No time to write much but these plots are of the Northern Hemisphere sea ice above and below 72 latitude. The split between the area is 50/50. My anomaly timeframe is from about 1988 to present, I think I will address that in future posts.
Reader Chuckles linked to a register article which linked to a nasa website on a new multi-year sea ice paper by Comiso, a coauthor of Steig 09. It is worth checking out. I think we should take some time to look at the differences in the future. There are some interesting graphics at the site though which are worth checking out. He is reporting a perrenial ice area loss in the 15%/decade range. I don’t have the north of 72 total ice area handy but it is approximately 6 million km^2 so we are looking at a total loss of about 274,000/6,000,000 = 4.6% trend loss for north of 72 degrees. A far lower number than he is reporting. His numbers may be quite accurate but that would mean that there is a balancing gain in annual ice area for that same timeperiod.
If someone has access to the paper, please send it to me. We already reported that global annual sea ice trends didn’t reach 95% significance. His result would likely mean demonstration of no trend in global annual ice, offset by a decrease in perennial ice.
When someone sorts just the multi-year stuff, shouldn’t we be able to sort the annual and report its trends?
Global annually melting ice:
the Air Vent 
Thanks, Jeff, for more details on Earth’s sea ice.
Readers may be interested in this new NASA composite photo of our beautiful blue planet.
https://skepticalswedishscientists.wordpress.com/2012/02/29/beautiful-blue-ball/
I was unable to identify any sea ice – perhaps because sea-ice and white clouds look alike from space.
Jeff,
There is someting odd about your first plot which shows virtually no melt in summer (or freeze in winter for that matter) for the period 1996 to 2005. This period compares to the large oscillation for years on either side. Either it is a data thingy (?due to different satellites) or there was a profound change in weather patterns. Just curious
And yet more thoughts on such things
http://www.theregister.co.uk/2012/03/02/bromine_explosions_and_surface_mercury/
Greg,
It is an anomaly plot based on 1988 – present. This means that the shape of the average seasonal ice is removed. Since the average is likely closer to the middle of the anomaly range as you shift from one weather pattern to the next, you can get things like a flattened central region compared to the ends.
Jeff,
Seriously?
In the NASA article it says: “Multi-year ice “extent” – which includes all areas of the Arctic Ocean where multi-year ice covers at least 15 percent of the ocean surface – is diminishing at a rate of -15.1 percent per decade, the study found.”
You say:
“274,000/6,000,000 = 4.6% trend loss for north of 72 degrees.”
Can you not see your very basic misunderstanding? You have just assumed that all of the ice north of 72N is multiyear…It isn’t. There’s an awful lot of single year ice up there.
Nathan,
Not a regular reader I see.
We have spent quite a bit of time looking at the ice melt by latitude. It turns out that ice below 72 degrees is absolutely single year ice. Stuff North of that is both as you correctly point out.
All I did was note the difference which is greater than I expected it would be. You probably will like this post which explains better.
https://noconsensus.wordpress.com/2012/02/28/single-year-global-sea-ice-shows-minimal-trend/
Jeff,
You still made an error, regardless of how often I read here.
You claim “A far lower number than he is reporting.”
It is because he is talking anout multi-year ice. All you calculated was the total decadal loss. They’re different and incomparable things.
“It turns out that ice below 72 degrees is absolutely single year ice. ”
Ice that forms there may be, but you cannot claim that all ice below 72 degrees is single year. Ice is transported out of the Arctic Basin that includes multiyear ice.
Read the link Nathan. I have shown you that the ‘basically’ all ice goes by-by every year below 72 North latitude all the way to the Antarctic pole. There are tiny patches in the Antarctic which survive from one year to the next but not much.
Not my fault. If you don’t like it, call god.
As to the ‘error’, I don’t see it. You seemed to think I was saying my number was identical. Whereas, in recognizing the difference between the results, if said above that if I add in comiso’s new single year ice which replaced the perennial, I bet the trend in global single year ice may even be positive – though probably not statistically significant. I was pretty clear about that I thought.
Jeff… I don’t know why you think what you have done is proof that all the ice south of 72N is 1 year old. Multi year ice travels south, blown by the wind. So the ice that melts south of 72N INCLUDES multi year ice that has been moved south.
The ice is not stationary, so as 1st year ice melts from 72N south, it gets replaced by multiyear ice that gets blown south. Your melt number for south of 72N INCLUDES multiyear ice.
As to the statement “A far lower number than he is reporting.”
The point I was making is that this is meaningless. You cannot compare them as they are measurements of different things. It’s like saying my 6 oranges is less than your 7 apples. The number 6 is less than 7, but it doesn’t mean anything in that context.
Nathan,
You are not reading. Basically all of the ice south of 72 degrees melts every single year. Gone, byby, nada, near-zip.
Very few patches are left in the peak melt.
That’s it. I differentiated the two results in my post above in that I would like to use Comiso’s result data to isolate trends in annual vs perennial ice completely.
As to the comparison being meaningless, that is your own interpretation. I don’t think it is. I think it is an important fact that demonstrates annual ice isn’t declining nearly as quickly as multi-year. In fact, globally, annual ice probably isn’t declining at all. Weird eh?
You can still run around with your sky is falling sandwich board on though. Nobody said anything about that.
Jeff
You are not thinking
“Basically all of the ice south of 72 degrees melts every single year.”
Yes, but that includes Multiyear ice that moves into that region.
“I differentiated the two results in my post above in that I would like to use Comiso’s result data to isolate trends in annual vs perennial ice completely. ”
BUT your data includes multi-year ice that moves into the region. Your data is not just first year ice.
Jeff
“In fact, globally, annual ice probably isn’t declining at all.”
The ACTUAL issue is ‘Arctic Amplification’ – nothing to do with Antarctic sea ice, which is not expected to change any time soon…
“I think it is an important fact that demonstrates annual ice isn’t declining nearly as quickly as multi-year.”
BUT it will change the Albedo of the Arctic.
“You can still run around with your sky is falling sandwich board on though.” What is wrong with you Jeff?
“BUT your data includes multi-year ice that moves into the region. Your data is not just first year ice.”
None of that ice lasts through the summer Nathan and I don’t see why your differentiation matters in any way at all. I compared ice which vanishes every year to a region of ice which doesn’t.
“The ACTUAL issue is ‘Arctic Amplification’ – nothing to do with Antarctic sea ice, which is not expected to change any time soon… ”
I do love people who repeat the latest global doom genuflection. You can have your opinion and preach the party line, but I’m very much unconvinced that it is ANY issue at all. The “actual” issue you repeat came about when it was “discovered” that the Antarctic wouldn’t cooperate with models. That’s what makes you sandwich board guys so much fun. Don’t you ever wonder just how much albedo change the Arctic experiences when the solar angle is so little?
I also wonder if you guys ever realize that warm is good, ’tis cold that is bad.
And you are the one asking what is wrong with me!
Jeff
“Don’t you ever wonder just how much albedo change the Arctic experiences when the solar angle is so little? ”
so calculate it! Why is it that the Arctic shows the greatest warming anywhere? Because it is experiencing the greatest albedo change.
You stated that the ice south of 72N is one year ice. It isn’t. It is all sorts of ice.
“I don’t see why your differentiation matters in any way at all.” Because it’s the case you made. You decided to compute the annual loss of single year ice, and your calculations don’t do that. They mix single and multi-year ice.
The problem Jeff is that your analysis is stupid and pointless. I don;t have to point to the Blogger who has already showed you why.
“I also wonder if you guys ever realize that warm is good, ’tis cold that is bad.”
This is also a dumb thing to say. I live in a pretty hot part of the world, making is hotter won’t make it better.
Jeff, it appears Nathan was the first to comment on Tammy’s blog re your findings. From that I conclude Nathan is a sock puppet for Tamino/Grant Foster or he tipped him off to your article.
If you care: http://tamino.wordpress.com/2012/03/02/how-fake-skeptics-fool-themselves/#comments
CoRev,
Thanks for pointing that out. I don’t read Grant’s blog. It should be fun to check out later
Nathan,
“I don;t have to point to the Blogger who has already showed you why.”
Actually, you do. I don’t read Tamino’s blog because he is as dishonest as they come. I will get some time later.
Edit note;
“all ice goes by-by ” That would be “bye-bye”. But definitely not “buy-buy”.
‘Bye now!
😉
It’s amazing what “believers” like Nathan spout. He said: ““I also wonder if you guys ever realize that warm is good, ’tis cold that is bad.”
This is also a dumb thing to say. I live in a pretty hot part of the world, making is hotter won’t make it better.”
After saying this: “so calculate it! Why is it that the Arctic shows the greatest warming anywhere? Because it is experiencing the greatest albedo change.”
Albedo is causing the Arctic warming? Near equatorial temps will rise to dangerous levels? Is this what you really meant to say, Nathan?
Thanks Brian.
CoRev
Nathan says a lot of stuff which doesn’t make sense. His comment on albedo shows a lack of reading.
Equatorial parts of the world aren’t expected to warm much from what I understand. Polar amplification and all that. It would be a sorry thing if the desert dwelling ancestors of Nathan had to move north in a hundred years though because of a 1C change.
Funny stuff at Tamino’s thread. I left this, which will absolutely be clipped – he always does.
What a riot. You guys are a lot of fun.
The 72 degree mark was the northernmost divider for non-annual ice. It
was identified by using data. Shame that. As you know, but managed to
fail to point out, the Antarctic ice melts almost completely every
year. Adding them together shows a picture of how regions which
don’t support multi-year ice are reacting to global warming. In other
words – Most of the ice on Earth. I found it interesting to see a
minimal trend and concluded nothing much from it. As I told Nathan,
nobody is taking away your end-of-the-world sandwich boards gentlemen.
I already believe the ice melt is vastly over-hyped and this didn’t
change my opinions at all. Like warming, it hurts nothing. Unlike
this group, I’m not an extremist bent on shutting down capitalism in a
fake effort to change the weather.
Recently Comiso came out with a paper which manages to divide
multi-year ice out by itself. Considering that this is a very small
fraction of Earth’s average ice, I wonder if that somehow makes sense
to you?
I thought so.
I honestly don’t understand what point Nathan is trying to make with his statement in #5 :
<
You have clearly indicated that conditions below 72N are such that ALL ice melts completely each summer. It apparently does not matter what type of ice that is. Only first year ice forms in the winter in that region.
It seems to me that the fact that other ice may float in to die from elsewhere is irrelevant to any calculation involving the change in multi-year ice because the loss of that ice has already been accounted for in the region from which it came. To count it again in its new location would be double-dipping.
Tamino didn’t take my comment. Surprise!!
Perhaps I’ll have to pound on him a little then.
Jeff, my bet is that he will take it. He is just mulling over his response before he posts it.
LL,
He did let it through but didn’t reply. I suppose we can just wait to be enlightened further.
Jeff,
You are still saying that all the melt south of 72N is single year ice. This is simply not true.
“Adding them together shows a picture of how regions which don’t support multi-year ice are reacting to global warming.”
How does it do this? This is the crux of your argument, explain how this analysis shows how this area is reacting to global warming. Why does it matter that this anomaly shows no real change? Your south of 72N anomaly has an input of ice from the north, driven by wind, that will periodically increase the anomaly. It’s not the amount of ice melting.
RomanM
“It seems to me that the fact that other ice may float in to die from elsewhere is irrelevant to any calculation involving the change in multi-year ice because the loss of that ice has already been accounted for in the region from which it came. To count it again in its new location would be double-dipping.”
He’s calculating the anomaly, not just the end of season tally. If the ice moves from one area to the other you HAVE to account for it. Jeff is not calculating how much ice melts, he’s calculating the difference bewteen ice area at any particular time to the average area at that particular time. Given that in the north the flow of ice into his zone from the north is dictated by winds (and is essentially random) and in the south it melts more or less in place and that the poles melt at opposite times of the year it’s no suprise that his ‘anomaly’ shows no trend. Additionally because he’s cut out a large portion of the Arctic the anomaly he calculates will be dominated by the Antarctic signal.
Jeff needs to show why the anomaly south of 72N is superior to the whole of Arctic anomaly.
Jeff
“Equatorial parts of the world aren’t expected to warm much from what I understand. Polar amplification and all that. It would be a sorry thing if the desert dwelling ancestors of Nathan had to move north in a hundred years though because of a 1C change.”
I live at 32S, hardly equatorial. No one will be moving north of it gets hotter. I don’t have any desert dwelling ancestors. I don’t live in a desert… Yet.
“You are still saying that all the melt south of 72N is single year ice. This is simply not true.”
I have repeatedly said nearly all. In which instance am I wrong?
#21 Nathan
If what you say is true and multi-year floats south on the wind on a regular basis, then I don’t see how that would affect the anomaly. Also, if that is a major reason for what is causing the disappearance of the aforesaid ice, then I don’t see how GW is necessarily the cause so it becomes a different issue.
As well, why does Jeff have to “show why the anomaly south of 72N is superior to the whole of Arctic anomaly”? Why is there just a single measure that gives a complete picture of what is going on? Showing what various aspects of a situation look like is often the key to a better understanding of what is going on.
Jeff, I did some calculations in R this afternoon. If no serious errors were made, the total ocean area above 72N is around 9.886 million km2 out of a total global (land and sea) area of 12.510 km2. It is interesting that the Cryosphere Today graphs show a maximum area (not extent) of over 15 million km2 in late 1979. This would require the equivalent of all the available ocean down to about 63N to be frozen over. Was it that extensive? Ther mind boggles…
Apparently, as part of the Tamino gaggle, I’m evil because a minus sign got dropped in a post a week ago. I finally heard about it today and figured out what happened. Read the update here.
http://wp.me/p7y4l-eR2
I find it hilarious they couldn’t figure this out themselves, they were so sure they had me for “purposely” deleting a minus sign.
And still doesn’t change anything. As the authors say there’s no statistically significant trend.
#25 Nathan
Nathan, how are random changes in the flow of multi year ice to south 72 lat going to have any affect on the trend of the time series for either the arctic or global ice south of 72?
In fact, since the total pool of available multi year ice has been diminishing over time, does it not stand to reason that the amount of multi year ice flowing out of the high arctic would also be diminishing over time? In this case the rhetorical effect of Jeff’s graph’s would be enhanced as part of the downtrend south of 72 would be due to diminishing multi year ice flow.
Ah yes! The old delete the minus sign trick.
.”Zat is vhy I have alvays failed vhere others have succeeded”…
Jeff
“I have repeatedly said nearly all. In which instance am I wrong?”
No, you keep saying ‘all’ and you graph is of ‘annual ice’ – it is not the case.
“No, you keep saying ‘all’”
Ok Nathan, you are right.
Let the record show that I hereby replace all instances where I inadvertently said “all” to mean ‘nearly all’.
We good now?
Roman M
“If what you say is true and multi-year floats south on the wind on a regular basis, then I don’t see how that would affect the anomaly. Also, if that is a major reason for what is causing the disappearance of the aforesaid ice, then I don’t see how GW is necessarily the cause so it becomes a different issue.”
Because the rate at which multi year is is diminishing is increasing. So, it is getting transoprted out faster, so it travels into Jeff’s 72N region. It pumps up the anomaly by travelling into that region; something it didn’t do as much in the past.
The reason it is travelling out in greater volume is because there is nothing to lock it in, the ice shelves in the north have all disappeared. it’s far easier now for ice to be moved about by wind because there’s nothing locking it to land and it is thinner.
Jeff it’s not good enough to say ‘nearly all’.
You haven’t bothered to investigate how much IS annual sea ice. All you can do is say that the sea ice anomaly south of 72N hasn’t changed much. but this is a pointless thing to say as you haven’t looked at the inputs into that region.
Nathan,
You have just inserted yourself into the central argument which started this recent effort on sea ice. I made exactly your point to Carrick that ice was flowing out due to weather pattern changes, he stated it was due to warming. I viewed the video endlessly with the intent to plot the flow out of the drains and see if there was going to be something measurably useful. Unfortunately for me, I think Carrick may have been right but I don’t have proof either way. I’m planning to overlay sea surface temperatures on the video as well as wind direction in the hope that something will jump out. If I’m lucky some stats will parse the difference and we can make a conclusion.
I wonder if you can back up your statement that ice is being transported below 72N more than in the past?
Either way, Roman is right, there is no real effect by such a thing as the amount of ice is very, very tiny and gone in the summer anyway.
Nathan,
“You haven’t bothered to investigate how much IS annual sea ice.”
Yes I have.
https://noconsensus.wordpress.com/2012/02/18/warning-anti-science-data-video/
Jeff, the ice is more mobile beacuse it is thinner and not attached to land anymore.
It is warming AND wind patterns. However, you would never had this much mobile ice if the ice had not thinned due to warming.
I’ll find info on the rate of ice moing out of the Basin.
I don’t have access to the paper but:
http://www.agu.org/pubs/crossref/2007/2007GL032043.shtml
It is refenced in this article:
http://nsidc.org/cryosphere/sotc/sea_ice.html
Which also states:
“So the acceleration in the sea ice decline since the mid 1990s may have been partly triggered by the strongly positive AO mode during the preceding years (Rigor et al. 2002 and Rigor and Wallace 2004) that flushed older, thicker ice out of the Arctic, but other factors also played a role.”
and
“Since the mid-1990s, the AO has largely been a neutral or negative phase, and the late 1990s and early 2000s brought a weakening of the Beaufort Gyre. However, the longevity of ice in the gyre began to change as a result of warming along the Alaskan and Siberian coasts. In the past, sea ice in this gyre could remain in the Arctic for many years, thickening over time. Beginning in the late 1990s, sea ice began melting in the southern arm of the gyre, thanks to warmer air temperatures and more extensive summer melt north of Alaska and Siberia. Moreover, ice movement out of the Arctic through Fram Strait continued at a high rate despite the change in the AO. Thus warming conditions and wind patterns have been the main drivers of the steeper decline since the late 1990s. Sea ice may not be able to recover under the current persistently warm conditions, and a tipping point may have been passed where the Arctic will eventually be ice-free during at least part of the summer (Lindsay and Zhang 2005).”
Nathan,
Your claim was that there was more ice drifting below 72 degrees. These articles discuss the melting process, which I agree with in general. They aren’t saying what you are though. This melting mechanism is what Carrick and I were discussing. I told him that the weather pattern change was responsible for pushing ocean water into the region, he said temperature was the main culprit and someone made me look up some references. Now admittedly it wasn’t that black and white for either of us but it started me looking into the data.
Look at one of my sea ice videos, if there was a positive trend below 72 degrees in the 2002-2007 timeframe, it didn’t last long.
I’ve seen a distinction made elsewhere between “perennial” ice and multi-year ice, with perennial being basically the 2nd yr. stuff. It has somewhat different patterns.
Another suggestion was that ‘warm’ westerlies open up water by pushing ice (which is continuing to form, as ‘warm’ is still very cold at those latitudes) up against the east shore, where it piles up and accumulates — until it spreads over the open ocean. So, warm westerlies cause more multi-year ice to form!
🙂
Nathan, let me try to understand you argument/concern over Jeff’s article.
Your contention appears to be semantical in that Jeff used the term “annual sea ice” versus “multi-year ice.” That the multi-year ice floats into that 72 area Jeff was studying, and accordingly it can float outside that same area.
If it is a semantical argument then the names change of that multi/annual sea ice to icebergs/flows. So if it is a strictly semantical argument, then let’s be a little more precise.
If its not a semantical argument, then what the heck are you trying to say? You have been all over the area here and at Tammy’s.
Jeff, let me apologize for even mentioning that crowd of scallywags. Clearly most of them have not read your articles.
Nathan, I seriously doubt that the time variations of drifing multi year ice into an area south of 72N would have any material impact on what Jeff is showing here. Since 100% of this drift ice melts south of 72N, there would have to be a tendancy for this drift ice to increase with time.
Your reference discusses possible flushing of multi year ice due to tendancies to positive AO for a few years in the 90’s – shown here. But since multi year ice does not accumulate south of 72N it is a strech to suggest that this is an ongoing issue that somehow biases the trend south of 72N. Circumstances are a little different north of 72N however. It would take years to recover a significant (non AGW related) flush of multi year ice and this would obviously confound the attribution of multi year sea ice decline to AGW.
LL,
I’ve worked on some more code based on individual sat data over the weekend. I want to incorporate your breakpoint analysis code into it, although I fully intend to unintentionally mess it all up. In the meantime Tamino is distracting me. It is just too much fun playing with them. It’s my weakness I’m afraid. He messed up on his original critique and has moved on because of it.
Yeah I should get back to literature review again myself. Not sure what you are working on but one fact I found interesting in my reading to date, is that NSIDC uses different SSM/I channel combinations to derive sea ice data in areas deemed to be annual ice (channels 19V and 37V) from areas of multi year ice (channels 37V and 37H). The offsets and calibrations are carried out for each channel independantly. I found this interesting because I originally questioned how statistical breaks could be satellite related when I saw that break timing differed when looking at north and south of 72N.
Jeff
“These articles discuss the melting process, which I agree with in general.”
No, they specifically talk about the transport of ice south (where it does melt).
Layman Lurker.
Jeff is analysing the melt anomaly south of 72N, he has to account for additional ice being transported into taht region, as time has gone on more ice is moing south from the Arctic Basin, which as it enters Jeff’s region will increase the anomaly. This is a really simple concept.
“But since multi year ice does not accumulate south of 72N it is a strech to suggest that this is an ongoing issue that somehow biases the trend south of 72N.”
but MORE ice is moving through there than in the past.
Since the 1990s more ice has moved into the region that Jeff marked out. He was saying that looking at the total sea ice between the 72 degree latitudes shows us something about global warming.
I say it doesn’t meananything at all. Jeff cannot articulate what it means, so Iguess I am right.
CoRev
“accordingly it can float outside that same area.”
It doesn’t float back north.
Jeff’s anomaly doesn’t account for transport of ice south, which has been increasing over the time period. He cannoyt say that this is the ‘annual sea ice annomaly’. He thesis makes no sense
Jeff
I cannot understand how you cannot understand why this discusses the increased transport of ice by wind
“Moreover, ice movement out of the Arctic through Fram Strait continued at a high rate despite the change in the AO. Thus warming conditions and wind patterns have been the main drivers of the steeper decline since the late 1990s.”
Can you see it now? Do you know where the Fram Strait is?
CoRev
Give that his post was entitled
“Single Year Global Sea Ice Shows Minimal Trend”
and at the end he says :
“Below, I have combined what amounts to nearly all of the annually formed sea ice into a single global ice anomaly (Figure 7). We have a negative trend which to my surprise, fell just short of the 95% significance estimate. We know the planet is warming so this was a little puzzling.”
It’s not surprising that I am concerned that he is counting other than first year ice.
His hypothesis is flawed because he isn’t actually looking at just first year ice.
If I made a hypothesis about Chimapnzees and included Gorillas in the data, people would say that was wrong too.
Nathan, if you can not understand or agree with the import of what Jeff has said, then instead of carping from the sidelines rephrase his findings. What he has done may or may not be significant, but you and the Tammy crowd are seriously over reacting. Why so defensive of his efforts?
I see you still want to split hairs: “CoRev
“accordingly it can float outside that same area.”
It doesn’t float back north.”
Really?!?
CoRev
No one is being defensive. Stop attributing a psychology to this when you don’t know who we are.
“What he has done may or may not be significant”.
No, what we are saying is that Jeff’s analysis is stupid and meaningless. He can’t even explain what it means.
Jeff lets his ideology dictate his science, which is ironic as that’s what he goes around accusing others of doing.
“rephrase his findings. ”
It’s not a matter of rephrasing them. He hasn’t done anything useful.
You could say something like “the sea ice anomaly between the 72 degree latitudes shows no trend” but this is a pointless thing to say.
It doesn’t say anything about any physical process or about global warming.
It’s like saying my bank balance on Wednesdays shows no trend.
As I said before, his premis was that he was looking at annual sea ice. If he can’t actually deal with annual sea ice then his premis fails. It’s not splitting hairs.
Nathan, all you have done is wave your arms about increasing flows of multi year ice to south of 72 lat. Check out figure 2 from KC(12). The green line shows the trend of multi year ice export over the last 10 years.
Layman
The paper I posted above is over a longer time frame; 1982 onwards.
“all you have done is wave your arms about increasing flows of multi year ice to south of 72 lat.”
So, no, I haven’t just ‘waved my arms about’
Click to access stroeve_etal_seaice_synthesis_2011.pdf
And that is a superior analysis of the state of the Arctic, than what Jeff did.
Nathan, maybe you could help us a bit and point us to the table or graph or hard numbers that takes your claim out of the realm of arm waving.
OK so how would Jeff incoprorate this data
http://journals.ametsoc.org/doi/pdf/10.1175/2008JCLI2819.1
into his analysis of annual sea ice loss from 72N south?
He subtracts the multi year ice contribution from his anomaly… And he needs to do it from all export gates.
My guess is he’ll end up with an even bigger pile of nonsense.
Nathan,
Please help me to reconcile these two statements from your references regarding Fram Strait:
From http://nsidc.org/cryosphere/sotc/sea_ice.html
“Moreover, ice movement out of the Arctic through Fram Strait continued at a high rate despite the change in the AO.”
From http://www.arcus.org/files/projects/supplemental/674/stroeve_etal_seaice_synthesis_2011.pdf (page 2455)
“There is no statistically significant trend in the Fram Strait area flux”
How can it be a continued high rate from the mid 90’s on if there is no trend between 1979 and 2008? High rate compared to what?
I believe this is the full version of the Cosimo study that Jeff is citing.
Click to access 20110008253_2011008656.pdf
Kan
Look at the data.
ice moves out faster during one phase of the AO.
Jeff
“When someone sorts just the multi-year stuff, shouldn’t we be able to sort the annual and report its trends?”
I suppose you can, but you just need to make sure you’re reporting first year ice.
#54 Nathan,
You have to read more here. Thank you for the paper, this represents exactly where I was going with my work. I wanted to section the straits on the drain and measure the outflow to see if it confirmed my point that warming is not the primary cause of ice loss but flow. This is readily apparent in the daily sea ice videos I’ve produced but the problem is that in more recent years, it appears visually that there has been less outflow with decreasing ice. “death spiral”
The paper makes my point from before, but doesn’t answer all of my questions.
The paper you referenced specifically says:
“Due to the lack of ice thickness data after the year 2000 we have not been able to update our time series volume flux. Our best estimate of the mean annual volume flux using satellite ice drift (between 1991 and 1999) remains at ~2200 km3/yr. This short record, during a period of positive AO and NAO, does not allow us to asses the volume flux during years when the AO/NAO remained relatively neutral.”
The fact may not maybe that ice moves out faster during one phase AO, but this paper does not address it – due to a lack of data as of 2008.
The fact maybe….
Nathan, you have claimed and made a big todo out of supposedly increasing export flows of multi year ice to south of 72 lat. You have offered nothing to substantiate that. Anyone can link to a paper. I have seen nothing you have linked to supports your contention. Not surprising that you haven’t even made a weak attempt to point to something specific in these links. In fact, Kan has begun to point out how your papers actually undermines your position. All your links do is help confirm that export flows of multi year ice to south 72 lat would do nothing to bias Jeff’s analysis.
Layman
Look at the data in the Kwok paper, match the peak outflows (fig 4c) against the higher anomaly years that Jeff showed in his graph. Jeff was charting the loss of annual sea ice, to compare with the loss of multiyear ice. He needs to subtract out the multiyear ice he has included in his graph. Staright away I can see that more ice was exported out during 2007 than during 2003… This would bring his graph ‘down’ and increase the negative anomaly.
I also note you are now quiet on how Jeff was posting about Annual sea ice.
At least now we can agree that the melt area south of 72N does include melt of multi year ice.
Jeff
“…confirmed my point that warming is not the primary cause of ice loss but flow. This is readily apparent in the daily sea ice videos I’ve produced but the problem is that in more recent years, it appears visually that there has been less outflow with decreasing ice. “death spiral” ”
Well the paper certainly shows that there is a lot of transport out of the Arctic, but doesn’t make any explanation of why. NSDIC seem to have good info on why; less resistance to movement out due to thinner ice and no ice sheets holding the ice in place. The last little ice sheet broke up in 2005.
Seeing as Jeff has based his analysis on Comiso 2012, here are the conclusions of that paper, which pertains to various discussions in this thread.
Click to access 20110008253_2011008656.pdf
[caveat – this may be an earlier, pre-publication draft. don’t know]
Nathan, my first comment was #29:
I have been arguing this point all along. Not the best on fact checking are you?
I’ll post the many reasons why figure 4c does nothing to support your claim later when I have time.
Correction, Jeff is not basing his analysis on Comiso.
Linkee above. They didn’t glean multi-year ice by masking everything south of 72N, by the way, but by satellite-borne passive microwave sensors.
Jeff, you compared the decline rates. I’m not sure if it makes much of a difference, but they are meausring to 30% minimum coverage for multiyear ice, and you are measuring area (all the ice), yes?
After pointing out that incessant arm waving is not valid support to his claim (that multi year sea ice export has been increasing and would bias time series trend for data south of 72 lat), Nathan finally referred specifically to Figure 4c of Kwok (08). Nathan is completely wrong in suggesting that 4c supports his claim.
1) After admonishing Jeff in #54 that he would need to account for multi year ice export at all three export gates (Fram Strait; Svalbard / Franz Josef Land; Franz Josef Land / Severnaya Zemlya ) what does he do? He points to the 4c graph of interannual variability from a single gate: Fram Strait singling out the singling out the increased export of roughly 250,000 sq k between 2003 and 2007. If we follow Nathan’s own advice and account for the export flows at the other two gates from 2003 to 2007 what happens? FJL-SZ = – 150,000 sq k; S-FJL = -120,000 sq k for a net change in export flow of -20,000 sq k considering all three gates.
2) Since we are interested in how the area south of 72N changes with time and more specifically the trend during the satellite record, then correspondingly any suggestion of bias must have implications on this trend. K08 makes it quite clear that there is no significant trend in export flow in any of the three gates over the satellite record.
Layman
Don’t be dumb. You can’t just say because there’s no trend in increased flow, that there’s no impact on Jeff’s ‘anomaly’. His anomaly is based on monthly(?) data so summing the data from Kwok is not a useful metric. Especially over the short time period you propose, given the anomaly Jeff is calculating is from 1979. And before you do it, no, summing the entire time period is not appropriate either.
I cannot understand why you wouldn’t want ot remove the multi-year sea ice component from Jeff’s anomaly, considering that he wants to measure single year ice.
If Jeff cannot subtract the multi year ice (especially seeing he has the data now) then he can’t use that to compare to multi year ice.
Barry,
Give up now, Jeff CANNOT be wrong. Apparently all ice south of 72N is single year ice… Oh and all that multi year ice just doesn’t matter because if you sum it between 2003 and 2007 is sums to not very much…
#69 nathan
You are the one who is claiming that export flows of multi year ice have biased the south 72 lat sea ice trend. You are the one who cited figure 4c in support of your claim. Now all of a sudden “His anomaly is based on monthly(?) data so summing the data from Kwok is not a useful metric.” All I have done is point out how your logic failed.
Maybe I didn’t make myself clear in #68. It is the author of K(08), not me, that has explicitly stated that there is no statistically significant trend in the ice outflows at any of the gates over the entire satellite record (not sure where you are getting the “short time period” stuff from). One might think you hadn’t read the paper or something.
Layman
“You are the one who is claiming that export flows of multi year ice have biased the south 72 lat sea ice trend.”
They have, as Jeff was saying it was single year ice.
It doesn’t matter if over the period there is no statistical significance. You need to check if the peak outflows of multi year ice match the up ticks in Jeff’s single year ice anomaly. If that happen to coincide (for example in 2007) it will drag his anomaly down.
Just use the data and check, Jeff was trying to work with single year ice, at least he could actually do that.
Nathan, I have approximated the sea ice export numbers for each gate from Kwok 08 and summed to approximate the net annual sea ice export. Since the SJL-FZ data is runs from to 1992-2007 (the other two gates run from 1979 to 2007) the approximation of net sea ice export is also limited to this period. This now allows a closer look at how export ice would impact on the south 72 sea ice data.
Since I am working with sea ice concentration data (which is dimensionless) I have normalized both time series, then scaled and offset the data to reflect the proportion of annual sea ice export to the south 72 annual sea ice maximum (which IIRC is circa 1:10). March is the typically the month of the maximum for this region. An illustration of the export sea ice drawn to approximate scale along side the south 72 March sea ice concentrations is here.
Now, here is 1992 – 2007 segment of the normalized March sea ice maximums, this time showing a segment in red where the effect of the approximated sea ice export has been removed. The corresponding ols trends are shown as well. I made the graph big enough that if you squint hard you can distinguish the two trends. 😉
Not sure how you could continue to argue your claim that export sea ice biases the south 72 lat sea ice data. You can choose to continue to argue your claim if you wish – I think readers can judge for themselves that your claim is a fantasy.
LL,
That is very interesting. As I too often have asked, I wonder if you would mind writing up a summary of your data based on this article. Show the gate locations, transport plots and any(or no) conclusions you think are interesting. You are way ahead of me on this paper and it is very worthwhile. I don’t really like that it is evidence that I’m wrong but it isn’t up to date enough to prove it so I might still wiggle out by luck. Either way, I find it extremely interesting.
Rewinding a bit,
How much annual ice is found North of 72? Is it possible that including it in the analysis would change results?
How de we satisfy the notion that this sub-section of ice behaviour is representative of the whole of Arctic sea-ice? (Or is that not the point?)
Barry, not speaking for Jeff here but:
I would suggest that the amount of annual ice north of 72 has likely been increasing. As multi year ice diminishes it is displaced by annual ice.
Maybe a little perspective might help. Jeff put this post up a while back initially using the arctic circle (64N lat I think) as a cutoff. In a follow up post he stated the following:
. The answer to your question on whether including annual ice north 72 lat would change the results is likely yes, because if such ice were included in the analysis, there is a good chance that it would be related to polar cap melt (as annual ice displacement of diminishing multi year ice). At least this is my opinion.
Jeff was posting using the arctic circle as a cut off. I got curious where the actual latitude cutoff was so I went to KNMI where they make it easy to dissaggregate gridded sea ice data. I posted a comment on yet another post (there were a bunch related to this that are all worth reading through) when I saw very easily where the effective latitude cutoff was for very near 100% annual ice melt.
The point of all this in my mind is nothing more then exploring data and discussing. During discussion threads that follow the posts perspectives are shared and ideas get floated out and sometimes interesting insights emerge and sometimes they don’t.
Sorry for getting long winded on you Barry. I just thought it might help you and others who might have come to the middle of this little kerfuffle what some of the background was.
LL, no I appreciate you taking the time and that all makes sense and seems very reasonable. It seems pretty clear to my eyeballs (a poor meter, I know), that there is and has been for the past 34 years, a great deal of ‘annual ice’ North of 72, and judging from NSIDC reports, it is increasing (49% of the sea ice at the end of the melt in 2009 was first-year ice).
Therefore, I am not able to make the leap with Jeff when he posits that he has plotted “all of the annual sea ice.” Your comments would also be at odds with the notion that to “mask the North pole by something other than latitude” would likely change nothing.
Therefore we can’t go along with the notion that a combination of this data and the Antarctic sea ice is somehow representative of a global trend, and that the lack of statistical significance for the trend is not only worthy of attention, but suggestive of a trendless state in global sea ice.
Jeff is transparent about his political predilections on the climate change issue, which is refreshing in a way, but what he describes as “interesting” appears only to be those results, however gleaned, that speak against or downplay AGW (to use shorthand). I’m not sure I quite accept your characterisation of the discussion as “nothing more than exploring data and discussing,” particularly as Jeff has elsewhere tried very hard to find sea ice trends that are not statistically significant.
I realize the value in determining trends that are not statistically significant, but I’m not sure why one should deliberately seek that result. Can you explain?
The Arctic refreeze is constrained by land, and it is possible that the same area could be covered wherever temps regularly fall to below zero while the summer melt dwindles at a rapid pace. We are already seeing sea ice area being more sensitive to seasonal change, suggesting a thinner cover melting out and refreezing. Should the summer ice completely melt out along the way, there will still be winter refreeze. I don’t think using changes in refreeze anomalies is a sound proxy for the Arctic (or global) sea ice change.
“Therefore, I am not able to make the leap with Jeff when he posits that he has plotted “all of the annual sea ice.”
Barry, it seems to me you are playing games with words. If somewhere I have not been clear enough, point me to it and I will change the wording but this conversation is stupid. Your accusations about ‘trying to find trends’ are insulting and also stupid. That ain’t how this works.
If I was interested in trends which were not significant, I would drop the AR1 model now and could widen the CI’s dramatically (and more defensibly). You are statistically ignorant for not realizing this. There is nothing wrong with ignorance, but there is something wrong when you accuse someone of doing something you are absolutely ignorant about.
Of course you could just listen to Tamino and ignore the fact that I redid the same 2 1/2 year old post you linked to only 30 days ago.
https://noconsensus.wordpress.com/2012/01/28/sea-ice-2012/
Don’t go around insinuating anything about my honesty you know-nothing [snip].
Jeff, I take it back. I thought there wasn’t that much here, but I must be wrong. Enough mathimagic roaches have crawled out of the wood work, there has to be pizza and beer somewhere around.
The most telling is that you are “mashing” things together. Every time I read this I chuckle. This is from a group that allows data where the sign of the relationship wrt known relationships can be abitrarily changed, and mashes pollen, ice, precipitation, speleo and tree rings from all over the world, and not complain of this mashing is just too funny. They even go so far to complain just questioning it makes you part of that devil brood of deniers.
Just too funny.
John,
It’s a rough morning at work. I’ll be offline for a while. I’m not sure if there is much here actually but that is how it was presented. These people have no stats chops are dropping by critiquing it and then saying completely statistically ignorant things. Frustrating.
I wonder how the half the human race puts their pants on sometimes.
Barry, reconsider what I showed you in my post. Jeff started with the arctic circle as his cutoff. It was me who showed him the 72 lat cutoff. Are you suggesting that he would have eventually represented south ac as 100% of the annual ice? He went on to post 2 or 3 articles (which I didn’t link to) that displayed brilliantly the changing nature of annual ice north and south of 72. Funny behavior for someone intending to hide that fact don’t you think. The commenters at tAV were following these posts and new exactly what the south 72 number represented when Jeff put that post up. There are too many at tAV who look at the data ourselves for Jeff to get away a misleading claim as you suggest. Finally, what would the point of all this be? In order for anyone explore this issue properly, you necessarily have to pick the area which which is known not to support multiple year ice through the record. If Jeff was trying to make a misleading claim about trends on annual ice, why wouldn’t he lump in all of the growing annual ice north of 72? Then his trend would be even less would it not? You can’t have this criticism both ways.
You may have missed the hyperlink to this:
“Now I’m sure that people will call this exercise “cherry picking” but I don’t think that is a fair critique having picked all of the annual sea ice.”
https://noconsensus.wordpress.com/2012/02/28/single-year-global-sea-ice-shows-minimal-trend/
I think it’s pretty clear that there is plenty of annual sea ice North of 72.
LL agreed that there is annual sea ice North of 72N, that it has probably increased and that including that extra annual sea ice may change the results of your analysis. This is directly at odds with your comments in the post I linked. I’m not competent enough to join in, but is that something worthwhile for you two to explore and discuss?
———————————————————————————————————————————————————-
“Of course you could just listen to Tamino and ignore the fact that I redid the same 2 1/2 year old post you linked to only 30 days ago.”
I linked to a post of yours dated Feb 28 2012 and quoted you from it ^. I got the link from your first reply to Nathan upthread, which you provided to him as background for this post.
I know I am pretty ignorant on many matters. Was there nothing sensible about my comments on using refreeze data as a proxy for the whole of Arctic sea ice? Or is that not what you have been doing?
LL,
“He went on to post 2 or 3 articles (which I didn’t link to) that displayed brilliantly the changing nature of annual ice north and south of 72. Funny behavior for someone intending to hide that fact don’t you think.”
This would seem to be at odds with his statement that changing the masking for annual sea ice (from 72N to something else) would not much change his results. Would you mind linking to those articles?
“If Jeff was trying to make a misleading claim about trends on annual ice, why wouldn’t he lump in all of the growing annual ice north of 72? Then his trend would be even less would it not? You can’t have this criticism both ways.”
Why should I care whether the trend is less or more? I really don’t understand this point. Could you elaborate on that just a bit?
I do see where I have gone wrong on one point: Jeff isn’t suggesting that the area he is using (below 72N and Antarctic) is a proxy for global sea ice. He is using it as a proxy for total annual Arctic sea ice, combining it with Antarctic sea ice and calling it ‘global’*. Do you agree that this is probably not warranted considering the significant changes in annual sea ice North of 72?
* “…isn’t it interesting that global single-year sea ice has not shown a trend which is easily differentiable from noise in the past 34 years?”
Barry,
Ok, if you insist, I will assume you are still incapable of sorting the thread to figure out what I mean.
“I think it’s pretty clear that there is plenty of annual sea ice North of 72.”
Yes there is plenty, of course it would be stupid to use every pixel of single year data as it would unfairly bias the trend and make it LESS significant. See, what I have done is to chose the MOST significant single year trend I could find without extensive work. NO, i didn’t mean to confuse anyone that I had used every scrap of sea ice on the blue marble. That would be a mistake. Ironically, it would have been the same mistake Tamino incorrectly critiqued me for making.
MOST significant. More than other trends.
More than Tamino.
I even attributed it to warming…….
See the difference.
==============
So you just came by and accused me of dishonestly looking for “insignificance” on a thread where I gave you the MOST significant single year sea ice trend. Even worse, I used an AR1 stat which has one of the narrowest CI.
Apologies?
#83 Barry:
Link. I don’t see that any statement is at odds with anything. Please quote the statement and someone will explain it to you. If you go north of 72 and count pixels, in some years some of those pixels will have annual ice and other years those same pixels will have multi year ice. We don’t want to include this type of ice, period. We only want those pixels characterized exclusively by annual ice over the entire record. That is why I pointed out to Jeff that he could go futher than the arctic circle to 72N for his dissaggregation cutoff.
Consider the following simple case. You have a circle which represents the set of NH sea ice from 1979 to 2012. A midline divides into two subsets, on one side of the line is the set which contains *only* annual ice during the record. On the other side contains the set of sea ice which was *not only* annual sea ice during the record. Intuitively, moving the cutoff line (slightly) within the south 72 area would not likely have a material effect. Moving the cutoff line to the north of 72 cannot be done because it falls outside the solution set.
Jeff’s question “…isn’t it interesting that global single-year sea ice has not shown a trend which is easily differentiable from noise in the past 34 years?” is not inconsistent with this in any way shape or form. By definition, the area of “single-year ice” over a given time period must exclude those areas areas which have contained multi year ice. In hindsight, a better choice of language would have been less confusing, but what possible reason would Jeff have for wanting to mispresent this? Answer: there is none.
You have suggested that Jeff misrepresented his analysis because he excluded annual ice north of 72. Presumably, the only point of the misrepresentation would be to set up his presentation of global single year ice downtrend as being statistically insignificant (unless you have some other reason?). Just one problem with this. This supposed act of misrepresentation – excluding annual ice north of 72 – actually creates a steeper downtrend.
Jeff,
getting you to state simply that you have not plotted all the annual sea ice has been difficult, but we’re there. It’s not my fault that you wrote a sentence that said the opposite. It’s good to clear it up.
In another sentence the area analysed is named ‘global annual sea ice.’ It is still not clear to me if you are suggesting the data is a proxy for ALL annual sea ice or not.
Let me analogise my point here: If I masked off the poles from 60N and 60S, developed a temperature record from what was left over, and then called it the “global temperature record,” it would be fair to expect some criticism, wouldn’t it? Here, again, is what you said;
You also infer that your masking method gives useful results for global annual sea ice in comments here;
Consider that it might not be that I’m nit-picking,but that you have been unclear, and the confusion is genuine, not malicious. Your emphasis is clearly that you have found a fair proxy for total annual sea ice via your masking method. I would like to see if LL maintains that including the annual sea ice above 72N would change results, and whether you two find your apparent difference of opinion worth discussing.
LL asked where you said you doubted that changing the masking on annual sea ice would change results. Here’s the quote;
“Now I’m sure that people will call this exercise “cherry picking” but I don’t think that is a fair critique having picked all of the annual sea ice. Perhaps I could spend even more time to mask the North pole by something other than latitude but I doubt it would change anything.”
——————————————————————————————————————————————————
You say that you get stronger statistical significance for annual sea ice trend using your method than by using all the annual sea ice data. Ok. If you combine ALL data for Arctic and Antarctic sea ice cover, including multiyear, leaving out no months of the year, is the statistical significance of the trend poorer than what you’ve derived with your choices?
Nathan is right when he says that multiyear ice also melts out south of 72, and that this will therefore be included in your analysis of ‘annual sea ice’. I imagine this would make a difference, particularly if less and less multiyear ice has been transported out each year, for example. The question is, have you attempted to account for that dynamic – to see if it would change your results?
As the Antarctic is relatively thermally isolated from the rest of the planet, due in large part to the large expanse of ocean surrounding and circumpolar winds, and that the radiative effects of ozone flux may also differentiate the region – because the Antarctic and the Arctic are by no means meant to behave in lockstep as the planet warms, shouldn’t there be appropriate caution and caveats when speaking about ‘global sea ice’?
I think the physical basis of your analysis is too simplistic, meaning even the results that are only “interesting” are overinterpreted.
“getting you to state simply that you have not plotted all the annual sea ice has been difficult”
Bull.
Not as accurately worded as it could have been, I admit. But non-morons know that my discussion of masking the North pole to achieve more single year ice area, implicitly states that there is more available single year ice area in the Arctic than I had masked. The problem was that had I included it, it would be difficult to PROVE to morons that I wasn’t intentionally biasing the result.
Not that I am calling anyone who doesn’t get it a moron, just that if I had done as some asked, the morons would be right and I would actually be a moron. Being a moron is not a good sales pitch for a technical blogger – in case anyone wondered 😀
Nathan is equally as right as those who cannot find that Manhattan sized pixel in the ice video data. If 200,000 cows fart in Nebraska the spring air in Iowa still smells fine.
If you have any evidence to support this nonsense, please feel free to link it.
LL,
forgive me for quoting you, but I feel we’re losing the thread at times. You said
Do you agree that Jeff has said the opposite where I quoted it in the post above?
If the annual sea ice North of 72 trends differently to that South of the border, then you have a problem with assumptions, no? And you have said that there likely will be a difference. so it would be misleading to say that the trend derived from the masking is an indicator for “global annual sea ice”.
I get the masking rationale. I just think it’s too simplistic and will give spurious results if intepreted beyond its own boundaries. More accurate terminology would be “annual/multiyear sea ice cover South of 72N.” Using that as a proxy for global annual sea ice is problematic.
You’re wrong, or you put what you mean the wrong way. Jeff was clear about the masking and the reasons for it. I implied that the method suited his predilections, not that he misrepresented it. I have said that I think he overinterprets his results. It’s a slight difference, but I’m a stickler for clarity (and I need to be better at it, too).
Are you saying that the steeper a slope is, the greater its statistical significance? inevitably? I am not a stats expert or anywhere in the vicinity, so I hope you’ll fill in some ignorance here. I would have thought that statistical significance is NOT dependent on the angle of the slope, and would still be mystified as to why I should care about steepness.
(I notice you have said, “his presentation of global single year ice downtrend.” I’m looking forward to some clarification on that from Jeff)
I’ll ask the same question I asked of Jeff – is the trend for combined N/S sea ice data for all months and all types (total cover) more or less statistically significant than the trend derived by Jeff’s method?
Wow, you don’t know about this? That explains a little.
1. Thermal isolation (relative to other egions) from circumpolar currents – http://www.sciencedirect.com/science/article/pii/S0031018211001799
…for example. It’s well-studied and well-known. Would you like more references?
2. Role of ozone depletion on Antarctic climate – http://www.atmos.colostate.edu/ao/ThompsonPapers/Thompson_etal_NatureGeoscience2011.pdf
Role of ozone depletion re Antarctic sea ice – http://ruby.fgcu.edu/courses/twimberley/EnviroPhilo/Turner.pdf
The measure of uncertainty on these metrics is no bar to the caveats they import to your analysis.
3. Two poles not meant to be in lockstep – you consider such a notion nonsense? Based on what? The Earth is not a smooth billiard ball. There’s no reason to expect the whole globe will behave uniformly everywhere. The IPCC has projected slower changes in the Antarctic than the Arctic, and there is nothing to indicate that timing of (any climate) events must be the same.
For example –> “In 20th- and 21st-century simulations, antarctic sea ice cover is projected to decrease more slowly than in the Arctic” [AR4 10.3]
But that’s only the concisest quote. There is more in AR4 and, of course, the studies that underpin it. Bottom line, the Antarctic and Arctic regions are strongly differentiated geophysically. It is a mistake to assume that their response to global climate change is somehow equivalent. It is overinterpreting to combine indices from each and suggest that the result is a direct response to global warming.
Stats is most useful here when it has a strong physical basis, no?
Not at all. When Jeff says that he has picked all of the annual ice he is saying he has picked all pixels (by latitude) characterized by *only* annual ice over the period 1979 to 2012. Annual ice north of 72 were not *only* annual ice for the period. Hopefully you understand this now.
By definition, we cannot include any latitudes which have included multi year ice during the period. IOW no ice north of 72. Ice to the south of 72 has (effectively) *never* contained multi year ice and that is how we have defined our question. This is the way we can investigate the question: “What effect does warming have on areas which have not supported multi year ice in the period of record.” This has been explained over and over and I’m not sure why you are still confused.
Slope is not statistically distinguishable from 0 trend. IOW it is a slight downtrend that could be accounted for by statistical noise. There will be a threshold (steeper) slope value that would allow us to say that such a downtrend is not likely to occur by chance.
I’m really not sure about that off hand as I haven’t looked at that data myself. I believe that the graph of 100% global ice would have a downtrend which is not as steep that is a guess.
[edit]
Barry,
Maybe I don’t understand. Are you claiming that the Antarctic is thermally isolated in a manner which would affect a 30 year ice trend or are you claiming that the difference in ocean currents means that we cannot combine sea ice globally?
We don’t need to discuss the ozone stuff. That is a popular theory to try and correct inaccurate model predictions. It ties together the old IPCC BS ozone story with AGW. Good money, iffy science.
Barry,
I don’t see why differences in how the ice is formed (or maintained) in one hemisphere over another should make any difference in this analysis.
Does albedo behave differently in the south vs. the north?
If the ice melts in the south instead of the north, will the sea level changes not happen, or maybe happen in some significantly different way?
Please help me understand why it matters. Since I think we can all agree that there are regional and even more local affects all over the globe, why are metrics like ‘global temperature’ ok when global ice is not? I suspect deserts and rainforests have different climates – should we separate those from other regions when reporting temperatures? Is “mashing” them together dishonest in some way?
Do your objections not apply to averaging temperatures? If not, why not?
Ken,
I’m confused too. I don’t know what he is saying anymore because the ozone paper supports long term trend changes – I don’t consider it good science yet. The other paper he linked to is about something else entirely.
Then there’s this:
Manabe et al 1990.
http://journals.ametsoc.org/doi/pdf/10.1175/1520-0442%281991%29004%3C0785%3ATROACO%3E2.0.CO%3B2
p.795 figure 11.
Model runs show an interesting result for growing ice in Antarctica for 100 (!) years due (indirectly) to increased CO2. At the same time it predicts arctic ice to all but disappear.
Tamino and others cite this paper as proof (see the post ironically titled “It’s the Trend Stupid” at “Open” Mind for more) that antarctic ice growth was predicted. The unfortunate problem for him and others is the paper also predicts this trend to continue for 100 years. You see, they need to say that the recent increase in Antarctic ice was predicted but ignore the fact that the *same paper* says the increase will continue unabated. That allows them to simultaneously accept the IPCC’s prognostication that the antarctic will soon begin melting.
LL,
Ok.
1. Because Jeff has used this metric + Antarctic as a proxy for global (total) annual sea ice, saying that the resultant statistical significance of the combined trend is “interesting.”
2. I don’t know why the question is being asked. That’s what is confusing. If it is because it is supposed to represent ALL annual sea ice, then I think you have a problem, as you yourself have suggested – including the annual sea ice North of 72 will probably change the resulting analysis. Otherwise, I don’t understand the utility of the analysis.
That doesn’t answer the question, does it? I’ll put it like this:
* Is it possible that a steeper trend can have less statistical significance than a shallower one? With the same number of data points?
I’m guessing the answer has to be yes (dependent on data variability). In which case I would set aside the comments you have made on this and ask someone to do the math (I can fiddle about with a linear regression in Exel but I’m way underquealified).
But I’d be curious how you answer *.
Jeff,
On the former, yes strongly possible. On the latter, no that’s not what I’m saying. There other ways that the Antarctic is relatively thermally isolated. Land heats faster than sea surface, Arctic warming is supplemented by heat transported fron the surrounding landmasses. The Antarctic is surrounded by ocean. The Amtarctic continent is a titanic ice cube,
That paragraph is hand-waving bullshit. The ozone ‘story’ was well underway and the Montreal Protocol ready for signing before the IPCC was formed. “Iffy science,” for Pete’s sake.
Question: Is the Antarctic sea ice trend statistically significant? Is the slope statistically distinguishable from zero?
If the Antarctic’s (relative) thermal isolation means it will respond more slowly to global warming, then that suggests that internal variability will dominate any trend for a longer period where the response to global warming is faster. And it’s not just ozone’s effect on the local climate that potentially complicates the issue, but how sea ice is formed. How much sea ice comes from calving off the continent, for example? Salinity is changing around the Antarctic differently to the Arctic – how does this affect sea ice growth?
Ken asked, and you demurred to his query, “Do your objections not apply to averaging temperatures? If not, why not?”
I can’t believe statisticians don’t see the potential problem here. I’m a noob and I can see it!
No, it doesn’t apply.
The global temperature record rests on combining many data sets. Jeff is combining just two data sets. For the global temperature data, a single data set has little influence on the result. For Jeff’s analysis, either data set has a huge influence on the result. In which case, each data set has to be well-understood, and the factors contributing become vastly more important than for individual stations in the global temperature record.
Finishing a sentence from above (and fixing typo)
“The Antarctic continent is a titanic ice cube, providing strong thermal inertia against warming”
The general point being that processes and quantitative physics are very different for each pole and must be carefully considered when combining data to make ‘global’ indices.
Oh, one straight question;
Jeff, are you using the masked annual sea ice as a proxy for total annual sea ice?
Barry:
If I understand it right, the models are saying that has nothing to do with the Antarctica is not warming as fast as the rest of the globe. (Otherwise this would explain why the Antarctic was warming from 1960-1980 when the rest of the land masses were cooling slightly.)
You really need to be writing to NSIDC and the other ice data centers and alert them to their “blunder” in combining North and South Hemispheric ice series. Jeff is a minor player here compared to them. They are the ones who get quoted in the press after all….
I can answer this:
That’s what I view him as doing, and have viewed him as doing from the day he introduced that series.
By the way, you guys that get so hot and bothered by Jeff using masks to compute series also need to write the people who do the ENSO series, because they do the same thing.
Sheesh. The question is “What effect does warming have on areas which do not support multi year ice?”. It is the presence or lack of multi year ice which determines the latitude cutoff point. Since north 72 does support multi year ice during the sat record, it is excluded.
The question itself defines these constraints. Just because you are confused doesn’t mean there is a “problem”.
That doesn’t answer the question, does it? I’ll put it like this:
* Is it possible that a steeper trend can have less statistical significance than a shallower one? With the same number of data points?
That wasn’t what you asked. And yes I did answer your question: “Are you saying that the steeper a slope is, the greater its statistical significance?”
I’ll reword slightly: “No that is not what I am saying. I am saying that the slope is not statistically distinguishable from 0. IOW it is a slight downtrend that could be accounted for by statistical noise. There will be a threshold (steeper) slope value that would allow us to say that such a downtrend is not likely to occur by chance.
I’m not going to explain it any further. Please take a stats refresher if need be.
Sorry I left out the blockquotes for this part:
That wasn’t what you asked. And yes I did answer your question: “Are you saying that the steeper a slope is, the greater its statistical significance?”
Carrick,
Could you provide links where NSIDC “combine” the data – and then talk about “global sea ice” trends undifferentiated between North and South? The graphics that I’ve seen there didn’t combine data, but showed them independently. I may well have missed something…
Also, NSIDC discuss the differences between Antarctic and Arctic, eg: http://nsidc.org/asina/faq.html#antarctic
Cryosphere today do a global asea ice chart, but AFAIK they’ve made no claims about it. It’s just the data graphically represented. Do you have a link to reputable site where they talk about global sea ice as if it’s undifferentiated between the two poles?
LL,
I understood the masking parameters from the first. I don’t understand the purpose of the investigation. You run a trend analysis of all sea ice south of 72N (where the ice melts out every year and *no* multiyear ice taints the data) in order to…. what?
You have tried to reassure me that Jeff’s masking choice provides a steeper slope, and therefore is more statistically signficant than (some other method). But, as I understand it, another choice (like using all coverage data) that comes up with a shallower slope could possibly have stronger statistical significance. Am I wrong?
But we can put some of this conjecture to bed by working out the trend and statistical significance of total sea ice, no masking.
Barry,
My understanding is that the Ozone hole was not blamed for cooling until recently when the Antarctic failed to follow models. This is different from the doom of the whole earth from UV due to ozone lost – which WAS bullshit. There may be an old paper which says otherwise about temps, dunno, but the point is that models showed more warming than what has been measured, ozone has now been conjectured (with enough handwaiving to fly a penguin) in the literature. I am not an expert on this matter and can’t even stand reading the articles on this subject because the ones I have read have been so poor.
So, ozone aside, you have now stated that the Antarctic is insulated enough from the rest of the globe, by a magic wall, that it would affect the 30 year trend. The abstract you link discusses a pressure gradient in ocean currents which makes it difficult for the ocean to flow directly to the Antarctic, Its abstract does not support your contention. Got anything else?
Barry,
I guess not NSIDC,
but cyrosphere does it as you point out.
I can think of things that adding them together tells you something about… global mean albedo for one. Why wouldn’t that be interesting?
Anyway, you’ve kind of demonstrated you’re coming at this from a biased view, it appears you are arguing that you shouldn’t combine them because you don’t like a conclusion that can be made from combining them. That’s not how you should do science.
Start with arguing over whether it’s reasonable to do a certain operation (it is reasonable to combine them) and then to ask what interpretations are valid (some interpretations are valid, others are not).
You’ve lost me here Barry. How many datasets is enough? 3, 4? What are your criteria?
Presumably the satellite temperature measurements are really only one dataset (consisting of thousands of measurements), so they should be thrown out completely if we follow your logic. Or maybe I don’t understand how you are demarcating “dataset.”
If I divide the global thermometer measurements into two “datasets”, NH and SH, can I average them or is that a no-no? (hint: it doesn’t change the answer if you break it into 1,000 datasets or one)
Barry,
Yes that would be interesting to compare against.
But remember we were in this to try and tease out signals for warming/cooling of temperatures versus Ekman transport.
So far it seems to me, that the temperature in both hemispheres has been the driving force here, for annual ice. And I think that’s what Jeff has found.
For multiyear ice, I believe that mechanical loss should play an increasingly dominant role: And it really isn’t rocket science.
We’re using the criterion “present or absent” here. So for “absent” to be true, the amount of ice remaining has to drop below the threshold of detectability for ice. That is much less likely for multiyear ice that is remains above 72° N than for annual ice below 72°N.
Even when looking for signals like effect of temperature on annual ice, if you select only ice below 72°N, you still have some work to do in designing the right statistical tests.
And again conformation bias is playing a role here for Jeff’s critics, rather arguing that it’s not a reasonable thing to do (it is) and calling him a Steve Goddard for doing so (h/t toto), you should be addressing how one improves the analysis. What features should be present, what should you test for.
After all, annual trend is kind of a first brush sort of analysis, it doesn’t tell you very much that is interesting about complex phenomena like this.
“And again conformation bias is playing a role..”
Or perhaps *confirmation* bias. 😉
Barry,
Until you answer this yourself, you should not use the words ‘statistical significance’ in a sentence.
You are getting closer. North 72 doesn’t “taint” anything. North 72 supports multi year ice therefore does not meet the selection criteria for disaggregation.
LOL. In order to answer the question we have asked.
First of all this statement tangential and obfuscates the real issue. Once the frame of reference has been determined, then masking at 72N is not a “choice” it is defined.
As for your tangential concern, let me ask you a question: Do you understand that it is possible for a time series of random, trendless, autocorrelated noise to have a linear trend slope? If you do then you should be able to understand what I have attempted to explain to you and no need to troll any futher. If you answer something like: “Wait a minute LL, you just contradicted yourself; how can trendless random noise have a linear trend slope?”, then go study some stats and stop asking obfuscating questions?
Jeff, do you see the annual sea ice derived from your masking method as a proxy for total annual sea ice?
If you can answer that straightly, I have a follow-up.
LL,
The purpose of asking this question was to answer it. I am clear on what you think of the excercise. Thanks.
My issue is that Jeff has projected the answer to this question into saying something about the trend of ‘global annual sea ice’. I naturally assumed that there was a bigger idea behind the asking of the question.
Yes. Linear trend analysis comprises the bulk of statistical discussion on the general topic of climate change (in the popular and semi-popular literature – ie, media & blogs). Even a maths dunce like me can at least understand the concepts.
Barry,
No.
My points were far more pertinent – and educated. I would appreciate it if you got your head around those.
“My issue is that Jeff has projected the answer to this question into saying something about the trend of ‘global annual sea ice’.”
No, your belief that I am bad has projected your answer onto my question. You are wrong, and you don’t have the stats chops to make a point either way.
“Jeff, do you see the annual sea ice derived from your masking method as a proxy for total annual sea ice?”
“No.”
You said in the top post here,
Maybe it’s a semantic issue, but it sure seems to me like you are employing your 72N analysis.as a proxy for global (all) annual (first-year) sea ice.
Let’s try caps with an exclamation point.
NO!
How is that…. Care to address the serious stuff yet doc?
Barry, I don’t think that you get it at all.
What Jeff is calling “annual” ice occurs (or sometimes may not occur) in a specific area each year. The key feature is that due to the environment it is in, it will melt completely each spring and summer and I would describe this type of ice as “transient”. He has found that in the available data, pretty much all of the ice below 72N has satisfied this condition and, for simplicity’s sake, he has chosen to look at the ice in this region. This ice is always first year ice.
On the other hand, ice above 72N may or may not melt, depending on a variety of factors. The ice that subsequently forms in that region is also first year ice, but it is NOT annual ice. Constant reference to that ice as if it belongs to the population in question is simply inaccurate.
Furthermore, your answer to Carrick above also indicates a lack of understanding of statistical issues (bold mine):
Carrick has asked whether you understood that for due to the nature of certain types of data, false trends may be evident when in fact they do not exist. Your answer is way out in left field and does not address the point of his question.
I genuinely feel that you truly do not understand most of what you write.
Thanks for #117 Roman but my name’s not Carrick. 🙂
Hello Roman,
I understood this from the outset.
It IS ‘annual ice’. It’s just not annual ice according to Jeff’s definition.
I asked LL if he thought that including the transient ice north of 72N would change the analysis, and he said yes (I agree). If Jeff had at the outset said that he was not using his analysis as a proxy for ALL transient sea ice, then I would not have persisted with that line of enquiry.
Which now leaves open the question of whether or not the trends of sea-ice South of 72N describes anything meaningful.
My answer was “yes”. I understood the question.
Statistical significance testing estimates the probability that a given result occurred by chance. A linear trend can be derived from any old data, but that does not mean that the trend is ‘real’.
“false trends may be evident when in fact they do not exist” *
What I lack in stats chops I make up for in comprehension and reasoning. I can spot tautologies *, contradictions and irrational arguments with ease.
Carrick
I am glad that this miscomprehenson has been just now cleared up by Jeff.
Sorry LL. My bad. I was in a hurry.
Barry, when I read nonsense such as “It IS ‘annual ice’. It’s just not annual ice according to Jeff’s definition,” I shake my head and wonder what the point is in trying to explain something to people who either don’t have the background or worse who are really are incapable of grasping a simple concept..
“What I lack in stats chops I make up for in comprehension and reasoning.”
No Barry, you don’t. You stomped around here, made dozens of errors far worse than any I have been accused of, claim to understand, and now claim that it was all just good thinking!
You avoided my tough questions, can’t grock simple math, abused statistics to the point where it hurt my brain to understand and now you want a pass for your reasonableness.
Dolt.
Barry, it’s better to admit when you don’t know what you know that to posture as if you do.
That strategy works at parties while trying to impress chicks, it doesn’t work so well in technical forums where people don’t mind speaking frankly.
What I see is somebody who not only doesn’t understand statistics and can’t keep his own arguments straight. Not so much on comprehension or reasoning for that matter.
It’s too bad you came in on this thread as it may give you the misimpression that we are a group that just support each other regardless of what they say. That’s hardly the case here…
Even with respect to the sea ice, Jeff views what he is doing differently than I view it, but as long as what he is doing is sensible and moving the dialog forward I don’t see anything to criticize him for.
You are going to have to do a better job with maintaining coherency in your arguments or you’re just going to get chewed up and spit out. A little more structured reasoning, a little less sophism please.
Barry said in #103:
“I understood the masking parameters from the first. I don’t understand the purpose of the investigation.”
Barry,
Sorry to butt in, but maybe you should stop talking for awhile and see what unfolds?
In the meantime you could go back over to Tamino with your concern about refreeze anomalies not being good proxy for anything. I see that you failed to communicate this apprehension when he used it as a rebuttal point.
Then head over to Cyrosphere Today and explain to them (even though they make no claims) that combining Arctic and Antarctic sea ice together is stupid and should never be done. Please take the Ozone paper with you.
Then come on back and see if there is anything new here.
Carrick,
“Even with respect to the sea ice, Jeff views what he is doing differently than I view it, but as long as what he is doing is sensible and moving the dialog forward I don’t see anything to criticize him for. ”
I doubt we disagree about the ‘proxy’ point. It seems impossible to get some people to realize the nuance of the study because I/we are not painted in a bad light then.
If you said this is a proxy for how single year sea ice responds to temperature, I would agree.
If you said this is a proxy for how single year sea ice area is trending globally, I would say no.
There are all kinds of nuance in that range but we all know why already so it isn’t worth going into the detail. Barry has pre-determined his opinion (Tamino say Jeff bad) so I fully expect him to continue to confound the issue.
Apart from neglecting to corroborate the well-known thermal isolation of Antarctica, and letting go the ozone issue as a waste of time, what did I miss?
And regarding the above:
http://www.ipcc.ch/ipccreports/tar/wg2/index.php?idp=595
There is a great book looking at climate change over geological times that describes the evolution of Antartica and says;
Do you need more? How many hours should I waste googling up links to something that is well-established just because you are ignorant of it?
I see that announcing my own definciency early on really made an impact. But to what do you refer exactly?
I did no statistics. I asked some questions, and received waffly replies.
Reasoning and reasonableness are two different things. You are careless with words.
Because you are careless with words, Carrick had the same impression I did on your using the 72N masking method as a proxy for global annual sea ice. If he and I came to the same wrong conclusion, and he maintained this impression from “the day [you] introduced that series”, then I don’t think that the confusion is my problem, Jeff.
Roman,
Jeff’s selection criteria and subsequent masking is very straightforward. It is the interpretation of results that I have taken issue with. People redescribing the parameters of the analysis to me have missed, or avoided, the point.
Carrick, a regular and ‘friendly’ here read Jeff’s interpretation the same way as I did. This is not about predisposition, as some would like to think, but clarity.
“It is the interpretation of results that I have taken issue with.”
Which interpretation did I make incorrectly?
The claim I took issue with was that you had found a trend (that failed statistical significance) for global annual sea ice. As you had also described the sea ice you were looking at as single-year ice and first-year ice, and there were glraphs labeled as global sea ice with no ref to 72N, I could not see but you had used your analysis as a proxy for gobal sea ice. Reading amongst the threads, and of course, Tamino’s page, I see that quite a number of people also had the issue that I did.
Since you have stated clearly that it was not a proxy for global, I think the issue is resolved. The subsequent posts are backchat.
If you have the time for a yes or no answer on something very basic…
It is my conviction, against something LL implied, that between two trends derived from the same sample size (but different data), the steeper trend could potentially have a lower statistical signficance value than the shallower one (if the steeper trend has ‘noisier’ data).
Yes or no?
“The claim I took issue with was that you had found a trend (that failed statistical significance) for global annual sea ice. ”
I thought we had cleared that up. Tamino’s thread has a different purpose or he would have cleaned it up.
“It is my conviction, against something LL implied, that between two trends derived from the same sample size (but different data), the steeper trend could potentially have a lower statistical signficance value than the shallower one (if the steeper trend has ‘noisier’ data).”
Barry,
You really need to answer this question yourself before using the term ‘statistical significance’. Personal understanding first Barry. I’m sorry not to spoon feed you on this, it won’t help.
That was an answer.
LL had said;
My understanding is that statistical confidence depends also on the variability of the data (noisiness), not just the steepness of the slope, as LL’s comment implied.
Am I wrong?
(Sample size is not a factor in this case)
Barry,
You have made claims about this post which rely on your ability to answer the question, your repeated asking means you were unqualified to make your statements. If you canhttp://wordpress.com/my-stats/ answer your question, including the proper nuance, we can talk more about statistical significance.
This comment reveals your depth. And after all, conservatives are stupid, biased, and anti-science. It shouldn’t be too hard.
Please try to read a little math before commenting on my math or its political bias. It seems a reasonable request right?
Yes, you clarified. Rehashes ensued, unfortunately.
Bollocks. My chat has been about categorization and physical processes. I don’t need stats skills to talk about that.
It would be, except that you are completely misrepresenting me. I have never commented in any way on your ‘math’. Nor could I, nor would I. Where I have said anything about stats analysis, I have asked questions (until prompted by you to provide my own answers).
Are you trying to reframe what I’ve said to make an easier target?
Your link goes nowhere useful.
The answer could have been as the following and provided enough content for the point:
“Yes, barry, a shallow linear trend can achieve a statistically significant result where a steeper one may not, depending on sample size and/or signal to noise ratio.”
My opinion on your political bias is not based on your math, but on what you have clearly written yourself at this site and elsewhere. My impression is that you do the maths honestly, and that you are disposed to seek out and/or emphasise analyses, deferring to the “pressure for a simple unified message,” in order to win “the public “policy” battle outright” on AGW. I quote your own words in an otherwise technical post at tAV. I assume those words were honest. It’s good, in a way, that you are so transparent about it.
Barry “It’s good, in a way, that you are so transparent about it.”
It’s always good when people admit their biases (which always exist) up front.
The rest is more sophism though. Boring.
Barry,
I’m still not spoon feeding it to you. You have again expressed an opinion that my interpretation is based on winning an AGW argument. This only represents to me that you don’t understand the stats. I therefor see your bollux and rase you one.
You pre-assumed what I am about from Tamino, and then you made your conclusion. Now you are trying to get the stats to fit your version of reality – about me.
Say what you want, I am honest in my presentation and reasonable in my interpretation. Reasonable being different from necessarily right. Frankly, having a newbie with a demonstrable lack of stastical chops accusing me of misrepresentation for political purpose is more than a little offensive. I/ WE have been arguing for AGW for almost an entire month and you didn’t even figure out on your own what was really written in this thread. To say it mildly, you are pretty weak on interpretation Barry. What’s more is that you seem intelligent enough so it reads more like Taminoesque, intentional misunderstanding for a political purpose.
So while you are working away at figuring out what I did wrong – and still not finding it – I’m thinking this accusation of bias for an agenda belongs on you.
I also need to address this bovine scatology : “Carrick, a regular and ‘friendly’ here read Jeff’s interpretation the same way as I did. ”
Carrick and I have it out regularly here. There is a thread about multi-model means where we battled out some suff for quite a few comments. It was about stats though but had you read that, you probably wouldn’t call him a ‘friendly’. We are different people with different ideas and have enough chops to discuss stats. I also think we had the same interpretation of this post, he just didn’t like the absolute I gave you on a proxy for single year ice – just a guess though. This whole string of sea ice posts started from another discussion – which was his fault as well.;)
You certainly won’t find that sort of discussion in tamino’s rock box threads.
Barry keeps talking about his mad skillz at “categorization and physical processes”.
Haven’t see them yet.