As some of you are aware, while the NH sea ice extent has been hovering above the record minimums for this time of year, the area has been tracking below the record minimum. My script generates 7 day averages for the area and I blinked when a saw the black trace had was now above the minimum:
This may be temporary. There is still time to lose area, but we are seeing the area oscillating increasing some days, decreasing others. Who thought it could be so exciting to watch ice melt?
Welcome to the club, Lucia! 😉
OMG! We’re all gonna dieeeee! Arctic Ice area has reached the 2007 level. Anyone else noticed that the dire predictions/commentary has diminished after the NH ice melt began to go above it earliest path.
Anyway, just had to test whether I was a 1st or 2nd rate/time commenter. 😉
CoRev– I’ve been turning that plugin on and off. But it’s wise for you to add comments in case I turn it back on.
But as for the “death spiral”, because of conversation with Romm, I read some of his sea ice predictions. He recently put some short term ones here:
http://thinkprogress.org/romm/2011/08/14/295625/arctic-sea-ice-record-volume/
Last week was the third week of August. The extent did not fall to its lowest for the date.
.
That was a perfectly reasonable assessment from Masters at the time. Like I wrote in the SIE update around the same time:
.
“I agree with Masters that we are currently seeing this high and according to the ECMWF weather forecast maps it will more or less stay put in the coming 5 days.
.
Towards the fifth day, however, we see the high extending all over the Siberian coast and a low forming over the Canadian Archipelago and Beaufort Sea. This is the exact opposite of the ideal Dipole Anomaly set-up and if it comes about, it will probably halt the above average extent decrease we are going to see in the coming days (as shifting weather patterns are prone to do).”
.
Right after that we had a day with a huge extent decrease reported (-127K km2). Had those weather patterns stayed put for 4-5 more days instead of that high shifting westwards along the Siberian coast, the difference with 2007 would definitely have been more than halved.
.
.
.
Lucia, I wouldn’t be too eager to take on Romm if I were you. You can think of him (and his bets) whatever you like, but his assessment of the current situation in the Arctic is very sharp and accurate. He’s the best blogger out there when it comes to this subject, and he will definitely stay so if things continue on this path. You need to catch up a bit more if you really want to have a dig at some detail or other that Romm might exaggerate. There’s more to sea ice than statistics. Much more.
Neven–
Masters’s forecast appeared perfectly sound at the time. But Masters also didn’t include or surround the forecast with words like “death spiral”. Master’s also didn’t highlight words like … to set a new low extent record with less extreme weather conditions. in bold to draw the eye and set a ‘certain’ tone.
Oh? How can you tell Romm’s assessment is sharp or accurate? For that matter: How can you tell what his assessement– especially for anything to do with precition– actually is? And by this– I mean is in quantitative terms.
I know he uses words like “death spiral”, but as far as I can tell, if you try to find out what constitutes “death”, it could be anything. So, what exactly is his assessment?
Real questions by the way.
As far as I can tell Romm is
a)being entirely non-quantitative with long term predictions
b) picks alarming nouns and adjectives which generally make it impossible to assess whether he is right or wrong because he studiously avoids saying anything testable.
There are no details to dig at because there are no details provided. I’m sure he will decree himself correct no matter what happens.
But as for “take on”? I’m not “taking on”, I’m just observing what he wrote, that he surrounded a forecast with a tremendous around of frothing at the mouth rhetoric, and it turns out the forecast was wrong.
Sure. I never said otherwise. But what does this have to do with Joe surrounding what is know with the sorts of wall or word histrionics he likes to spew forth?
.
I have learned quite a bit about the Arctic sea ice in the past two years. Not that I know everything, far from it, but I know enough to separate the BS from the factual stuff. And Romm writes very little – if any – BS about Arctic sea ice. As simple as that. Because that’s how the situation in the Arctic currently is. Things might shift back to some pre-2007 situation, but as long as they don’t, it makes perfect sense to prepare for the IPCC to be completely wrong on this.
.
.
This is what I mean when I say you need some catching up to do (it’s not hard, even I could do it). What on Earth do you think the ‘death’ refers to in connection to the Arctic sea ice?
.
Why are you so eager to bash Romm?
Neven–
I get that you somehow think what he write is not “BS”. But what I’m saying is that it is just gush that permits plausible deniability.
Could you translate “death spiral” into quantitative terms? Once you do, we can figure out if what Joe says is BS or not. Otherwise, we can’t.
I’m not saying otherwise. What I’m asking you– who seems to believe what Joe says seems to be “not BS”– is this: Can you quantify what “death spiral” even means?
Catching up on what? In the comment I mentioned what Joe wrote. He happened to pull out a forecast, surrounding it with hot air and drumbeats of alarm, and he happened to be wrong in this instance. I assume you are not suggesting that any catching up I have to do will retroactively make my observation wrong.
As for what Romm actually predicts: If you think I have “catching up” to do and you think you know what he predicts, could you quantify it for me. Because as far as I can tell, no one can really know what he predicts. It is couched in the language of “plausible deniability”.
Look: I’ve said that I do not know precisely what Joe means by “death spiral” — and in particular in quantitative terms. If you claim to know and understand, please tell me what you think it means rather than trying to defend it by asking me a rhetorical question.
Huh? I don’t think I am “bashing” Romm. I don’t think merely observing facts is “bashing”.
Did I, in a comment, I point out histrionic language? Yes. He does. If you think otherwise, say so.
Did I– in a post– point out that his bet to Revkin was not clearly stated. That is: what the bet was could not be figured out by reading the post. (It seems you could infer by going back to 2007 posts, or guessing things Joe things must be obvious or….)? Yes. I did this– and my motivation was to figure out what the heck the bet was and possibly tease out what Romm is intimating about the ice in his post.
Do I say I don’t know what “death spiral” means? Yep. I don’t know what it means.
I don’t see why I can’t observe these things about a widely read blogger.
So: real question: What, precisely, do you think constitutes “bashing”?
Most forecasts will look like they make sense when they are made. Then they turn out to be wrong later.
Andrew_FL–
That’s why most meteorologists don’t include words like “death spiral” in their forecasts. If they did, they’d sound like pro-wrestlers touting their prowess before a fight.
.
That’s the whole point. You go on a rampage, without pausing for a moment that it is perhaps you who lacks some basic knowledge. And instead of doing some research and learn a bit for a while, you go and accuse people of ‘histrionics’ and ‘gush that permits plausible deniability’, etc. More and more words are wasted, and for what? Nothing really.
.
As usual, this has everything to do with you, and not so much with the matter at hand. You are trying so hard to be rational, but are at the same time so unconscious of your own lack of competence in this area, that you show your true subjectivity. Like I’ve always said: Lucia is very much a lukewarmer when it comes to the technical stuff, but as for the rest, she is 100% karma point people.
.
Now, you go out and find out what ‘death spiral’ means in connection to Arctic sea ice, where the expression came from, and how folks like buddy Anthony Watts have spun it to ridicule a most serious matter. Google is your (real) friend.
.
I’m outta here. I’m tired.
Neven
Rampage? I don’t think it is a “rampage” to observe that “death spiral” is not quantitative, that I don’t know what it means and asking you what it means. You seem to have your knickers in a twist that someone would suggest that this is opaque. But it is opaque and if you know what it means, I’d appreciate if you would tell me.
After you defend Romm, yes, I characterize his rhetoric involving words like “death spiral” as “histrionics” and “gush”. His posts are full of histronics and gush. Specifically, he uses words like “death spiral” and doesn’t quantify what we are supposed to understand by that.
I’ve asked you to translate these his language and statements into quantitative statements, but you haven’t done so.
Whether Anthony “spun” this emotion-laden gush term into ridicule is neither here nor there.
The present issue is: Romm used it. It is an emotion laden gush term that creates a alarmist tone while saying nothing quantitative. Romm’s use of this term with no accompanying detail to permit the reader to determine what “death spiral” makes it impossible to know what he is really saying is going to happen.
And stop with the childish whining about Watts or suggesting that what Romm means by “death spiral” would somehow become quantitative if I used Google. If you can quantify in words what quantitative statement Romm is making using the word “death spiral”, I’d appreciate it.
I’ve asked you directly several times. Quite honestly, I think your “answering” with rhetorical question and telling me to google, or including an irrelevant dig to Anthony suggests that you know that “death spiral” cannot be quantified. That it is what I say: unquantitive, untestable gush, used in a histrionic way to set an atmosphere of alarm while maintaining plausible deniability.
Go ahead and call that observation “a rampage” if you like. But as far as I can tell, it’s just an observation of a fact that you find uncomfortable.
Have a nice night. Sweet dreams. Etc. Maybe tomorrow you can reveal the quantitative statement made by the term “death spiral”.
Lucia, Tammy and Romm acolytes can not be convinced. They do provide some comic relief.
Has anyone noticed that DeSmog Blog has opened its comment edits? Some more actual dialog going on there, versus the above mentioned. Maybe they’ve learned a lesson.
CoRev–
I rarely read DeSmog blog. I only visit if I a blog I do read drops an interesting seeming link.
CoRev:
Hasn’t DeSmog always been open comment?
Since we’re bringing up the “arctic death spiral” and hearing from Neven how Anthony is blowing it way out of proportion, here’s a link to arctic death spirals that seems to be engaging in histrionics:
Arctic Death Spiral: Sea Ice Passes De Facto Tipping Point Thanks to Deniers, Media Blow The Story, Again
I personally think this particular story is worthy of ridicule, and if I had more devil in me, I’d send it to Anthony’s suggestion box.
(Edit: I jumped in the middle, ridiculing this nonsense is what got Neven so bent. My bad.)
ok…remind me
is the Aug sea ice bet for “area” or “extent”?
I thought is was for “extent”, but now not sure with the comments on “area”.
I also think Neven, not being an expert on sea ice in spite of two years of studying it, has a lot of catching up to do.
First to realize seasonal sea ice loss is predominantly driven by regional weather patterns (primarily due to mechanical sheer from wind), not climate (e.g., mean temperature in the arctic), and since Arctic weather is notoriously unpredictable, seasonal forecasting is practically impossible.
Nor, in my opinion, is there any evidence that long-term change in mean sea ice is linked at all to the seasonal variations.
The hypothesis is that since recent warming has lead to a shift in the ratio of “new” to “old” sea ice, that the new ice should be more vulnerable. This should show up as an increase in vulnerability of the sea ice, but that has as yet to show up in the annual variability of sea ice (in a statistically significant sense).
IMO, if we were reaching a true “tipping point” it’s not the DC offset (secular trend) that we should be looking to as evidence, it is an increase in the annual variation.
Carrick (Comment #80637)-One thing I find interesting is the lack of strong seasonality of sea ice trends compared to the strong seasonality of temperature trends. I analyzed the NCEP reanalysis for the Arctic region, and the mean temperature near surface air temperature of the annual coldest day was warming about sixty times faster than the warming of the annual warmest day. To the extent that air temperatures are reflective of the seasonality of ocean temps, at least in terms of trends (perhaps not at all?) this would mean that ice is much more sensitive to slight changes in summer temperatures than it is to winter temps. Which is, I guess, not entirely surprising.
Well, either that or the reanalysis is misleading in that respect. Could be. The next step is to examine some data for water temps. The records might be unsatisfactorily short, however.
“Lucia, I wouldn’t be too eager to take on Romm if I were you. You can think of him (and his bets) whatever you like, but his assessment of the current situation in the Arctic is very sharp and accurate. He’s the best blogger out there when it comes to this subject, and he will definitely stay so if things continue on this path. ”
WTF? Neven the commenters at your site are miles beyond Room when it comes to being SHARP. Nobody is leading the pack in being accurate, least of all Romm who rarely says anything quantifiable.
Maybe we can narrow this thing down:
#####
Source:http://www.grist.org/article/exclusive-new-nsidc-director-explains-the-death-spiral-of-arctic-ice-brushes
Death Spiral means ice free summers.
Summer in the arctic starts when? June 21st and ends in sept.
So, ice free ( zero extent) through the entire arctic starting on June 21 and then sometime in september ice comes back.
“Very close” to ice free summers means, ice free starting june 22 2020.
That’s how I read the sharp and accurate Romm. of course experts can disagree about the sharp and precise meaning of the word “very” but its undeniable that june 22 is very close to june 21.
Ice free september is different in meaning from ice free summers.
A death spiral of course implies that lazarus ice is not in the cards.
Blockquotes and link inserted by admin– lucia
The Arctic is a capricious old girl, given so sudden changes in direction. No single year can tell us much about what is happening. I suspect none of us really have that much of a handle on what is going on at the moment. I strongly suspect changes in the age of the ice áºhich change its salt content and hence its albedo and density are having as much of an impact as anything else, also the thinner ice is easier for the wind to move around. But changes in the salinity and temperature of the water are very liable to change the currents flowing in and out of the Arctic.
It is my position that the increase in temperature over the past 40 years has thinned the sea ice till it is far more susceptible to other factors and these are liable to be unpredictable.
For example an open Hudson Bay in December\ January may produce enough moisture for a deep snow pack that may cool NH springs.
Or changes to the salinity on the Arctic may change sea currents and stop warmer water flowing in.
Wally Broeker famously stated that we were poking an bear with a stick when we were playing with climate change. I think the bear is begining to wake up but as Wally also stated, the younger Dryas, a very cool period, was probibly started by the Holocene warming. This is why I did not vote. I feel that we have now set in motion changes that we cannot predict.
So even the Death Sprial theory is one that I am not 100% sold on.
Still we are watching the results of a planet wide science experement. Grab a beer, it could be interesting.
Andrew_FL:
If you look at the frequency content of the ice loss, there is a strong annual component (which is usually anomalized away), and other than a slow secular loss, really no evidence of influence from long-period atmospheric-ocean fluctuations.
See e.g. this.
There are two possible explanations I can come up with (or a mix of these two):
1) Large effective thermal mass of the arctic sea-ice system.
2) Decoupling of the Arctic Ocean from the other ocean basins. (This could either be because of the role that sea ice plays in Arctic Ocean dynamics, or simply there isn’t that much ocean mass exchange between the Arctic Ocean basin and the other ocean basins).
The melt since mid-July has been 17.98% below the average. That is a large number in the context of the sea ice trends.
So where are people getting death spiral and new record low from? Its certainly not from the facts.
dolormin:
If it is becoming more susceptible, shouldn’t this show up as a measured increase in susceptibility? (This is back to the comment that it is easy to make data do anything you’d like them to do, as long as your descriptors remain qualitative.)
I would argue that the sea ice extent data don’t show this increased susceptibility right now.
I’m also arguing that looking at a change in the annual amplitude of the sea-ice-extent/area fluctuation is a much better measure of “tipping points” than nonsense like using an exponential function to extrapolate future ice loss, as was done by the “very sharp” Joe Romm.
As to “unpredictable”, I guess that’s the argument I’ve been making on this thread…as long as it is tied to the accumulated annual effect of weather systems (i.e., the ice loss is mostly via Ekman transport), don’t expect to see any patterns in the short-period fluctuations in sea ice extent. If we see a shift where the amplitude of the annual fluctuations starts growing rapidly, that would be my measure of reaching a “tipping point.”
If Neven thinks that a left wing partisan, rather than any scientist is “the best blogger out there when it comes to this subject” I rather think he has some bias himself. No blatantly political person should be considered so esteemed, especially not when politics is essentially his entire focus and reason for being.
Carrick (Comment #80635)
August 21st, 2011 at 12:11 pm
Since we’re bringing up the “arctic death spiral†and hearing from Neven how Anthony is blowing it way out of proportion, here’s a link to arctic death spirals that seems to be engaging in histrionics:
Arctic Death Spiral: Sea Ice Passes De Facto Tipping Point Thanks to Deniers, Media Blow The Story, Again
I personally think this particular story is worthy of ridicule, and if I had more devil in me, I’d send it to Anthony’s suggestion box.
Agreed, this is a particularly ridicule worthy article. I guess the words “de facto tipping point” instead of [plain ordinary] “tipping point” give the author an extra big helping of plausible deniability. Ohh, I didn’t really mean to imply an actual literal tipping point, only that that the circumstances gave the appearance of the characteristics of an actual tipping point because deniers were manipulating the media.
Re: Carrick (Aug 21 15:50),
For Cryosphere Today Arctic area, the amplitude is increasing. Annual difference in the maximum and minimum area has an OLS linear slope of 0.0285 Mm2/year with a p value of 0.0016
lucia,
The smoothed area loss rate anomaly (Cryosphere Today data) for 2011 is still below average, barely. If there is an average loss in area from this date, there will be a new record low area. The area rate anomaly for 2007 averaged -0.02 Mm2/day for the last six days while 2011 has been close to zero. 2007 has four more days of negative rate anomaly before it goes positive for about 10 days averaging close to 0.01 Mm2/day. In other words, 2007 had a very flat bottom, reaching near minimum early and starting to freeze up again late. If 2011 had continued to track 2007, then the minimum would likely have been far below 2007. Right now it looks to be only slightly below, 2.88 Mm2 compared to 2.91 in 2007. OTOH, area loss from this date has averaged greater than average for the last few years.
There is a chance the area will stabilize sooner rather than later. Concentration–and thus area–has been rising north of the 80th parallel. The loss happening now is driven by the edges in the lower latitudes. It’s not necessary that the ice stop melting at the extreme latitudes of the arctic; its the balance that matters to the overall statistic.
Re: Laura S. (Aug 21 23:25),
That always happens this time of year. The expected date for minimum area is September 9, which is less than three weeks away. The sun goes below the horizon at the North Pole for six months shortly thereafter at the autumnal equinox.
DeWitt, I did a multiyear calculation, corrected for autocorrelation, etc comparing 1972-2006 versus 2007-2010. While it’s true 2007-2010 has a slightly larger variability, it’s not statistically significant compared to the historical average variability.
I suspect your p < 0.0016 didn't correct for autocorrelation, because I certainly don't see anything approaching that strong of an effect in these data. [I got something around +1 sigma.]
This is minimax, which it sounds like the approach you used. This uses a more sophisticated method (read less noisy: OLS fit to Fourier coefficients.)
Not cyrosphere, but I’d be surprised if the numbers are that different, but I’d be interested in seeing your numbers in more detail and if you accounted for autocorrelation or not.
It’s also pretty obvious that most of the long term variability is explained by the offset shift, not any putative change in seasonal amplitude.
Carrick, I found DeSmog as open as RC and Tammy with less editing of comments. They seemed to prefer banning similar to Little Green footballs.
Lucia, I also stopped going to DeSmog after some ill treatment (lost comments and dropped comment privileges.)
As you see, I still lurk here. 🙂
O HO HO HO HO! I believe Dr. Cadbury, phd. informed everybody there would be no record broken!
@Neven
“Lucia, I wouldn’t be too eager to take on Romm if I were you. You can think of him (and his bets) whatever you like, but his assessment of the current situation in the Arctic is very sharp and accurate. He’s the best blogger out there when it comes to this subject, and he will definitely stay so if things continue on this path…”
No Neven I’m afraid he doesn’t know what he’s talking about and you don’t either. It’s just too cold up there Neven, it’s really ashame I know how important it is to you. Do you want a source, Neven? My source is my brain, telling me how stupid Joe Romm is. Also, my brain tells me anybody who thinks Joe Romm is smart is a really big idiot so I probably should never listen to you.
The bet here is for extent. I think extent is generally a more popular metric for betting everywhere. It’s the one I’ve consistently used mostly because there is better more prompt access to data and has been for a while.
dorlomin (Comment #80642)
Whether I agree with you depends on what “more susceptible” means. I would say that I anticipate ice is more likely to continue to decline than not. I’d also say that if someone predicts the arctic ice extent could have 0 area in at least 1 september before 1931, they might very well win their bet. On the other hand, I think Romm will lose his bet against Annan, Schmidt and Connoley. It is highly unlikely that the September NSIDC extent will fall below 10% of the 1979-2000 average in at least 1 September between now and 2020, That said: I don’t think it’s impossiblefor this to happen– I only think that the odds strongly favor Annan, Schmidt (brian), and Connolley.
At the risk of Neven once again accusing me of going on a “rampage”: I can’t even begin to say whether I’m “sold” on it because I don’t know what the theory is.
steven mosher (Comment #80641)
I added blockquotes to better separate what Romm (CP) wrote, what Serreze said and what you are saying.
I could also add that the result of “research” using “google” — as recommended by Neven suggests that at least sometimes Serreze’s definition of “death spiral” is for ice melt that is more dramatic than merely having ice free “summers”. This article reports:
So, if we read this, it appears Serreze considers “ice-free Arctic summer ” by 2030 but he also does not consider this rate to be “a death spiral”. It’s just “on a fast track.”
I can also find stories where Serreze’s use of “death spiral” only seems to apply to an individual summer. That is: He’s describing what people were watching right then. The use didn’t seem to convey anything bout what would happen in 2030, 2070 etc. (I can also find stories where the journalist mixed quoting and paraphrasing in a way where I can’t tell what Serreze meant by “death spiral”. This was not Serreze’s fault. I’d say Serreze is generally more clear and certainly more willing to be quantitative than Romm.
Mind you: This is Serreze, not Romm. Death spiral is a metaphor, and everyone can mean a different thing each time they use a metaphor. So, observing that Serreze does so is — in my opinion– in no way a slam on Serreze. But what it does mean is that even if I could be identify one thing that Serreze meant every single time he used the metaphor, it would still not tell me what Romm really means whe he uses the metaphor.
To the extent that Romm communicates in emotional terms using metaphors and then adding modifiers like “virtual” to “ice free” and “de facto” to “tipping point etc. I don’t think anyone can really be sure WTF Romm is really saying will occur. The metaphors and meaning-cancelling modifiers make Romm’s “predictions” unquantifiable gush.
Re: Carrick (Aug 22 01:54),
I’m looking at the difference between the annual maximum and minimum extent. I’m calculating an ordinary least squares linear trend. The residuals from the fit aren’t significantly autocorrelated according to the acf function in R. They’re barely significantly autocorrelated at lag 1 before detrending. Doesn’t that mean the autocorrelation comes from the trend? If that’s the case, then the significance of the slope doesn’t need to be corrected for autocorrelation, or am I missing something?
Lucia:
It’s my understanding that it’s easier to accurately measure extent than area too. Smaller systematic and other measurement errors makes it more attractive to use to study long term climate change than area-based metrics. (For the pre-satellite data, there wouldn’t be any proxy you could use for area either, which is another problem.)
DeWitt, perhaps it’s a difference in the data sets for one thing? I’m looking at daily data, which has an extremely high correlation coefficient. What period are you using for your slope by the way? I’m having trouble matching up with your stated OLS-derived slope.
My predilection would be to partition the data, and look at differences in the averages over different periods, as I described above… taking a longer baseline would tend to dilute any potential effect of AGW on sea-ice vulnerability if you are just fitting it to a straight line.
(I also don’t like using the max-min approach for this test, because it’s noisy, and you don’t generally end up with normal-like data, so getting a p value from this is problematic.)
Re: Carrick (Aug 22 10:00),
I would expect the daily data to be highly autocorrelated, which is why I don’t use it. I’m taking the maximum area for each year and subtracting the minimum area for that year. That gives me 32 data points for 1979-2010. I ran auto.arima on the residuals from the linear fit and the result was ARIMA(0,0,0) with zero mean. The Jarque-Bera test fails to reject the null of normal distribution of the residuals. That probably doesn’t mean much with only 32 data points.
Re: Carrick (Aug 22 09:53),
The systematic error for extent is going to be smaller (more accurate), but the precision will be worse because mechanical variation in extent from weather (wind and currents) can give you a wide range in extent for a given total area.
http://rankexploits.com/musings/2011/7-day-average-nh-sea-ice-area-not-at-minimum/#comment-80652
Dewitt:comment 80652:
Are you using 7 day smoothing– or something else? Are we using the same CT file? Or do I have a problem using R to convert days?
I think I’m ok. But here’s what I get
I’m using CT today data from here
Today it reads:
2011.6191 -1.8659449 3.3909633 5.2569079
2011.6219 -1.8012291 3.4372973 5.2385263
2011.6246 -1.8534746 3.3502688 5.2037435
2011.6274 -1.8466347 3.3217022 5.1683369
2011.6301 -1.7933171 3.3418646 5.1351819
2011.6329 -1.8430079 3.2694285 5.1124363
2011.6356 -1.8732083 3.2275105 5.1007185
My information is the 3rd column is the raw areas. The final is the baseline and the 2nd the anomalies.
In km^2, the most recent 7 values for areas are:
3390963 3437297 3350269 3321702 3341865 3269428 3227510
Using R, I convert that last day to 8 20 2011 which is the ‘232’ day of the year.
The mean is 3334148. This is the centered smooth value for day “229” of the year.
Using the same conversion tool in R, I get that from 8/13-8/20 =, 2007, the CT areas were
3344271 3302150 3226199 3229104 3168237 3073220 3070472
The mean is: 3201951
3.20 is less than 3.33.
So, I’m getting that currently, the 7-day smooth area is greater than on a similar day in 2007.
Dewitt–
I see I misunderstood and then turned myself into a pretzel. I thought you were telling me the area was still below the minimum. I spent time double checking (which is probably a good thing…) . But my questions seems to be going off on a tangent relative to what you say.
Yesterday’s post was not meant as a prediction. Only an observation. I would also observe the reported area change is starting to oscillate up and down from day to day. It does that as it approaches the minimum– and it’s now doing it. I mostly don’t focus on 1 day losses for area because it really is very noisy as we approach the minimum, but I do observe it’s now oscillating.
DeWitt:
Then you’re replacing one error source (correlated but normal) for another one, that is probably uncorrelated but larger and harder to characterize in terms of its PDF, so I don’t know how you’d reliably pull a p-valule from it. (I have ideas, but that’s all at the moment.)
Re: lucia (Aug 22 11:07),
Cryosphere Today needs one more significant figure in their date. When I back calculate the day of the year from the decimal fraction, there are several days that aren’t unique. Mildly annoying. Not to mention having to use 366 days for leap years and 365 for all the rest.
Actually, it’s oscillating all the time. It just isn’t very obvious when the slope is large compared to the oscillation.
Yes, the columns are date, anomaly, area, average area
DeWitt:
Did you mean to say “precision”? Don’t you mean “variance”?
DeWitt:
That was one of the things I did with the combined GSFC-JAXA data set—resample at 1/365 year^-1.
Carrick–
I just use whatever R does, and I think the day number starts with january 1. At anytime, there is a problem making “day N” in year X potentially differing from “day N” in year Y. For purposes of my blog plots, I don’t worry about it. Some years have 366 days.
For true physics not getting the days perfectly aligned by point where the earth really is as it’s going around the sun, but for what I’m doing here, I think it doesn’t matter.
Out of curiosity… do you make sure the equinoxes match or anything like that? Seems to me that if it’s important to align numbers, that’s what you might want to do.
Re: Carrick (Aug 22 11:36),
Yes. That’s what I wrote at first and then changed it to precision for no good reason.
Lucia:
I didn’t really worry about sidereal versus solar time. On years with 366 days, I could have easily dropped day 366. I just needed a consistent sampling rate.
I recall a study which looked at the details of changes in diurnal cycles of temperature and aligned days to one another by the local solar noon for that point of the year. At an hourly resolution this leads to perhaps a half hour’s imprecision in timing. Not bad. Similarly aligning annual cycles could perhaps be done by aligning to the summer solstice perhaps? Equinoxes would work too. One is still left with the problem of irregular lengths of years, unless for some “years” copy a day from the next or previous year over.
When I did a “trend by day rank within year” analysis didn’t bother to try and use non calender year bounds, I just copied over the mildest day of the year in non leap years. If the biggest effect of leap years might be focused around February, that could possibly explain why my thirty coldest days (mostly in January) warmed so strongly, but the next several coldest days, which I would bet are mostly in February (the next coldest month in the US IIRC) actually cooled. I think I might revisit my analysis with some method that better accounts for leap year effects (my original approach was pretty crude).
Andrew, I believe the temporal resolution of the “daily” series is closer to 7 days (effective Nyquest interval). So it doesn’t make sense to worry about temporal errors smaller than about 3 1/2 days.
The main reason I went from monthly to daily was to be able to fully resolve the measurement noise and seasonal cycle effects.
My cartoon bubble whenever I read a post by Neven:
http://tinyurl.com/2bvzwpx
TerryMN, what’s the Vikes outlook this year? WHO DEY
Andrew
Hey Andrew – probably better than last year, but I’m not expecting great things. Good luck to your Ain’ts 🙂
“I wish we had one of those Deathspirals”
General “Buck” Turgidson
Lucia,
I think we are in dire need of a baby sea ice story at this point. Can you work your magic?
SteveE–
The baby ice stories were Jeez! Weeeee neeed Jeez!
Agreeed! Weeeee neeed Jeez!!
Carrick (Comment #80681)-Heh, good point I suppose. I was under the impression it was a 7 day moving average or something though.
Incidentally, I believe an analysis done with some long temperature records from Europe suggested that the anomalistic year rather than the tropical year, although like the sidereal year it is slightly longer than the average Gregorian year length. There are some ultra-high resolution data in the tropical oceans and I am interested in calculating “exact” seasonal cycles (some bouys have data at ten minute resolution I believe) So I wanted to determine the best length to use for seasons. I think I’ll use whatever gives me something close to a whole number anyway. After all the distinction between anomalistic and tropical years is four days in three hundred years.
DeWitt:
My feeling was to raise the suggestion that the minimum will be reached before September 9–i.e., earlier than usual. Typically, the concentrations are higher north of 80, so the freeze vs. melt dynamic favors melting for a longer period of time (the first areas to freeze are usually already iced over). That’s not the case this year, the ice is diffuse. So the coldest areas have a chance to freeze and contribute new ice to the ice area earlier than usual.
say his name three times and I will summon the great jeez
Andrew_FL:
I think this is right & it’s pretty much what I said in so many more words 😉
(That is the effective sampling interval is 7-days, so 1/2 a sampling interval is pretty much the minimum resolvable shift… at least for episodic events like leap years.)
Re:#80695
Awww, I was hoping for Bender.
Tim “Bender, Bender, Bender. . .: W.
Re: Laura S. (Aug 22 23:27),
I’m not sure it works that way. 2005 reached its minimum area on 8/31. If you compare satellite images, it looks to me like there was more high concentration ice at high latitude than there is now. The time series for 2011 doesn’t look like it’s flattening out yet while 2005 started to flatten on about 8/16. The loss rate is still below average. Based on that, a minimum on 9/9 or later looks to be more likely.
Carrick (Comment #80696)-I see, it’s the “effective” time resolution which is 7 days.
Incidentally I am giving up for now on the high resolution buoy data analysis. I looks like I’d have to order the data, it’s not readily available.
“That’s the whole point. You go on a rampage, without pausing for a moment that it is perhaps you who lacks some basic knowledge. And instead of doing some research and learn a bit for a while, you go and accuse people of ‘histrionics’ and ‘gush that permits plausible deniability’, etc. More and more words are wasted, and for what? Nothing really.”
I would like to present this comment as a prime example of BS in that it generalizes and says absolutely nothing specific or in detail that can be discussed. Neven, you must be tired – get some rest.
How to restore arctic ice.
http://i149.photobucket.com/albums/s54/hurricanedude/001.jpg
@ Kenneth Flinch
Who was that comment directed at, Flinch boy? If the comment was directed at Lucia, the good doctor must step in and take you to task. If it correctly directed at Neven, then kindly disregard the following inflammatory language.
All that needs to be done is for somebody to look at the temperature. The problem with this is, Lucia wants to keep things interesting, so she isn’t going to provide the info. Neven, on the other hand, wants everybody to think that the ice caps are going to be gone tomorrow, so he isn’t going to present the temperatures. That’s another thing Neven, apparently the voice of Romm supercedes temperature data to you. Romm could draw a circle with a blue crayon, tell Neven it represents the arctic, draw a big black X over it and Neven would believe him.
Conclusion: In the record, 1979-present, it is very obvious there is almost no temperature change. If there were, somebody would be presenting the information.
Lucia-I seem to recall at some point you showed a scatter plot of sea ice area against sea ice extent. I was wondering if there might be a chart of Anomalies of area against extent anomalies.
Michael Topbis says:
The only way to simplify ourselves out of the present mess is by cutting our population 80%, unfortunately.
Individual actions are well and good, but as Gore said in his Noble lecture, and as Obama sed about his lightbulbs, they aren’t enough. Not even close.
http://www.stumbleupon.com/su/1ZXrmg/initforthegold.blogspot.com/2009/01/can-computation-help-solve-new-economic.html
just shocked…a few weeks ago I contributed to a post on Tobis’s site about economics textbooks because his crowd is clueless about anything other than what they can do…do not ask me what that is. I look back and I cannot work out which thread I responded to and, in the meantime, he has turned into a mindless Connolley clone.
Where did I go wrong?
Graeme
Did he say how quickly?
Lucia,
While he has said much the same thing on other occasions, I don’t think he has ever said how quickly, nor how exactly to go about making that draconian reduction in population. In any case, whatever the timing or means he envisions, I suggest Micheal volunteer to be first in line. Maybe Micheal watched Soylent Green one time too many, and it has kind of, well, influenced his thinking a bit.
The truly bizarre thing is that he appears to really, truly believe what he says about reducing population… which he freely admits is not going to happen. How depressing it must be for him to get up every morning knowing that humanity is utterly doomed and all his activities are without meaning or consequence.
Graeme,
“in the mean time”? Over the past couple of years I have never been able to tell where one ends and the other begins. I have not noted any changes.
aha…just founjd the nthread…lots of nonsense from Hank bRoberts and Neven,…am surprised that Mashey was not there too
Dhogaza obviously felt out of his depth.
However, it would be good to see the “scientists”…..the people who “know”, defend their position publically.
http://rankexploits.com/musings/2011/7-day-average-nh-sea-ice-area-not-at-minimum/#comment-80713
sincere apologies for derailing your thread
Graeme (Comment #80717),
Can you provide a link? I need a good laugh.
If you want a good laugh, you can have a look at chart neven posted up in May 2011 (made by a poster FrankB) modeled on a 2nd-order polynomial fit to the PIOMAS ice volume model.
There will be Zero ice volume …
… in 2030
… in the height of the winter, that is
… and Zero in the 2016 September minimum,
http://neven1.typepad.com/.a/6a0133f03a1e37970b014e885c65ac970d-800wi
Same fit methodology was applied by Tamino recently (although at least he noted we shouldn’t extrapolate the method too far in the future).
I fit a 6th-order polynomial to the same data and, shockingly, we are in a major ice age by 2030 with ice all the way down to Chicago (although the 5th-order fit says the Arctic will be at warm bath temperature in 2030).
Bill Illis,
The real hoot is that the ice volume is the LEAST confident of the measures of sea ice…. easy to exaggerate the volume loss when the data is flimsy. I sure wish they would offer some physical rational for the function they choose to fit to the (flimsy) data. I mean, there ought to be a reasoned connection between a proposed physical mechanism and the function chosen.
The JAXA sea ice area appears now to be in a period of rapid decline ( http://www.ijis.iarc.uaf.edu/seaice/extent/AMSRE_Sea_Ice_Area.png ). The cryosphere today area has also begun dropping again and is approaching 2007.
SteveF (Comment #80720)
August 23rd, 2011 at 6:22 pm
“The real hoot is that the ice volume is the LEAST confident of the measures of sea ice…. ”
————————————————–
Perhaps not for long – the Cryosat 2 data is coming in ( http://www.esa.int/esaLP/SEMAAW0T1PG_LPcryosat_1.html ) and will soon be able to confirm or reject the PIOMAS modeling results. My sense is that PIOMAS will be shown to be a pretty good model.
SteveF,
I have a suggestion for a kind of – sort of – not quite – nowhere near really physical model that creates a parabolic decline in sea ice. However, it becomes linear once there is low ice area/extent in summer.
If the temperature rise in the Arctic is linear (physical ish) and the heat is evenly distributed (not physical) then the ice decline is parabolic. You can model this by simply holding the amount of ice that forms in the freeze each year steady and increasing the amount that goes in the melt season by a set tonnage.
When there is low ice area/extent in summer, large amounts of the excess heat are pumped into space, meaning that the decline for the rest of the year becomes linear.
If you have, for example, 30,000 cubic kilometres of ice at the peak of the melt season and 20,000 cubic kilometres at the bottom of the melt season and there was equilibrium, 10,000 melts each year and 10,000 forms each year.
To take an extreme example, if you melt an extra 100 each year, there is no ice at the end of summer in 20 years, while there is 11,000 at the end of the freeze – which in my model would take another 110 years of linearly increasing temperature to disappear.
Not really physical, but the linearly increasing temperature appears to be about right. And ocean mixing may well spread the increase in energy relatively evenly geographically. But of course the ice is not spread evenly geographically, and the heat needs to go where the ice is to keep the melt linear.
Owen–
I was aware of the recent relatively large losses. Chatters at Neven’s are pretty excited about them.
The CT areas still seems to be above the 2007 area. I added some more spaghetti to my script and I’m looking for records etc. So far, I haven’t found any records broken.
David:
I’m really not sure I follow why it would go as the square of temperature rise. Am I missing something simple? Seems to me, it should be proportional to temperature difference, regardless of how the heat gets distributed.
Couple of other problems, if you’re referring to and trying to justify the figure reprinted by Romm, it actually fits to an exponential (Joe Romm got it wrong). Of course one doesn’t extrapolate using polynomial or exponentials unless you have a physical model to justify these. The original author admits this, but he used the extrapolation to predict the point of total ice loss anyway. Fuzzying thinking going on there, not “sharp” at all.
Secondly, most of the ice loss isn’t from direct melting, it’s getting broken up and transported by wind further south before experiencing death by heat. There’s a couple of references I gave on another thread that support this, I can dig these up for you if you’re interested. The phenomenon in question is called “Ekman Transport” and it even has an “Ekman spiral “associated with it. 8D.
I did a study on the seasonal amplitude of the ice, while I found the ice amplitude numerically increased, it didn’t meet p < 0.05 (closer to p < 0.1). DeWitt did it slightly differently and find p < 0.001. Obviously one of us is wrong. Naively you'd expect the amplitude to increase as the ratio of new to old ice increases, that's why I did the study of amplitude of seasonal ice to start with.
Carrick,
I know that it is not direct melting. But that is not really the issue. If increased temperatures are leading to an increased ability for ice to be transported and thus melted elsewhere, the ice loss is still a function (of some kind) of temperature.
Regarding proportional to temperature difference, I am not sure that that is anything different to what I am saying.
As a toy example of what I mean:
Arctic anomaly/Melt
0/10,000
.1/10,100
.2/10,200
.3/10,300
etc.
And I am not sure what Joe’s plots were: the ones that I use for looking at PIOMAS are fitted to a second-order polynomial – parabolic – not an exponential function. That is why I started investigating how it could be possible for a linear increase in temperature to cause such a thing. The very first model plot that I did gave me a parabolic decline. And it took me a bit of fiddling around to work out why that was so mathematically. Not physical, but still interesting (to me, at any rate).
I think Jeez is still around the blogs, but I suspect his artistry cannot be forced or summoned. That was god awful funny stuff he wrote about baby ice. He almost had me inquiring where I could adopt some baby ice for a safe residence in my frig.
Make that fridge.
I regret that I missed Jeez’s baby ice post — where can I find it?
A taste of it is here:
http://rankexploits.com/musings/2008/baby-ice-fighting-for-its-life/
I think jeez had some more stuff on Air Vent, can’t quite recall.
Kenneth
I think Jeez used to write for Jacques Cousteau. That suggests his artistry can be summoned. That said, we aren’t forking over the big bucks Jaqques would have had to fork over!
I should have added that Jeez’s pieces provided great imagery and were purposely overdone, I thought, to tweak any of us taking the Arctic ice follies all too seriously.
Lucia, I thought the words simply flowed from Jeez’s keyboard with no greater effort than that from an inspired poet spilling his words to nobody in particular and certainly not in expectations of a pay day.
Lucia, I think I have noticed a little hyping of a down-to-the-wire contest from your gambling emporium of late with comments like “a real horse race”.
Kenneth–
Of course it’s hyping. Someone is going to win the qualtoos!
Jeez is busy. Yes he did write for Jacques Cousteau.
Things are a bit hectic what with friends coming into town to perform a weekend gig at a local venue. So, patience.
Lucia, I think your modeling example/tutorial should add interest to the ice betting, but does anyone reading here have any good links to papers that might model Arctic ice. I do not mean minimum for any given yearr, but what effects increasing regional and global warming and other factors might have on the long term levels of Arctic ice at its winter peak and summer minimum.
I read somewhere that the albedo effects are not straight forward and that decreasing ice and increasing clouds can be counter balancing.
David Gould:
I’m willing to buy this statement, but why would it imply the relationship is quadratic? For the third time in a week, I’m going to point out that as long as you keep it in words (don’t apply math), you can make physics do anything you want it to do.
I’d suspect a spectral-based model would have more chance of being right. Polynomials are terrible for extrapolating with (end point effects).
Carrick,
The quadratic function falls out of the linear increase in temperature model; in other words, if temperatures are increasing linearly and if ice melt is a linear function of temperature then the September ice minimum will decrease quadratically.
If you have a linearly increasing yearly melt while the freeze amount is stable, the summer ice declines parabolically.
Let m be the amount by which melt is increasing linearly.
Then the total ice will decline according to the function:
Ice volume = (-m/2)t^2 – (m/2)t + k, where k is the ice volume that you started with and t is the time in years that have elapsed.
Create the graph of a linearly increasing melt with a stable freeze amount, and you will see this to be true.
It works this way because the amount melted is a sum function.
After 1 year of melting, the ice has dropped by m.
After 2 years of melting, the ice has dropped by 3m (the first year’s melt plus 2)
After 3 years of melting, the ice has dropped by 6m (the first two year’s melt, plus 3)
After 4 years of melting, the ice has dropped by 10m (the first three year’s melt, plus 4)
And so forth.
And this function is the one I wrote above.
Remember: the extra ice melted each year *does not freeze back in the winter*. Freeze is held constant. So each year you are melting from a lower base.
As I have said before, this model is not physical except in some broad outlines. If – for example – the ice transport is not linearly related to temperature, then that would mean that the assumptions are not correct.
I found the following articles with references to modeling Arctic ice with my summary below each:
http://web.mit.edu/newsoffice/2011/arctic-ice-melt-0810.html
Thinning ice and mechanics of breaking into small pieces which melt faster than larger ones. Cracks in ice as a negative feedback due to opening sea to freezing and creating more ice.
http://www.sciencedaily.com/releases/2011/08/110811113956.htm
Arctic ice could increase or decrease over periods of a decade and models indicate the faster than predicted decreases over the recent years can be attributed one half to natural causes and one half to AGW. Evidently it was the natural causes that the models erred on.
http://psc.apl.washington.edu/wordpress/research/projects/arctic-sea-ice-volume-anomaly/
Ice volume modeling.
Are there links to other Arctic ice modeling?
Jeez, Jeez, Jeez…
Mosher,
I still think we are in need of his calming influence. I hope the weekend gig goes well, but as Lucia said, “Weeeee neeed Jeez!”
I think Kenneth Fritsch is right, “Lucia, I think I have noticed a little hyping of a down-to-the-wire contest from your gambling emporium of late with comments like “a real horse raceâ€.”
After all, Baby Ice is at stake! Who are we to gamble for or against its future?