Analysis of NSIDC August 4 News

I don’t think temperatures of around freezing above the surface of the ice (in summer) are the most important. I think that what goes on below the ice gets the upper hand in summer. It melts away from below.
Just a hunch.

Billy Liar
Please propose a physical mechanism whereby the water heats underneath compact ice.

Same post as before until Dr. Meier answers:
Also quite misleading is Fig. 1 Part1 of ice age distribution: 1985 jumping to 2008, 2009 and 2010. Over the past 3 years the variability of distribution is quite significant. Therefore juxtaposing a snapshot at 1985 out of context, without showing the variability then simply occult the processes that lead to yearly transformation of the Arctic ice.
Let’s notice also that the multiyear ice on these images is considered 5y + and not 20y or 50y. Therefore, if Arctic ice was so stable in the past, where has 50y ice gone Dr. Meier? It is likely it has been melted and recycled in an ever dynamic Arctic. Thus presenting 1985 as a supposed representation of the Arctic Ice of the past -i.e. stable- and the evolution since as a abnormal situation -i.e. AGW induced changes- is misleading at best.

When your mandate is to give nothing but an alarmist perspective, you give nothing but an alarmist perspective, right or wrong!

David Davidovics
I have always found NSIDC personnel to be extremely responsive and helpful.

stevengoddard says:
August 5, 2010 at 6:37 pm
Billy Liar
Please propose a physical mechanism whereby the water heats underneath compact ice.
In the same way that air moves over the surface, (warm) currents move underneath the ice?

Billy Liar
The Arctic Ocean is stratified. There is a thick layer of cold water underneath the ice, which does not mix with the warmer water below. If it did, there wouldn’t be any ice in the Arctic.

stevengoddard says:
August 5, 2010 at 6:37 pm
Billy Liar
Please propose a physical mechanism whereby the water heats underneath compact ice
steve, ever been in a mine shaft under ground? is it warm or cold?
Why is it everyone assumes it is cold water on the bottom of the ocean?
Is the center of the earth frozen? or molten?
just saying ……

It is getting so I don’t know who to believe anymore. Everyone’s data is so conflicted. DMI shows good 30% concentration. Ice area looks good. It’s just a mess, really.

Why the splitting hairs over pixel size and the fixation with 2007 as a single point of reference? Take a look at the big picture: 2010 is well below the ’79-2000 average. Its called ‘climate’, not ‘weather’ for a reason … the long period behavior (average over several years, trends) is the message.

Steve, you seem to be assuming that the NSIDC polar map is as a photo taken directly above the North Pole at an infinite distance, so that longitudes are displayed as shorter away from the pole. I would assume it to be a polar projection with static longitude scale, i.e., if it was extended to the equator, the equatorial lands would be far too large. If my reading is right, you are weighting the pixels in the wrong direction, thus doubling the error beyond if you did not weight them at all. That would indeed account for the gap from 3% to 7.5%, so I suspect this is what is happening.

stevengoddard says:
August 5, 2010 at 7:29 pm
RoyFOMR
I am well aware of how some people distort what I write, but the cool thing about the Arctic is that September is only a few weeks away.
Ultimately the visible satellite imagery will decide who is correct.
________________________________________________________________
Unless of course the satellite imagery goes off line until after November 15………

latitude says:
August 5, 2010 at 6:37 pm
At least they can’t claim the NWP opened up.
Someone will. After an icebreaker goes through they’ll start talking about how boats can make it through in 2010.

Tim L says:
August 5, 2010 at 7:30 pm
stevengoddard says:
August 5, 2010 at 6:37 pm
Billy Liar
Please propose a physical mechanism whereby the water heats underneath compact ice
steve, ever been in a mine shaft under ground? is it warm or cold?
Why is it everyone assumes it is cold water on the bottom of the ocean?
Is the center of the earth frozen? or molten?
just saying ……
_______________________________________________________
Because cold water is more dense and it sinks. And also because it has been measured. http://www-argo.ucsd.edu/
“…What is Argo?
Argo is a global array of 3,000 free-drifting profiling floats that measures the temperature and salinity of the upper 2000 m of the ocean. This allows, for the first time, continuous monitoring of the temperature, salinity, and velocity of the upper ocean, with all data being relayed and made publicly available within hours after collection….”
However if you are talking the temperatures around the ring of fire or “plates” all bets are off.
http://oceanservice.noaa.gov/education/yos/multimedia/oceanexplorer.noaa.gov/oceanexplorer.noaa.gov/explorations/06fire/background/chemistry/media/eifuku_champagnebubbles.html
Page down for some really great photos:
http://www.barking-moonbat.com/index.php/weblog/more_bothersome_volcanoes/

jcrabb says:
August 5, 2010 at 7:57 pm
It is interesting that NSIDC sea ice extent data was acceptable back in early July
Would you cite why you are able to say that?

Marge says:
August 5, 2010 at 8:17 pm
Alternatively, when your mandate is to give nothing but an anti-warmist perspective
There are people in this comment thread who have been given this mandate? Or is there a better word than mandate, did you mean something other than mandate?

Marge says:
August 5, 2010 at 8:17 pm
Alternatively, when your mandate is to give nothing but an anti-warmist perspective, you give nothing but an anti-warmist perspective, right or wrong.
The skeptic questions, the anti-warmist accuses.
Alternately, when your are nothing but a warmist harp seal, you give nothing but harps and always wrong!
Harp on warmists!

muographer says:
August 5, 2010 at 7:49 pm
Why the splitting hairs over pixel size and the fixation with 2007 as a single point of reference? Take a look at the big picture: 2010 is well below the ’79-2000 average. Its called ‘climate’, not ‘weather’ for a reason … the long period behavior (average over several years, trends) is the message.
_________________________________________________________
Fine then look at the WHOLE Holocene:
Greenland Ice Core graph – 10,000 years:
http://jonova.s3.amazonaws.com/graphs/lappi/gisp-last-10000-new.png
And then couple it with this:
Temperature and precipitation history of the Arctic
“..Solar energy reached a summer maximum (9% higher than at present) ca 11 ka ago and has been decreasing since then, primarily in response to the precession of the equinoxes. The extra energy elevated early Holocene summer temperatures throughout the Arctic 1-3° C above 20th century averages, enough to completely melt many small glaciers throughout the Arctic, although the Greenland Ice Sheet was only slightly smaller than at present… As summer solar energy decreased in the second half of the Holocene, glaciers reestablished or advanced, sea ice expanded, and the flow of warm Atlantic water into the Arctic Ocean diminished. Late Holocene cooling reached its nadir during the Little Ice Age (about 1250-1850 AD), when sun-blocking volcanic eruptions and perhaps other causes added to the orbital cooling, allowing most Arctic glaciers to reach their maximum Holocene extent…”
You might also want to read:
Lesson from the past: present insolation minimum holds potential for glacial inception (2007)
The 20th century the sun has been very active according to this paper and NASA But as we enter the new century according to the Solar Dynamics Observatory Mission News that no longer holds true.
If I were you I would be petitioning Congress to emit as much CO2 as possible!

Amino Acids in Meteorites says:
August 5, 2010 at 8:21 pm
Mark says:
August 5, 2010 at 6:51 pm
When your mandate is to give nothing but an alarmist perspective, you give nothing but an alarmist perspective, right or wrong!
I’ve seen government employees do such when the government does that mandating.
_____________________________________________________
If you are a scientist and your employer – corporate or government – tells you to lie, you lie or you quit. Those or your only two options in the USA unless you are a certified Professional Engineer.
That is straight from the lawyer I consulted on a case of certificate falsification…. I quit.

muographer says:
August 5, 2010 at 7:49 pm
Why the splitting hairs over pixel size and the fixation with 2007 as a single point of reference? Take a look at the big picture: 2010 is well below the ’79-2000 average. Its called ‘climate’, not ‘weather’ for a reason … the long period behavior (average over several years, trends) is the message.
Tell me how, even in a recovering arctic ice pack, you could have expected any of the years immediately following 2007 to get close to the 79′-2000 average. Thats simply an absurd proposition and that it hasnt does not prove that a long term downward trend leading to the 2007 loss has continued beyond that point.
Your entire argument is disingenuous. 2007 and the years leading up to it, for whatever reasons, changed the rules through extensive loss of multi-year ice. 2007 began with a lot more multi-year ice than 2008, 2009 and 2010 so the result for 2008 to 2010 must be seen in that context. What would the minimums have been in 2008 to 2010 if they had started wth the same amount of multi-year ice as 2006 did (or as 1979 even)?
Measuring 2008-2010 against the long term average loses some context in this case. Weather conditions that might have had very little impact on a “normal” arctic ice pack will see different results through the course of a melt season on an ice pack that has lost much of its multi-year ice.
For some time we will see only modest increases in the minimum extent due to the lack of multi-year ice and probably still some years that it falls. But I certainly haven’t seen anything in the past 3 years to indicate things are getting worse. Rather there are indications that the ice pack is gradually starting to recover.
BTW looking at current JAXA data, I’m still with Steven on a September minimum of close to 5.5 million sq km which would be 250,000 sq km above the 2009 result. Perhaps more if DMI temp data is right and the temps drop below zero in the Arctic basin within the next week (as they did in 2006).

In terms of the stratification of temperatures in the Arctic, this is the cross-section across the whole Arctic produced by the US Navy (Aug 1, 2000).
The temperatures are around -2.0C for the first 100M or so. If the ice is only 2 or 3 metres thick, the temperature below this is really just above the melting point of salty ocean water so would be -1.9C for example. Remember the least dense form of ocean water is frozen sea ice, it will always float, and then the next least dense water will be the warmest layer of water available. After that, the colder the water, as long as it stays above the transitional freezing temperature, will be the most dense and will sink to the bottom (or more accurately the temperature profile will reflect the density).
So below 100M, the temperature then rises to reach a peak of about 0.6C or so and this goes down to about 1000M. Below that, it gets colder and the coldest water is roughly at the bottom at 4000M, and is about -0.8C (Antarctic waters are similar but they don’t get above 0.0C in the warm 100M to 1000M layer).
There is a seasonality trend as well where the whole profile gets just slightly colder in the winter months. This cold dense -0.8C water will seek out the lowest levels possible and thus will actually flow out of the Arctic (and Antarctica) wherever the geography of the ocean bottom allows it and will become the deepest water in the deep oceans. (This also means the deep ocean will not really warm up until there is no sea ice anymore).
http://a.imageshack.us/img62/7945/arcticdeepoceantemp.gif

RockyRoad
The Arctic Ocean is stratified with layers of different densities due to the amount of dissolved salts. The more dense water stays near the bottom, and the less dense water stays near the top.

Eric Flesch says:
August 5, 2010 at 6:50 pm
The phenomenon of limb foreshortening is well understood, being practiced daily by those who measure sunspot area.
eg: g 2010 08 04 14 27 32 11093 15 137 23 222 10.35 353.40
where sunspot 11093 was near the limb (foreshortened) and was measured in appearance as 15 Umbra and 137 whole spot.
Upon correction for foreshortening, the umbra was calculated at 23 and whole spot at 222 (if seen from directly above, which it was not).
Therefore, the same pixel that represents X area at the north pole is going to represent much greater area as the radius from the pole increases, while the image is centered above the pole.
A simple empirical experiment with a sphere and a marking pen will suffice to drive the point home.

I have seen little discussion of the relationship between the Beaufort Gyre and the 2007 record Arctic sea ice minimum. It seems possible that the record ice loss was connected to a breakdown in the Beaufort Gyre – whether an anomaly or something that is a 30/40/50 year cycle remains little known. It takes approximately 4 years for ice normally to circulate completely in the Beaufort Gyre – and the map above of multiyear ice suggests that the Beaufort Gyre is well along in reestablishing multiyear ice on its circumference – and will complete the cycle in 2011. Once that occurs, it would seem probable that the enclosed area within the gyre will see a resurgence of multiyear ice, as well.

Steven Goddard:
“Billy Liar
Please propose a physical mechanism whereby the water heats underneath compact ice.”
Apparently you haven’t figured out that liquid ocean water is WARMER than ice. The NATURAL FLOW OF HEAT is from warmer to colder. That is, water HEATS ICE!!! The current continually brings water from outside the area against the edges of the ice and underneath, or, are you going to deny than ANY ice melts every year??
Come on man, get REAL!!

Steve, you seem to be assuming that the NSIDC polar map is as a photo taken directly above the North Pole at an infinite distance, so that longitudes are displayed as shorter away from the pole. I would assume it to be a polar projection with static longitude scale, i.e., if it was extended to the equator, the equatorial lands would be far too large.
Even if you are right, that does not explain the discrepancy. At 75°a polar projection with static longitude scale will only introduce a small error (cos (15°) = 0.966). The difference noted is far greater. Those running NSIDC can presumably do elementary trigonometry and should not be using this as an excuse.

NSIDC predictions and analysis true or false? In less than one month the actual facts will be known, the minimum is near now and all the projections and predictions and models will be tested by actual reality.
Who will be right and who will be proven wrong, who was accurate and who was guided by ideology. NOAA has a north pole cam taking pictures and they used to have an external temperature reading which has disappeared, the website although still alarmist in nature seems to have been edited and information that was available at the start of the melt season is no longer there(can anyone confirm?).
I do not believe that NOAA and NSIDC would stoop as low as to deceive and lie and cheat, yes they might garnish the truth and allow themselves to view reality through the lens of their beliefs but outright fraud? I think not.
However, IF their corporate predictions are proven to be wildly inaccurate and misleading then as with the UK met office their reputation will take a battering, a battering which will take some time to repair. I find it hard to believe that the NSIDC would throw away their reputation by fiddling a graph which may be proven wrong in such a short time.

rbateman
A photograph of the earth taken from above the north pole at a long distance away would approximate the distortion as 1/sin(latitude.)
At 75N, the distortion would be only 3%. At 30N, the distortion would be 200%. At 1N the distortion would be 5700%.

My apologies to Eric Flesch — I misspelled his name on my previous post.

JAXA shows 2010 10% higher than 2007. The same number I measure from NSIDC maps.
So what’s up with the NSIDC graph?

rbateman says:
August 5, 2010 at 9:20 pm
Eric Flesch says:
August 5, 2010 at 6:50 pm
The phenomenon of limb foreshortening is well understood, being practiced daily by those who measure sunspot area.
eg: g 2010 08 04 14 27 32 11093 15 137 23 222 10.35 353.40
where sunspot 11093 was near the limb (foreshortened) and was measured in appearance as 15 Umbra and 137 whole spot.
Upon correction for foreshortening, the umbra was calculated at 23 and whole spot at 222 (if seen from directly above, which it was not).
Therefore, the same pixel that represents X area at the north pole is going to represent much greater area as the radius from the pole increases, while the image is centered above the pole.
A simple empirical experiment with a sphere and a marking pen will suffice to drive the point home.
==============================
Repeated here for effect.

RockyRoad says:
So you’re saying the cold, dense, heavy water stays in the upper level, while the warm, expansive, lighter water stays in the lower levels? And both are moving subject to tides and earth’s rotation yet stay in equilibrium? Does not your model defy that common force called gravity? All it takes is one small vertical perterbation, and the two horizons start to trade places. After that, there’s no stopping them. Admittedly I have no data to support my idea yet logic can be a magnificient, overpowering force.
Water under 4 deg C becomes LESS dense as it approaches 0 deg C, sea water with it’s high salinity goes a few deg below 0 before freezing, as well. This means that the LIGHTER coldest water will naturally float on top, near the ice.

I’ve also looked at the saltiness issue. It doesn’t make much difference at all, there is no stratification according to salt content, it is all close enough to the same that it makes little difference. Temperature is, by far, the dominant factor.
Ocean water is going to have a minimum salt content and, after a certain threshold is reached, it no longer becomes an issue. There are larger differentials where fresh water is strongly flowing into the ocean as in the Amazon outflow, but other than this particular area, saltiness does not have to be taken into account.

There is no longer a large Antarctic area/extent
http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/seaice.recent.antarctic.png
Also, I wonder if the northern route on the NW passage will be open this year?
http://ice-glaces.ec.gc.ca/Ice_Can/CMMBCTCA.gif
Andy

Eric Flesch says:
August 5, 2010 at 8:02 pm
Steve, you seem to be assuming that the NSIDC polar map is as a photo taken directly above the North Pole at an infinite distance, so that longitudes are displayed as shorter away from the pole. I would assume it to be a polar projection with static longitude scale, i.e., if it was extended to the equator, the equatorial lands would be far too large. If my reading is right, you are weighting the pixels in the wrong direction, thus doubling the error beyond if you did not weight them at all. That would indeed account for the gap from 3% to 7.5%, so I suspect this is what is happening.
—–
I was not sure of NSIDC’s map and whether it was the same as the map in the post, so I compared the one in the post to the NSIDC map found on the Sea Ice Page by overlaying and aligning them on Photoshop and they are essentially the same. (I assume that Steve is using this NSIDC map for his pixel counting.) Next, I compared the NSIDC map to Google earth by carrying a print screen of the approximate Google Earth center axis that matches the presentation axis of the earth for the NSIDC map (it isn’t the earth’s axis) and overlaid the NSIDC map. The coastlines pretty much matched up.
Now, with earth in the same orientation in Google as the NSIDC, I measured the distance between latitudes using the ruler tool and no matter which two I chose, whether from 90 to 85 through 55 to 50, I was essentially given the same distance — approx 345 miles.
From this, it’s pretty certain that a pixel nearest the axis of the NSIDC map is going to represent a smaller — and more accurate — area of the earth than a pixel at the furthest reaches of the map. Now I can probably investigate further and determine a a table of adjustment factors for pixels counted between certain latitudinal intervals to adjust for the fact the NSIDC map is not equal projection, as NSIDC stated, but I’d have to brush up with books I haven’t used in 35 years, so suffice it to say a pixel at the furthest reaches from the presentation axis of the earth used on the NSIDC map represent a larger portion of the earth’s surface than ones next to the axis.
If my original assumption is correct, I’d also note that I’ve seen, over time, many comments arguing the area differences based on the earth’s axis as the general starting point reference for accurate representation of area. This is not applicable with NSIDC, since the earth’s axis is not their map center so the pixels nearer, say, the Chukchi Sea is further out of proportion by an axis shift of at least 700 miles or about 10 degrees of latitude.
This increasing area representation might be better understood if one envisions a globe of the earth with a sheet of paper with a grid of pixels perpendicular to any axis of the globe you choose. The area of the earth each pixel represents is defined by extending the rectangle vertically down to where it intersects the globe. As you move away from the axis on the paper the area represented by each pixel increases.

AndyW says:
August 5, 2010 at 10:55 pm
Better have a 2nd look at that graphic of Antarctic Sea Ice Anomaly.
The average is the brown number. The present measurement is the blue number.
The present measurement is greater than the average, or positive.
The red number is the positive anomaly reading, currently rising once again at 0.893M km2.

stevengoddard says:
“Pixels near the pole are perpendicular to the “camera” – so are seen as their correct area. As you move towards the equator, the earth curves away, each pixel area covers a much larger area of the earth’s surface.
So pixels further from the pole represent have to be weighted more than pixels near the pole.”
Here is a link to the projection used for the NSIDC sea ice data:
http://nsidc.org/data/polar_stereo/ps_grids.html
It’s a polar stereographic projection. The projection plane they use is 70 degrees latitude- not the pole!
That means there is no area distortion at 70 degrees. North of 70, areas are compressed (so each pixel covers a bigger area than actual). South of 70, areas are expanded (so each pixel covers a smaller area than actual).
Depending on the exact shape of the ice, and the relative proportions of ice above and below 70 degrees, counting pixels can give you very different results for the exact same ice area.
Dr. M is correct- Counting pixels just does not work.

Chris V
Everyone counts one form of pixels or another – and in fact NSIDC teaches image pixel counting to their students. Pixel distortion is not the cause of the discrepancy.
REPLY: Pixel counting is used in many areas of science where imaging is involved. For example to estimate biomass:
A chromaticity-based technique for estimation of above-ground plant biomass
Tomasel, F.G.1,3*Paruelo, J.M.2,3Abras, G.4Ballarín, V.4Moler, E.4
1Departamento de Física, Facultad de Ingeniería, Universidad National de Mar del Plata, Juan B. Justo 4302, 7600 Mar del Plata, Argentina; 2IFEVA and Cátedra de Ecología, Facultad de Agronomía, Universidad de Buenos Aires, Av. San Martín 4453, 1417 Buenos Aires, Argentina 3Research Scientist of the Consejo National de Investigaciones Científicas y Técnicas (CONICET) 4Departamento de Ingeniería Electrónica, Facultad de Ingeniería, Universidad National de Mar del Plata, Juan B. Justo 4302, 7600 Mar del Plata, Argentina
ABSTRACT
Abstract. This paper presents a new and simple technique to derive quantitative estimates of green or dry biomass using colour information from digital pictures. This pixel-counting technique is based on the association of particular plant material with a representative region on a two-dimensional colour space, and applies to cases of non-overlapping canopies. The efficacy of the method is demonstrated using sets of samples obtained from both field and laboratory studies. It is shown that application of the proposed approach results in a highly linear relationship between pixel count and foliar area for both green and non-green material [r = 0.99 (p < 0.001)]. Analysis of images from a short-grass steppe shows a high correlation between pixel count and measured values of green biomass [r = 0.95 (p < 0.001)]. The method outlined here allows for a substantial improvement in the speed of sample evaluation to estimate biomass both in the field and in the laboratory. It also provides a non-destructive alternative to monitor plant cover and biomass in open canopies.
================
Weather Forecasting System Based on Satellite Imageries Using Neuro-fuzzy Techniques
(3) The Institute of Systems Science, National University of Singapore, 25 Heng Mui Keng Terrace, 119615 Singapore
Abstract
We have built an automated Satellite Images Forecasting System with Neuro-Fuzzy techniques. Firstly, Subtractive Clustering is applied on to a satellite image to extract the locations of the clouds. This is followed by Fuzzy C-Means Clustering which operates on the next satellite image, seeded with the cloud clusters of the previous image. With the matching of cloud clusters across successive images, cloud cluster velocities are deduced. Using a Generalized Regression Neural Network, we interpolate the cloud cluster velocities over the whole area of interest. Finally, the linear forecasting scheme then moves each cloud pixel in that satellite image according to the velocities of the past hour.
============
Navigation-related structural change in the hippocampi of taxi drivers
Image Analysis Method 2: Pixel Counting.
The three-dimensional images from the 16 taxi drivers and a precisely age-matched sample of 16 normal controls taken from the 50 used in the VBM analysis were submitted for region-of-interest-based volumetric measurement of both hippocampi by using a well established pixel-counting technique (12, 13).
======================
In all cases, the imager provided numeric data that became an image, the image was then pixel counted. It is no different with a satellite data image. Pixel counting is an accepted technique in many areas of science, engineering, and even in manufacturing product quality control. – Anthony

There is a lot more ice in 2010 than 2007. Period.
http://climateinsiders.files.wordpress.com/2010/08/nsidcaugust032010vs2007.png

“Sadly, DMI stopped updating their graphs two days ago – so I am no longer able top do comparisons between DMI 30% concentration and NSIDC 15% concentration” – I sent a mail to DMI this morning, asking for the reason why their 2010-plot hasn’t been updated since 3. August(?) No reply yet.

NSIDC data is in a polar stereographic projection…it is true 25km around 70 latitude

Anthony –
Nobody’s arguing against pixel counting – however, one must be sure to do the pixel counting accurately, while keeping units in mind.
In essence, Steve is reporting a 7% difference in the counts of square pixels. But he’s comparing this with the difference from the NSIDC plot which is in square km. So if he were to accurately convert square pixels to square km (not a 1:1 conversion), he will likely find the percent differences from the area projection and NSIDC plot will be much closer.

Isn’t it the water temps that really count?
AFAIK, icesheets melt from the bottom up, rather than from top down.
Do we have a reliable reading of the actual water temperatures in said areas, and can we compare them to the water temps there in former years?

Don’t know if this has been mentioned yet, but at the bottom of the NSIDC August 4th update is a link to the July SEAC update;
“The July Outlook for arctic sea ice extent in September 2010 shows some notable adjustments from the June Outlook, with both downward and upward revisions from last month.
Downward revisions reflect in part rapid ice loss observed during June together with the presence of the Arctic Dipole Anomaly (DA), which promotes clear skies, warm air temperatures, and winds that push ice away from coastal areas and encourages melt. Upward revisions reflect the slowdown of ice loss during the first two weeks of July and a change in atmospheric conditions to cooler, cloudier weather.
The July 2010 Sea Ice Outlook Report is based on a synthesis of 17 individual pan-Arctic estimates using a wide range of methods: statistical, numerical models, comparison with observations and rates of ice loss, composites of several approaches. Two contributors to the outlook represent “public” contributions.
Including all contributors, the individual Outlook values for September 2010 range from 1.0 to 5.6 million square kilometers, with a mean of 4.6 +/- 1.10 million square kilometers. Excluding the outlier of 1.0 million square kilometers by one of the public contributors gives a range of 4.0 to 5.7 million square kilometers, with a mean of 4.8 +/- 0.62 million square kilometers. This is below the 2009 minimum of 5.4 million square kilometers and just slightly above the 2008 minimum of 4.7 million square kilometers. Only three of the Outlook contributions give a value equal to or above the long-term linear trend line (5.6 and 5.7 million square kilometers, respectively). All of the estimates remain significantly below the 1979-2007 average of 6.7 million square kilometers, and six estimates indicate a new record minimum.
The spread of Outlook contributions suggests about a 29% chance of reaching a new September sea ice minimum in 2010 and only an 18% chance of an extent greater than the 2009 minimum (or a return to the long-term trend for summer sea ice loss). 53% of the Outlook contributions suggest the September minimum will remain below 5 million square kilometers, representing a continued trend of declining sea ice extent.”
http://www.arcus.org/search/seaiceoutlook/2010/july

Something fundamental has changed.
http://wattsupwiththat.com/2010/06/01/the-ice-who-came-in-from-the-cold/
The strong annual signal that Willis noted.
As we both agreed, if real, the NULL hypothesis is challenged.

“And no mention of the record high ice extent in Antarctica.”
Why should they? These people have an agenda. The media hardly ever reports of the ‘loss’ in the Arctic then balances it with record ice extent in Antarctica, i.e. nothing to shout about. Globa sea ice is unchanged over the past 30 years. Yet we get manufactured alarm about something that has happened numerous times in the past.

A question for Dr Meier: WHY are your sea ice extent maps not equal area? Doesn’t extent = area?

This is the 5th Arctic post this week, though one was about James Hansen’s Arctic temps and not ice. Close to 700 comments altogether. I suppose if you thought Arctic ice wasn’t worth blogging about you would see by now that people are paying attention to global warming predictions, Arctic ice being one of them.

Sorry, that equation should be 100m*2C*specific heat/latent heat (=100*2*4000/200,000=4).

“Looks like DMi has updated” – indeed. I was informed a couple of hours ago by DMI (Gorm Dybkjær who was previously (favo[u]rably) mentioned here?) that a minor technical ‘glitch’ had been corrected this morning.

The pixel distortion argument is nothing but a distraction and a diversion. The effect is minor at high latitudes and does not account for the large discrepancy. Something else is going on.

Has anyone revised the 1979-2000 average?, could you get an older average?

Steve,
How do you know that – have you calculated it? You may want to brush up on your geometry if you think correction for a geographic projection is not important in a comparison. If you’re not willing to do it, send me your pixel data (lat/long) and I’ll calculate it. By my estimation, the scale factor of pixels at 70 degrees latitude is 3%. There’s your difference right there!
Or you can read about it yourself: http://nsidc.org/data/polar_stereo/ps_grids.html, keeping in mind that:
map scale factor = ( 1 + sinΦ_0 ) / ( 1 + sinΦ )
where Φ_0 is the map projection latitude, and Φ is the latitude of interest.

Matt
Look at the map. There is a lot more ice in 2010 than there was in 2007.
http://climateinsiders.files.wordpress.com/2010/08/nsidcaugust032010vs2007.png?w=510&h=344&h=344

No, but this is science. If you want to claim that there’s something wrong or dishonest about the 3% difference in NSIDC ice extents, you’re going to have to do better than telling me to ‘look at the map and eyeball it.’

Matt
If you believe that pixel distortion is significant, feel free to do the math. You are the one who is concerned about it.

Arctic-ROOS, Jaxa, DMI, and the actual extent maps all agree there is significantly more ice in 2010 than 2007 (on the order of 10%). The NSIDC graph is the outlier amongst all the rest of the data, so clearly it is not a “pixel” argument or “you can’t just tell us to eyeball it”. Its pretty obvious there is a solid consensus amongst all the data with the one lone exception being the NSIDC graph.
However, if you look at today’s NSIDC graph, it appears to be diverging from 2007 again, though still closer than the other data sources.

Thanks Richard M. Much appreciated.

Steve, you aren’t using their raw data, you’re using a processed data product as input.
The bigger distortions probably come from the manipulation that they do to get the raw data onto pixels on that projection. Yes, their data is manipulated in that image. You’ve caught the fact that anyone taking raw data and creating a graphical projection of it like that inherently manipulates the data.
Pixel counting images like this is a stupid trick to try to impress the masses.

Matt
The excess ice in 2010 is (of course) at lower latitudes. So any correction would increase the weighting of those pixels and therefore increase the difference between 2010 and 2007.

stevengoddard says:
August 5, 2010 at 10:37 pm
“Anu,
As usual, you forgot to use my complete quote : “Please propose a physical mechanism whereby the water heats underneath compact ice.”
When the ice is very compact (as it is this summer) it is difficult for the water below it to warm – because no sunlight reaches the water. “
This factoid is one of the reasons that it is so difficult to predict the final outcome of sea ice extent at the end of the melt season. A period of strong winds tends to facilitate the fracturing of the edges of the sheet, while clouds tend to prevent the sea warming. There are too many non-linear processes at work to enable accurate prediction even just a few weeks before the next re-freeze.

Steve –
So if you were pointed out to be wrong yesterday, why are you still continuing to make posts claiming NSIDC is wrong.

Matt
You are going to believe what you want to believe.
I can see that logic is not relevant. Please don’t start a discussion about mathematics if you don’t intend to complete it.

Several posters on this thread have mentioned heat from below, which is mostly radioactive decay. Please allow me to add numbers to this assertion. The planetary thermal gradient is about 25C per kilometer depth in the crust. But the quantity of heat (not temperature) is low. Heat from the bottom of the sea is roughly 0.1 W/m2, which is about 1/10,000 the amount of heat from solar insolation.
Heat from below is tiny, even at the poles where solar insolation is small.

stevengoddard says:
“The excess ice in 2010 is (of course) at lower latitudes. So any correction would increase the weighting of those pixels and therefore increase the difference between 2010 and 2007.”
No- you’ve got it backwards.
Say 1 pixel = 1 km2 at the pole.
With the scale factor (squared, because you’re calculating areas), at 70 degrees, 1.07 pixels = 1 km2. So 1 pixel = .93 km2.
So by assuming all pixels cover the same area (1 km2 in this example), you are OVERESTIMATING ice extent away from the poles, with the amount of overestimation increasing the further away from the poles you get.

Dan in California
Actually, no place on earth receives more summer solar insolation than the poles.
http://earthobservatory.nasa.gov/Features/EnergyBalance/page3.php

As previously alluded to by Gilbert K. Arnold, given that the purpose of the NSIDC map is to show ice area, it seems misguided to use a projection that preserves angles rather than area. Why not use a Lambert azimuthal equal area projection?

I seriously doubt NSIDC has suddenly changed the way they compute ice extent from the ice concentration data. They are a transparent institution who makes their data and their processing codes available. Steve should have downloaded the data and made his own calculations of ice extent if he was concerned processing recently changed. Counting pixels in a non-equal area grid is just plain wrong.

As the total Arctic and Antarctic sea ice extent seems to have remained relatively constant for the last 30 years, I wonder if a term describing the interaction between the two regions like “Trans-Arctic Oscillation” might one day enter the lexicon of Meteorology?

JJB
No method is perfect, but counting pixels is more than adequate for this exercise.

JJB
I am not sure where you are getting your information from, but my summer 2010 prediction is right on track.
In fact, odds are my prediction will turn out to be too conservative.

stevengoddard-
No, there is distortion along both latitude and longitude. There has to be- the stereographic projection is CONFORMAL. It preserves angles; the only way to do that is to have distortion in both directions.
And you don’t understand the concept of map scale.
On a stereographic projection, scale INCREASES away from the center of the map (in this case, the pole).
That means that means that the size of a 100 km2 land mass gets BIGGER ON THE MAP as you go from the pole (90 degrees) to 70.
Since it’s bigger ON THE MAP, it covers more pixels.
But the size on the earth of that 100km2 land mass hasn’t changed- it’s still 100km2.
So more pixels cover the same EARTH AREA at 70, compared to the same earth area at the pole.
That means that each individual pixel at 70 covers a SMALLER area ON THE EARTH, so you have to DIVIDE, not multiply, by the scale factor(squared-because we’re talking area) to correct for the distortion.
Here are a couple of good links that have info on map projections and scale:
http://egsc.usgs.gov/isb/pubs/MapProjections/projections.html#stereographic
https://www.e-education.psu.edu/natureofgeoinfo/

HaroldW
Also, your interpretation of that document is incorrect. A projection that is intended to be accurate at 70 degrees is also going to be fairly accurate at 75 degrees. The equator is in the opposite direction.
The ice is at higher latitudes than 70 degrees, not lower latitudes. Comparisons of distortion at the equator are meaningless to this problem.

stevengoddard says:
“We don’t know what projection they are using and it doesn’t make any difference”
Yes, we do know- I posted a link earlier to the NSIDC website that shows what projection they use. Polar stereographic on the secant at 70 degrees.
And it does make a difference. There is a reason why there are many dozens of different map projections; you choose the projection to based on what you want to use the map for.
If you want to measure areas accurately over large areas of the earth, you need to use an equal area projection- that’s why they exist. The projection used by NSIDC is not equal area, so you can’t use it to accurately measure areas.
And since you’re talking about differences in ice extent of only a few percent, that accuracy is important.
I have no idea how the Aug 2010 ice extent compares to 2007 (and i really don’t care). But I do know that you can’t measure it accurately by counting pixels on a stereographic projection.

Chris V,
You said “Polar stereographic on the secant at 70 degrees”
You do realize that that the ice is north of 70 degrees?

Amino Acids in Meteorites says:
August 5, 2010 at 8:31 pm
“It is unfair to compare a year of data (2010) to a time period that began when Arctic ice was larger from the 1945 to 1976 cooling in the earth. To get a true picture of climate we should see what climate has done for 1000, or more. ”
Its been said ‘Climate is what you expect, weather is what you get.’ It doesn’t take 1000 years to define climate. But… if you object to using a single year (2010) as a comparator, stop using 2007 for min ice extent or 1998 for max global temperatures. And stop using the ice extent on a month-by-month basis to draw any inference at all.

Matt,
The NSIDC graph shows 2010 almost the same as 2007. Look at the overlay of their maps below. Their graph is simply not credible.
[youtube=http://www.youtube.com/watch?v=kvatHj3T-hA]
Their graph does not agree with DMI or JAXA. How can you continue to ignore this blatantly obvious fact?

Sorry, last message was directed at Matt – not Chris V.

stevengoddard says:
“There are quite a few different projections they could be using. You are talking about a different projection from what we were talking about earlier. ”
No- I (and Matt, and Harold…) have been talking about the same projection since the beginning the one NSDIC does use. What map projection were you talking about?
“You said “Polar stereographic on the secant at 70 degrees” You do realize that that the ice is north of 70 degrees?”
Yes I do- that doesn’t change how the scale varies across the map. It just means that distortion is smallest at 70. North of 70, areas are reduced, south of 70, expanded. But there is still 6% distortion between the the pole and 70 degrees, and about 2.2% distortion between 75 and 70 degrees.
Your calculated difference was what, 4.5%? The distortion across the map is greater than that.
“You are claiming that the green region is distorted.”
Not a claim- it’s a fact, based on the map projection. Heck- the entire map (like every map) is distorted.
When choosing a map projection, you (should) choose one that distorts the stuff you’re not interested in.

Now the NSIDC maps show a leveling off in Sea Ice. I’m going to trust DMI from now on, instead of the NSIDC!
http://nsidc.org/data/seaice_index/images/daily_images/N_stddev_timeseries.png

JJB says:
August 6, 2010 at 12:53 pm
“This posting appears to be another attempt to pretend that the ice is not shrinking in summer. The ice in 2010 did not recover as Steve has predicted all year, and in fact it has continued the amount of ice loss seen the last few years. I do not understand all this effort to try to hide the decline.”
LOL. I think your calling things a little early there.
If we match the ice loss of 2006 for the remainder of the season, this years minimum ice extent will beat 2009 by a good margin. The average melt for the remainder of the season based on JAXA data has been as low as 25,000 sq km per day in some years and as high as 50,000 sq km in others. The years that have higher melt rates late in the season do so because they lost their first year ice later so the higher loss rate is seen late in the season.
This year the first year ice was lost very early early. I dont consider conditions particularly favour rapid loss of multi-year ice this year. I’d be expecting a repeat of 2003 or 2006 ice loss rate for the remainder of the melt season bringing this years result above 2009.
Not bad for the “warmest” year on satellite record with an ice pack that lost a lot of multi-year ice in 2007. In fact, a little mystifying if you believe AGW is destroying the Arctic Ice Pack.

…squared to adjust areas…

Steve, NSIDC takes the sea ice concentration data on the polar stereographic grid and derives the ice extent using the “actual” area per pixel. I certainly wouldn’t publish any numbers assuming a constant 25 sq-km per pixel.
The idea that NSIDC is suddenly manipulating the data to show a steeper decline in the 2010 extent is wrong. If you do the extent calculations accurately (i.e. using the actual area per pixel for each pixel which you can get from NSIDC’s ftp site, together with the sea ice concentration data) you will get the values that we record. Any differences in extent between institutions is a result of using different data sources, different algorithms, different thresholds for ice, different weather filters, etc.

Steve Goddard says:
“Those who think the map is distorted at lower latitudes should write to NSIDC and tell them that their 2007 ice loss maps exagerated the magnitude of the loss.”
The map they use isn’t designed to be used for calculating ice areas directly.
When NSIDC calculates ice areas, they calculate each pixels individual area, based on the map scale at that point:
“Each pixel’s area is calculated individually, and is obtained by multiplying the nominal pixel size (625 km²) by the square of the map scale at the center of the pixel. Pixel areas range from 382 to 664 square kilometers for the Northern Hemisphere and 443 to 664 square kilometers for the Southern Hemisphere, under the polar stereographic projection and grid used for the input data sets.”
From here: http://nsidc.org/data/docs/noaa/g02135_seaice_index/index.html#doc_info
NSIDC does it the correct way.

JJB says:
August 6, 2010 at 12:53 pm
This posting appears to be another attempt to pretend that the ice is not shrinking in summer.
Really? Wow, that’s wild, man!!

muographer says:
August 6, 2010 at 3:03 pm
if you object to using a single year (2010) as a comparator, stop using 2007 for min ice extent or 1998 for max global temperatures. And stop using the ice extent on a month-by-month basis to draw any inference at all.
Ok. I wasn’t using them for the things you say anyway.

Marge,
You know, I may be misunderstanding what you are saying.

Julienne
No one said anything about “manipulating” data.
But there is a large difference in ice extent vs. 2007 which is not reflected in NSIDC graphs, but is seen in the DMI and JAXA graphs.

R. Gates, yes that certainly is possible, especially given all the open water areas in the interior of the pack. Ice divergence tends to cause more openings within the ice pack as the ice spreads out. Having thin ice also doesn’t help. In 2008 ice loss in August was the fastest recorded during the satellite data record, reflecting the fact that the ice cover started to run out of thickness. The same thing may happen again this August…

MJW
NSIDC maps currently show a 10% increase since 2007, which agrees closely with JAXA .
The current NSIDC graph shows about 6%. I expect it will come back into line in a couple of days. What bothered me is that the very low graph made it into the NSIDC newsletter. It gave a wrong impression about the state of Arctic ice.
July had the slowest melt in the JAXA record.

R. Gates
I’m sure you have considered the fact that the sun is now getting low in the sky in the Arctic. It will set at the pole in six weeks.

It would no doubt be best, in the best of all possible worlds, if Steve Goddard weighted his pixel counts to correct for the projection distortion. However, as Julienne Stroeve says, “While this increases the distortion at the poles by six percent, the latitude of 70 degrees was selected so that little or no distortion would occur in the marginal ice zone” (my emphasis). The projection effect cannot possibly account for the difference Steve finds, because the gains and losses are almost entirely within the low distortion zone.

Paul,
Thanks for that. All of the discrepancy lies between 70 and 80N, so there is no reason to believe that pixel distortion has any meaningful effect on the area measurement.

stevengoddard says:
August 7, 2010 at 6:30 am
MJW
July had the slowest melt in the JAXA record.
_______
Steve, why do still insist on equating sea ice extent with total melt? In truth, no one knows what the total melt of Arctic sea ice was in July 2010 (as in the total loss of volume) and we won’t really have a good idea on total melt until CryoSat 2 data starts to come on line. The point being, the sea extent trend is NOT a good gauge of the rate of melt during certain shorter periods, especially when the ice is spreading or diverging from the passage of weather fronts, etc. During these times, the ice is still generally melting just has fast (because much of that melting is coming from the water anyway) and so the total volume is decreasing just as fast, but the divergence makes the sea ice extent not drop so fast, so you get these little bumps. Moreover, the divergence sends more ice over open water leading to a potential acceleration of melt later on.
I think myself and others have spoken to this before, but it seems to be a key point you often miss. Thankfully, with accurate CryoSat 2 data, we might eventually have an accurate total sea ice Volume measurement (not models) and this will be the best guage to see how the ice is actually behaving. These future sea ice volume charts should be far smoother and less prone to the the short term bumps in the sea ice extent that when see when weather systems cross the arctic and cause divergence, even though the sea ice continues to melt at a constant rate during these times driven mainly by heat flux from the water.
Also, in regards to your unisys map for SST’s…I prefer the much higher fidelity and accuracy of the NOAA chart and hold by my statemennt that the majority of Arctic SST’s are warmer than average this year, have been for several months, and have shown a longer term (multi-year) trend of increasing (due mainly of course to more open water for longer periods of time in the summer months).

JAXA is taking a turn that is bad news for the rotted/alarming thinning hypothesis. What is happening in the real world has always been bad news for ‘global warming’ because the real world isn’t doing what global warming says it will do.
So, how many are now ready to conclude the rotted/alarming thinning hypothesis is wrong?
Or are going to continue arguing about pixels, and the NW Passage, and warming currents that don’t exist, and frozen or not frozen melt ponds on top of the ice, and if Barrow had an open shore, and DMi accurate cooling data vs. James Hansen’s GISS fudged warming, and if PIPS is really used by the Navy or not, etc., etc., etc.?

R Gates,
Any place that “normally” has ice cover, but doesn’t this year is of course going to have a positive SST anomaly. That tells you nothing about the water temperature under the remaining ice.

The projection effect cannot possibly account for the difference Steve finds, because the gains and losses are almost entirely within the low distortion zone.
I’ve been arguing the distortion effect is low, but it certainly is not low for this reason.
Any percentage changes are measured as a percentage of the whole. If the count total for the whole is artificially low – which it is – then the percentage change will be wrong. Even if the area being counted is the part with no distortion.
Rather the effect is low because the portion most distorted – the pole at 6% – is also the area with the least volume.

R. Gates
Cryosat won’t tell you anything about historical ice thickness. There will be no frame of reference to compare against, but I am sure that won’t stop people from making things up.

R. Gates says:
August 7, 2010 at 2:33
I would suspect that you’d gladly get many a skeptic to take you up on making that conclusion…which means they know nothing about the long term trends or dynamics of sea ice.
Ya, I know, not right now, but 40 years from now…..look out, the global warming monster will come out from under the bed.
By the way R. Gates, should anyone listen to you since your prediction about Arctic ice this year is so bad? You insisted for weeks that there was going to be dramatic ice loss in August. But the data in the shows you are wrong. With that in mind, should we trust your long term predictions when your short term ones are so bad?

I don’t see a long term trend of ice loss. I do see a trend of ice loss since 1979. But that is not a long term trend. It is a short term. To just use data from 1979 to 2010 to establish what is normal in Arctic ice is bad science. We cannot establish anything from such a short data set.
But some are doing bad science and are thinking they have established what is normal in Arctic ice from that small data set.

No, R Gates
you said there was going to be dramatic ice loss in August. But, as I thought would happen, you change what you said now. You revise, as always happens in ‘global warming’. People don’t have to look very far to find your predictions of dramatic ice loss in August.

stevengoddard says:
August 7, 2010 at 6:45 pm
R. Gates
What possible logic could there be behind your forecast of 2.5 million?
REPLY: Yes, I agree. Even NSIDC thinks there will be at least 4 million, probably 5 million. I had an email exchange with Walt Meier that confirms this. 2.5 million is just unsupportable. – Anthony
________________
I highly respect Walt Meier, and appreciate your private exchange with him, but I stand by my forecast of a 2.5 million sq. km. minimum during some summer by the 2015 Sept. minimum. (I hope no one thought I was talking about this summer, as I’ve never made that forecast). I’m hardly an expert like Walt nor like Julienne, but I’ve got what I feel are very good reasons for my longer range forecast. I’ve stated for many months that I thought the long and deep solar minimum of 2008-2009 had the effect on Arctic Sea ice of merely slowing down and leveling the accelerated decline we saw in 2007, but hardly mounting any sort of recovery. I feel quite strongly that upcoming solar max in 2013, combined with a what I think will be a strong El Nino in 2012-2013, will combine the general global warming trend of the last few decades to push the sea ice down to this level. I also don’t see any let up in warmer Arctic SST’s, and I also see the positive feedback increasing Arctic Dipole events combining with the above mentioned solar and ENSO mentioned events combining to push the sea ice down to the 2.5 million sq. km. range, somewhere in the summer of 2012, 2013, 2014, or 2015. Exactly which year will be subject to the actual weather of those summers, but I have confidence it will happen.
Just as shocked as everyone was initially at the low extent in 2007, I think this new low will be equally shocking (at least for some), but will set the intial stage for an ice free summer Arctic by 2030, though it could linger in a virtually ice free zone (1 million sq. km. minimums) for a period of time in the 2020’s.

R. Gates
Arctic ice has been tracking a little below 2006, just as I said it would in May and have said several times a week since.

stevengoddard says:
August 7, 2010 at 9:45 pm
R. Gates
Arctic ice has been tracking a little below 2006, just as I said it would in May and have said several times a week since
_________
So you are still good with your 5.5 million sq. km. minimum extent though, right?

Amino Acids in Meteorites says:
August 7, 2010 at 6:09 pm
“I don’t see a long term trend of ice loss. I do see a trend of ice loss since 1979. But that is not a long term trend. It is a short term. To just use data from 1979 to 2010 to establish what is normal in Arctic ice is bad science. We cannot establish anything from such a short data set.”
***
Well some can, how about “Global cooling since 1998”
****
But some are doing bad science and are thinking they have established what is normal in Arctic ice from that small data set.
***
Bad science? Take it up with Dr WHO, borrow his time machine and launch satelites from say 1800? Further back? Your choice and may I suggest using the latest version Cryosat. Let me know when its done, can’t wait to see a few hundred years of arctic ice minimums and maximums. What a trend that will be.

R. Gates
Have I given any reason to believe that I have changed my forecast?

Steve,
I just did some calculations of my own using the binary sea ice data available on the NSIDC website, and from a simple pixel count, I get the same % difference between 03 August 2007 and 2010 as you’ve calculated from the NSIDC timeseries — 3%. Not sure where you’re getting your 7.5% from.
Here’s the numbers I got from a simple pixel count (pixels with concentration > 15%):
2007 = 9578 pixels
2010 = 9895 pixels
I can send the code if anyone wants to verify.
Conclusion: NSIDC plots are correct. There must be an error in Steve’s pixel count.