We’ve all been waiting patiently while Roy let life and finicky technolgy interfere with prompt reporting of the UAH TLT anomaly for December. (Roy, congratulations on your status as grandpa! Hope your daughter and grandchild are well!)
Now that Roy is back to the grindstone, he announced the UAH TLT anomaly: 0.123C. I’ve plotted values since 1980 along with the least squares fits obtained using a variety of statistical models:
Quatloos!
This month, the #1 winner nailed the anomaly to 3 significant figures. Congratulations John Norris! Congratulations to Tamarra and Dennis who took “place” and “show” and Arfur Bryan who, in 4th place nevertheless netted quatloos.
Alas, after that, the pot was empty. So, you’ll have to try again next month. You can see how you did relative to others below:
| Rank | Name | Prediction (C) | Bet | Won | |
| Gross | Net | ||||
| — | Observed | 0.123 (C) | |||
| 1 | John Norris | 0.123 | 5 | 80.366 | 75.366 |
| 2 | Tamara | 0.118 | 5 | 64.293 | 59.293 |
| 3 | Denis | 0.128 | 5 | 51.434 | 46.434 |
| 4 | Arfur Bryant | 0.115 | 5 | 25.606 | 20.606 |
| 5 | Ray | 0.135 | 5 | 0 | -5 |
| 6 | nzgsw | 0.137 | 5 | 0 | -5 |
| 7 | Joel Heinrich | 0.143 | 5 | 0 | -5 |
| 8 | lucia | 0.1 | 5 | 0 | -5 |
| 9 | Nyq Only | 0.1 | 1 | 0 | -1 |
| 10 | enSKog | 0.099 | 5 | 0 | -5 |
| 11 | Freezedried | 0.09 | 3 | 0 | -3 |
| 12 | Vlasta | 0.09 | 5 | 0 | -5 |
| 13 | anton | 0.081 | 3 | 0 | -3 |
| 14 | MikeP | 0.079 | 5 | 0 | -5 |
| 15 | ob | 0.074 | 1 | 0 | -1 |
| 16 | ivp0 | 0.068 | 5 | 0 | -5 |
| 17 | Eddieo | 0.067 | 5 | 0 | -5 |
| 18 | Don B | 0.063 | 4 | 0 | -4 |
| 19 | Pavel Panenka | 0.186 | 3 | 0 | -3 |
| 20 | Eds | 0.056 | 5 | 0 | -5 |
| 21 | Robert Leyland | 0.053 | 4 | 0 | -4 |
| 22 | hmcard | 0.051 | 5 | 0 | -5 |
| 23 | LC | 0.051 | 4 | 0 | -4 |
| 24 | DocMartyn | 0.05 | 5 | 0 | -5 |
| 25 | plazaeme | 0.05 | 1 | 0 | -1 |
| 26 | Lance | 0.203 | 4 | 0 | -4 |
| 27 | denny | 0.206 | 3 | 0 | -3 |
| 28 | Peter | 0.04 | 5 | 0 | -5 |
| 29 | AMac | 0.029 | 2 | 0 | -2 |
| 30 | Les Johnson | 0.01 | 5 | 0 | -5 |
| 31 | Hal | 0.01 | 5 | 0 | -5 |
| 32 | John F. Pittman | 0.01 | 3 | 0 | -3 |
| 33 | MarcH | 0.01 | 5 | 0 | -5 |
| 34 | mccall | 0.003 | 5 | 0 | -5 |
| 35 | Fred N. | 0.002 | 3 | 0 | -3 |
| 36 | BenjaminG | 0.001 | 5 | 0 | -5 |
| 37 | andrew | 0 | 5 | 0 | -5 |
| 38 | feabqtcqy | 0 | 1 | 0 | -1 |
| 39 | MDR | -0.005 | 4 | 0 | -4 |
| 40 | Adam Soereg | -0.01 | 4 | 0 | -4 |
| 41 | Anteros | -0.017 | 5 | 0 | -5 |
| 42 | Chuck L | -0.024 | 4 | 0 | -4 |
| 43 | Pieter | -0.032 | 5 | 0 | -5 |
| 44 | Cassanders | -0.043 | 5 | 0 | -5 |
| 45 | Anthony | -0.05 | 3 | 0 | -3 |
| 46 | Steve T | -0.05 | 3.7 | 0 | -3.7 |
| 47 | Roy Weiler | -0.05 | 5 | 0 | -5 |
| 48 | George Tobin | -0.05 | 5 | 0 | -5 |
| 49 | RobB | -0.056 | 5 | 0 | -5 |
| 50 | Bob B | -0.1 | 3 | 0 | -3 |
| 51 | janama | -0.11 | 1 | 0 | -1 |
| 52 | Jon P | -0.113 | 5 | 0 | -5 |
| 53 | KÃ¥re Kristiansen | -0.18 | 4 | 0 | -4 |
| 54 | Chris Brown | -0.98 | 5 | 0 | -5 |
The net winnings for each member of the ensemble will be added to their accounts.
To Chris Brown: -0.98? For now I’m assuming that was a typo. Please reassure me!
Jeff had another visit from FOIA it looks like most of the AR5 drafts.
When will NASA report, or have I messed up my cache again?
MikeN– I don’t know when NASA will report. I haven’t gotten an alert yet. 🙂
I won 75.366 Quatloos? I might even be out of the red now. Even a blind squirrel finds an acorn every now and then.
I am going to have to go back and look at which one of my 6 top secret algorithms I used for November and get a win streak rolling here.
BTW hearty congratulations to Dr. Spencer!
I am smack in the middle, 27th of 54. But my bet happens to be the highest number of all bets. What does this tell you about the betting cohort? Obviously, that most were looking for a much smaller or even negative number. Is this due to climate sceptic’s bias? You tell me.
denny – I picked a lower number because I am an idiot 🙂
Cangrats John Norris!
Drat, just outside the Quatloos again!
I have done a quick calculation and there seems to be 0.001 Quatloos unaccounted for in the prizes.
Can I have that?
Is Lucia getting rich on rounding errors?
The NASA/GISS data does seem to be on the late side this month, although it isn’t unheard of.
Actually I am having some difficulty in connecting with their site this morning. Not sure if it is the site or my connection, but everything else seems o.k.
NASA and HadCRU are often relatively late.
Ohhh!!!! I better go spend that on donuts. 🙂
Woo hoo!!
Quatloos for Christmas shopping!
RE: denny (Comment #87354)
“I am smack in the middle, 27th of 54. But my bet happens to be the highest number of all bets. What does this tell you about the betting cohort? Obviously, that most were looking for a much smaller or even negative number. Is this due to climate sceptic’s bias? You tell me.”
More like a classic case of chasing past performance. If you look at Ch 5 during Oct. it was heading straight for the basement. In Nov. it leveled off, denying those of us who like to forecast trends from past data. In Dec. it is moving down again. Will global temps continue to fall or reverse course throughout the rest of Dec?? That is the million quatloo question.
ivp0 –
Can you tell me the relationship between Ch 5 and the overall anomaly? Or rather, is the anomaly merely a mean of all the currently available channels? Are they weighted?
I only ask because my state-of-the-art algorithm is doing significantly worse than chance at the moment – I’m thinking of asking my cat for a January prediction..
ivp0 –
I suppose an example of my confusion is that channel 9 has shown a rising trend through Dec….
I’m aware that the amounts are tiny, but I’d appreciate an end to my consternation [on this one matter..]
Anteros,
My explanation of the relationship between UAH and AQUA Ch 5 might leave you more confused than you are now. Your cat might provide far more precise climate information. I use Ch 5 purely to determine if temps in the troposphere are trending up or down. Add in a little Nino 3.4 and I place my bet. (so far the house is winning)
Dr. Spencer has a good comparison discussion here:
http://www.drroyspencer.com/2009/01/daily-monitoring-of-global-average-temperatures/
I find it interesting that, while we have been in the throes of a 19-month, double-dip La Nina of sizable proportions (http://www.esrl.noaa.gov/psd/enso/mei/), the UAH temperature series has shown significantly higher LT temperatures than smaller La Ninas have produced in the recent past (http://www.woodfortrees.org/plot/uah).
This anecdotal evidence is in line with recent findings by Foster and Rahmstorf (http://iopscience.iop.org/1748-9326/6/4/044022) that the warming signal (composite of three surface and two tropospheric methods), corrected for ENSO events, volcanic activity and solar intensity, shows an unabated increase during the past 32 years, with the two warmest years from their analysis being 2009 and 2010.
ivp0 –
Thanks. If it is horribly complicated I may well end up more confused, but that seems hard to imagine at the moment! I checked Roy’s site again and he says they use Ch5 for estimates of mid- tropospheric temperatures…. Seeing as Ch4 is not functioning at the moment, do they use a GISS-esque extrapolation estimate for the lower troposphere? In a way I’m happy to leave my quatloos behind [I’m a serially ‘unlucky’ gambler..] but I would like to be able to make more sense of the AMSU page – particularly how the various measurements relate to the single figure monthly anomaly.
P.S. My cat is a worse gambler than me.
Owen –
It’s a fair point, even though I think F & R overdo it. I think it is especially pertinent for those [of us..] who can be seduced by short term trends and are tempted to believe that the radiative properties of carbon dioxide molecules can magically change, or can be deduced from short observational records. I’m not sure there is much supporting evidence for either of those hypotheses..
Owen (Comment #87406),
I have not read the paper by Foster and Rahmstorf, so I can’t comment on the quality of their efforts. However, it seems pretty clear that an underlying secular warming trend pretty much has to be visible if it is possible to accurately account for natural variations. Two questions stand out: 1) is there a longer-term oscillation (or pseudo oscillation) which continues to influence the recent trend, but which is not included in the Foster and Rahmstorf analysis, and more importantly, 2) what does the underlying trend tell us about Earth’s climate sensitivity?
.
I suspect an underlying trend of 0.13C-0.16C per decade (if that is in fact correct!) may be consistent with anything from quite low to quite high climate sensitivity, depending on how much ‘aerosol offset’ and ocean heat uptake is assumed.
I am going to place a bet today, mainly as a test and in case I forget tomorrow.
I will probably place a revised bet tomorrow, if I remember.
Drat, wrong blog!
SteveF,
The F&R paper, by the way is downloadable from the link I provided. I did read the paper, but cannot vouch for the statistical methods used (my sense, however, is that Foster is a pretty good statistician). WRT your questions, no longer-term oscillation is identified or fit to the data by F&R, and I don’t believe they comment specifically on sensitivity except that they demonstrate the unabated secular trend with a given slope.
The question of feedbacks also does not enter in, but recent interest and findings in methane release from tundra (http://nsidc.org/news/press/20111130_permafrost.html) and from arctic ocean clathrates (http://neven1.typepad.com/blog/2011/12/arctic-methane-russian-researchers-report.html) may portend increases in the slope of the warming trend.
Owen, Foster and Rahmstorf reinforce something that I’ve been arguing for all along…that most of the variation between series is explained by differences in the amplitudes of short-period climate fluctuations in their reconstructions. This in turn can be explained by differences in how they perform the global averaging.
Averaging over the surface of the Earth is a form of “wavenumber filter”, where, if you had an unlimited number of sensors, you’d receive only the zero-frequency spatial component upon computing the average. Because there is a finite coverage of the sensors, and because the coverage is not uniform spatially and contains temporal gaps, there is leakage of shorter-wavelength fluctuations through the global averaging process, and these shorter-wavelength fluctuations have commensurately shorter time-periods associated with them.
It turns out that for periods of less than about 30-years (and possibly longer), there is a substantial influence of exactly how you compute the global mean average, given incomplete spatial coverage and temporal gaps in the data, on your resulting average.
For the last decade, some CAGW types have preferred GISTEMP to the other series because, for the last decade, it is running warm, and have railed against HADCRUT because it has been running cool. It may come to pass in the coming decade or these two series switch places, with GISTEMP running cool and HADCRUT running warm… and all of this due to the nefarious effects of short-period climate fluctuation on the series.
Of course picking which series to use based on whether it confirms your prior beliefs is a mistake by itself, neglecting to inform your belief by the very real manifestations of short-period climate fluctuations when comparing temperature series is another (most will admit that these fluctuations can lead to a short-duration “flattening” of warming, the people favoring GISTEMP due its current high-ball values don’t admit these fluctuations can lead to short-period incongruences among the various surface reconstructions).
Whether F&R have done it correctly is a fair question, but I think the general methodology is a good idea. But if confirmed by others, I do think it’s the clearest statement yet of the problems with people dismissing the influence of short-period fluctuations on temperature measurements.
SteveF,
For kicks, I pointed out that some results of F&R11 implicitly support the oft-criticized argument of Spencer and Braswell with respect to measuring climate sensitivity:
http://troyca.wordpress.com/2011/12/12/foster-and-rahmstorf-2011-lends-support-to-spencer-and-braswell/
Can anybody else provide me with the missing information? [comment 87407]. I’d like to know how the lower tropospheric anomaly is calculated given that Ch5 measures mid tropospheric temperatures (and Ch4 is temporarily defunct). Are the other channels factored in somehow? Is extrapolation used?
Anteros,
If I understand correctly, ch5 is used to construct both mid AND lower tropsopheric temperatures. Ch5 simply refers to the frequency (~53.5 GHz) the receiver is sensitive to (although AMSU technically uses two frequency bands). AMSU will monitor that frequency at a number of different view angles, and since the temperature at different altitudes will affect the observed radiation differently for each angle trhough the atmosphere, they attempt to isolate the temperature at these rough levels (TLT and TMT) based on differences in the observations at these various view angles within channel 5.
Here’s the description for RSS, although I think the basics apply to UAH as well:
http://www.ssmi.com/data/msu/support/Mears_and_Wentz_TLT_submitted.pdf
Isn’t this pointless being a month by month bet? The month by month noise is larger than the warming signal, so even if you get the trend right, you will lose to people who guess correctly, since no one can predict the noise.
Re: bugs (Dec 16 16:52),
> Isn’t this pointless being a month by month bet?
Yes.
This is strikingly similar to the reason that there is no market for betting on individual games of football, basketball, and so forth.
And why Las Vegas is a sleepy cow town.
Still… if your understanding is subtle enough, you just might become a quatlillionaire… pretty tempting!
🙂
“Isn’t this pointless being a month by month bet?”
bugs is so ate up with Global Warming, he doesn’t even know how to have fun anymore. 🙁
Andrew
I think lucia should earn a vig AMAC,
maybe even loan people quatloos.
Troy CA @87431
Thanks – much appreciated. The link is very informative.
Mosher – I’ve tried to persuade Lucia to ‘advance’ me some of my future winnings – strangely reluctant..
Mosher–
I don’t know what rate of interest I should charge for advancing quatloos. What do credit card currently charge? We always pay off our balance so I don’t keep track of that. Also, how do I investigate whether people have sufficiently good credit ratings? Does some sort of inter-galactic credit ratings agency exists? I don’t want to lend quatloos at 2% a month if it turns out Anteros is roughly as credit worth as Greece without Germany or France to back him up.
The Foster and Rahmstorf paper is a reasonable effort but they did not include the AMO in their reconstructions of natural variation.
Foster, of course, did try out the AMO in his pre-analysis work for the paper but then decided to discard it because it resulted in too low of a warming residual afterward.
Tamino even produced this chart in Jan, 2011 so why wouldn’t Foster have included it – pretty obvious that it should have been included.
http://tamino.files.wordpress.com/2011/01/sst_giss.jpg?w=500&h=325
I note that the UK Met Office is now using the AMO as one of the components in its medium-term climate model (which is the only accurate one I imagine). They describe the AMO as “highly statistically significant”.
UK Met at the 2011 AGU.
http://vimeo.com/33332398
Hello Anteros,
Sorry for the delay. We are having a fierce windstorm in SoCal today and my cable internet connection has been down more than up.
Regarding Ch 5, as I understand it Ch 5 takes a measurement slice at 14,000 feet elevation. This is 600 millibar region and it is here that weather mostly happens. A keen meteorologist could probably fill in the blanks but by measuring here we get a picture of changes in the mid-lower atmosphere which is a good indicator of weather and climate patterns. Other channels take a slice at other altitudes and give us different information about our atmosphere. I hope the helps somewhat. Here is a link to the site:
http://discover.itsc.uah.edu/amsutemps/execute.csh?amsutemps
ivp0 –
Thanks. Troy CA chipped in with this link which has a lot of background info’. http://www.ssmi.com/data/msu/support/Mears_and_Wentz_TLT_submitted.pdf
@Bill Illis
Tamino explains exactly why you are wrong. Can you read what he says.
http://tamino.wordpress.com/2011/01/30/amo/
Your next link is just as bad. They are talking about predicting the temperature for the next year, not the temperature trend for the next century. The year to year warming signal is smaller than the year to year variation.
Bugs,
I find the following statement by Tamino to be confusing:
“global warming is the cause, not the effect, of much of the changes in N.Atl SST anomaly”
Does any measure of “global warming” not include SST anomaly? If it does, then how can the “output” be the cause for the “input”?
Also what is the proccupation with the lowpassed “non-linear trend” which also appeared in the Coumou and Rahmstorf paper? Can I just make up my own definition for “trend” by lowpassing my data below some arbitrary boundary? Wouldn’t it make more sense to call the “nonlinear GW trend” the part which has some (lag)correlation to forcing and the rest the low-frequency variation around that nonlinear trend?
If Tamino wants to stick to his definition of a nonlinear trend, then the following opening of that blog entry:
“I regularly get comments claiming that ocean cycles are the cause of global warming. They couldn’t be more wrong.”
seems almost trivially wrong doesn’t it?
Luica,
25% is a good rate
I will make up a silly name for him if he falls behind on payments, like breaking his legs, only virtually.
bugs (Comment #87445)
December 16th, 2011 at 7:44 pm
———————
The Foster and Rahmstorf paper was about removing the impact of known factors of short-term variability on temperatures from 1979 to 2010, so that we could see how much global warming forced by GHGs actually happened.
The AMO index is detrended so that it does not include a global warming trend. Like the ENSO, the AMO index has ZERO trend from 1856 to today. Tamino’s post did not explain it properly.
The AMO does very closely match temperature trends over any time scale we can look at. So, it either just reflects how those temperatures are changing over time, or it is one of the big factors affecting those temperatures over time.
But the temperature record and the AMO both have matching long-term cycles up and down and short-term rapid spikes up and down (unrelated to GHGs or volcanoes or solar influences). ie They exhibit Natural Variability.
So, either the AMO is just a really good indicator of the other (so far unknown) factor which is driving those long-term cycles and short-term natural variability, or it is the driver.
I don’t recall typing that! How embarrasing. Was probably meant to be -0.098, but it was a while ago.
Oliver,
Yes, it is trivial that *raw* sst should correlate with global warming. Please explain that to bill so he stops posting a comparison between sst and temperatures as if it were somehow relevant.
Bob’s point is not about sst but about AMO. It is more subtle. AMO is detrended in order to remove (long term) warming signal. IIUC Bob notices that AMO is nevertheless strongly correlated with temperature, especially over the last two decades. Since this cannot be caused by the global warming signal which is removed by the detrending. Therefore, he concludes, AMO must be a strong driver of natural variability, AND almost overwhelms global warming in influencing temperatures over the last 20-30 years.
IIUC Tamino points out that while AMO detrending is linear, the global warming signal isn’t. Therefore you can’t claim that detrending removes global warming, and you can’t say that AMO causes recent warming much more than co2. The claim is nonsensical, precisely because AMO does include a co2 driven, global warming signal. Which is what Bob doesn’t acknowledge.
Disclaimer: the above is my understanding of what Tamino claims about what Bob says. So it’s pretty indirect. Grains of salt not provided.
Toto:
Reference please? And not just tamino, he’s well known for making authoritative sounding statements that have zero content.
Well on this particular case he does have content, and I found it pretty convincing.
If you accept that co2-induced global warming did not occur as a pure linear trend from 1856 to present (which is how AMO is detrended), his argument seems pretty straightforward.
I may have missed something. Feel free to follow bugs’ link and point out any flaw.
You don’t have to use linear detrending to show that AMO amplitude is increasing over time, Fourier based analyses show the same effect. So the linear versus nonlinear detrending argument is just a red herring.
Regardless… if you have a leakage between two measures, who’s to say which causes which? It seems perfectly plausible that AMO modulates climate just as other atmospheric-ocean oscillations affect climate do (and drive long-period natural variability).
Saying that it affects climate doesn’t say that it is responsible for global warming, but that it affects climate is not only a plausible statement, it is one supported by climate science.
It is even possible that part of the global warming signal we’ve seen may be associated with AMO, but I think you would need a GCM that could accurately model AMO in order to be able to assess this sort of effect though. So I also think it’s a mistake to just look at AMO and other natural sources of variability without considering changes in secular forcing over time (e.g., what’s wrong with Tisdale’s arguments, IMO).
Toto,
Here is the Raw Unaltered AMO SST data versus the IPCC AR4 Multi-Model Mean hindcast.
I think it is pretty clear, the AMO is not responding to GHG forcing or Aerosols or Solar forcing (although there might be a volcanic signature in it).
It is about as different as it can get.
http://img853.imageshack.us/img853/2836/rawamoandar4hindcast.png
Please tell Tamino he screwed up.
Bill Illis,
Nice graphic. It is pretty clear that the de-trended AMO index is highly correlated with variation of the temperature history around an underlying (causal GHG forcing?) trend. Whether it is cause or effect is not clear, but Tamino is whistling past the graveyard with arm-waves about “non-linearity”. That is just silly. The AMO is widely discussed in peer reviewed literature… Tamino is certainly not the last word on this subject, and whatever he contributes can safely be discounted will little or no loss off understanding. At least Tamino is consistent: he will always generate some kind of excuse (no matter how far fetched) to try to discount data which conflicts with the “story line” of extreme future warming. Many, many grains of salt need to accompany any of Tamino’s many excuses for conflicting data.
Be careful when comparing the published AMO index to global or regional temps over the last 30 years or so, as the century-scale linear detrending won’t properly remove accelerating anthropogenic forcing: http://rankexploits.com/musings/2011/the-atlantic-multidecadal-oscillation-and-modern-warming/
Zeke,
Of course the underlying trend may be less than linear!
However, even linear de-trending should remove much (and maybe most) of the GHG forcing signal. If you look at the rest of the oceans (say, the HADSST3 trend) you see that the rest of the world’s oceans do not so closely follow the AMO trend. It is not clear what is cause and what is effect here, but it is clear that the correlation of the de-trended AMO with deviation from the secular trend is very strong. There are plenty of peer reviewed papers which suggest the de-trended AMO index provides useful information about a natural pseudo-oscillation. Please clarify: do you really think all that literature is mistaken? (honest question)
SteveF, Bill Illis, or anyone else,
What physical meaning does the detrended (linear or otherwise) AMO have, and why might it be expected to correlate with deviation from the secular trend? With ENSO, we can readily envision a giant heat exchange mechanism across the Pacific through which heat is exchanged in an (almost) cyclic manner between ocean and atmosphere. It makes physical sense. But what physical sense is associated with AMO?
Todo,
Thanks for that constructive contribution.
Yes, I read both Tamino’s nonsense and Zeke’s much more measured post from earlier this year. Since you take the liberty to question my comments, let me question yours: Do you ever actually think about the substantial conflicts between measured temperatures and the prescribed monotonic warming from climate models? I rather suspect not.
.
If you can rationally remove natural variation from the historical data, that should help you understand the underlying trend (which is a very good idea!), which can then be related to GHG forcing and climate sensitivity. If you instead choose to ignore/discount a rather obvious oscillation like the AMO (widely discussed in the literature) because you find it inconvenient to promote your POV, then you and Tamino are only fooling yourselves.
Owen,
I think Wikipedia lays out the issue pretty clearly:
http://en.wikipedia.org/wiki/Atlantic_multidecadal_oscillation
It is not clear how much is cause how much is effect, but there sure seems to be at least some theoretical support for this process. I note that Lake Okeechobee in Florida is at a very low level for this time of year, just as the detrended AMO index would suggest it should be.
Whether the AMO is cause or effect, it strongly suggests some kind of underlying oscillation. If you use Zeke’s analysis from earlier this year, there is STILL a +/-0.1C oscillation that remains. Yes, that does suggest some of the rise in temperatures from the late 1970’s to now was due to a natural oscillation, and yes, that conflicts with the 100% GHG driven version of the story.
SteveF, my own Fourier-based analysis suggests something closer to ±1°C for the AMO. My suspicion is that his detrending subtracted off part of the AMO signal.
There’s a way of checking this—you do a simultaneous fit to the AMO amplitude + secular trend. It’s not self-consistent to fit to the trend first then analyze the residual for the AMO…
(The way I get around this is to do a window-by-window detrend, which prevent over-fitting by the polynomial of the oscillatory portion of the signal.)
Hmm… my bad…wrong units: that was a power spectrum.
Here’s the amplitude spectra for three different series. The amplitude of the “AMO” is between 0.12 and 0.15°C.
Sorry for any confusion that caused.
Re: Owen (Comment #87483)
A link to the meridional overturning circulation in the Atlantic has been discussed in the literature. I see that such a connection been mentioned in the Wikipedia article linked by Carrick, but the Wiki gives no references, so see Knight et al. (2005), also Delworth and Mann (2000).
Essentially, strong phase of meridional circulation -> more poleward heat transport -> warmer Northern Hemisphere (big part of AMO).
The AMO is not responding to “forcings”. I showed that in the chart above.
In 1877-88, the Raw AMO SST data went to +0.46C. It has only been matched in a few months since then. Last month, it was only +0.12C.
It is going up but it has its own long-term trend and one that is best described as “linear” with a +/- 0.55C long-term and short-term oscillation about that trend. That trend is only 0.0245C per decade, less than half of Hadcrut3’s trend over the period.
Obviously, the linear trend is tied to the general warming of the planet. But the warmest AMOs were the early 1850s, 1876-88, mid-1940s, 1998 and mid-2000s. Those periods are not GHG forcing periods (if anything, they are high El Nino periods).
oliver (Comment #87493)
Thanks so much for your response. So, the AMO reflects a low frequency cyclic change in thermohaline circulation in the North Atlantic and therefore affects ocean-based heat transport from the equator northward? If so, would this not produce more of a localized northern hemispheric effect with no net change expected in an average global temperature?
Oliver or Bill Illis,
One more thing, if I may. Regarding the AMO, I assume that the “warm phase” (usually shown in red and above the line) means that North Atlantic SSTs are running warmer during those periods and vice versa for the cool phase. Is this understanding correct?
bugs (Comment #87445) says: “Obviously they’re strongly correlated. Bob Tisdale (and others) simply can’t wrap their brains around the fact that global warming is the cause, not the effect, of much of the changes in N.Atl SST anomaly.â€
Obviously bugs “simply can’t wrap his brains around the fact that†Tamino can make major errors in his analyses, like with his critique of my post. Tamino used the wrong SST dataset for his comparison with GISTEMP. Refer to my response to Tamino’s post here:
http://bobtisdale.wordpress.com/2011/02/03/comments-on-tamino%e2%80%99s-amo-post/
Toto (Comment #87461) says: “Therefore you can’t claim that detrending removes global warming, and you can’t say that AMO causes recent warming much more than co2.â€
Actually I can. The multi-model ensemble mean of the CMIP3/IPCC AR4 climate models, which basically represents the forced component of the models, shows a linear trend for the North Atlantic of approximately 1.6 deg C per Century. But the observed rise in the SST anomalies of the North Atlantic is approximately 2.5 deg C per Century.
http://i55.tinypic.com/vfinn5.jpg
The graph is from this post:
http://bobtisdale.wordpress.com/2011/04/19/492/
Toto continued, “The claim is nonsensical, precisely because AMO does include a co2 driven, global warming signal. Which is what Bob doesn’t acknowledge.â€
Looks more like your comment is nonsensical.
Owen,
Since AMO is just an observed mode of variability, I guess it would be most accurate to say that it reflects whatever underlying reasons cause the pattern (e.g., changes in the large scale circulation and therefore the heat transport, or whatever other reason given to explain the AMO). Once you have SST changes across the N. Atlantic it wouldn’t be too surprising for the atmosphere to propagate this state information to other regions (but in what way, I don’t think anyone could precisely say). What effect this would have on average global temperatures would not be clear.
You might imagine a redistribution of heat poleward leading to a higher average temperature at (near-) equilibrium, just due to the T^4 scaling of Stefan-Boltzmann. This comes from the difference between averaging T^4 and T. I will have to think about that some more in the morning when I am a bit more coherent. 🙂
Oliver,
If the AMO reflects fluctuations in the thermohaline-driven north Atlantic currents, it probably also reflects oscillations in the oceanic/atmospheric heat exchange (along the lines of ENSO) associated with the gulf stream warming of northern latitudes. If so, it may well have a global impact. In it’s warm phase, for example, we would expect more ocean to atmosphere heat transfer, the temperature effect of which would then be averaged into the global mean temperature.
Owen, the AMO shows up in precipitation data too.
Thus it could be modifying cloud patterns (and show up via albedo changes). It could also influence the agricultural sector of the global economy (for example), which in turn could show up as a modulation of economic productivity.
This reference has a good review of the topic:
Carrick,
The bit bout the extent of Arctic sea ice might make a certain climate scientist, who has bet some serious money on the sea ice trend over the next few years, more than a little nervous. 😉
.
So, is it safe to say you disagree with Tamino about the unimportance of the AMO WRT the rate of warming since the late 1970’s?
.
The hubris fairly well drips from Tamino’s posts; it’s hard for me to imagine something more satisfying in climate science than seeing his pompous prognostications turn out flat wrong.
SteveF:
It’s safe to say “I don’t really know.” 😉
Tamino has this bad habit of making proclamations…appeal to self as authority. And this bad habit of threading in insult and innuendo at anybody who disagrees with him on his threads. It makes what he says almost unreadable to me. I almost have to block diagram it on part to not get distracted by the “used car salesman” gig he does.
Carrick,
Thanks for the link to the Dijkstra et al. paper. The AMO is clearly a well-established phenomenon linked to climate effects in North America and Europe. I find it interesting that two seminal papers involving proxy data from Michael Mann (http://www.gfdl.noaa.gov/bibliography/related_files/td0004.pdf) were used to extend the time scale backwards for a duration long enough to confirm and define the AMO as a legitimate long-term cycle. Of course, in this case, Mann’s proxy data must be solid (as opposed to his proxy data as related to global temperature anomalies).
Owen, there’s no question that, in general, tree ring proxies measure climate, the question is what. Probably what most of what tree rings measure correlates with precipitation and cloud cover.
I’ve actually found analyzing the underlying tree ring data to be pretty interesting, without necessarily trying to connect them to a disparate physical quantity pretty interesting. (The later requires the leap of faith that over the entire interval of the proxy that it exhibited temperature limited growth, which in paleoclimate terms means picking trees that are currently near growth boundaries.)
Nonetheless, interesting paper.
Thanks as always for the informative graphs.
A day late and a dollar short, here’s a histogram trends for 2002-2052 from the the A1B scenarios:
Figure.
I thought this was the best way to display the differences between the models during the “validation period”.
The distribution is obviously very non-normal, as well as being multi-modal (peaks near 0.2 and 0.3 °C/decade).
Carrick,
Nice graphic, but my eyeball estimate is a mean of about 0.27C per decade, not the 0.2C per decade the IPCC seems to throw about.
.
At least one model seems to have the trend (0.15) close to right! 😉
Pretty darn close, SteveF!
The arithmetic mean for that period is 0.264°C/decade (stddev of 0.046°C/decade), median is 0.271°C/decade with an interquartile range of 0.229-0.298°C/decade.
If you’re interested in that sort of thing, the geometric mean is 0.260°C/decade (technically it’s more appropriate here than arithmetic mean).
For the ten year period 2002-2012, the arithmetic mean of the trends is 0.19°C/decade, with a standard deviation of 0.15°C/decade (the error of the mean, probably meaningless, is around 0.02°C/decade).
Carrick,
Humm.. 0.19 +/- 0.15 (one sigma!) for 2002 to 2012. I guess we can be pretty certain the variability in the models is high enough to include most any measured warming trend. And since there is run-to-run variability for individual models, at least some model run of one of the models will always be close to right (AKA the models are consistent with measured trends). I love this kind of science. Kind of like a class in art appreciation, but with fewer course requirements, and no grades.
Of course it makes sense that the trend should be increasing as you go forward in time from the present because A1B assumes that CO2 forcing is increasing over that period.
Here’s the model means and error of the means as a function of calendar year.
Carrick,
Any idea why the trend jumps by 0.07C per decade between 2017 and 2020? Hard for me to see how that big a jump in decadal trend can take place over just 4 years.
I think it’s because they assume a sudden rapid increase in CO2 starting around 2010.
Figure.
Here’s a revised version of the other figure showing the 95% CL interval (shaded green area).
Figure.
I’m showing the points for the temperature trend for a 10-year interval centered on each five year increment. The CL bounds however are shown for yearly increments.
Carrick,
Or maybe because 2021 is far enough away to be safe from refutation, while 2014 or 2017 are too close for comfort. Naw, that is too cynical.
Carrick,
Well in any case, if the warming rate doesn’t reach at least 0.22C per decade by 2021 people are really going to start asking questions.
SteveF, I guess the trick is to look at how the real CO2, aerosol emissions look versus time versus the assumed ones.
If I remember right, CO2 emissions are higher than A1B. My memory could be faulty though.
I’ve not found a text version of the A1B forcings, and have better things to do than google for one at the moment.
If somebody can point towards that, it would be appreciated.
Carrick,
Yes, that would be the trick, but aerosols (and especially secondary aerosol effects) are not really known, nor will they be in the near future. Chinese sulfate emissions are already being invoked to explain lower post 2001 warming. Climate science is a beautiful thing… almost like an art form.
I never really have a good answer to that…
I suppose you would put in the entire range of allowed aerosol emissions and use that to bound the radiative forcings. I suspect the uncertainty is big enough (especially when you’ve included the poorly modeled indirect aerosol effects) that a wide range of sensitivities is possible.
On another level, you can argue that, “Based on these premises, this is the result you get.” For climate science is “these premises” already get them into plenty of trouble (like no detectable AGW warming prior to 1970), so I’m not sure it’s absolutely necessary to step back and demand the full range of uncertainty be prominently displayed…