Bishop Hill On the limits to climatology
Dec 11, 2013 Climate: Statistics 
This is the abstract of Professor Leonard Smith's lecture at the AGU meeting currently taking place on the other side of the pond. There are salutary lessons for climatologists. If anyone can point me to the video, that would be helpful.
Over the last 60 years, the availability of large-scale electronic computers has stimulated rapid and significant advances both in meteorology and in our understanding of the Earth System as a whole. The speed of these advances was due, in large part, to the sudden ability to explore nonlinear systems of equations. The computer allows the meteorologist to carry a physical argument to its conclusion; the time scales of weather phenomena then allow the refinement of physical theory, numerical approximation or both in light of new observations. Prior to this extension, as Charney noted, the practicing meteorologist could ignore the results of theory with good conscience. Today, neither the practicing meteorologist nor the practicing climatologist can do so, but to what extent, and in what contexts, should they place the insights of theory above quantitative simulation? And in what circumstances can one confidently estimate the probability of events in the world from model-based simulations?
Despite solid advances of theory and insight made possible by the computer, the fidelity of our models of climate differs in kind from the fidelity of models of weather. While all prediction is extrapolation in time, weather resembles interpolation in state space, while climate change is fundamentally an extrapolation. The trichotomy of simulation, observation and theory which has proven essential in meteorology will remain incomplete in climate science. Operationally, the roles of probability, indeed the kinds of probability one has access too, are different in operational weather forecasting and climate services.
Significant barriers to forming probability forecasts (which can be used rationally as probabilities) are identified. Monte Carlo ensembles can explore sensitivity, diversity, and (sometimes) the likely impact of measurement uncertainty and structural model error. The aims of different ensemble strategies, and fundamental differences in ensemble design to support of decision making versus advance science, are noted. It is argued that, just as no point forecast is complete without an estimate of its accuracy, no model-based probability forecast is complete without an estimate of its own irrelevance.
The same nonlinearities that made the electronic computer so valuable links the selection and assimilation of observations, the formation of ensembles, the evolution of models, the casting of model simulations back into observables, and the presentation of this information to those who use it to take action or to advance science. Timescales of interest exceed the lifetime of a climate model and the career of a climate scientist, disarming the trichotomy that lead to swift advances in weather forecasting. Providing credible, informative climate services is a more difficult task. In this context, the value of comparing the forecasts of simulation models not only with each other but also with the performance of simple empirical models, when ever possible, is stressed.
The credibility of meteorology is based on its ability to forecast and explain the weather. The credibility of climatology will always be based on flimsier stuff. Solid insights of climate science may be obscured if the severe limits on our ability to see the details of the future even probabilistically are not communicated clearly.
Over to you Professors Walport and MacKay.
Reader Comments (49)
Bish; Anthony Watts is at that meeting (as you are no doubt aware). One of his posts contained this link to live streaming and video on demand which might have what you are looking for:
http://fallmeeting.agu.org/2013/virtual-options/live-stream-video-demand/
So what is new about this? Did Walport or Mackay ever say models are infallible or that we should accept what they tell us without question?
Actually yes, Chandra- witness the partisan IPCC reports, Stern Report and the resulting crass stupidity of the Climate Change Act- all largely based on output from such models.
"So what is new about this? Did Walport or Mackay ever say models are infallible or that we should accept what they tell us without question?"
They DID, however, say we should accept them with a 97% certainty. And that if we don't we are akin to a holocaust or tobacco lung cancer denier. They did (and do) say that we should remodel our, entire society, at the cost of trillions, on the basis of the models.
Is that definite enough for you?
Actually yes, Chandra- witness the partisan IPCC reports, Stern Report and the resulting crass stupidity of the Climate Change Act- all largely based on output from such models.
Dec 11, 2013 at 4:48 PM | Don Keiller
"Partisan" "crass stupidity"
It decreases the credibility of a scientist when he descends to using emotional trigger words in a scientific discussion.
EM
Surely one can (rightly) describe the Climate Change Act as crass stupidity as it is an act of parliament and therefore political and not scientific in the slightest?
Actually yes, Chandra- witness the partisan IPCC reports, Stern Report and the resulting crass stupidity of the Climate Change Act- all largely based on output from such models.
Dec 11, 2013 at 4:48 PM | Don Keiller
=======================================================
Don - please give the CCA its full title, being
Climate Change (Destruction of the Economy) Act 2008.
Brought to you by the clown Miliband Minor.
Stuck-record
Indeed you are. This straw man is brought out regularly by those like yourself.
Read AR5, the Stern Review or any scientific paper on climate change. You will find the uncertainties laid out as probabilities and confidence limits. Uncertainy is part of a scientist or forecaster's normal working pattern.He is aware of its existance, and its limits.
I never see probablilities or confidence limits in the comments of the sceptics. This is because you are certain that you are right. Belief does not need confidence limits.
EM
Since the IPCC is a political entity not a scientific one and the Climate Change Act is by definition a political action what is this "scientific discussion" you are talking about?
Chandra's usual occupation of trying to derail a thread has got itself two replies both aimed, quite rightly, at the jugular (or some other part of his anatomy perhaps). Your attempt to pontificate on what is and is not "credible" in a scientist simply puts you in the same category as him.
In any event it is not the science that is relevant to mention of Walport and Mackay and to pray in aid the Stern Report simply reinforces the accusation of crass stupidity.
I haven't heard any mention of "uncertainty" from the scientists-turned-political-activists who are "advising" government. I look forward to the day.
Entropic Man, You said "I never see probablilities or confidence limits in the comments of the sceptics. This is because you are certain that you are right. Belief does not need confidence limits."
I suggest that that is a very silly comment. It is of the essence of a sceptical scientific approach to a technical subject matter to take into account the uncertainty of life. An intelligent scientist specifically looks for the falsifiable nature of a statement or hypothesis. It is the cult warmists like yourself who have a religious belief in the pseudo science of climatism, not those who criticise them and look for hard evidence instead of unsubstantiated claims of, for example, the idea that trivial increases of anthropogenic carbon dioxide in the atmosphere will shortly lead to irrevocable and extreme global warming. If you actually possess this evidence, then document it. Everything else is hand waving.
Tony.
Don Keiller, no you don't understand. Show me where Walport or Mackay ever said models are infallible or that we should accept what they tell us without question. Or are you saying they wrote the IPCC reports the Stern report and the CCA?
The bottom line is that the models bear no relation to reality. There has been no increase in temperature this century but the models show rapid warming. These are facts. Everything else is speculation.
Mike Jackson, Anthony Ratlliffe
This is a thread discussing the science of climatology, The Climate Change Act is the political side issue. I dont know why Keiller raised it
The science is documented. There are a very large number of scientific papers out there, covering many aspects of climate change.
Together they fit into a coherent paradigm describing and explaining how the climate would be expected to behave under natural conditions. They also document how our own actions as a civilization, particularly our CO2 emissions,
are distorting that behaviour.
This paradigm is falsifiable if you have the evidence. Do you?
You'll have read your Popper and heard of Russell's Teapot. If AGW were an unfalsifiable hypothesis such as "God exists?" I would expect the burden of proof to rest solely on the theists. This is not the case. There is considerable evidence for AGW and it is falsifiable.
To falsify AGW you would need to present counter evidence, including supplying alternative axplainations for all the parameters which are changing as the AGW paradigm would expect. Anything less is dodging the issue.
Schrodinger's Cat
The bottom line is that the models bear no relation to reality. There has been no increase in temperature this century but the models show rapid warming.
These are facts. Everything else is speculation.
The bottom line is that the models bear no relation to reality. There has been no increase in temperature this century but the models show rapid warming.
These are facts. Everything else is
speculationPolitics.There - fixed it for ya.
I have some issues with Smith's choice of words in the abstract.
What is a computer simulation if it is not automated-theory in action?
Extra computing power certainly helps if a given problem is made tractable by brute force, but there are simpler systems than climate that are refractory to this approach. He sounds like a man who wants a new computer.
"Together they fit into a coherent paradigm describing and explaining how the climate would be expected to behave under natural conditions"
This is not the case. The models are very poor at simulating natural climate drivers. The climate scientists do not understand the natural climate. That is the big problem. When the computing power and modelling techniques became available they had already decided that CO2 dominated everything and the rest didn't matter, so they have spent all this time modelling CO2 driven warming in ever greater detail.
Unfortunately for them, correlation is not proof of cause and effect. They backed the wrong horse when they picked CO2 concentration. They should have been looking at ocean cycles, solar cycles and cloud coverage. They have very little understanding of these factors and they are miles away from being able to model them.
I'd have thought that if EM actually read the abstract above he might pause and contemplate it's meaning.
If not, he may prefer to consider Nic Lewis's current post at Climate Audit which rather nicely shows that the scientists that EM is so keen on say one thing in one chapter of the AR5 WG1, but another chapter says something else. Whatever else this state of affairs may be, it is most definitely not a "coherent paradigm".
EM at 6:31 suggests skeptics ought to give their counter-proposal science for similar attack.
This suggestion misses what many of the skeptics are saying.
What these skeptics are saying is not that we have better science, but that the establishment science is not yet good enough to justify astronomical investments. If some guy in a bar shows you a poor map to the lost gold mine, it is not your job to produce a better map to the lost gold mine. Rather, it is your job to hang onto your wallet like grim death until convincing evidence is shown.
Many people outside the ivory towers of science understand this principle perfectly, and live it daily. It is an important part of OUR life, whether or not it is part of yours.
There is a proper place for the proposals and counter-proposals. But that does not dominate the conversation here. This conversation also includes the real-world consequences of banking too early and too heavily on those proposals.
EM seems to know the falsifiability criteria for AGW. I'd be pleased to hear them all and know the peer-reviewed papers which set them out.
EM
You're looking through the wrong end of the telescope again.
You can't just come along and say, "I've got this wonderful idea that CO2 is going to cause the earth to overheat and kill us all and if you don't agree with me you are in the pay of Big Oil and you can shut up and **** off because I am a scientist and therefore I am right."
Because if you do I shall say "show me some evidence" and when you point me at your Play Station in the corner I shall repeat "show me some e-v-i-d-e-n-c-e; not your computer games" and then I shall duck while you throw your toys around.
I may have over-simplified the situation slightly but not by all that much.
If you are going to try to prove that "this time it's different" when the evidence all around is that "no it ain't" then I am not likely to be sympathetic.
I have been asking for the empirical evidence for the last 20 years. I am still waiting.
EM
[...] Together they fit into a coherent paradigm describing and explaining how the climate would be expected to behave under natural conditions.[...]
And isn't that the very core of the problem when modelling 'climate'?
The fact that you have no idea what 'natural conditions' might actually look like (no baseline). The fact that you have no idea what 'natural processes' (all those new unknowns that have been recently discovered) may impact 'natural conditions' (that one can't identify anyway).
I see a modelling problem. You have no baseline, no idea what may or may not influence any identified element and you dodge these 'inconveniences' by playing around with 'ensembles'.
I claim to know everything about Horse racing by betting on every horse in the race - I always win for some reason.
In my long life, I have learned a number of things very well. These include:
bigger, faster computers assist in the making of bigger mistakes faster
civil servants making demands on my wallet are always totally intemperate
politicians can always justify whatever they imagine requires justification
men really do go mad in crowds and recover their sanity one at a time
intelligent trolls are an oxymoron
Enthalpic Man, I note that you manage to come over sanctimonious without using any trigger words at all. Well done!
PS You're obviously unaware of the 'existance' of the spell checker in the comment box.
Slightly off topic--are the pictures on the blog now knitted or is it my computer?
One of the biggest problems with AGW is that it is a form of Catch-22. Few theories can be proven, they can only be disproven; an idea that AGW protagonists can hide behind. However, it is not possible to prove a nothing; just as in a court of law it is rarely possible to prove you are innocent of a crime you are on trial for (hence the “innocent until proven guilty” idea), AGW does not exist, so it is not possible to prove it does not exist; however, the onus is on the sceptics to show the falsity in an unfalsifiable theory. The final irony is that any change in the general weather patterns for an area can be used as “proof” of AGW (witness the Cook Islands representative in – where? Berlin? – in June this year: “It is so cold here because of global warming.”)
AGW is perhaps like UFOs – no-one can prove UFOs do not exist, but the vociferous UFOlogists seem unable to prove that they do.
EM
Keenan wrote a report on AR5 indicating there is NO observational evidence found by the scientific establishment that
there is warming.
You tell us what's wrong with his statistical analysis on AR5, or shut up.
As far as I can ascertain Prof Leonard Smith (and it is a picture of him I thnk) has a paper in press:-
The myopia of imperfect climate models: the case of UKCP09
Frigg, Roman, Smith, Leonard A. and Stainforth, David A. (2014) The myopia of imperfect climate models: the case of UKCP09. Philosophy of science . ISSN 0031-8248 (In Press)
with the following succinct abstract
http://eprints.lse.ac.uk/54818/
The United Kingdom Climate Impacts Program’s UKCP09 project makes high-resolution forecasts of climate during the 21st century using state of the art global climate models. The aim of this paper is to introduce and analyze the methodology used and then urge some caution. Given the acknowledged systematic errors in all current climate models, treating model outputs as decision relevant probabilistic forecasts can be seriously misleading. This casts doubt on our ability, today, to make trustworthy, high-resolution predictions out to the end of this century.
That is quite a devastating indictment.
Perhaps a little off-topic (but when has that ever stopped me?!): while not a Twit, myself, but found this in a side bar on another site:
I suspect that meeting must have been a barrel of laughs! “Democratic” energy? Curious as that concept may be, I have little doubt that what energy they will be championing will not be affordable or sustainable.
Pity the frozen pensioners this year, with folks like this on their side.
Helpful but as long as there is a financial incentive to produce results that support a particular conclusion there will always be a willing cadre of deluded, innumerate or just plain dishonest "scientists" ready to oblige.
Only the passage of time (and the observational evidence it inevitably produces) is powerful enough to break the spell and expose the various hues of charlatanism which are currently at work in the climate debate.
Entropic man you ask someone for a lot of money you better have a much better reason than 'because I want it ' if you hope to get it . Right now that is all climate 'science' has and they not just a great deal of money but massive changes , often inspired by a political not science based approaches .
Your 'faith ' in the rightness of 'the cause ' no matter how strong means nothing if you cannot 'prove ' why others should share it.
'The trichotomy of simulation, observation and theory which has proven essential in meteorology will remain incomplete in climate science.'
I'm puzzled by claims that climate theories are not test-able. Why can't the climate modelers show that their theories and models predict the MWP, the little ice age, or the changing rain patterns in the American Southwest around 1300? (based on the conditions a priori, of course).
Normally, the clinically insane who insist on harming their fellow human beings are placed in secure mental hospitals under guard.
Now they are recruited by the IPCC to run Climate Alchemy, and DECC to run energy policy.
EM. It scientists cannot foretell the future state of the climate, therefore doing it with uncertainties and confidence limits is a futile, not to say deceptive, activity. I suggest you tool over to IPCC TAR3 WG1 14.2.2.2 written before the scientists became under the thrall of the politicians and you'll see it state unequivocally that the Earth's climate is a coupled non-linear chaotic system and therefore its future state cannot be predicted.
That doesn't say there isn't global warming, or that humans have caused it, it simply says that whatever's happening we cannot say what it will be like in the future because small changes in input can make huge changes to the predictions.
So please stop with the propaganda and get back to the science.
If you believe the power of computers can solve everything, then I suggest you think of a computer program that can look at the results of the throw of a die and make a prediction about the next number from the next throw with less than a 6 to 1 chance of being right. Don't bother you can't so some things are impossible.
It looks as if this thread is being taken over by the trolls and comments to their utterings. As always, they bring no further insight, just waste of time and annoyance. I wish they would keep to the Telegraph comments, these have become a hopeless resonator of mostly nonsensical viewpoints. I do hope this will not happen to the comments at Bishop Hill. Normally they are mostly quite valuable, I think it is the objective of Mr EM and Ms Chandra only to stick spokes in the wheel. I strongly believe the best method is to ignore them completely.
At 2.30 this morning I finished a submission to the House of Commons Climate & Energy Committee review of the IPCC & AR5.
Quotes:
"for all the IPCC claim to know about the climate, the truth is that they know diddly squat!"
"It is warming – we (public sector academics) don't know why most of the change occurred, but we can't think of a better reason than all that man-made CO2 from private business and so we think they should stop."
@Albert "I think it is the objective of Mr EM and Ms Chandra only to stick spokes in the wheel. I strongly believe the best method is to ignore them completely."
Agreed.
Chandra uses the typical handwave - that nobody expects the models to be perfect - but the real issue is that the models are not even adequate. Nevertheless current policy is based on the most pessimistic (and hence less likely) of these model outputs. I use and create models all the time: For some things I know they are very accurate - by testing against reality but for many more things they are just as inadequate as climate models and for exactly the same reason; nature is too complex!
The second handwave usually used is that the models "are all we have". But they aren't; We can make a decent prediction that mankind is not dominating the climate because we have already injected a large plug of CO2 into the system and the planet stopped warming, the stratosphere stopped cooling and the oceans have flatlined since accurate records were available. That was a pretty good test of the models - or rather their underlying assumptions - and the result is that the hypothesis failed the test. So while it was plausible that we may have affected the climate - the results are now in and they say that our contribution from CO2 is so minimal that it is easily overpowered by other factors. And those factors are in the parts of nature that the IPCC admit they know little about.
I suspect the only reason McKay and Walport and so many apparently clever people are convinced that manmade warming is dangerous is primarily because of the fashionable faux-green dislike for cars, roads, large houses and oil companies and so are easily convinced by any argument that denigrates them.
The only thing that moved the climate catastrophists away from their previous ice age panic (from the selfsame fossil fuels) was that the planet did the opposite of the predictions. The same is happening again and so eventually everyone will have to admit that the skeptics were right ot be skeptical. The only reason it is taking so long is because of the amount of people involved in the climate pessimism industry.
What climate doomsday scientists need is a computer the size of a planet. Not a brain, mind you, but a computer the size of a planet. Then, and only then, shall we know exactly when the climate doomsday will be upon us. Until then the uncertainties and the precautionary principle means we should act as though the climate doomsday is already upon us. This is a peer-reviewed fact that climate doomsday deniers just don't get.
JamesG, I like your 2nd paragraph. Occam's razor and all that.
(your other paragraphs are fine too!)
I dare not speculate how much computer power has multiplied since 1979, but I dare say, as does Willis, that climate doomsday science hasn't moved forward an inch since then. There was a doomsday scientist at the Met Office ( sorry, I can't remember the name but it doesn't matter because they are all Julia Slingo clones) who once asked an even more powerful computer to the super-duper one that had just been installed. His argument was that it would only cost a fraction of the cost of flood prevention measures in Thames river in case of a climate doomsday. Next, the Met will say an even better computer will cost only a fraction of moving to another planet in case the Earth goes kaput.
JamesG - very well put.
The only thing I would add (referring primarily to your third paragraph) is the strength of scientism in the circles in which the likes of Walport & McKay travel. Warmism comes decked out in all the paraphernalia of academic worthiness: how could it possibly be wrong?
This reminds me of the old saying about family firms: rags to rags in three generations. Warmism is a good example of "capital," built up by earlier generations, being squandered by feeble or feckless successors.
There have been such episodes in science before (eugenics, etc.), but never so extensive or so serious. I believe that the greenshirts WILL be defeated politically, because our survival depends on it (although the price of that victory may be terrible). But how will all the institutions that have been corrupted - including the enterprise of science itself - recover their reputations?
JamesG (Dec 12, 2013 at 10:30 AM):
Sorry to disappoint you, but they already have this covered – in October 2003, the U.S. Pentagon released a controversial report, “An abrupt climate change scenario and its implications for United States national security,” that explored how global warming could lead to rapid and catastrophic global cooling. (U.S. National Research Council, Critical issues in weather modification research (Washington, DC, 2003); Peter Schwartz and Doug Randall. “An abrupt climate change scenario and its implications for United States national security,” 2003, available at http://www.environmentaldefense.org/documents/3566_AbruptClimateChange.pdf (2 Jan 2006).)Now, whatever happens – if Singapore is disappears under snow, or the Arctic ice expands until Kent is buried under a mile of ice, it is all “proof” of global warming. Remember that South African delegate to a German AGW seminar:
You never know, perhaps those tin foil hats will be useful!Computer modelling and simulation is climate science's Enron, providing merely the illusion of scientific advancement, but with no meat on its bones, no actual substance and nothing empirical upon which to build.
There is no need to falsify AGW. But the purported impending catastrophes reside solely in the flawed extrapolations of computer simulations and can't be trusted further than you can throw Ken Lay's headstone.
Len Fiske
I wondered about the 'knitting' too, but I assume that Bish is putting the photos through a 'watercolour' filter to avoid having to dig out attribution details. Anything that fuzzes out Davey and Gummer's facial details (later post) is worth doing, IMO.
RE Shrodinger's Cat:
"They should have been looking at ocean cycles, solar cycles and cloud coverage. They have very little understanding of these factors and they are miles away from being able to model them."
That covers most of my views on GCM's. If we have little understanding of the role of clouds, running GCM's becomes a waste of time. I am coming to the conclusion that continuing to run a large number of models is an indication that the maintenance of funding tops the list of objectives of the organizations and scientists involved with them. I have yet to get a good answer when I ask why hasn't the focus of climate modeling research shifted to improved regional forecasting.
RE Shrodinger's Cat:
"They should have been looking at ocean cycles, solar cycles and cloud coverage. They have very little understanding of these factors and they are miles away from being able to model them."
That covers most of my views on GCM's. If we have little understanding of the role of clouds, running GCM's becomes a waste of time. I am coming to the conclusion that continuing to run a large number of models is an indication that the maintenance of funding tops the list of objectives of the organizations and scientists involved with them. I have yet to get a good answer when I ask why hasn't the focus of climate modeling research shifted to improved regional forecasting.
Wow, the true believers are desperately parsing words. Notice that when confronted with their own position: that their claims about AGW should be taken as 100% accurate, they run and hide. And yet still call skeptics wicked evil liars.
In 2010, Leonard Smith and co-authors argued that 'the lack of current predictive capacity on the relevant scale is a strong argument for why we must both control greenhouse gas emissions and prepare to adapt'. That is, I don't think you will find him saying that the shortcomings of the models diminish the need for mitigation, but rather the reverse.
http://www.jstor.org/stable/info/10.1086/657428
Ruth
Let's consider the usefulness of climate models first. If we conclude that their output is worthless then we can consider things in a new light.