I was aware of this story yesterday, but I didn’t like the original plot, (see at the end of this post) since use of straight line linear trends doesn’t accurately reflect the reality of the observation data. While it is often hard to find any reality in climate models, linear trend lines mask the underlying variance. Today, Dr. Spencer has produced a graph that I feel is representative and very well worth sharing, because it does in fact convey an EPIC FAIL speaking directly to the accuracy of an ensemble of climate models. – Anthony
Dr. Roy Spencer writes:
In response to those who complained in my recent post that linear trends are not a good way to compare the models to observations (even though the modelers have claimed that it’s the long-term behavior of the models we should focus on, not individual years), here are running 5-year averages for the tropical tropospheric temperature, models versus observations (click for full size):
In this case, the models and observations have been plotted so that their respective 1979-2012 trend lines all intersect in 1979, which we believe is the most meaningful way to simultaneously plot the models’ results for comparison to the observations.
In my opinion, the day of reckoning has arrived. The modellers and the IPCC have willingly ignored the evidence for low climate sensitivity for many years, despite the fact that some of us have shown that simply confusing cause and effect when examining cloud and temperature variations can totally mislead you on cloud feedbacks (e.g. Spencer & Braswell, 2010). The discrepancy between models and observations is not a new issue…just one that is becoming more glaring over time.
Read his essay here: http://www.drroyspencer.com/2013/06/still-epic-fail-73-climate-models-vs-measurements-running-5-year-means/
==============================================================
Here is the linear plot from Dr. Spencer’s post yesterday. He writes:
Courtesy of John Christy, a comparison between 73 CMIP5 models (archived at the KNMI Climate Explorer website) and observations for the tropical bulk tropospheric temperature (aka “MT”) since 1979 (click for large version):
Rather than a spaghetti plot of the models’ individual years, we just plotted the linear temperature trend from each model and the observations for the period 1979-2012.
Note that the observations (which coincidentally give virtually identical trends) come from two very different observational systems: 4 radiosonde datasets, and 2 satellite datasets (UAH and RSS).
If we restrict the comparison to the 19 models produced by only U.S. research centers, the models are more tightly clustered:
Now, in what universe do the above results not represent an epic failure for the models?
Just playing around with the UAH TMT Tropics dataset and found something very interesting.
The TMT tropics (UAH and RSS) is totally dominated by the ENSO. There is no global warming tropical hotspot from increased water vapor here. The ENSO runs this region with as much as a +/- 0.7C impact very directly one-for-one. (other influences like volcanoes and the AMO are also clear). And there is no long-term trend in the ENSO so one would expect there to be no long-term in the TMT temps either.
http://s8.postimg.org/4oi49v1sl/UAH_TMT_Drivers.png
So there is NO global warming signal here at all. In fact, the regression coefficient for Ln(CO2) is Zero. UAH TMT can be modelled very closely without a global warming trend at all.
http://s11.postimg.org/3q98mgpab/UAH_TMT_Modeled.png
The climate models have to necessarily be very inaccurate about this important level.
Richard M says:
June 7, 2013 at 7:48 am
Alex, instead of complaining, why don’t you go to Spencer’s blog and ask him. The link is in the article. I’m sure he’ll be happy to provide you with the specifics. Then you can report them back here.
——
What is there to ask?
If Roy cares about his reputation, he should do it himself and not post something that makes no sense,
After his funny post about
http://wattsupwiththat.com/2013/04/24/spencer-slays-with-sarcasm/
and the even funnier present one, I have a strong suspicion he went mad (Alzheimer?).
As I said, the models say: the top of troposphere cools down, the bottom warma up.
This is absolutely in-line with observations.
Any reasonable person may guess there is “a middle of troposphere” with a flat temperature trend.
So, why he claims any “failure of models”, only a doctor may clarify.
alex says:
“If Roy cares about his reputation, he should do it himself and not post something that makes no sense” <— [makes no sense to Alex.]
There, Alex. Fixed it for you.
Now if you still don't understand, go ask Dr Spencer. He has a lot of experience compasred with you. You could really learn something from him. And it is your reputation that is at stake, not Dr Spencer’s.
Or you could carry on, and continue to snipe from the sidelines.
alex says:
June 8, 2013 at 9:42 am
“If Roy cares about his reputation, he should do it himself and not post something that makes no sense,
After his funny post about
http://wattsupwiththat.com/2013/04/24/spencer-slays-with-sarcasm/
and the even funnier present one, I have a strong suspicion he went mad (Alzheimer?).”
Well, you’re a warmist troll, so you surely think we are mad, the enemy, insane, paid by Big Oil or whatever canard you need to block out inconvenient information. That is no surprise.
given that we have a dozen of your ilk on this thread, it looks like Spencer has really hit you right where it hurts. Usually only two members of the warmist cult are delegated for each WUWT thread.
alex says:
June 8, 2013 at 9:42 am
As I said, the models say: the top of troposphere cools down, the bottom warma up.
This is absolutely in-line with observations.
Any reasonable person may guess there is “a middle of troposphere” with a flat temperature trend.
So, why he claims any “failure of models”, only a doctor may clarify.
————————————
TMT is measured right in the middle of where the big tropical hotspot is supposed to be.
Hotspot climate model prediction (for 2XCO2) courtesy of RealClimate – see 600 hpa to 200 hpa
http://www.realclimate.org/images/2xCO2_tropical_enhance.gif
So “failure of climate models” is proven. Its white, not deep red.
Beautiful, RGB.
Does an “attractor” nullify chaos?
The comments from “Phitio” are downright creepy.
Only a natural force as powerful as organized religion stands a chance of arresting and correcting the spreading corruption of university & government sponsored modeling “science”.
May God bless the human race.
Even more creepy, the authorities are repeatedly messing with access to a crucial landmark paper that illustrated far more than it addressed in text. After being inaccessible for an extended period of time, public access had recently been briefly restored, but the article is now once again inaccessible. During the extended period of inaccessibility the link was giving the following strangely worded error message:
“The NASA technical reports server will be unavailable for public access
while the agency conducts a review of the site’s content to ensure that it
does not contain technical information that is subject to U.S. export control laws
and regulations and that the appropriate reviews were performed.
The site will return to service when the review is complete.
We apologize for any inconvenience this may cause.”
Fortunately, there’s an alternative way to access the article:
Dickey, J.O.; & Keppenne, C.L. (1997). Interannual length-of-day variations and the ENSO phenomenon: insights via singular spectral analysis.
I advise everyone to not assume that the article will remain accessible via this alternative. Should the alternative access vanish, I’m sure the authorities realize that this will just come across as looking yet more creepy. Creepy seems to be the image authorities are going for these days, so let’s keep a curious eye on this alternative copy of the paper to see if the authorities seize this easy opportunity to look yet more creepy by ordering that the alternative copy be removed from the web. Monitoring accessibility provides an opportunity to gain information.
gymnosperm (June 8, 2013 at 11:31 am) asked:
“Does an “attractor” nullify chaos?”
Attractors are a feature of chaos — a feature that’s easily completely overlooked by standard time series analysis methods when they’re based on patently false assumptions, which are the cultural norm. (details another day…)
Reblogged this on Climate Ponderings and commented:
GREAT READ!
As Glenn Tamblyn and I have pointed out, there are a number of issues with the data, including the fact that averaging the UAH and RSS data sets hides the fact that there is considerable disagreement between them and the fact that the TMT data channel, as shown here, introduces a cooling bias because the weighting function vs altitude has a tail going into the stratosphere where it has been cooling.
So, here are the trends that can be gleaned directly from the data:
UAH TMT for tropics: 0.029 K/decade
RSS TMT for tropics: 0.090 K/decade
Average for tropics: 0.060 K/decade (This is what is presumably plotted as blue squares.)
RSS TTT for tropics: 0.127 K/decade (The TTT is a product using multiple channels to effectively reduce the weighting function to zero in the stratosphere.)
So, if the RSS TTT is correct, the data has more than twice as steep a trend as is shown here. And, as Glenn points out (although I haven’t investigated, the STAR/Nesdis group ( (http://www.star.nesdis.noaa.gov/smcd/emb/mscat/index.php) gets about a trend of 0.124 K/decade for TMT, i.e., even without trying to correct the TMT channel for stratospheric contamination.
The summary of all this is that there are simply too large uncertainties in the data to say with confidence that the discrepancy is a failure of the models.
And, furthermore, as has been pointed out in the literature, the models and data agree quite well for amplification seen in response to fluctuations in temperature on monthly- to yearly-timescales. It is only when looking for amplification of the multidecadal trends does one run into problems. But the multidecadal trends are exactly where the data are least reliable because of artifacts that can affect the long-term trends. Finally, it is hard to understand how the physics that leads to this amplification on monthly- to yearly-timescales could fail to lead to the same amplification for the long-term trends.
This is why the U.S. Climate Change Science Program report several years ago on this subject concluded that issues with the data are the most likely source of the discrepancy. [And, they have already been proven at least partly right in the interim since an error was found in the UAH data set that significantly changed the trend in the tropics, making it agree somewhat better with both RSS and the models.]
As I and another have pointed out at Roy Spencer’s blog, there are a number of issues with the data, including the fact that averaging the UAH and RSS data sets hides the fact that there is considerable disagreement between them and the fact that the TMT data channel, as shown here, introduces a cooling bias because the weighting function vs altitude has a tail going into the stratosphere where it has been cooling.
So, here are the trends that can be gleaned directly from the data:
UAH TMT for tropics: 0.029 K/decade
RSS TMT for tropics: 0.090 K/decade
Average for tropics: 0.060 K/decade (This is what is presumably plotted as blue squares.)
RSS TTT for tropics: 0.127 K/decade (The TTT is a product using multiple channels to effectively reduce the weighting function to zero in the stratosphere.)
[For UAH data, see http://vortex.nsstc.uah.edu/public/msu/t2/ , specifically, http://vortex.nsstc.uah.edu/public/msu/t2/uahncdc.mt . For RSS data, see http://www.remss.com/msu/msu_data_description.html , specifically http://www.remss.com/data/msu/monthly_time_series/RSS_Monthly_MSU_AMSU_Channel_TMT_Anomalies_Land_and_Ocean_v03_3.txt for TMT and http://www.remss.com/data/msu/monthly_time_series/RSS_Monthly_MSU_AMSU_Channel_TTT_Anomalies_Land_and_Ocean_v03_3.txt for TTT.]
So, if the RSS TTT is correct, the data has more than twice as steep a trend as is shown here. And, the STAR/Nesdis group ( (http://www.star.nesdis.noaa.gov/smcd/emb/mscat/index.php) gets about a trend of 0.124 K/decade for TMT, i.e., even without trying to correct the TMT channel for stratospheric contamination.
The summary of all this is that there are simply too large uncertainties in the data to say with confidence that the discrepancy is a failure of the models.
And, furthermore, as has been pointed out in the literature, the models and data agree quite well for amplification seen in response to fluctuations in temperature on monthly- to yearly-timescales. It is only when looking for amplification of the multidecadal trends does one run into problems. But the multidecadal trends are exactly where the data are least reliable because of artifacts that can affect the long-term trends. Finally, it is hard to understand how the physics that leads to this amplification on monthly- to yearly-timescales could fail to lead to the same amplification for the long-term trends.
This is why the U.S. Climate CHange Science Program report several years ago on this subject concluded that issues with the data are the most likely source of the discrepancy. [And, they have already been proven at least partly right in the interim since an error was found in the UAH data set that significantly changed the trend in the tropics, making it agree somewhat better with both RSS and the models.]
Paul Vaughan,
I eagerly await the day. I don’t know what attractors are except what may be gleaned from the name. Very interested in the relationship between chaos and entropy. Suspect many will say they are entirely separate. The third law does not preclude organization in closed systems, yet in a real sense both chaos and entropy are enemies of organization.
joeldshore:”So, if the RSS TTT is correct, the data has more than twice as steep a trend as is shown here. And, the STAR/Nesdis group ( (http://www.star.nesdis.noaa.gov/smcd/emb/mscat/index.php) gets about a trend of 0.124 K/decade for TMT, i.e., even without trying to correct the TMT channel for stratospheric contamination.
The summary of all this is that there are simply too large uncertainties in the data to say with confidence that the discrepancy is a failure of the models”
I would say that your summary is a bit incomplete. Even assuming you are right about TTT being the most accurate (which I’m pretty sure is debatable), temp observations are (by my eyes) are still substantially below all the predicted trends in the models for 70 out of 73 models. IOW, we can say with confidence that the *vast majority* of the models out there are failures at reproducing the temperature behavior of the last 30+ years. However, some of the (presumably lower sensitivity) models still do a reasonable job with reproducing the real world temperatures as reflected in the TTT dataset.
Cheers, 🙂
Bill Illis says:
June 8, 2013 at 11:13 am
alex says:
June 8, 2013 at 9:42 am
As I said, the models say: the top of troposphere cools down, the bottom warma up.
This is absolutely in-line with observations.
Any reasonable person may guess there is “a middle of troposphere” with a flat temperature trend.
So, why he claims any “failure of models”, only a doctor may clarify.
————————————
TMT is measured right in the middle of where the big tropical hotspot is supposed to be.
Hotspot climate model prediction (for 2XCO2) courtesy of RealClimate – see 600 hpa to 200 hpa
No it isn’t, it’s a weighted average from the surface to 25km (1,000 hPa – 25 hPa), peak weighting at ~4km.
“However, some of the (presumably lower sensitivity) models still do a reasonable job with reproducing the real world temperatures as reflected in the TTT dataset.”
+++++++++++++++
Or let’s put it another way. The models that the IPCC do not use for the Summary for Policy Makers are not so bad. Or you could say the models that are least promoted and least useful to the IPCC are pretty good. Or you could say that the models that show similar results as what is happening, actually were supposed to represent what would happen if we made drastic reductions to fossil fuel such that the CO2 level went down to where it was a decade ago.
I’m Just sayin’
joeldshore says:
June 8, 2013 at 5:46 pm
———————————
Here is UAH and RSS Tropics TMT.
There are almost identical over the whole timeframe (although there is slight difference in the trends which seems to start at about the time of the 1997-98 El Nino).
That changes nothing. There is NO global warming trend in the Tropics TMT level which is supposed to catch the hotspot. It is controlled by the ENSO (like water vapor is, cloudiness, OLR, and just about every important climate feature).
http://s8.postimg.org/71v03r9w5/UAH_vs_RSS_TMT_Tropics.png
RCP8.5 is the most extreme set of assumptions used in these models. There are others. All Dr Spencer has shown is that the model runs with the most extreme assumptions aren’t the closest to reality. That isn’t “EPIC FAIL”, that is just chartspam. I’d also recommend the breakout of each of the observational datasets, as joeldshore mentioned above.
Here is a blog with criticisms of Dr. Spencer and his blog post.They think his satellite data is suspect and that he is not providing full coverage of the topic.
http://blog.hotwhopper.com/2013/06/how-much-out-of-whack-is-uah.html
REPLY: This is one of my blog children, banned from WUWT due to multiple policy violations. It is run by a clueless female eco-nut type in Australia who calls herself “Sou”. As usual, she denigrates anonymously, since she has no honor like most of her ilk. – Anthony
http://blog.hotwhopper.com/2013/06/how-much-out-of-whack-is-uah.html
The usual complaints about Dr. Spencer’s chart and UAH temperature dataset as their justification on why they dislike his blog post.
As David vun Kannon said, all of these models use the RCP value of 8.5 (worst-case-scenario). Not only does the source of Dr. Spencer’s information offer the actual historical values of RCP (http://climexp.knmi.nl/selectfield_cmip5.cgi?id=someone@somewhere), the CMIP states that the predictive RCP values should not be used before a date of 2006 and that historical values should be used for modeling of the recent past (see page 126: http://cmip-pcmdi.llnl.gov/cmip5/docs/standard_output.pdf). So, Dr. Spencer took the worst-case-scenario predictions for the future value of a variable for which he could have obtained the actual value, plugged that into models starting several decades back, extrapolated the models, then compared their predictions to historical observation. Since the RCPs are predicted future values, it is literally impossible that the models could match observation.
I did not realize it was banned from your site.
I agree that it is a terrible response to Dr. Christy and Dr. Spensers presentation.
I got this website link at facebook posted by one of the clueless warmists who make it clear he expects members of the discussion group chose hotwhopper over your blog on this topic.I didn’t and what is more I plan to show why he is only making a fool of himself doing it since hotwhopper objections are crazy.
Phil. says:
June 8, 2013 at 9:05 pm
Bill Illis says:
June 8, 2013 at 11:13 am
TMT is measured right in the middle of where the big tropical hotspot is supposed to be.
Hotspot climate model prediction (for 2XCO2) courtesy of RealClimate – see 600 hpa to 200 hpa
No it isn’t, it’s a weighted average from the surface to 25km (1,000 hPa – 25 hPa), peak weighting at ~4km.
—————————
Thanks, Phil, but still I don´t understand.
Much worse it is – we have to guess (!) what Roy actually meant.
And this is no good.
David vun Kannon says:
June 9, 2013 at 7:43 am
RCP8.5 is the most extreme set of assumptions used in these models. There are others. All Dr Spencer has shown is that the model runs with the most extreme assumptions aren’t the closest to reality. That isn’t “EPIC FAIL”, that is just chartspam. I’d also recommend the breakout of each of the observational datasets, as joeldshore mentioned above.
+++++++++++++++
No David: It’s an Epic Fail. Do you know which charts are used for the IPPC’s “Summary for Policy Makers?” THE MOST EXTREME models. Stop being in denial of reality. I’m tired of the games your ilk play, hoping the naive will fall in line.
Eustace Cranch just axed, “Why doesn’t the observed temp line show the recent leveling-off? Just askin”
Cause the linear trend line encompassed the period 1979 through 2012. If it just covered the last 13 years, there probably wouldn’t be any positive trend. Do you understand linear regression and the idea of statistical significance?