I've pondered how to approach this post, but decided just to go through what I've done in a linear fashion, storywise, in the hope that people can follow my reasoning. Futhermore it's a very graphic heavy issue. However I don't like animated gifs because you can't pause them, so I've decided to put up with the risk of a slow loading page. All of the graphics are plotted with the same range of colour-values. I've presented only a small subset of the graphics I have viewed at NCEP/NCAR, after a point collecting them became superfluous.
I used the period June to August as the summer, this is common to all the images here, as is the pressure scale. I was interested in whether the UK had a preponderance of low pressure systems during summers since 2007, so I looked at sea level pressure (SLP) from 2007 to 2011, and just for starters subtracted the previous 5 years, 2002 to 2006. This gave me a composite that was the difference in SLP between two equally lengthed periods, pre and post 2007. This is the result.
The general pattern is an area of higher pressure in the mid Arctic and Greenland, with a suggestion of a surrounding band of lower pressure. When I say higher/lower I'm referencing to the pre 2007 period. This is the pattern I looked into further...
I wondered if this pattern was just due to general global warming: Would it be repeated in earlier periods? So I chose some 5 year periods at random and subtracted their mean SLP pattern from the mean SLP pattern for an adjacent 5 year period.
I saw year to year variations in weather, whichever way round I did it, no suggestion of a trend since 1950 that could account for the pattern I'd seen in the five years around 2007. What interested me then was just how robust this pattern around 2007 was.
So I took the period 2007 to 2011 and deducted from that some of the five year periods used above.
Frankly, the results amazed me! That the pre/post 2007 pattern appears robust when compared to disparate sets of years since 1950 was not what I was expecting at all! This pattern appears robust in the context of these randomly chosen periods, and more that aren't shown. This implies it is something new in the context of the second half of the 20th Century.
However the period of 5 years used for the above is rather a clumsy instrument: When did this change actually happen? Instead of using means I decided to revert to anomalies from the climatology and look at all the years between 2000 to 2011. This should show when this change happened precisely.
Obviously there's a lot of weather variation, as is to be expected. But in the post 2007 period there seems to be a tendency for an Arctic high anomaly ringed round with low anomalies.
Nothing I've read has lead me to expect such an apparently strong summer atmospheric response to sea ice loss. As 2007 was the year of the greatest crash so far it is tempting to point the finger of suspicion at areas of open water due to the anomalously low summer ice extents since (and including) that year. However IARC/JAXA extents for August 1st are (million km^2):
And in this respect the step-drop starts in 2005, with 2006 being matched with 2008.
I'm reminded of Stu Ostro's work, he has been drawing a link between ridge/trough formations in the atmosphere around the Arctic. His lecture slides are here, warning - huge 21 Mbyte pdf, page 51 and onwards shows what he's talking about.
But before I speculate any more I'll leave it at that and wait to see if anyone can point out the obvious explanation for all this that I've missed.
PS. On a parochial note - notice how the UK has been in low pressure anomalies since 2007? Now I see it I no longer care.