It looks like the “skeptics” (if that’s what you call them) were right again. Steve McIntyre pointed out a paper in print at GRL that was similar in content to some of his an Ross McKitrick’s own work (MMH10)demonstrating again that climate models run generally higher than observations. This is particularly true in the case of the tropical warming in the upper troposphere covered in this paper – On the warming in the tropical upper troposphere:Models versus observations (free). The paper was written by Qiang Fu, Syukuro Manabe, and Celeste M. Johanson (FMJ11).
Abstract:
[1] IPCC (Intergovernmental Panel on Climate Change) AR4 (Fourth Assessment Report) GCMs (General Circulation Models) predict a tropical tropospheric warming that increases with height, reaches its maximum at ∼200 hPa, and decreases to zero near the tropical tropopause. This study examines the GCM‐predicted maximum warming in the tropical upper troposphere using satellite MSU (microwave sounding unit)‐derived deeplayer temperatures in the tropical upper‐ and lower‐middle troposphere for 1979–2010. While satellite MSU/AMSU observations generally support GCM results with tropical deep‐layer tropospheric warming faster than surface, it is evident that the AR4 GCMs exaggerate the increase in static stability between tropical middle and upper troposphere during the last three decades. Citation: Fu, Q., S. Manabe, and C. M. Johanson (2011), On the warming in the tropical upper troposphere:
My bold of course. This is a key factor in the global warming debate as the missing hot spot is an indicator that feedback mechanisms are not properly quantified in climate models. The results of FMJ11 are rather stunning to see in print these days as there was um …. considerable review difficulty experienced in rebutting the Santer 08 paper claiming that models do match observations. It is worth noting that Santer has a new paper out on the topic which I will discuss in another post here.
Now this is a slightly different claim than Santer and MMH10 which looked at simple trends of the atmosphere, but the result is perhaps more important in that climate science is now recognizing in print that key differences between observed vs modeled feedbacks DO exist and that modeled trends are overstated. I have yet to see any of the best known climate scientists admit these unquestionable discrepancies. The methods of the paper seem reasonable enough and are not difficult to follow except that you do need to understand the data in question is the difference in the temperature trends between tropical upper‐ and lower‐middle troposphere (T24‐T2LT) as defined and discussed in the paper linked above. How much upper vs lower troposphere warming is observed is a result of CO2 heat capture and feedback in real world and modeled climate systems.
Below is the key plot of modeled vs observed temperatures:
As we have seen so often in blogs and print, the models run hot in comparison to observed trends both individually and in aggregate. Again, the difference here is that this paper analyzes the heat buildup region of the tropical troposphere which is a key prediction of climate models. This is different from the result of MMH10 which shows that the model trends are generally high. Instead, this is a clue to the reason as to WHY they are running high.
In the text, however, a simple statement is made which corroborates MMH10, the numerous posts by Chad Herman at TreesFortheForest and Steve McIntyre on Briffa 08, Lucia and others and contradicts the general conclusions of Santer08 that climate models are accurately representing observation.
Tropical surface temperature trend from multi‐model ensemble mean is more than 60% larger than those from observations (Table 1), indicating that AR4 GCMs overestimate the warming in the tropics for 1979–2010. Thus larger T24‐T2LT trends from AR4 GCMs are partly caused by GCM overestimation of tropical temperature trends.
The models are running too hot. Either that or the thermometers need to be replaced. There has been a lot of denial in the climate scientist community on these basic facts perhaps we should begin collecting quotes for the future retractions as the main stream will soon need to address this reality. One can’t help but wonder if a change in predicted warming magnitude will affect the overall anti-industrial, anti-capitalist message of the climate community.
The conclusion with my bold:
One of the striking features in GCM‐predicted climate change due to the increase of greenhouse gases is the much enhanced warming in the tropical upper troposphere. Here we examine this feature by using satellite MSU/AMSUderived deep‐layer temperatures in the tropical upper‐ (T24) and lower‐ (T2LT) middle troposphere for 1979–2010. It is shown that T24‐T2LT trends from both RSS and UAH are significantly smaller than those from AR4 GCMs. This indicates possible common errors among GCMs although we cannot exclude the possibility that the discrepancy between models and observations is partly caused by biases in satellite data.
IPCC AR4 GCMs overestimate the warming in the tropics for 1979–2010, which is partly responsible for the larger T24‐T2LT trends in GCMs. It is found that the discrepancy between model and observations is also caused by the trend ratio of T24 to T2LT, which is ∼1.2 from models but ∼1.1 from observations. While strong observational evidence indicates that tropical deep‐layer troposphere warms faster than surface, this study suggests that the AR4 GCMs may exaggerate the increase in static stability between tropical middle and upper troposphere in the last three decades. In view of the importance of the enhanced tropical upper tropospheric warming to the climate sensitivity and to the change of atmospheric circulations, it is critically important to understand the causes responsible for the discrepancy between the models and observations.
With so many papers coming out in support of a less severe warming effect, the climate community might do well to accelerate the release of AR5 before they are forced to recognize these problems. As an aside, it isn’t the message that is causing public opinion to change, it isn’t the fact that the public doesn’t understand, it is the regular discovery of the truly uncertain nature of the climate future which cuts into the message.
the Air Vent 
The link to FMJ11 doesn’t work for me.
An alternate link is http://www.atmos.washington.edu/~qfu/Publications/grl.fu.2011.pdf
(appears to be Q. Fu’s site at U of Washington).
Pielke Senior noticed this about a month ago:
http://pielkeclimatesci.wordpress.com/2011/07/08/new-paper-illustrates-another-failure-of-the-ipcc-mullti-decadal-global-model-predictions-on-the-warming-in-the-tropical-upper-troposphere-models-versus-observations-by-fu-et-al-2011/
It looks like this paper is actually downplaying the size of the discrepancy between models and observations. For instance, this statement:
“While strong observational evidence indicates that tropical deep-layer troposphere warms faster than surface”
Appears to be flat out untrue. The “best estimate” of the ratio of lower tropospheric warming to surface warming in the tropics is .8±.3 and the ratio for mid tropospheric warming to surface warming is even lower, around .4:
Click to access r-358.pdf
I cannot find the files for T24 on the UAH website. It appears to be a metric used in this study but not published by UAH themselves. It is some sort of suspicious combination of stratospheric series and mid-tropospheric series, given the authors it probably involves a weighting scheme that violates the laws of physics rather blatantly to enhance the trends.
So what we are left with is even if you cheat you can’t reconcile the satellite observations with the models.
#2, You have to genuflect if you want to please the lords.
The T24 combination seems reasonable to me. It adds to 1 and has a weighting function set higher in the troposphere so it is hard to see where the violation of energy would be. There are a lot of these types of analyses in print now so AR5 is going to have difficulty ignoring them – well it should anyway.
Not a good name for a paper.
Fu Et Al
Although … F.U. et Al might be the emotional response for warmenistas.
“I have yet to see any of the best known climate scientists admit these unquestionable discrepancies.”
Syukuro Manabe is the daddy of them all. Manabe and Wetherald 1967 was a pioneering modelling paper.
It’s great to see him still publishing.
Nick,
I should have written “most popular” or something. These three are all known.
Jeff the link to FMJ11 didn’t work for me either. I got:
Sorry, we are unable to retrieve the document for viewing or you don’t have permission to view the document.
Alternatively, you can view the plain HTML (automatically redirecting).
and then a quick jump to a page saying my gmail account is temporarily unavailable (which is untrue – had no trouble accessing it).
Thanks for the working link Charlie A
Nick#5
Well, I guess with a guy with that level of ‘cred’ he is unlikely to get ‘pier reviewed’.
Steve,
It works fine for me. Hmm.
Jeff,
Reality has a way of winning all arguments, at least in the long run. Still, I think you can count on a few hastily prepared papers coming out prior to AR5 which claim the result of Fu et al (11) is less certain than it really is….. politics is terribly important you know.
It is most likely that AR5 will (again) reduce the “most likely” warming rate a bit (my guess: a bit under 3 C per doubling, maybe 2.7 to 2.9), while widening the uncertainty range sufficiently to maintain some fig leaf of possibility for the most frightening of catastrophic warming scenarios. Vast uncertainties in aerosol estimates will be maintained to lend credibility to even the most outrageously wrong models, and I bet that ocean heat accumulation (where the uncertainty really is much less than even 8 years ago) will be claimed to be more uncertain than it really is. We know the script already, since Hansen is already doing an ad-lib.
FMJ11 reference list:
2 x Fu et al
2 x Christy, Spencer et al
1 x Spencer & Christy
2 x Santer et al
1 x Hansen et al
0 x Dessler et al
0 x Trenberth et al
Although atmospheric deep‐layer temperature
time series derived from different research groups still give
diverse trend results, the MSU/AMSU atmospheric deep layer
temperatures represent one of the most reliable data
sets for the trend analyses.
Clearly they forgot to induct this good old boy into the very inner circle of the Lodge…oops
3-I find negatively weighting on some of the stratospheric temperatures an odd, unphysical thing to do. Well, I suppose one can say, they did the same with the model temperature profiles, so it’s apples to apples, at least, even if they are both poison apples. Considering, also, that this measure has the effect of reducing the discrepancy between model and observed amplification ratios substantially, I have to wonder even more why they chose to do the comparison this way. Other comparisons not involving suspicious weighting schemes show larger differences.
I seem to recall we had some discussion here about “reverse thermometers” in the Antarctic recons, which strikes me as what Fu et al. do in the higher parts of the lower stratosphere.
Conflicting information on energy that powers the Sun, changes Earth’s climate, and sustains our lives [1-3] is a grave danger to National Security.
The National Academy of Sciences was “established by an Act of Congress . . . signed by President Lincoln on March 3, 1863, . . . which calls upon the NAS to “investigate, examine, experiment, and report upon any subject of science or art” whenever called upon to do so by any department of the government” [1].
Now that we have a Republican House, I ask that you please join me in asking Congressional representatives to request a FULL and CANDID evaluation from NAS of conflicting information in [1-3], BEFORE approving more funds for government research.
References:
1. UN’s IPCC reports (1990-2007)
http://www.nasonline.org/site/PageServer?pagename=ABOUT_main_page
2. Super-fluidity in the solar interior: Implications for solar eruptions and climate”, JFE 21, 193-198 (2002)
http://arxiv.org/pdf/astro-ph/0501441
3. “Neutron repulsion”, in press
http://arxiv.org/pdf/1102.1499v1
The time has come to stop arguing and for Congress to ask NAS to fulfill its responsibility to find out if experimental data and observations have been misrepresented.
Sincerely,
Oliver K. Manuel
http://myprofile.cos.com/manuelo09
Some explanation of why I think the T24 measure is erroneously reducing the actual discrepancy between models and observations:
First, what are the implications of weighting large areas above the tropopause negatively? The cooling in those layers is turned into warming in the overall result. Now, in the lower areas near the tropopause, the weighting is positive, so that if the stratosphere were cooling perfectly uniformly with height, there would be total cancellation. In point of fact, however, the trend becomes more strongly negative with increasing height. This by itself is not a problem. However, if you look at the radiosonde profiles (see Douglass et al-this was not a point they were criticized on), the observations don’t just show that there is not a strong maximum of trends at about 10 km in the tropics, they also show that the trends drop off more rapidly with height to become negative much lower in altitude than models, and more negative in general. Now, what do you think will happen, if you take a region that cools more than models say, and an area that warms less, and average them with negative weight applied to the area that cools too much? It cancels part of the discrepancy of the other region with too little warming. The models look better because their higher altitude trends are even worse. That’s messed up.
TTCA,
It isn’t stratospheric temperature being negatively weighted but rather an entire profile of the atmosphere. I agree that it is odd, but if you want to maximize a vertically regional response it makes sense. Imagine the response curves are exact matches above 1000m and one curve steps to zero below 1000m. If you want the below 1000m response only you take the difference.
Not the same as reading a thermometer upside down IMO.
Steve F #11,
There are a lot of Steve’s in blogland these days. 😀 We need more Jeff’s.
Unfortunately, I think you have nailed the worst case correction for them. If they can’t make it go away, your answer is the most expedient for the politically motivated types. Expand confidence intervals, tweak averages minimally, same results. Even reason can’t fight the blob.
#15 TTCA,
I get it. So we have a reduced quality of non-match between models and observations. The questions become: Is there some way that the authors could have avoided this bias and made a more accurate point?, Is it a simple matter that they should have disclosed this problem more clearly?
Re Nick Stokes, Re Syukuro Manabe:
Good catch, Nick. It appears Prof. Manabe may still be visiting at Princeton,
http://www.princeton.edu/aos/people/faculty/manabe/
— and his 80th birthday is coming up Sept. 21,
http://en.wikipedia.org/wiki/Syukuro_Manabe
Happy birthday, Prof. Manabe!
18-I’ll be honest, I am probably being too suspicious of them: they may not even realize the reason why their way of weighting the data is giving what appears to be an erroneous improvement of model-data relationship. Given their willingness to point out that the models are not matching reality very well, they perhaps aren’t even aware that they are reducing the actual difference. To me, however, it seems that, if one wants to have a net zero weight above the tropopause, one might as well just have all heights above the tropopause weighted zero. It isn’t a nice smooth weighting function, but at least it isn’t turning stratospheric cooling into tropospheric warming, which, in net, their algorithm appears to do, because the cooling is stronger where they weight negatively. I do understand why they specifically want a record of temperatures from the upper troposphere, one which has little stratospheric influence: That is where the signal is theoretically (but not in reality, apparently!) maximized. However, given that their amplification ratio is higher than for both the middle and lower troposphere records (both less than one, although lower troposphere not significantly so) it makes one wonder how they could get such high upper tropospheric trends, given the data from which they are derived. The reason appears to be overcompensation for stratospheric influences…
Meantime, while the resident self appointed Mormon theologian-climatologist from BYU e.g.:
http://maxwellinstitute.byu.edu/publications/review/?vol=12&num=1&id=332
http://www.tektonics.org/af/bickmore02.html
http://bbickmore.wordpress.com/about-this-blog/
is busy as a little bee calculating how much juicy Apostasy he can squeeze out of those pesky sceptical scientists to balance on the pin of his head:
http://bbickmore.wordpress.com/roy-spencer/
http://bbickmore.wordpress.com/the-church-of-monckton/
where the rubber on the tires of the mighty CAGW bandwagon (which poor BB so desperately wants to reconcile/integrate with the teachings of Joseph Smith, Brigham Young etc., etc.,) meets the highway of reality…..
alas, the famed oil slick in the shape of Jesus (of which the latte day saint Carl Hiassen wrote in his frothy missive to the faithful ‘Lucky You’), awaits:
Click to access mmh_asl2010.pdf
Click to access grl.fu.2011.pdf
Jeff writes : “Even reason can’t fight the blob.”
But from the movie…
Lieutenant Dave: At least we’ve got it stopped.
Steve Andrews: Yeah, as long as the Arctic stays cold.
oh noes
Once frozen, the blob’s brittleness will crack itself.
=======================================
“As an aside, it isn’t the message that is causing public opinion to change, it isn’t the fact that the public doesn’t understand, it is the regular discovery of the truly uncertain nature of the climate future which cuts into the message.”
Jeff ID, I am not at all certain how much the “public” understands the nuances of uncertainty, but for the thinking and involved public, the uncertainty issue surrounding climate change cannot be brought forward often enough.
What I also find most interesting and curious is that it sometimes takes papers of an opposing POV (Douglass et al) that might not be entirely correct to get members of the consensus to reply and show evidence of the truly uncertain nature of model results and even those from observations. I put the recent Lindzen paper in the same catergory. It is almost as though some of these sticky and uncertain issues are better waved away than really studied by the concensus scientists – until a paper comes along and puts some skepticism into the data and evidence.
I think that some of the temperature reconstruction papers have been left off the hook after the initial critical MM papers and probably because those papers involve questions of methodology. Certainly Mann (2008) deserves some very heavy criticism for the use of temperature measurements for proxies and non scientific truncation of the Schweingruber series and adding on a mysterious series to the end. After some study on my own I am convinced that the results of Mann (2008) can be attributed nearly entirely to chance – and chance of a proxy, having long term persisitence or even the proper Arima model response, and having series ending relatively long increased responses. I am not saying that the temperatures have long term persistence, since I am not also saying that the proxies are responding to temperature without a lot of noise, but rather that the proxies used do have LTP..
My current project is looking at the USHCN temperature trends as related to CRN ratings and again the first striking feature is the uncertainty involved in these analyses.
Time for my classic shaggy dog story, with appropriate spelling change:
Anonymous,
“Once frozen, the blob’s brittleness will crack itself.”
Just means you have more blobs when the pieces thaw!!! 8>)
Referring to Figure 2; the the key plot of modeled vs observed temperatures, I find it curiouser and curiouser that (Australia’s) CSIRO Mark 3.0 GCM matched the RSS and UAH-measured tropical (T24 – T2LT) value in K/decade (average ~0.010) so well within error bars and yet their presumably later much improved Mark 3.5 model had ‘drifted’ right back to (essentially) match the GCM ensemble mean (~0.100, although both took into account ozone depletion.
I would be absolutely fascinated to find out just what might have been the key difference e.g. with respect to tropical cloud handling between the CSIRO Mark 3.0 and Mark 3.5 models?
There might be a powerful clue therein.
Any idea Nick?
Steve #28
I’m not familiar with CSIRO’s recent models. There is a report here dealing with the differences between 3.0 and 3.5. They have improved the absorption of LW radiation – they mention only minor changes to cloud models.
ps I tried to reply at CA, but it’s in moderation.
Sorry Nick, I’m getting ‘Error 404 – Not Found’ for your link. Please re-try. Thanks.
Steve,
Hope this works.
Nick.
Beaut. Thanks.
If you look at table 1, the “model ensemble mean,” although certainly running “hot,” is within the uncertainty of the “observations,” with the exception of UAH T24.
As to the existence of enhanced upper tropical tropospheric warming, below is a compilation of papers that have detected the “missing hot spot.”
“Insofar as the vertical distributions shown in Fig. 3 are very close to moist adiabatic, as for example predicted by GCMs (Fig. 6), this suggests a systematic bias in at least one MSU channel that has not been fully removed by either group [RSS & UAH].”
Click to access sondeanal.pdf
“In the tropical upper troposphere, where the predicted amplification of surface trends is largest, there is no significant discrepancy between trends from RICH–RAOBCORE version 1.4 and the range of temperature trends from climate models.”
Click to access i1520-0442-21-18-4587.pdf
“The observations at the surface and in the troposphere are consistent with climate model simulations. At middle and high latitudes in the Northern Hemisphere, the zonally averaged temperature at the surface increased faster than in the troposphere while at low latitudes of both hemispheres the temperature increased more slowly at the surface than in the troposphere.”
http://www.nature.com/ngeo/journal/v1/n6/abs/ngeo208.html
“We find that tropospheric temperature trends in the tropics are greater than the surface warming and increase with height.”
Click to access grl.fu.2005.pdf
“At middle and high latitudes in the Northern Hemisphere, the zonally averaged temperature at the surface increased faster than in the troposphere while at low latitudes of both hemispheres the temperature increased more slowly at the surface than in the troposphere.”
Click to access VinnikovEtAlTempTrends2005JD006392.pdf
“In an Intergovernmental Panel on Climate Change (IPCC) project on model comparisons, Santer et al. (2005) found that the tropical atmospheric temperature trend was increasing vertically from the surface to about 300 mb for all 19
models used in the study, which was consistent with those predicted by the moist adiabatic lapse rate theory suggested by Stone and Carlson (1979). The T23 and T24trend results from this study and RSS dataset support these model results, although the magnitude of the amplification is different. For the global ocean means, the tropospheric amplification also occurs for temperature trends of this study, but is not obvious for the RSS and UAH datasets.”
Click to access Zou.2009.ErrorStructure.pdf