Concentration vs. Extent

MODERATOR :
There is supposed to be a youtube video below the last sentence. Please add this text:
[youtube=http://www.youtube.com/watch?v=762CMOmZ4jE&feature=player_embedded]

For the period 2003 to 2009 based on JAXA data, the greatest loss on a daily basis occurs during the month of July. It was during this time that 2007 saw the worst of its loss.
I think the next 4-6 weeks will pretty much tell the story of this years minimum extent but my money is on an improvement over the 2009 result. I think we got a lot of late ice at the end of winter that melted (or broke off) very quickly leading to the falls in extent we saw in May. I think the thicker ice will be a bit more resilient.
Look at the daily ice loss on a 15 day moving average and you can already see the rate of loss levelling out when the long term trend seems to be for the rate of loss to increase until late July or August.
Having said this, my understanding is that localised conditions can always override global ones in the short term so there will always be a fair degree of uncertainty in any prediction for a single locality for a limited timeframe.

Of course, I think Steve and everyone else knows what I (and the NIC) think of his PIPS 2.0 data, but he continues to weave in throughout his interesting tale of the current Arctic conditions. Let’s take a look at his PIPS 2.0 map for June 10, 2010 for a second, to see how severely flawed it is. Look at this map:
http://www7320.nrlssc.navy.mil/pips2/archive/retrievepic.html?filetype=Thickness&year=2010&month=6&day=10
And then compare it to these two very detailed maps:
http://arctic.atmos.uiuc.edu/cryosphere/NEWIMAGES/arctic.seaice.color.000.png
http://iup.physik.uni-bremen.de:8084/amsr/arctic_AMSRE_nic.png
And you see some major melting going on now in the Beaufort sea and out into the Arctic Basin and over into the Chuckchi Sea. Yet the PIPS 2.0 model doesn’t seem to even recognize this break up and rapidly lowering concentration of ice. Though the PIOMAS model seems to show that melt occuring:
http://psc.apl.washington.edu/zhang/IDAO/seasonal_outlook.html
In short, the PIPS 2.0 model would seem to show ice that is anywhere from one to 2.5 meters thick, where there is actually concentration falling down into the 50% range, meaning that there could even be open ocean, and in fact, if we look at this high res satellite image of the region, we can see that in fact there is a large break up of ice that is not the “shearing” that Steve was pushing as the cause of such a rapid ice extent loss just a week or so ago:
http://ice-map.appspot.com/?map=Arc&sat=ter&lvl=7&lat=73.190674&lon=-137.618661&yir=2010&day=161
Clearly this ice is melting, and concentration is falling, yet the PIPS 2.0 model misses a great part of that.
WUWT? Maybe another sign that PIOMAS has it right, and the Navy and the NIC are right in not putting much credance in the PIPS 2.0 model data?

R Gates
The PIPS map shows the same region of ice being blown out of the Beaufort Sea as the UIUC maps. PIPS2 is the best available source of thickness data. Your claims are inaccurate.

Actually, looking at June 1989 in comparison
http://igloo.atmos.uiuc.edu/cgi-bin/test/print.sh?fm=06&fd=10&fy=2010&sm=06&sd=10&sy=1989
the ice is less concentrated in 1989, however that had a very large positive AO of 1.679 3.106 3.279 for December, January and February and so that is puzzling. 1990 has lower AO, but is less concentrated. An interesting topic in itself.
Andy

I don’t believe there has been any significant melt in the Arctic Basin so far this year, as temperatures have been too cold. The areas of open water are primarily the result of wind.

“Watts Up With That?” has been included in this weeks Sites To See. I hope you like the image I featured, and I hope this helps to attract many new visitors here.
http://asthecrackerheadcrumbles.blogspot.com/2010/06/sites-to-see_11.html

Anu says:
June 10, 2010 at 11:42 pm
It is natural to think that high resolution will always give you more information than low resolution, and it is also a sign that you do not have much experience or any of quantitative image analysis in either 2D or 3D. We are talking about measuring differences between two images / datasets. What you find in the real world, as I have discussed earlier with R Gates, is that, somewhat counter-intuitively, if you take images of two samples (or groups of samples) with fine spatial structure, and image them at a set of different resolutions, the ability of image analysis so measure differences between the two samples is surprisingly well preserved with decreasing resolution. In some cases the ability to measure difference is oddly strengthened by decreased spatial resolution. This is related to signal to noise ratio.
As I said before, I’m not arguing that low is better than high – everyone involved in image analysis (which is my day job at present) likes the highest resolution and best images. But it definitely is not a general truth that lower resolution means decreased or no ability to see differences. Often the reality is very little change.
What adds to this is if the structures in question have fractal pattern – similar architecture repeating at different spatial scales. A snowflake is a fractal object. Many coastlines are also, and Arctic ice probably is also fractal to some degree.
So this resolution argument wont work as a way of refuting the conclusion from the images posted above that the Arctic ice currently does have a high concentration.

Why not wait a few months and try not to get too excited?
Yes, what’s happening to ice is of continuing general interest, especially the teasing out of all the reasons for ice cover changes – winds, water currents, insolation – with the attendant difficulty of reading long term climate signals from ice cover.
But whether the earth has warmed a bit/is still warming/may be on a plateau/is beginning to cool, doesn’t speak directly to the “CA” bit of CAGW/CACG. Well, I suppose cooling would! But I’m always seeing/arguing about confusion between climate shifts as such, and “warming from man made CO2”: to accept there may have been a little warming isn’t to accept that it’s all CO2 driven via positive feedbacks. Or that warmer = catastrophe.
But, darn it, I forget that the Media don’t care about logic so loads of ice would help off set the poor furry wurry poley bear argument…
Thanks to ALL you informed types who blog here, it really helps us less informed ones, and to Anthony & mods for running it.

Excerpted from: Anu on June 10, 2010 at 11:42 pm

Looking at the exact same data in 1296 x 1296 resolution…

Whoa now, that would be some trick. The entire UIUC image is only 1709 x 856 pixels, same as the copy used here. If you’re blowing it up from an 856 height to a 1296 then you’re “creating” data where none exists. Are you using the James Hansen method of image processing?

The importance of high resolution:
Low resolution: picket fence looks almost solid
http://tinyurl.com/2fajczo
High resolution: hey, you can see right through it
http://henbogle.files.wordpress.com/2008/05/picket-fence.jpg

The importance of using the same data. That is not the same fence, the slat and space widths are not known, we don’t know if the ratios are the same… They weren’t even taken from the same angle or distance. Heck, if you want to see through a picket fence all you have to do is get close to it, and the second picture is clearly taken closer than the first. If that second picture had been taken at that distance but at the same resolution as the first, you still would see lots of detail right through it.
Now let’s see some real analyses of the same data. If you don’t mind, of course. 🙂