Sunday, 15 March 2015

Mid March 2015 Status

So far the sea ice extent maximum was set on 22 February 2015, however as of 13 March extent is only 0.14M km^2 below, so it is too early to claim a maximum. That noted there are good grounds to suspect that the 22 February maximum may hold. PIOMAS data is covered at the end of the post.

This post uses data to the 13 February 2015 for NSIDC Extent, Wipneus's calculation of extent (hereafter known as Extent) and PIOMAS gridded data for February.

There are 36 years from 1979 to 2014, considering the date of maximum for NSIDC Extent:
  • 28 years (78%) had maxima on or before 13 March.
  • 8 years (22%) had maxima after 13 March, of which 4 happened in the week following 13 March.
  • 7 years (19%) had maxima in February, 1987 and 1996 had joint earliest maximum on 21 February.
  • The latest maximum was in 2014, on 31 March.
So from the most simple considerations it seems likely that we have already passed the maximum. However what counts for an individual year are ice and weather conditions.

As previously I have amalgamated sea ice extent data for the Cryosphere Today regions into larger sectors.
  • Pacific sector: Okhostk & Bering Seas.
  • Alaskan Sector: Beaufort & Chukchi Seas
  • Siberian Sector: East Siberian & Laptev Seas
  • Atlantic Sector: Kara, Barents, & Greenland Seas
  • Central Sector: Central Arctic & Canadian Arctic Archipelago
  • Canadian Sector: Baffin Bay, Gulf of St Lawrence, Hudson Bay
The Alaskan, Siberian, and Central Sectors are not factors in determining this years minimum, all three have maximum extent and will not decline until later in the year. So we are left with Pacific, Atlantic, & Canadian. The data for these three sectors are plotted below.

The Pacific Sector has flatlined and has been low all year, the February extent is the lowest since 1979. 

I do not expect this pattern to change during the remainder of March, the PNA remains neutral in the ensemble forecasts (which are well grouped).

The Atlantic Sector has declined from a maximum 2.36M km^2 on 15 February 2015, the decline has been 17%. Most of the decline in the Atlantic sector is due to the Barents Sea, Barents has declined 33% from its maximum on 2 February 2015. It is not unusual for large declines of this nature to occur in late winter and such declines can be followed by increases of a similar magnitude, but any such increase will probably be fighting reductions in the Pacific and Canadian Sectors over the coming weeks.

As can be appreciated from the above plot in terms of the post 2000 period the summer decline starts from around mid April. Considering the 1981 to 2010 average seasonal cycle (NCEP/NCAR anomaly baseline period), it can be seen that the Atlantic plays relatively little role in determining the overall maximum of extent.

It is the peaks in the Pacific and Canadian Sectors that play the dominant role in setting the maximum around March.

From the plot of Sector Extent it can be appreciated that Canadian Extent is still rising. This is totally due to Baffin Bay, Hudson Bay is flatlined at maximum extent, and the Gulf of St Lawrence has risen to a plateau in March, showing little change throughout March. However Baffin Bay is likely near peak extent.

The highest annual maximum extents in Baffin Bay are around 1.8M km^2.

From the behaviour of Post 2000 extents in Baffin Bay I consider it very unlikely that extent can reach 1.8M km^2 over the rest of this month, it seems too late in the season for that to be a realistic prospect. 

In terms of the two datasets the Wipneus dataset and NSIDC Extent differ slightly. This was the reason I gave up processing myself and it is due to a problem with masks. However there is a new version of gridded sea ice data now available which might help to rectify the discrepancies, more about that in due course. That noted, in this instance the agreement is reasonable. The following table shows the date of maximum, maximum extent, extent on 13 February, and drop from maximum to 13 February, for this year (2015). All extents in million km^2.

Date of Max Max Extent Extent on 13  Feb Drop
Wipneus 22-Feb 14.36 14.22 0.14
NSIDC 22-Feb 14.54 14.40 0.14

0.14M km^2 is an achievable increase considering that the Pacific Sector has shown a slight recent upturn, the Atlantic has levelled, and the Canadian Sector continues to increase. So I am not quite willing to call the maximum at 14.54M km^2 on 22 February 2015, not yet anyway. If the 2015 maximum has been set on 22 February it will be the lowest on record, however weather has played a large role in this, as discussed previously with regards the Pacific Sector.

Now for the PIOMAS February data. 

PIOMAS shows a substantial thickening of ice in Beaufort.

HYCOM and the Drift Age Model both show that this is part of a large export of thicker multi-year ice (MYI) into Beaufort, Chukchi and the East Siberian Sea (ESS). The following week is for early February (Week 5 of 2015).

This intrusion of MYI can be expected to play a strong role in the coming melt season.

Volume is higher than previous years in the Central Arctic region, note that in the above plot the scale for the Central Arctic is four times higher than for the other regions. This localised source of the 2015 volume increase shows up in the difference between February 2015 and February 2013 (immediately after the 2012 record).

Looking at the breakdown of volume attributable to five bands of grid box effective thickness, this year is more similar to 2008 than any of the intervening years. However due to the areal restriction of the thicker ice to the Central Arctic volumes attributable to ice below 2.9m thick remain similar to previous years.

The peak of extent and area will happen shortly, if it has not already happened. However because of continued thickening within the Arctic, where thickness is greatest, the peak of volume will not happen until April.

The volume gain from 2013 and 2014 will largely carry over into 2015's summer season.

The above graph is different from my previous post because this time I have used the twelve months after the 2012 record minimum as a baseline. As happened in 2014 following a peak of the increased volume in late 2014 the volume increase is falling off. Whilst this is in part due to export of ice through the Fram Strait, it is also a result of first year ice thickening thermodynamically (and increasing volume as a result) faster than the thicker ice that survived the previous summer. March and April will show a further reduction in the volume difference because of more vigorous late season thickening in 2012.  

The volume increase from 2013/2014 will largely carry over into 2015, I consider it likely that even if it reduces through spring it will re-emerge in late summer and it virtually removes the possibility of a sea ice free state in Beaufort, Chukchi & the East Siberian Sea by September 2015.


Nightvid said...

I was with you until the last paragraph!

Given that the volume on March 8th (the end of the series) is essentially the same as in 2007, how can you conclude that those seas won't melt out completely (they did in '07)?

Chris Reynolds said...


If we get a repeat of the 2007 weather we should expect lower end of season extent than in 2007. 2007 was a 'black swan event' and I don't factor such wildcards into predictions, however informal.

But total Arctic volume isn't the best metric. Gridded grid box effective thickness, the drift age model, and HYCOM all suggest a lot of MYI being transported through Beaufort and into the Chukchi Sea. This is similar to last year and 2010 and I think we'll see persistence of ice leading to higher extent from Beaufort to the East Siberian Sea in September as a result.

What kept last year's extent down was the state of the Laptev Sea, that was because of very specific weather conditions. Such a situation is not likely to recur this year.

Nightvid said...

We have some sort of circulation of MYI into those seas virtually every year though, and it didn't survive in 2007, 2008, or 2012. I think it's a hit or miss proposition - that's 3 out of the last 8 years.

Chris Reynolds said...

The latest plot from the Drift Age Model is for week 5 of 2015.

Have a look at week 5 for 2015 back. The intrusion of MYI in 2015 is more like 2014 and 2007 (also the years before and much greater extent than years like 2008, 2009, 2010, 2011 and 2012. At least that is what I see in that data.

Every year there is export into Beaufort due to the semi-permanent Beaufort High, but some years the export is stronger than others.

If there is a relationship between the strength of the Beaufort Gyre and export into Beaufort/Chukchi over winter - seems reasonable but I haven't explicitly examined it.

However using the box, 70 to 80degN by 210 to 230 degE, and calculating the anomaly of pressure for Dec/Jan/Feb from a 1951 to 1980 baseline I get the following numbers of pressure index - a rough back of envelope indication of the strength of the Beaufort Gyre.

2000 -3.3
2001 1.3
2002 -3.2
2003 -1.4
2004 3.7
2005 -1.1
2006 -6.3
2007 -3.4
2008 -1.2
2009 -0.6
2010 1.6
2011 -0.2
2012 -5.4
2013 2.4
2014 2.6
2015 1.0

Note that 2006 2007 and 2012 were notable loss years overall, all had strongly negative indexes. But 2010 had a persistent mass of ice in Chukchi/ESS by late summer, 2013, 2014 also had the same behaviour. All those years had a positive pressure index over Beaufort, as does 2015.

Nightvid said...

Interesting. Maybe worth another blog post :)

Chris Reynolds said...


You read my mind.

But I do too much hand waving vaguery over those drift age model plots. There is binary gridded data from which I should be able to get regional breakdowns etc. Last night I started reading up on EASE grids and the format of the data. :)