Bishop Hill From the least absurd models
Jan 20, 2014 Climate: Models A few days ago, I mentioned a paper that had looked at climate model simulation of clouds and found that the ones that did the best job of this narrow task produced the highest predictions of temperature rises. As I noted at the time this means that one can paraphrase the findings as "best cloud simulators are worst temperature predictors" but, as is normal in these circumstances, the headlines were all about global warming being "worse than we thought".
Yesterday Nature published a similar paper, this time looking at the El Nino phenomenon. It seems that if you take the models that best simulate extreme rainfall they predict that extreme Ninos will take place much more frequently in future, with all the floods and droughts and the like that accompany them.
The team identified 20 climate models — half of those available — that were capable of simulating extreme rainfall. They then used the models to compare the occurrence of extreme El Niños in a control period, 1891 to 1990, versus a warmer period extending from 1991 to 2090. Although the total number of El Niño events decreased, 17 out of the 20 models predicted more major El Niño events, with the average frequency increasing from once every 20 years to once per decade.
As an aside, it's notable that only half of the climate models were considered capable of simulating extreme rainfall, so one can reasonably wonder why we should base public policy on computer simulations that can't get even basic phenomena like heavy rain correct. One can also wonder at the paper's use of CMIP5 model runs, which don't incorporate the IPCC's latest estimates of aerosol forcings, and a perturbed physics ensemble, presumably the Sexton one, which has problems of its own. This is speculation though and I'll need to confirm when I get hold of the paper.
Of more immediate relevance to my point is the similarity in approach to the earlier paper - I think what we have here is a case of "the least absurd rainfall simulators predict a devastating increase in El Nino". Now you would think that, this being the case, the writers would have exercised a little restraint, perhaps mentioning uncertainties here and there. But not a bit of it. Here are a few of the headlines:
- Nature: Frequency of extreme El Niños to double as globe warms
- Guardian: Unchecked global warming 'will double extreme El Niño weather events'
- The Australian: 'Twice as many extreme El Ninos'
- Business Standard: Brace yourself for extreme El Ninos every 10 years
- New Scientist: Devastating El Niño events to double this century
Reader Comments (120)
Is this what Judith Curry calls "'climate model taxonomy'"?
In her interview here (audio and transcript) http://www.econtalk.org/archives/2013/12/judith_curry_on.html
she is scathing in her criticism of this type of "research",
" They get a press release and there's a lot of funding in this area. And I call this climate model taxonomy because I don't have confidence in climate models on regional spatial scales, so I think the whole impacts game related to climate models is rather pointless"
Russ Roberts is a fine interviewer and his economic podcasts are often worth listening to.
Apologies if this interview has been pointed out before.
Climate Model "Porn"?
/facepalm
When will these folks admit that we have not the foggiest idea about many of the factors involved with global warming and climates? Why can they not remember the butterfly effect, when one small change in the system can create chaos further away, and there are gazillions of small changes occurring at every moment? In other words, it is a chaotic system, and we simply do not have enough information to disentangle the chaos – or even if we ever will be able to!
Nope. In minds more tightly shut than the vaults of Gringott’s Bank, these folk are determined to prove that it is all the fault of CO2 – and not just any old CO2, but only the human-generated CO2!
It has to make you wonder about the mentality of those who actually believe what these people say.
Readers should be aware that the IPCC is a science-free zone and its UK outlets are exercises in applying the propaganda locally, e.g.: http://www.tyndall.ac.uk/people/Asher-Minns
I am left asking myself:
Of the 20 selected models that could "correctly simulate extreme rainfall", what proporion of those have correctly predicted - or at least not been falsified by - temperature trends. i.e. out of sample temperature predictions for those models.
Which models are at the intersection of "best climate model simulation of clouds" and "best climate model simulation of extreme rainfall" ?
I despair that Nature makes such unscientific claims. What has happened to this, once respected journal? New Scientist was on the way down years ago: and one has never expected better from the Guardian. We wait now for the BBC 's take on the work.
There was discussion some time ago (I think at Lucia's) about why they don't simply discard all the models that run way too warm, and Nick Stokes defended the modellers with the argument that the models don't just model temperatures.
His argument was that some of the models that get temps all wrong get the rainfall about right, and vice versa, so if you simply select on temps you get dreadfull emsemble result on other metrics. Dr. Brown at Duke has interesting things to say about ensemble results ;)
I personally think that Nick's argument means that all the models should be rejected as none get the climate right. To paraphrase Tamsin, "All models are wrong, so ditch them."
All the models are based on incorrect 'Forcing' physics. The definition of hot or cold is based on the extent to which they use incorrect cloud physics to offset the exaggerated temperatures.
Two wrongs don't make a right. Full Stop.
They must know that each of these pieces of "research" is showing that the models are even more divorced from reality than previously thought. By allowing that it to be said that their "research" shows that things are "worse than we thought", shows complete dishonesty.
They must know that each of these pieces of "research" is showing that the models are even more divorced from reality than previously thought. By allowing that it to be said that their "research" shows that things are "worse than we thought", shows complete dishonesty.
They must know that each of these pieces of "research" is showing that the models are even more divorced from reality than previously thought. By allowing that it to be said that their "research" shows that things are "worse than we thought", shows complete dishonesty.
What's going on here. Is it just me, or is there something strange happening at the site?
I see no scientific merit in selecting models on the basis that they may be able to handle one aspect of observed weather when they are clearly wrong on the other major trends.
This is a computer game that has endless possibilities. Just select the ones that generate the most alarmist headlines and wait for the funding to roll in. This is not science.
Most models were developed in the belief that carbon dioxide dominated the climate and nothing else really mattered. These models have been shown to be wrong. Their creators have been tweaking the parameters to simulate more closely the observed data, but distorting a wrong model to give a better output is a waste of time.
The reality is that climate scientists are just beginning to discover that the decades they spent being obsessed about radiative forcing have left them ignorant about the real drivers of our climate.
One of the authors of this paper is Mat Collins, Prof at the Met Office and one of the authors of the IPCC SPM. Yesterday on twitter he did not link to the paper, which acknowledges that other work does not support the El Nino claim, but tweeted a link to the hyped-up Guardian article, and retweeted Damian Carrington.
Mat Collins @mat_collins 15h
Unchecked global warming 'will double El Niño weather events' http://www.theguardian.com/environment/2014/jan/19/unchecked-global-warming-double-el-nino-weather …
Mat Collins @mat_collins 12h
Devastating El Niño events to double this century - environment - 19 January 2014 - New Scientist: http://www.newscientist.com/article/dn24889-devastating-el-nino-events-to-double-this-century.html
Damian Carrington @dpcarrington 13h
by me: Unchecked global warming 'will double extreme El Niño weather catastrophes' http://gu.com/p/3mxcy/tw #climate #elnino
@ Philip Bratby
"They must know that each of these pieces of "research" is showing that the models are even more divorced from reality than previously thought. By allowing that it to be said that their "research" shows that things are "worse than we thought", shows complete dishonesty."
I'm sure they must mean:
"Our research shows that things, that is our models, are worse than we thought!.
Phillip Bratby, it's just you ;)
"What i tell you three times is true."
The Hunting of the Snark
Lewis Carroll
Follow the link that HaroldW gave in “McItnyre’s mirror image”. All you need to do is read the caption under the photo: “…published fabricated data in 30 peer-reviewed papers.” to realise how tenuously attached to reality “peer-review” can be.
I know I can get very boring on this subject but can I remind everyone that the instigator of "The Climate is Out to Get You™" was Strong and the Club of Rome or at least their determination to roll back civilisation and reduce the population, an aim which could only ever be realised by the creation of a supra-national government based on the UN.
CO2 turned out to be a very useful means to this end especially since it brought all the eco-warriors onside with their opposition to "globalisation" and "development" and all things "chemical".
30 years down the line they're not making all that much headway but the useful idiots in the media (which includes the scientific press like Nature) are still happy to trot out any old rubbish so any old rubbish is what we get, and the scarier the better.
Professor Brown is right: if none of the models is accurate then all of them are wrong and averaging their output is also wrong even when it's right (on the stopped clock principle).
Schrodinger's Cat is partly right in that everything has been based on CO2 and only CO2 being what mattered but (I contend) wrong in his use of the word "belief". I don't think they believed it was CO2 at all but CO2 was the one thing where a drastic reduction in output could bring about the socio-political result desired.
It was never about the science; it was always about the politics. And still is.
There is a real and apparent need to relate ENSO with CO2. The only way that GAT is going to increase in the foreseeable future is with an El Nino kick. The predictions, forecasts, models, will be nullified when GAT fails to rise and falls out of the model uncertainty zone for good.
If ENSO is still defined as natural and cyclical and GAT increases after the next El Nino then where does the CO2 science stand?
Expect to see many more stories testing the understanding of ENSO.
Andrew, a weekly roundup of news stories on Energy Matters called Blowout. The story by Gerald Warner in the Scotsman should in particular entertain some of your readers.
Blowout week 3
The Arab Spring - Impact on Oil Production
The Scotsman:
Paul Matthews
This is a key thing to note. The paper (like the Sherwood one a couple of weeks ago) is useful in that it looks into the details of atmospheric processes as simulated by the models and as seen in reality, and it does of course acknowledge other work with different conclusions.
While it is an important conclusion and well-articulated in the paper, the media have, unfortunately, done their usual thing in hyping it up and ignoring uncertainties. It is very rarely correct to say "will" when talking about climate model projections - "could" or "may" are usually about as confident as you can get.
FWIW I think the Sherwood paper was a similar example of a good piece of work which was subsequently over-hyped by some. Although Sherwood et al did compare with observations (radiosondes) and reanalyses (essentially models "merged" with observations to better constrain quantities that cannot be directly observed) this was only really a small part of the paper. The most useful part of the Sherwood paper was identifying a major reason why the models differ so much in their estimates of climate sensitivity - so it's a good paper, but not the "proof" of high sensitivity that some seemed to think. I think it just reinforces the fact that uncertainty in climate sensitivity remains large.
Bishop Hill, I'd be intrigued to know what you think the alternative is to climate models, given that nearly everyone (including you) agrees that increased emissions of GHGs has at least some influence on climate, and that we are increasingly moving into unknown territory in the climate system which has not previously been observed (ie: CO2 concentrations well above those for a long time back in the past). We don't know what's going to happen, but surely it makes sense to at least try to estimate what might happen - how do we do this if we don't try to calculate it?
Richard
You are fine to carry on developing GCMs, so long as you don't try to use them as inputs into the policy process. I think you have to tell policymakers:
1. That the GCMs are inadequate for making predictions
2. That empirical studies suggest CS is low.
Compare to what the CCC told the HoC ECC committee a couple of weeks ago.
Bishop Hill, I'd be intrigued to know what you think the alternative is to climate models, given that nearly everyone (including you) agrees that increased emissions of GHGs has at least some influence on climate (...) We don't know what's going to happen, but surely it makes sense to at least try to estimate what might happen - how do we do this if we don't try to calculate it?
Jan 20, 2014 at 11:00 AM Richard Betts
Well, I'm not BH but here is my view.
- Yes it makes sense to try to calculate these things. And, as the Met Office has said, "climate models are the only way to predict future climate", so it makes sense to do research into how to construct and test models.
But, until there is some way of validating models, it is better to simply say "we don't know" than to rely on the output of models which have either not been validated or which have already failed in their predictions. If you rely on an unvalidated model, you are worse off than having no model at all.
I regard as rubbish the Met Office's claim that, since models can reproduce past climate, this validates them. This is the fallacy of 'testing on the training data'. (Plus, my understanding is that when current models are asked to reproduce climate from further back, they fail there too, as well as having failed to predict recent climate.)
One of the symptoms of the sickness of climate science is that it regards the output of a model as evidence, rather than as an illustration of a hypothesis.
Re: Richard Betts
The climate science community have, unfortunately, done their usual thing in allowing the media to hype it up without making any attempt to correct them.
Climate models are fine, just don’t present their output as anything other than conjecture.
"I suppose it is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail." - Abraham Maslow, The Psychology of Science, 1966
I am beginning to feel a little wistful. It would be so easy to leap on such papers and generate big newspaper articles out of them, each one forecasting a worse and sooner apocalypse than the last. What a great job it must be to behave in such a manner. Where did I go wrong?
Richard, thanks, your comment is welcome, but as Terry says, your colleagues are playing an active role in the hyping.
Here is a recent paper saying that "Overall, there is no evidence that there are changes in the strength, frequency, duration, location or direction of propagation of El Niño and La Niña anomalies caused by global warming during the period from 1871 to 2008."
The alternative to models is real-world data.
Richard Betts
1. I take it that you are not one of those who believes that readings taken of CO2 in the 19th century show levels higher than the climate science community claim,
For example: Slocum disputes Callendar's 292ppm claiming that his data were cherry-picked and that if all the data were included a more accurate figure would be 335ppm.
Beck (2006) shows anything but a sub-300 steady state followed by a rapid rise to the current ~400, with spikes of 440ppm in the early 19th century and again in the early 1940s.
2. In its first assessment report the IPCC itself admitted that climate was chaotic and therefore probably not capable of being modelled in any meaningful way. I don't see in what way the climate has "changed" in that way in the last 20-odd years.
3. I also take it that you are not a fan of Professor Brown but it seems eminently logical to me that if none of the models is right then an ensemble of the models cannot be right either. And given my point 2, why do we bother?
4. I actually agree with you about the media over-hyping. It would help if those in the scientific establishment would themselves be considerably more cautious in the language they use in press releases (or perhaps forswear issuing press releases altogether — what useful purpose do they serve?) and complain loud and long and in public when the media make them more scary than they really are.
A few of us have been looking at the long term PDO cycle which drieves the decadal frequency of El Ninos and La Ninas. Mike Jackson put up a graph.
http://www.woodfortrees.org/plot/hadcrut3gl/from:1880/to:1910/plot/hadcrut3gl/from:1880/to:1910/trend/plot/hadcrut3gl/from:1910/to:1940/plot/hadcrut3gl/from:1910/to:1940/trend/plot/hadcrut3gl/from:1940/to:1970/plot/hadcrut3gl/from:1940/to:1970/trend/plot/hadcrut3gl/from:1970/to:2000/plot/hadcrut3gl/from:1970/to:2000/trend/plot/hadcrut3gl/from:2000/to:2013/plot/hadcrut3gl/from:2000/to:2013/trend
This shows that the last two "warm" PDOs (1910-1941 and 1972-2002) warmed at about the same rate. The intevening "cool" parts of the 60 year cycle showed most cooling btween 1880-1910, and least cooling after 2002.
Perhaps the change in rainfall is not due to a strengthening effect from more frequent El Ninos, but a weakening of the cooling effect of La Ninas.
Mike Jackson
Think of it as spread betting.
1. Naively using climate model runs with no context and deeper understanding to decide policy is folly.
2. Ignoring information that you gain from climate models in policy is folly.
3) The above two statements shouldn't be separated :)
If you set yourself against a method of investigation because you don't like the answers it gives you, you'll make bad decisions.
I understand that there about 100 such models being developed.
Is there any reason to suppose that continuing to fund these unskilled creations won't be throwing good money after a great deal of bad?
What development plans do they have? What 'prototyping' has been done to show that these developments will increase their predictive skill?
But most of all
WTF have you all been doing for the last 15 years when the temperatures have been static? Did it never occur t any of you to look out the frigging window? Or do your models all exist in such a virtual universe that any contact with grubby 'reality' is beneath you?
Whatever the cause, I can see no reason at all for the failed models to be included in policy discussions. And no amount of technical bafflegab will conceal the sad fact that they don't do what they say on the tin.
> Think of it as spread betting.
Where the spread is win, lose or draw so the house (modellers) always win and the punter (public) always loses.
I would love a statistician or mathematician to work out how they are screwing with probability functions as they select subsets of models from a pool in order to run some "scenarios".
These papers produce probalistic statements such as "likely" or probably" etc. without know what influence their methodology is having on probabilities of getting the model output they are looking at.
Climate_Science™
.,.. it's all about the PR & less about the science
In my oft expressed (and hugely unpopular) opinion ^.^ NO model should ever be used in the formation of a scientific theory. I do think that models have a use in trying to predict the consequences of known science but even there great care should be taken.
Until someone shows me the error of my ways; I stop reading when I read the word "model" ^.^
Bishop Hill, I'd be intrigued to know what you think the alternative is to climate models
Richard Betts
I take that as a subliminal agreement that the models are not working but you will never come out and admit it. The trouble with this attitude is the model outputs are given much higher credence in some quarters than they deserve. The MET office's and you own continued silence on this problem makes you just as liable for the mess we are in as the mad warmists who push their own agenda.
Yes, he does indeed. The particular one that usually comes to mind is:
"One simply wishes to bitch-slap whoever it was that assembled the graph and ensure that they never work or publish in the field of science or statistics ever again."
Comparing a set of selected model predictions with another set of selected model predictions, is just predictions about predictions. When a modeler is able to actually make usefully accurate predictions about the real world, then rest assured; they will be the first to tell you. Until then.... back to the slapping.
I am writing to you on behalf on the Scottish Climate & Energy Forum, we are conducting a survey of those interesting in the climate debate which should be of interest to all involved.
The main focus is on the education and work experience of participants, but it will also assess employment and social factors for their relationship with views on climate.
We would be very grateful if you would take the time to complete the survey. The responses are confidential.
The url is: http://scef.org.uk/survey/index.php/868721/lang/en.
regards,
Mike Haseler
PS. Despite a glitch this morning (caused by all things trying to back up entries), we already have 341 respondents.
I agree with DougM's point 2.
The "information that you gain from climate models" is that they are not validated and have no predictive power. It would be folly to ignore that in policy making.
Less flippantly, I have to agree with Latimer. In what other field can you be consistently wrong for 15 years yet still keep plowing full steam ahead...and all with no accountability. Where is the brave man to call a halt to this expensive failed project (happens all the time in industry)?
In answer to Richard's question, no modelling is better than failed modelling. Just say we don't know, indeed can't know at this stage. Get someone to keep collecting the raw data (preferably someone who can keep and file accurate records with adequate change control and use Excel: a £15k/yr student should do) and go off and do something more useful (you keep telling us your jobs don't depend on the CO2 meme).
MikeHaseler
I object to the implication that my conclusions about climate change and energy are related to my work experience, education and social factors. I therefore refuse to take part in the survey.
David Rose:
I can't say for sure I assume you understood what history would make of such work. Future egg on face is worth avoiding even if it means being called silly names today.
Jan 20, 2014 at 11:47 AM
Another EM banality/naivety presented as a profundity.
Richard Betts
I know you're a busy man with a real job or two but I'd be interested in your response to The Bishop's and other's points in answer to your original question. When you don't respond it looks like another Mosher Drive By
This is because their theory claims that temperatures will increase with increased water vapour in the atmosphere. Unfortunately clouds do not provide a positive feedback but a negative one. Rainforests are cooler than drier deserts of the same latitude, ie., the same insolation.
Doug, Richard,
Point 2. is only valid if the information you gain has any basis in reality and that’s the bloody problem! Time and time again the models are shown to be wrong when measured against what actually happens. No-one objects to you working on them, just man up and stop pretending that the models are anywhere near being useful (useful as in useful to the man in the street). You’re kidding no-one but yourselves.
Breath of Fresh Air
If you have no accurate forecasting tools you have two options.
You can follow the precautionary principle and hope for the best while planning for the worst.
Or you can do nothing and wait and see. Anyone in a high-risk profession will tell you that wait-and-see is a good way to get oneself killed. Google "incident pit".
Did those people who consulted the oracle at Delphi make better decisions because they thought they knew what the future held? I'm sure that at the time they thought that flawed advice was better than no advice at all.
Entropic Man,
Exactly - we should plan for the probability of fairly frequent cold winters, and the possibility of occasional brutal winters, by ensuring that abundant supplies of cheaply-produced energy are available. People will die otherwise.
@MikeHaseler. Completed your survey and was very impressed with how well it was constructed.
I actually enjoyed taking part. A+