What caused the volume loss in 2010?

I've previously discussed why I see 2010 as a key event in the loss of volume in the Arctic as modelled by PIOMAS, and that there is some indication this has an accompanying impact in satellite observation of sea ice area. But what caused the loss of volume in 2010?


Key to answering the question of what happened to the thick ice is the question of when it happened. The following graphic is the volume series for years 2006 to 2012.

This shows that the years plotted fall into two groupings, years after and including 2010 are clearly below the rest. This happens from approximately day 115 (26/4/10) to 160 (10/6/10), when the plot for 2010 moves from one grouping to the other. This would suggest that whatever happened, it happened in May 2010. But the numbers show that the process was not as clear cut, yes the largest volume drop, creating the deviation from previous years to 2010 happened in May, but more happened before.

I have broken down the volume loss into 25cm categories, to see what finer analysis shows? The following graphic is from a spreadsheet and shows the interannual differences between 2009 and 2010, i.e. 2010 minus 2009. Rows are months, columns are thickness categories, figures in 1000 km^3.

Firstly I'll dispense with a detail that may look important but isn't actually relevant. From May to August in the categories 1.75m and 2.0m there are large drops. These are due to those categories being high in those months in 2009, relative to preceding years, as they aren't due to change between 2010 and the typical recent state of the ice, as typified by 2009 (an unremarkable year compared with 2010), I've discounted them from the following considerations. It may be thought that it would be best to use anomalies, however I've not found these bring out the details I'm concerned with, and as I'm looking at the cause of the volume drop between 2009 and 2010 directly comparing these years seems the relevant consideration.

I've emboldened the categories I see as relevant, and in the final column is the sum of the emboldened figures. That sum does not add up to the month to month net volume change because of offsetting from some volume gains, but again; I'm concerned with the losses that seem to account for the net interannual volume loss.

Firstly it is necessary to look at conditions prior to the above table. What is notable is that from December 2009 there appears to be a failure of the regrowth of ice over 3.5m thick, this condition persists throughout the winter and is visible on the thickness maps as a lack of the darkest ice area which is typical off the Canadian Arctic Archipelago (CAA). The most reasonable explanation for this is the effects of the Warm Arctic Cold Continents (WACC) pattern of Winter 2009/10.

The winter of 2009/10 was characterised by a pattern of warmer than average temperatures across the Arctic with colder than average in a band across the mid latitude land masses of Europe, Asia and North America. This is seen most strongly in February. This was due to the reversal of the polar vortex and influx of mid latitude into the Arctic, with cold outbreaks of arctic air at lower latitudes, a pattern known as Warm Arctic Cold Continents, background here. These warmer temperatures would not cause melt, the anomalies are from a baseline well below zero degrees Celsius. However ice thickens by accretion of new ice at the ocean/ice interface causing sea water to freeze onto the base of the ice, this process is driven by heat flux through the ice, if the surface temperature rises, that heat flux reduces and the ice thickens less.  The following graphic shows NCEP/NCAR surface temperature anomaly for February to April 2010, a similar pattern of warm anomalies over the Arctic ocean and CAA was evident back through the winter. The averaging of this graphic omits the large warm anomalies that were present in the Arctic ocean off the coast of the CAA, during occasional months.

However it is not until March that we see substantial drops of volume relative to previous years typical volumes. From March to June the losses of volume relative to 2009 increase. NCEP/NCAR surface temperature plots for February, March and April all show large warm anomalies over the Arctic Basin and CAA.  May is warm, but this is not as marked, and as the summer begins the anomalies reduce in the Arctic Basin as melting ice 'pegs' the temperature at near-zero preventing warm excursions. The role for a warm anomaly from February to April should not be misunderstood as implying that thin ice wasn't lost; Thickness anomaly maps show that from March 2010 a massive negative anomaly in excess of 2m down from baseline (1980 to 1999) developed off the CAA from February. This corresponds with a lack of thickening in that region in February and March, which I'm attributing to the relatively warm conditions. However it seems to me that the warm anomaly preventing thickening off the CAA explains the loss of the thickest sea ice categories by June.

Going back to the interannual differences table above, most of the interannual drop in volume from March to June is accounted for by increasing loss of ice over 3.50m thick. As the ice volume of over 3.50m hits zero in July the losses spread down the thickness categories, arguably this is happening in the 2.75 to 3.50m thickness categories in June. This spreading process occurs until September, after which regrowth occurs.

So it seems to me that the volume losses of spring 2010 were caused as follows: The WACC pattern of the winter 2009/10 primed the ice by retarding the formation of the thickest ice, leading to low volumes over the winter. Then warm conditions in the spring caused further loss of thicker ice, these warm conditions were centred over Greenland and the CAA, which may explain why thicker ice in particular was lost to a greater degree than in recent years preceding 2010. These losses of volume spread down through the thickness profiles as the summer melt proceeded.

However as the previous post shows, this does not mean that the ice then recovered in the following years. Again we see a similarity with 2007. Both 2007 and 2010 are weather driven events that leave what turns out to be an indelible impact on the sea ice.

In the first graphic of my previous post I showed how 2007 and 2010 are together with 1981, 1993 and 1995 as a set of years with notably high volume loss. It might seem that this suggests the recent episodes aren't unusual. However apart from 2008, every year since 2002 has shown loss of volume from the previous year, so while large volume loss episodes are not unusual, the earlier volume losses were followed by recoveries. That is not the case for 2007 and 2010.

Leaving aside the possibility of atmospheric feedbacks on sea ice loss, as appears to have been the case in 2007: To dismiss 2007 and 2010 as mere weather driven events is to ignore the crucial fact that the forcings driving sea ice loss are unremitting and are not giving the ice time to recover. As the ice gets thinner open water formation efficiency goes up and weather driven loss events become more likely.