Thursday, 23 July 2015

Mid Summer 2015 Status

Taking the Arctic Ocean summer season as June to the sea ice minimum, around now is the middle of the summer. In this post I look at the state of the ice mid summer, and offer a guess of what the minimum will be.

The first three graphs show the situation on 22 July in the context of the range of previous years. The red triangles show the anomaly for 22 July 2015, the black lines show the range of anomalies from 1979 to 2015 to give an idea of where we are relative to the variation of the past. All values are shown as difference from the 1981 to 2010 average, otherwise known as anomalies.

Extent is towards the lower end of the 1979 to 2015 range from Chukchi, along the Eurasian coast to the Barents sea. This leads to the peripheral region (Beaufort round to Laptev) being low relative to the range of past extents. However all other regions of the Arctic Ocean are about average. Extent outside the Arctic Ocean (Extra Arctic) is low at this time of year, but relative to past years is pretty average, this will be discussed shortly.

Area shows a similar pattern to extent, it is shown for completeness.

Compactness is is defined as area/extent, a compactness of 0.5 implies that within the ice covered region, as defined by extent, there is actually 50% of open water, either between leads or melt ponds. Compactness is potentially significant if weather remains favourable for melt. Low compactness implies open water within the pack which allows absorption of insolation to be increased (over that for ice), due to dark water between ice floes. This has the potential to get the ice to a state where floes melt out later in August.

Compactness is low in the same regions for which extent is historically low. Note that Okhotsk and Gulf of St Lawrence contain 'phantom ice' that will almost certainly go when the data is cleared up before being added to the long term dataset, hence the high compactness (so those regions can be ignored).

Switching view from compactness to open water, (Open Water = 1 - Compactness), here are the open water fractions for the ten day average from 13 July to 22 July together with the long term average.

So if weather conditions persist, there is ample scope for above average extent losses through August, which means that September extent could feasibly meet or exceed the 2007 record minimum. I still doubt that 2012 is possible due to the distribution of multi year ice (see the final two images of this post).

To get an idea what has happened since the start of June it is worth looking at extent for the whole Northern Hemisphere over the course of this year.

2015 (this year) is shown in red together with the other post 2007 years. This year started off as one of the lowest winter extents, with March extent the lowest on record (NSIDC). What followed was a poor ice retreat which led to the 2015 line crossing the pack and entering July as one of the higher extents since 2007.

As I have previously shown, using data from within the Arctic Ocean did not support the conclusion indicated by the winder northern hemisphere data of a poor melt this year. What was not apparent in late June was what would follow in July, the re-emergence of the Arctic Dipole (discussed here).

Updating the data in that post I have plotted the sea level pressure (SLP) and surface vector winds for the ten day period 12 to 21 July 2015 (up to the most recent NCEP/NCAR data).

The bull's eye pattern of the low index AO summer pattern is seen in the plot of SLP, and this is a particularly exemplary instance of the pattern. Data NCEP/NCAR.

Within the Arctic Ocean the pattern of high pressure across the ocean, and lows along the Eurasian coast drives winds from the Bering Straits along the coast of Eurasia. This is seen in the vector winds, annotated as below. Note that in the top right quarter winds driving export through the Fram Strait are rather strong, which means Greenland sea extent is up.

This has resulted in strong declines across the Eurasian coast, the Peripheral Seas are the seas from Beaufort round to Chukchi, shown here, Extra Arctic is all regions outside the Arctic Ocean, and the Arctic Ocean is Beaufort round to the Greenland Sea, Central Arctic and the Canadian Arctic Archipelago (Alaskan, Siberian, Atlantic and Central in the above linked map).

Total extent (blue) had been declining at a leisurely pace, mainly due to below average loss rate in the Extra-Arctic (green), however with the emergence of the Arctic Dipole in July the Peripheral Seas extent (purple) crashed, resulting in increased decline rate in the total extent.

Changing perspective to anomaly of extent, the difference from the average seasonal cycle for 1981 to 2010 (a period chosen as it is the same baseline used by NCEP/NCAR reanalysis):

Firstly the Arctic Oscillation index went negative shortly before 1 July 2015.

And at around the same time the Siberian sector went from slightly greater extent than average to below average beginning a severe drop into strongly negative territory. For the last ten days of June Siberian sector extent averaged 1.64M km^2, for the last ten days of data (13 to 22 July 2015) the Siberian sector has averaged 1.08M km^2, around 1/3 of the drop of 0.6M km^2 has been larger than average and can reasonably be attributed to the emergence of the Arctic Dipole.

Note how for all other regions the extent has remained below average, but loss has either been average (anomalies flat) or only a bit above average (anomalies declining).

The Siberian region is the combination of the East Siberian (ESS) and Laptev Seas. The loss has mainly come from the East Siberian Sea.

In a post entitled The Chukchi Rout I outlined the collapse of extent and area in the Chukchi Sea and suggested that this would open up the ESS, but due to multi year ice would probably have less of an effect on Beaufort. In actuality what has happened is that the Arctic Dipole anomaly has taken over as the driver. So what of Chukchi?

Compactness in Chukchi is 0.404 (average for the last 10 days), that is 6/10 open water. A rough calculation using average losses over the last ten days suggests that if current conditions continue Chukchi may be ice free by early August, not this month as I have previously suggested.

Beaufort however continues to resist. In the plot below, the first drop stalled when the ice edge hit the thicker ice shown in the US Navy's Hycom model, since then it has not picked up.

Without a serious turn in the weather drops of extent will come in Beaufort. Compactness is low, but as can be seen from the following anomaly plot it is in the middle of the post 2007 pack.

However it is now rather unlikely that Beaufort will melt out totally this year, and the most likely reason is the large export of multi year ice into that region over the winter from the Central Arctic.

The peripheral seas of the Arctic Basin are the seas in which minimum extent has been largely set post 2007. They are Beaufort, Chukchi, ESS, Laptev, and are shown in the following image, adapted from Cryosphere Today.

Whilst I have addressed issues with all four seas it is still worth looking at the overall picture.

First extent in the peripheral seas.

Extent is around halfway to zero, but as discussed above, Beaufort is not likely to melt out. We're tracking fairly close to 2007 and 2011, absent a severe turn in the weather to less favourable melt conditions those years seem like a reasonable expectation. Barents and Kara are nearly bottomed out, and the Greenland sea is dependent on export, more export means a greater extent, but that means less ice to melt within the Arctic Ocean (and we're tracking 2012 in the Greenland Sea).

What is significant for the peripheral seas is compactness, which is tracking 2012 and 2007. This suggests conditions are good for strong extent losses in August, maintaining the expectation of something around 2007 and 2011.

Meanwhile in the Canadian Arctic Archipelago (CAA) extent is approaching the fall off in August, same with the Central Arctic, we will have to see how they go. But compactness in the CAA and the Central Arctic are low, this may mean drops in extent in those areas, in the case of the Central Arctic I must add the proviso; if the ice edge can penetrate to it. And in that respect conditions from Chukchi to Laptev might just allow some penetration in August.

But everything hinges on the weather. My opinion on that might be changing from a straightforward expectation of persistence of the Arctic Dipole. My gut feeling is that the Arctic Dipole will continue to dominate through August. But I have little to back that up aside from studying the pattern in years between 2007 and 2012, it may wane, but in those years it often came back once it had established early in the summer. For example in 2010 the pattern arrived in June, but was largely absent through July before re-appearing in August, on a week to week basis (using Hovmoellers) the pattern can come and go. Currently the AO Index prediction ensemble is fairly tight and consistently low index, suggesting high pressure dominance. GFS reflects this high pressure dominance, but shows some parts of the pack dropping below freezing in early August. But that occurred in 2012, and the 850mb temperatures don't seem that difference.

However after 2013 and 2014, with some people talking about the AMO shift being relevant, a subject I don't have a strong opinion on right now, I am being cautious, the Dipole may fade and give way to conditions that are poor for melt.

If the weather holds I think there is a good chance of September being around 2007 or 2011 levels, if it doesn't it will be higher. But I tend to the opinion that ice now is such that a re-run of 2013 or 2014 is rather unlikely.


Nightvid said...
This comment has been removed by the author.
george_formerly_of _the jungle said...

Don't believe the AMO shift hypothesis. The amount of heat transported northwards out of the tropical north Atlantic according to AOML has increased over the past 5 years. That may lessen north Atlantic hurricane activity but there is not evidence there is going to be a decline in heat transport to the Arctic. -George aka FishOutofWater

Chris Reynolds said...


Well, I might be a little spooked by 2013 and 2014. I'm still prepared for more summers like those, but remain open minded.

Anonymous said...

The regional extent and area graphs are identical--something you might want to check.

I'm surprised that a blogger known for PIOMAS data assessment has omitted posting the details of the June data dump and doesn't mention it in the recap of the progress of the summer melt season. The expected rate of change to the area and extent is dependent on the ice age and thickness.

Transport is a large factor in the expected melt and, I expect, likely the critical driver. But, the dominance of the transpolar drift relative to the Beaufort Gyre must be persistent over an extended time frame and two weeks in July is simply not enough time to move the ice pack the long distances necessary to flush the MYI into the Atlantic. The daily wind plots don't show a persistent Dipole, yet. For now the Gyre remains strong and the ice is accumulating in that region.

Chris Reynolds said...


Thanks the area graph in question is corrected.

I didn't include PIOMAS because the emergence of a dipole followed the June PIOMAS data. However it might have been better to address it.

The dipole pattern shown in the past has been the dominant feature of the atmosphere through most of July.