Hakkinen et al have provided independent corroboration for the other findings of increased storms in their use of sea-ice speed.

What particularly interests me is the apparent jump in speeds during summer, seen in the latter part of their figure 1 (above), as this corresponds in timing with certain findings about the Arctic Dipole Anomaly. This jump may also be connected with the rapid thinning of the Arctic sea-ice that we've seen this decade. As is so often the case with papers things are moving so fast in the Arctic that I wish they'd produce updates to the studies, so that we can see how events like 2007 have played out up to the current year.

There is also a suggestion of an exponential to the summer drift speeds, given that this is already squared that suggests that winter's trend is somewhat logarithmic (the result being linear) - flattening in recent years. But summer's is somewhat linear, with an uptick at the end. Frankly I'd prefer to have seen those results without them being squared.

Hakkinen et al, 2008, "Sea ice drift in the Arctic since the 1950s."
PDF here.


At present there are a lot of cyclonic systems massing around the Arctic, there have been for the last two weeks. I watch the Arctic daily, using Crysphere Today and Bremmen images of the sea-ice. But a mainstay for my watching atmospheric and ice conditions has been ArcticIO's excellent overview page, I also use the University of Kolne's Arctic Synoptic plots. From watching those pages I've come away with the distinct impression of storm systems 'feeding' off the open water that now surrounds the pack. This is to be expected from first principles; the open water is relatively warm and should feed convection with latent heat (even though it may not be deep convection of the tropical type). From the scant detail of the ice-pack visible under the cloud it seems to me that these storms may be ripping up the edge of the ice. The AMSR images show a broad and persistent area of reduced concentration in the Beaufort/Chucki region. Of course it's hard to tell, not only is visibility poor but the perpheral ice tends to get ropey at this stage in the melt season anyway.

However cyclonic systems produce winds that circle them in a counter-clockwise manner. The coriolis effect means that the ice moves to the right of the wind driving it. Combine the counterclockwise winds with the coriolis effect and cyclonic systems have a divergent effect on the sea-ice beneath them. Of course a lot depends on the location of the centre of the system, but the movement of these systems around the ice edge should produce a lot of mixing in the peripheral ice, and mixing of the warmed ocean waters under the ice-edge.