In my last post I made a mistake in processing the data files through excel. I mixed the area and anomaly data around and accidentally posted an Ice Area anomaly. Something I have never seen before but is likely around.
While I am entirely at fault, my previous post isn’t entirely wrong because I was using the Cryosphere page from University of Illinois which shows Area anomaly below but shows the extent curves. Click on the picture to go to the cryosphere page.
This is the hemispheric ice area plot from UIUC.
Here is my bootstrap National Snow Ice Data Center plot of ice AREA.
There is a small difference between these graphs. If you look at the first year on record, the peak maximum in the NSIDC data (graph above) is lower than subsequent peaks. This is differennt from the peaks in the IUIC graph which have been adjusted to have a stronger downslope. I will request the data for this plot from the IUIC guys again but it clearly is another adjustment in favor of AGW,
Well what’s the difference between extent and area. From the NDSIC web page.
total ice extent is computed by summing the number of pixels with at least 15 percent ice concentration multiplied by the area per pixel.
Total ice-covered area is defined as the area of each pixel with at least 15 percent ice concentration multiplied by the ice fraction in the pixel (0.15-1.00).
So Area includes the amount of ice in each pixel, extent just includes the area of pixels with Ice. It seems to me that Area would be a better measure of the amount of ice in the hemisphere but the extent anomaly (which has a much stronger downslope) is often used.
There is a big difference between bootstrap and nasateam algorithms in total ice area but as far as extent goes they are almost exactly the same. So to keep this post shorter I will compare just the more modern bootstrap data recommended for publications.
Bootstrap ice extent, compare to ice area above.
Bootstrap area anomaly and extent anomaly.
Compare these to the anomaly from Cryosphere
A pretty close match for ice extent anomaly yet it is placed next to the AREA graphs. Its odd that they chose not to show the ice area anomaly and instead chose the extent anomaly. The titles are messed up also (1978-2000) so it might just be a bit sloppy but it led to some confusion on my part for sure.
It leaves me with some more questions though.
What is happening to total ice area on earth vs extent?
Wouldn’t area be a superior determination of ice melt trends than extent?
the Air Vent 
Jeff, one explanation I’ve heard is that “thin” ice can freeze and melt so rapidly that it is not a true harbinger of anything more than very short-term conditions, whereas using an ice extent will be a little more stable on a day-to-day and season vs. season basis. I think both have value, but have no doubt that there is also some cherry-picking going on.
I have contended, as I’ve seen the last 2-3 years cycle in the Arctic, that what we are seeing would be perfectly normal in an environment in the initial stages of cooling. There is little doubt that temperatures have increased in the Arctic and ice extent has trended down over time. But when the 2007 minimum recovered last winter, to me it seemed very logical that this “new” ice would melt quicker than realtively thicker “old” ice, but while some were saying that this still spelled out global warming, I disagee. That new ice extent, by virtue of its recovery, still extended the season. New or old, it takes energy to melt the ice, and the energy required at melt is substantial. This, overall, could be expected to lower heat energy melting as much ice in 2007, and that’s what we saw happen. We have now seen an even greater recovery in Arctic ice extent than in 2007, and I suspect we will probably see the minimum in 2009 still fall well below average, but higher than 2008. If 2009 winter ice then recovers, I think we’ll see some very strong ice in the following years.
But what do I know? I just use basic physics and common sense, and I’m an actuary. Surely, the experts have all the answers.
Still up eh?
I have heard he same thing about thin ice. There is something interesting going on though. I think, without finishing my calcs, that there has been some serious cherry picking going on here. I shouldn’t say things like that but some of the data looks pretty interesting. I plan to do some unique calcs and plotting tomorrow.
The sharp downslope in ice extent this summer, which led to all kinds of headlines about fastest ever loss of ice was followed by the fastest ever increase (I have read but not proven yet) in ice. One theory put forth was that the ice had fragmented into millions of blocks and therefore refroze quickly. I read a couple of days ago that the microwave sensors have difficulty detecting ice with water pooled on the surface, this would also cause the loss of area effect and still be consistent with AGW.
Jeff, an interesting read on Watts today on the Nansen data adjustments to ice extent. Check it out.
Jeff, The abrupt step in your area anomaly data plot around 1987 is not explained by the difference in area and extent. The extent data anomaly does not have that step change. I can’t imagine the void fraction factor that corrects one data set to the other would not make such an abrupt step from one decade to the next. There is little difference between the two curve afte 1990; why should there be such a consistent 0.5 m km^2 offset prior to 1887?
I noticed that the raw data plots of area and extent are “missing” data (omitted plot lines) around 1987 or 1988. Could the missing data be biasing your analysis in some way?
I think the anomaly tracks pretty well with the peaks of the trend in Graph 3. Am I missing your point? The missing data is removed by the software so it shouldn’t have an effect.
Jeff,
Your extent plot matches the CT extent plot well and the local peaks and valleys between the area and extent look fine. I’m looking at the area anomaly plot from 1978 to 1987 as compared to the extent plot. That entire decade of data appears to be offset down about .5 m km^2 from where you would expect it to be, using the remainder of the plot as a basis of the area to extent comparison.
This may be an artifact of the data; I see that it was from three different satellites (?) or data sets. Maybe different methodologies?
If you could plot the value of the difference between the anomalies, or a comparison of the area vs extent data I would expect the offset to be evident. (E-A)/E would be a comparison of the void fraction. My eyeball says that the first decade is markedly different from the second and third decade. What natrual phenomena would create such a discrepancy?
Or maybe I’m just crazy and forgot how to compare data plots!?
until the end of 1987 the daily data is not daily, but only for every second day. How did you handle that? Is that the problem?
Reply: I am having difficulty with wordpress. It seems to be confused about this post and I am unable to comment.
If we can take this discussion to my other post which is on the southern hemisphere I can continue to answer.
Jeff Id
https://noconsensus.wordpress.com/2008/12/14/sh-sea-ice-data-and-anomaly/
Hi Jeff,
Is there a data set for the Northern hemisphere Sea Ice Anomaly 1978 – 2009 that appears on Cryosphere that is easily accessible?
Kind regards,
Ian Edmonds
There’s a link on my latest ice post which can get you there. You have to be careful, there is a step in the area time series which is not corrected. The extent timeseries doesn’t have the problem.