Inconsistencies may undermine model’s reliability for projecting decade-to-decade warming and lead to misinterpretation of data
DURHAM, N.C. — A new Duke University-led study finds that most climate models likely underestimate the degree of decade-to-decade variability occurring in mean surface temperatures as Earth’s atmosphere warms. The models also provide inconsistent explanations of why this variability occurs in the first place.
These discrepancies may undermine the models’ reliability for projecting the short-term pace as well as the extent of future warming, the study’s authors warn. As such, we shouldn’t over-interpret recent temperature trends.
“The inconsistencies we found among the models are a reality check showing we may not know as much as we thought we did,” said lead author Patrick T. Brown, a Ph.D. student in climatology at Duke’s Nicholas School of the Environment.
“This doesn’t mean greenhouse gases aren’t causing Earth’s atmosphere to warm up in the long run,” Brown emphasized. “It just means the road to a warmer world may be bumpier and less predictable, with more decade-to-decade temperature wiggles than expected. If you’re worried about climate change in 2100, don’t over-interpret short-term trends. Don’t assume that the reduced rate of global warming over the last 10 years foreshadows what the climate will be like in 50 or 100 years.”
Brown and his colleagues published their findings this month in the peer-reviewed Journal of Geophysical Research, at http://onlinelibrary.wiley.com/enhanced/doi/10.1002/2014JD022576/.
To conduct their study, they analyzed 34 climate models used by the Intergovernmental Panel on Climate Change (IPCC) in its fifth and most recent assessment report, finalized last November.
The analysis found good consistency among the 34 models explaining the causes of year-to-year temperature wiggles, Brown noted. The inconsistencies existed only in terms of the model’s ability to explain decade-to-decade variability, such as why global mean surface temperatures warmed quickly during the 1980s and 1990s, but have remained relatively stable since then.
“When you look at the 34 models used in the IPCC report, many give different answers about what is causing this decade-to-decade variability,” he said. “Some models point to the Pacific Decadal Oscillation as the cause. Other models point to other causes. It’s hard to know which is right and which is wrong.”
Hopefully, as the models become more sophisticated, they will coalesce around one answer, Brown said.
###
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
In other words, “pay no attention to how badly these models had predictrd climate for the past decade or so, we still expect a climate apocalypse. Please send your donations to…”
The man is starting his career. Give him a break. He’ll find his way if he is honest with himself.
“””””……“The inconsistencies we found among the models are a reality check showing we may not know as much as we thought we did,” said lead author Patrick T. Brown, a Ph.D. student in climatology at Duke’s Nicholas School of the Environment……”””””
The sheer brilliance of this statement simply confounds me.
A candidate for a PhD would get further along towards that goal if he simply said:
“We never can know as much as we thought we did.”
I do hope that RGBatDuke doesn’t get wind of this discovery. Well I hope he’s not one of your kids Robert.
The notion that you can model the climate of this whole planet, without simply thousands of variables, astonishes me in the arrogance of that idea.
g And yes you can model some things; lots of things; but not this.
+1
Just because those guys are nasty doesn’t mean skeptics should emulate them. It’s not good strategy. After all, they’re losing, at the very least in the court of public opinion.
Unfortunately, he will face re-education and will learn to toe the line when it comes to grant funding etc. That is why every study that shows problems with AGW ALWAYS include the caveat “but this in no way
implies the earth is not warming due to man made CO2” etc. etc.
He warns readers not to “over interpret.” That is always a good idea.
But then, he goes on to say “Don’t assume that the reduced rate of global warming over the last 10 years foreshadows what the climate will be like in 50 or 100 years.” In other words: The apocalypse is still on! Rah! Rah! Rah!
“If you’re worried about climate change in 2100, don’t over-interpret short-term trends.”
And yet, when the spike occurred in 1998, the short term trend was enthusiastically over-interpreted to mean runaway global warming. They can’t have it both ways.
Bruce Hall,
Arguably the short-term trend from 1970-1998 was over-interpreted as runaway warming: http://www.woodfortrees.org/plot/hadcrut4gl/plot/hadcrut4gl/from:1970/to:1998/trend
Perhaps “they” are learning … unless “they” includes politicians and journalists …. in which case all bets are off.
“They can’t have it both ways.”
Yes they can, especially when the data is adjusted to fit……….
They’re doing a pretty lousy data fudging exercise if you ask me.
Somehow I never imagined this excuse for the pause. And neither did Ben Santer with his 17 years falsifies paper. So, the models work fine year to year and century to century, just not decade to decade? When the conclusion is simplified like that, it becomes clear that some of Duke’s grad schools do not teach clear thinking, or maybe even thinking at all. Decades pass one year at a time, amd each of those years creates an imcrementally greater discrepancy between CMIP5 and reality, despite best efforts by the likes of NASA and NOAA to distort that reality.
Wonder what RGBatDuke thinks about this new paper from his institution?
From an author with the same last name.
Brown in the report might just be RGB
Patrick T. Brown.
Nope, RGB has a different first and middle name.
Rud,
I think it’s more there a part of the explanation that’s missing, and it’s the long cycle ocean states, AMO, PDO, etc, and the size of the impact to surface temps when these change.
Mi Cro,
The party line is that GCMs get the frequency and amplitude of such events nearly correct but suck at exact timing. Depends on the model of course. Stuff like AMO and PDO don’t exist as input parameters so not even the hindcast portion of a CMIP5 run will get the timing of those wiggles correct.
Not if you take this as true
my highlighting
It’s what happens when you have outputs that feedback into the inputs, you get emergent behavior. Analog circuit simulations were always tricky, the simulator itself could oscillate as each node of the matrix tries to stabilize, and then you can start to step time forward, at which point the whole thing can oscillate. So here you have a field that loops in back on itself with a time delay. And while you can figure out how a cell will respond, as I wrote this I decided that you could easily have enough conditions on the signals coming into a cell that you might not even figure this out without running it in a debugger mode to see these conditions.
What you say is true. But there are also two other large factors. First, as I nelieve Akasofu was among the first to point out, CMIP3 completely and CMIP5 mostly have tuned parameterizations from a warming period now sandwiched between two periods without warming, so will run hot. Computational limits force grid scales much larger than the quintessial tropical convection cell, a Tstorm. So they cannot get the water vapor and cloud feedbacks right. That is why CMIP5 still produces a tropical upper troposphere hotspot when none exists in any of the observational data.
Covered in much more details in several essays in my new ebook.
It’s neither me nor any relation. However, I agree with the conclusions, and have said much the same thing in this venue, as have many others. This is also far from sufficient as a meta-analysis of the shortcomings of CMIP5.
But it’s a start.
And as for clear thinking, I didn’t see Patrick Brown asserting that predictions of total climate sensitivity from the models were EITHER high OR low — only that when one model produces CO_2-driven warming balanced against aerosols and another produces CO_2-driven warming balanced against decadal oscillations and another gets its results from something else entirely, and when all three models fail to have the right decadal power spectrum or variance, it is difficult to take any of their results seriously EVEN IF one thinks — as I think — that the preponderance of both the data and the theory indicate that TCS due to increasing CO_2 in the atmosphere is between 1 and 2 C per doubling. It is also not only difficult but outright impossible to make the assertions made in e.g. AR5 SPM that one can have some “confidence” that most of the observed warming over ANY given interval but especially over the last 30-40 years is due to anthropogenic or natural causes or any particular mix in between. CMIP5 models both attribute warming to different causes and mixes of causes and fail to reproduce observed reality particularly accurately and are inappropriately presented and collectively averaged as if the average is somehow a meaningful number without any regard to how well or poorly the models in question do compared to reality.
Why is it a bad thing or evidence of sloppy thinking to publish this? Who would even disagree with it?
rgb
It’s probably because they are hoping to see something along the lines of “CO2 is innocent of all charges.” Too many of us seem to think of science in courtroom or political terms these days rather than in properly sceptical yet speculative terms that actually convey nuanced degrees of scientific knowledge. Then too, many of us seem to have “made up our minds,” which we fail to realize means they simply haven’t closed the coffin lid on us yet.
Who would even disagree with it?
=====
Me for one…it’s a joke
The “models” have good consistency year to year…….when you’re dealing in 100th of a degree
….but fall apart in decades…when you’re dealing with 10th degrees…………ROTFL
And what….the temperature is supposed to jump way up to get back on that trend line sometime in the future!
With all due respect, I disagree with “If you’re worried about climate change in 2100, don’t over-interpret short term trends.” It’s a truism, appears directed at skeptics who interpret the “pause” as debunking the models, without recognizing the downstroke, that models over-interpreted to begin with. Maybe I’m too sensitive, but there’s been a steady stream of truisms tossed about by CAGW believers, the tone of this article reminds me of the end-of-world cults the day after the predicted (pardon me, “projected”) end doesn’t happen. That said, I re-read after seeing your post, and must admit that the doubts and criticisms expressed by Mr. Brown are a breath of fresh air from academia. Your post expresses these doubts much more clearly, thanks for that.
Rud Istvan,
I am hopeful that someday in the near future that Ben Santer will be paraphrased correctly. The actual quote is, “Our results show that temperature records of at least 17 years in length are required for identifying human effects on global-mean tropospheric temperature.”
http://onlinelibrary.wiley.com/enhanced/doi/10.1029/2011JD016263/
rgbatduke,
I answer your rhetorical questions straight: those who consider anything written about GCMs from a climate consensus perspective to be a priori wrong.
Here is something from SkS written in 2010. The years went by and we are still at a standstill.
When will they update this page? Sometime in the years to come.
Yes, that is his paper. You cite the last line of the abstract, but not the two before. Read together, the line you cite means that we now have (take your pick of method and dataset) anywhere from 15 to 18 to 19 to 26 years of observed no statistically observed warming. Which means the steadily rising CO2 didn’t cause any for more than his model requiste period. Which means that by his criteria, the models are busted. See above comment to Mi Cro for inherent reasons why this must be so.
Brandon Gates says: January 26, 2015 at 12:14 pm
”I answer your rhetorical questions straight: those who consider anything written about GCMs from a climate consensus perspective to be a priori wrong.”
Brandon, the author of the paper clearly states –
“This doesn’t mean greenhouse gases aren’t causing Earth’s atmosphere to warm up in the long run”
– an assumption that as an accepted priori prevents the author from ever truly understanding why the models fail.
Remember it is not just the AGW hypothesis but the foundation assumption of a net atmospheric radiative GHE that is unproven.
Adding radiative gases to the atmosphere will not reduce the atmospheres radiative cooling ability. Nor will adding radiative gases to our radiatively cooled atmosphere reduce its ability to cool the solar heated surface of our planet.
I am hopeful that someday in the near future that you will check your quotes and assertions. It has been quoted dozens of times on WUWT.
site:http://wattsupwiththat.com "Our results show that temperature records of at least 17 years in length are required for identifying human effects on global-mean tropospheric temperature"
Click the word site for Google site search / quote search results.
I beg your pardon,
but this staring at your navel and deciphering the message therein. Idiots with funding. Do they expect to be taken seriously?
Is Kermit a PhD candidate at Duke?
“The inconsistencies we found among the models are a reality check”
How very clever of them, weasel-wording around modelling reality. What they need to do is check reality!
Reality is there to be altered, not checked.
“Other models point to other causes.”
People do the pointing – the models follow.
Quote:
“This doesn’t mean greenhouse gases aren’t causing Earth’s atmosphere to warm up in the long run,” Brown emphasized.
Endquote.
Translation: “Please don’t reduce my grant…….”
Ralph
Well I agree with the student’s assertion: “This doesn’t mean greenhouse gases aren’t causing Earth’s atmosphere to warm up in the long run,”
BUT !! if that IS the state of knowledge of the situation, then an EQUALLY TRUE statement is:
That doesn’t mean greenhouse gases ARE causing Earth’s atmosphere to warm up in the long run.
You can’t have a factual third assertion.
So pray tell, what in this student’s mind cause him to select his assertion over the equally valid assertion, or why didn’t he simply specifically say both; that what they do know, is that they DON’T KNOW EITHER WAY.
Of course if you have a destination bias, then you might choose the one option that reinforces your pre-conceived notion about what you don’t know.
“WE DON’T KNOW”, is a perfectly acceptable scientific conclusion.
I suspect that his grants and PhD might be in jeopardy if he, in fact, stated his position as you have correctly pointed out that it should be stated. The reality he, and other young scientists, are working against is the institutional bias — to h3ll with what’s really going on in the world.
Nicely put. Precisley so!
No no George E Smith,
“We JUST don’t know”
“The inconsistencies existed only in terms of the model’s ability to explain decade-to-decade variability, such as why global mean surface temperatures warmed quickly during the 1980s and 1990s, but have remained relatively stable since then.”
So, the models work well over the warming period which they were tuned for, but they don’t work after that. They’ll probably work better for the recent decade after being tuned to match that also. Doesn’t mean that there is a basis in reality.
As Tamsin Edwards says (although she phrases it slightly differently) : “All models are wrong, and some are useless.”
Communication………
GCMs are worse than worthless except to help falsify the CACA hypothesis which they are designed to show.
That’s ok. We’ll just average all the output together and call it good.
See? Science ain’t so hard!
Jeff Alberts,
At first blush this paper appears to be arguing the opposite of simply taking an ensemble mean and calling it good.
I don’t have access to the paper (and might not even understand it), but my jab was mainly at the IPCC and their ensemble means. Only one of the models can be right (but probably none of them are), averaging bad output with good output can’t give you good output.
So let me ask this question: Is he asking what causes decade-to-decade variability in the OUTPUT of the models ?? OR Is he asking what causes decade-to-decade variability in THE CLIMATE (you know ; the thing that Mother Gaia keeps track of to see that it ALWAYS obeys the laws of Physics.)
If “other models” point to “other causes” and some models point to some causes; and neither one points to what the climate itself says is going on, then ALL of such models are wrong.
If the models are correct, then they must point to what causes the “decade-t-decade” variability in the climate, or they are just rubbish.
In my opinion, the reason for the wiggles in the Temperature is; as I have said on multiple occasions. Physical systems never respond to the average of anything, as they have no precognition. Physical systems only respond to NOW; instantaneous values and simultaneous values of ALL parameters or variables. Those “instantaneous” intervals can (evidently) be as short as 10^-43 seconds; or may be it’s 10^-34 seconds.
We are familiar (some persons are) with phenomena that happen in as short as 10^-18 seconds; attoseconds. (no not me; I’ve just read about them in the optics literature.)
g And like I said, that’s just my opinion. Patrick Brown better not rely on that for his PhD thesis !
Right.
The models don’t provide ANY explanations. They simply parrot the assumptions programmed into them.
I am fond of yelling, “Don’t do what I say, do what I mean” at a particularly recalcitrant piece of code. I oft see it written by climate modelers, “We don’t know what it will do before we run it.” And yet, AOGCMs only do exactly as they have been asked.
Sounds like we agree.
Notice the “as the atmpsphere warms” pro-forma genuflection to the climate gods. No introsection about *if* the climate actually warming. No questioning of assumptions.
Well, out of a child’s mouth comes the truth. So long Settled Science, Consensus et al.
Any “Climate Scientist” that can not create and read a 100 year graph is Too stupid to be allowed to voice an opinion on Climate. Weather conditions change over a 60 year cycle. Climate over 5 times that. pg
I think this is very good from a PhD student – they can hardly be expected to come out and say “the whole thing is crap” because they have to get their thesis approved by a committee of people who have a vested interest. At the same time, a proper analysis of the models, with the pretty clear statement that we cannot learn anything from them because they don’t match the actual data, is a big step forward.
The statement that we can’t say anything about the temp in 2100 from decadal long trends is exactly correct – that is the mistake that was made looking at the trend up to 1998. The fact that this is not stated explicitly in the paper is probably on the (wise) advice of the students supervisor who wants to see their student be employed some time in the future.
Well for the record, 65% of ALL USA PhD graduates in Physics, NEVER ever get a paying job in the field of “expertise” characterized by their Doctoral Thesis.
That doesn’t mean they are incompetent; or even substandard; nor does it speak to the competence of their supervisor or advisor.
It means NOBODY is going to pay ANYBODY to try to make a profitable business venture out of ” The Hermaphroditic Life Style of the Palolo Worm ”
Companies hire people who can help make money for them.
But all sorts of “institutions” take in Post Doc “fellows” to continue their fetishes, if only to help bring in more endowments from folks who are enamored with credentials.
But if you DO pick a productive study for your shingle then companies WILL pay you handsomely for your expertise.
I thought of getting a PhD in “ice cream making”, but it seemed that the idea was well understood already.
g
That’s just the problem, When a field is well understood, in order to get a PhD in that area, you must say that it is incorrectly understood, that black is white and white is black. With the right statistical techniques, you can prove that ice cream making is based on denial of newly discovered post-normal tenets. Having done this, you can get a job as Chief Physicist at a soda fountain.
Ah my, there goes his ambitions of getting a PhD. I think the lesson here for all aspiring Climate Scientists is that you should never, ever publish a study that is off-message. Experienced Climate Scientists know that the best way to get ahead in the world is to wait for a politician to give a speech that makes some kind of a poorly supported claim, and then to publish research that provides the needed support.
For example. New York Mayor Bill de Blasio will shortly be claiming that the New York Blizzard was caused by Global Warming. This is an opportunity for a PhD student to manufacture adjusted data to support this claim. You will become famous as the press eagerly reports your finding broad and wide (despite the fact that you are still a student). Then, for extra credit, if anyone points to the older studies that said that snow fall would become a rarity in New York, you can say that “such claims come from deniers funded by tobacco industry groups who are intentionally misrepresent the data for political purposes. In fact, the models have perfectly predicted increasing snow fall. Now we have reached the tipping point and we can no longer delay action.”
PhD granted!
and if the blizzard that has started yet turn out to be less than predicted, it will be global warming that made it not agree with the models.
So in spite of the evidence we just produced showing the inability to predict with any accuracy, It’s still going to happen. Perfect.
“… If you’re worried about climate change in 2100, don’t over-interpret short-term trends. Don’t assume that the reduced rate of global warming over the last 10 years foreshadows what the climate will be like in 50 or 100 years …”.
==========================================
Phew that’s a relief, the ‘pause’ had me worried that I had to stop worrying.
+10. Funny. On point.
‘ Don’t assume that the reduced rate of global warming over the last 10 years foreshadows what the climate will be like in 50 or 100 years.’
Or to paraphrase “better to use that long term trend 1979/1998”.
Hopefully, as the models become more sophisticated, they will coalesce around one answer, Brown said.
Definition: Sophistry
n. noun
1: Plausible but fallacious argumentation.
2: A plausible but misleading or fallacious argument.
Hmmmm – I believe Mr. Brown used the term correctly!
Hopefully, as the models become more sophisticated, they will coalesce around one answer, Brown said.
===========
They do that already. After billions of $ spent and unclounted years of effort, the models have coalesced around the wrong answer.
ferd,
Unfortunately it’s not the “wrong answer” to the faithful believers in CAGW or the money changers intent on profiting from carbon trading and renewable energy money geysers. It’s the answer they want.
The only way the models can coalesce around the same answer is if they all use the same parameters in the same way i.e. one model. Each model is based on different theories. It has alway puzzled me how averaging different theoretical models is meaningful.
Right on Mate. Leads to my observation that people who consider themselves “sophisticated” usually are !
The problem is not that the model are not sophisticated enough. It could be that they are too sophisticated already. If you read Monckton’s article last week, IPCC = Inferior to a Pocket Calculator Calculation..
The real problem is lack of data. Tim Ball’s recent article highlights this. Models can not manufacture data that do not exist, particularly if you are looking for accuracy to 0.01C degrees.
Given that climate models disagree THAT CLIMATE WIGGLES OCCUR (the pause), it’s not too big an intellectual leap that they also disagree on why wiggles occur.
We’re right back to “climate models are crap”.
Climate “science” is descending to the level of surreal comedy.
How stubbornly and irrationally are they going to go on denying intrinsic chaotic dynamics within climate?
Are they unaware what idiots they look as they cast about for for discreet external “forcing” explanations for ever tinier climate oscillations on an ever smaller and smaller, fractal scale?
Reminiscent of ever-increasing numbers of epicycles to bolster Ptolemaic cosmology.
Hmmm… no. It was Copernicus who had to increase the number of epicycles to maintain consistency with Ptolemy.
Might be interesting to know how those models are programed. My understanding is that a ‘greenhouse gas’ molecule in the atmosphere is a unique entity. It can radiate over 360 degree cubed directions. If said molecule happens to be several kilometres above the surface then rather more than 50% of the radiation will escape to space. Only molecules close to the surface might achieve the claimed 50% back radiation. Ah well, best to let the ‘experts’ muddle along. Hope you youngsters are prepared for the coming cold. Won’t bother me, I’m too old.
Actually the thickness of the atmosphere is such a small fraction of the earth radius that the curvature of the atmosphere barely comes into play in determining where a photon goes.
But the photon is more likely to escape freely going outward. Going inward, toward a thicker atmosphere, it is more likely absorbed sooner (blocked from descending). This creates a net flow vector outward, not downward.
My BS detector went off with this statement:
What seems clear is that the ENSO process strongly influences the up and downs in the temp anomalies of the past 18 years, .i.e. The Pause. And no one has been able to accurately model the ENSO process, and most have concluded it is chaotic. No GCM can tell us (so far) what the ENSO will do in the next decade, much less a century. (see below) And since as Mr. Brown noted, with his “decade-to-decade” statement, several decades is all we have so far that we have to judge the models against. And they are failing badly. So why should the suddenly be nailing it on longer intervals?
And the fact that they were tuned with past knowledge, saying they accurately forecast the 1880 to 1998 interval is disingenuous. It’s only what the modelers forecasted for the future is how the GCMs must be objectively evaluated. They knew the “past” when they tuned them.
A good discussion of the problems of ENSO modeling is here:
http://scienceofdoom.com/2014/12/26/natural-variability-and-chaos-six-el-nino/
Scienceofdoom discusses this problem of ENSO modeling, and uses a figure from a 2009 paper on modeling the ENSO to highlight how problematic ENSO is for the modelers.
And since we can’t model the ENSO process over a few decades, there should be no expectation that we can model the Climate over a century or longer, with any skill.
The unexplained variability of climate change in the form of an unexplained lack of variability of climate appears not so much decade-to-decade, but more decade-to-decade-to-decade.
The solar wind was faster during the multi-year La Nina of 1973-76, 1983-85, 1998-2001, and much slower during the bigger El Nino episodes of 1997/98 and 2009/10:
http://snag.gy/ppB3v.jpg
“This doesn’t mean greenhouse gases aren’t causing Earth’s atmosphere to warm up in the long run,” Brown emphasized. “ Right, keep your eye on the future grant money and the defense of your PHD thesis.