[correction – I originally said 18% but had copied the number over incorrectly from my spreadsheet]
Turquoise represents ice present in 2012 which wasn’t present at the 2007 minimum. Red shows the opposite. There is 14% more ice than there was at the 2007 minimum.
http://www.natice.noaa.gov/pub/ims/ims_gif/DATA/cursnow_alaska.gif
ims2007262_alaska.gif (512×512)
Very little change has occurred north of 80N over the last few days.
Hi Steve,
It looks like ice has been increasing these past two days according to DMI, with new scattered ice filling up most if the open sea in the Arctic.
http://ocean.dmi.dk/arctic/satellite/index.uk.php
It may turn out that we will assist to a very fast sea freezing soon.
Steve has already predicted a record recovery. that will be unprecedented. A record loss and record recovery. very abnormal
Cut the crap. I never said anything remotely like that.
Oh come on Steve, this can’t be true. You just drove up there in the middle of the night and painted that on there. Fooling the satellites again… sheeesh…
Watching the beginnings of the snowpack forming both in siberia and the canadien arctic islands…hoping for a dramatic re freeze up.
Why? What difference will a “dramatic re freeze up” make? Will that change anything, anything?
August 19th: 40% more
August 25th: 36% more
August 27th: 28% more
August 28th: 22% more
August 29th: 18% more
… to be continued …
P.S. Spotting a trend yet? I reckon you’ll be eating crow in about a week.
P.P.S. I’m not quite sure why your numbers are different from the ones Julienne gives (4.26 million today, 3.99 for 2007 minimum, i.e. a difference of 6.8%). I’d hazard a guess that it’s due to you counting from a downsampled low-resolution picture and/or not accounting for any effects due to the stereographic projection.
Possibly, but there has been almost no change north of 80N, which will have to happen to get to zero.
My pixel counting agrees very closely with NIC graphs.
Really? In that case you’ll be happy to overlay the last 10 days’ data (day 241 and 231) and put the 80 degree line on. There has been very substantial ice edge retreat above 80N.
take a picture with your cell phone and count those pixels.
How is that one million km^2 forecast coming along?
How are you calculating your 18% figure Steve?
Pixel counting the colors
thx Steve I thought so..
weird. I counted the pixels and got a different answer. Did you use your cell phone again?
You are hiding your data and methods better than Mann ever did. You are the Ice Mann.
Count how many pixels we are above one million km^2 and then get back to me.
Decline hider.
It would be helpful if you gave the link to the actual image for today; rather than the cursnow_alaska.gif it’s better to use the actual gif with the year and day of year eg. http://www.natice.noaa.gov/pub/ims/ims_gif/ARCHIVE/AK/2012/ims2012242_alaska.gif
As for pixel counts, Here is what I get comparing the above image with the Sept 22, 2007 image
http://www.natice.noaa.gov/pub/ims/ims_gif/ARCHIVE/AK/2007/ims2007265_alaska.gif
First you have to mask out the yellow box in the legend. Then using Photoshop to get an exact count of the Yellow pixels (RGB=255,255,0) you get:
ims2012242_alaska.gif = 20,172 pixels
ims2007265_alaska.gif = 19,296 pixels
difference = +4.54%
PS to get the exact pixel count in Photoshop: 1) crop or mask the yellow ice legend box 2) use Select>Color Range with Fuzziness=0 to select the Yellow Pixels, 3) use Window>Histogram and select Expanded View to show the exact number of pixels selected
Pretty sure Steve was looking at the previous day, but I’d be surprised to see that much of a difference between them. Can you check what the figures are like for http://www.natice.noaa.gov/pub/ims/ims_gif/ARCHIVE/AK/2012/ims2012241_alaska.gif ?
Sure, no problem.
ims2012241_alaska.gif = 20,763 pixels (Aug 28, 2012)
ims2007265_alaska.gif = 19,296 pixels (Sep 22, 2007)
difference = +7.60%
Did it myself with ImageMagick:
Sept 2007: 19,296 pixels
Aug 27: 21,939 pixels (+13.7%)
Aug 28: 20,763 pixels (+7.60%)
Aug 29: 20,172 pixels (+4.54%)
So, replicated with an independent piece of software, I think we can conclude Steve’s measurements are wrong. Steve, what’s your methodology?
Also, Steve – care to let us know how much you charge professionally for image analysis?
David, why have you not included turquoise and magenta pixels in your census? Isn’t that kind of, how shall I say, “important”? Thank you.
RTF
Richard: The original images only have yellow pixels, see the links David put in the post you replied to. He counted each year independently and divided the numbers. Steve’s done something different by recolouring them and overlaying them before counting – likely this step is where he introduced the error.
If I had to guess, I’d say Steve is counting the turquoise pixels and then (mistakenly) dividing by the number of yellow pixels. That is, instead of using the Sept. 2007 area as the denominator, he’s using only the areas in common between 2012 and 2007. That’s just my guess though, it’s impossible to tell what he’s done. I tried to pixel-count the jpg he posted, but the conversion to jpg (and lossy compression) has dithered the edges and made all the colours fuzzy instead of pure. I can count them, but essentially you have a situation where almost every pixel’s a subtly different colour.
I linked to the wrong 2007 image in the article
That explains why I couldnt match his results either. He’s hiding the decline.
You are being a complete idiot.
you post the wrong link.. and who is the idiot?
you cant replicate your own work.. and who is the idiot?
you dont understand a triple point.. and who is the idiot?
haha. decline hider
Not at all – the comparison is between the total amount of ice in the August 29 2012 NOAA image compared to the total amount of ice shown in the September 22, 2007 NOAA image. Both images show ice in Yellow. So you add up the Yellow in each image and calculate the difference.
The NOAA images are GIFs which have precise color values.
The Image posted above with Turquoise and Magenta ( which is a JPEG and not a GIF) simply shows that the ices is not all in the same place in 2012 as it was in 2007.
My apologies for failing to notice that.
RTF
My bad. I put the wrong link to the 2007 minimum image in the article. This might be the one I was working with, but I can’t check it now.
http://www.natice.noaa.gov/pub/ims/ims_gif/ARCHIVE/AK/2007/ims2007262_alaska.gif
OK thanks. Then that gives:
aug 28, 2012 20,763 pixels +14.3%
aug 29, 2012 20,172 pixels +11.1%
sep 19, 2007 18,161 pixels
Yup, that’s the one you used for the overlay – it matches the magenta. That changes the denominator to 18,161, with the following results:
Sept 2007: 18,161 pixels
Aug 28: 20,763 pixels (+14.3%)
Aug 29: 20,172 pixels (+11.1%)
I’ll grant that it’s closer, but I still don’t see where you got 18% from for the 28th. Ah well, it’ll be immaterial by the end of the week.
You are including the legend in your count. That is a no-no.
No, I’m not including the legend. Moreover, the legend is only 126 yellow pixels (6×21) and hence wouldn’t affect the ratio by more than about half a percent anyway.
Perhaps you should put a correction notice on your post. perhaps you should post your data and your method.
No, I’m not
Another minor point of detail Steve is that you have selected the Alaska images which as you can see on the right side of the image above cut off the Eastern side of Greenland and the Fram Strait.
The entire set of images is here: http://www.natice.noaa.gov/ims/gif_archive.html
The Northern Hemisphere images have the greatest coverage but cannot be used for pixelcounting since they have white latlong marks. However the Europe/Asia images do have complete coverage albeit at a lower resolution.
here’s the images
http://www.natice.noaa.gov/pub/ims/ims_gif/ARCHIVE/EuAsia/2012/ims2012241_asiaeurope.gif
http://www.natice.noaa.gov/pub/ims/ims_gif/ARCHIVE/EuAsia/2012/ims2012242_asiaeurope.gif
http://www.natice.noaa.gov/pub/ims/ims_gif/ARCHIVE/EuAsia/2007/ims2007262_asiaeurope.gif
and the pixelcounts:
aug 28 6105 pixels +14.7%
aug 29. 5862 pixels +10.2%
sep 19 5319 pixels minimum
not a huge difference, but a bit
Ice that isn’t included in that picture is all walking dead anyway.
I checked my spreadsheet and I copied the number over incorrectly. I actually calculated 15.8% difference, not 18%. I am checking now as to the source of the of the rest of the discrepancy