Bishop Hill New climate consensus: GCMs are lousy
Nov 10, 2014 Climate: Models The models are lousy and probably wrong. In which direction they are wrong is not clear at all, yet.
Eduardo Zorita's comments on climate models are interesting. Given the strong evidence that the models run too warm, I think he is being far too even handed, but perhaps more interesting is the way his words echo the terminology used by the UK's top climate modeller Brian Hoskins.
Interviewer: Tell us the war story. How bad were the climate models when you started out?
Brian Hoskins: Ah, they were pretty lousy, and they're still pretty lousy, really. They were terrible.
I think we could say that there is a consensus here, but rest assured that it a bad case of climate denial to suggest that GCMs are not suitable tools for policymakers.
Bishop Hill
Interesting to compare these comments to Keith Shine's description of the climate models to parliamentarians.
...state-of-the-art climate models, which are our embodiment of the laws of physics as applied to the atmosphere...
I was roundly criticised by some for suggesting that Shine's words were misleading.
Reader Comments (50)
If economic models were as bad as this, they wouldn't've predicted the global banking crisis. Er, wait a minute.
Robert Brown says it better
http://wattsupwiththat.com/2014/10/06/real-science-debates-are-not-rare/
for anyone who missed it
"In which direction they are wrong is not clear at all, yet"
I read that as wrongness everywhere, not just temperature. Or is that all they're used for these days?
They may be getting to the point where no amount of fiddling with aerosols, TSI and greenhouse gases can reliably give them the current pause and still match the past temperatures. It may be dawning on them that there's something fundamental they don't understand and they're not just getting it wrong, they're not including it at all.
Are the computer models reliable?
Computer models are an essential tool
in understanding how the climate will
respond to changes in greenhouse gas
concentrations, and other external effects,
such as solar output and volcanoes.
Computer models are the only reliable
way to predict changes in climate. Their
reliability is tested by seeing if they are able
to reproduce the past climate which gives
scientists confidence that they can also
predict the future.
But computer models cannot predict the
future exactly. They depend, for example, on
assumptions made about the levels of future
greenhouse gas emissions.
Met Office publication: Warming Climate change – the facts
Related, Nick Stokes has a post where he claims the models don't match reality because weather is chaotic. In chaotic systems, models initially coincide with reality but slowly diverge. The IPCC models diverge right at the beginning.
Over 20 years of basically unlimited funding and we don't even get a lousy t-shirt.
Think of the opportunity cost of the money and time squandered on climate obsession. And they have the chutzpah to dare pose as morally superior and demand even more money and power.
Everything they know is encapsulated in these models ... and yet they were completely unable to predict the climate even one year ahead and spectacularly failed to predict the pause.
Like a motorcar ... the most dangerous bit is the nut behind the wheel. Likewise climate models are perfectly fine ... until being driven by an eco-nutter.
Model or no model, when someone claims their work is "state of the art" you know they are about to try and sell you something.
And the Met Office just hit pay dirt to the tune of £97m for their new box. Why would they stop when the ploy carries on working?
The models have some value for extrapolating the behaviour of existing features of tropospheric flows over a few hours and days to help with weather forecasting. But for climate, they are like patients in intensive care who need to be monitored, have their fluxes adjusted, and generally be pampered and cajoled to keep going and stay within tolerable, or even desirable bounds. Unlike those patients, however, the models can be switched off, restarted, and redirected at will to get what is being sought: illustrations of warming following small and temporary reductions in heat loss while keeping the incoming energy unchanged. The dramatic political/media impact of the Limits to Growth computer models can scarcely have escaped the notice of the charmers in the Club of Rome, and anyone else who wanted to make a big deal out of CO2 emissions. They got their big deal for sure. They have been dramatically successful. Maybe now the models can be sidelined. de-emphasised, or whatever, now that some of their limitations are visible to all who care to look. The CO2 scare bandwagon has reached such mass and momentum that it can surely be kept going without them. Josh has helped us see the way here, for example http://www.bishop-hill.net/blog/2014/8/30/inspirational-betts-josh-289.html and http://www.bishop-hill.net/blog/2014/9/5/come-on-down-josh-291.html.
After all, they used rising temperatures to scare us, as they tried with falling temperatures in the 1970s, and now that the temperatures have not been rising for nearly 20 years, they tell us they don't matter much and we should be looking elsewhere, or perhaps even that we need to recognise the complex multi-faceted nature of climate. Just like Houghton did not do as he sat beneath a huge picture of the MBH Hockey Stick during a press conference way back in the day when those temperatures mattered very much indeed to him and his like. Way back, that is, to 2001. See Steve McIntyre's presentation in 2009 or the Hockey Stick Illusion in 2010 for more context..
'I was roundly criticised by some for suggesting that Shine's words were misleading...'
Oh, you restraint does you credit, Your Grace...
I would suggest that his words were - how can I put it - it begins with 'bol' and ends with 'locks'...
A Computer Model is a Computer Game. A Computer Model is a Computer Game. A Computer Model is a Computer Game. There I have said it thrice. What I tell you three times is true.
I recall reading a blog (not BH) over this year - the name escapes me now - and a commenter reckoned there were just over 200 climate models and all but half-a-dozen or so had failed. It worked out at 97% failure rate. I wish I could find that comment now. Maybe someone on here knows how many models there are - and how many are considered to be failures.
I am not qualified to have a serious opinion. This is the closest to my view from someone who is.
Petr Chylek
To blame the current warming on humans, there was a perceived need to “prove” that the current global average temperature is higher than it was at any other time in recent history (the last few thousand years). This task is one of the main topics of the released CRU emails.
Some people were soeager to prove this point that it became more important than scientific integrity.The next step was to show that this “unprecedented high current temperature” has to be a result of the increasing atmospheric concentration of carbon dioxide from the burning of fossil fuels.
The fact that the Atmosphere Ocean General Circulation Models are not able to explain the post-1970 temperature increase by natural forcing was interpreted as proof that it was caused by humans. It is more logical to admit that the models are not yet good enough to capture natural climate variability (how much or how little do we understand aerosol and clouds,and ocean circulation?), even though we can all agree that part of theobserved post-1970 warming is due to the increase of atmospheric CO2 concentration.
Thus, two of the three pillars of the global warming and carbon dioxide paradigm are open to reinvestigation.The damage has been done. The public trust in climate science has been eroded. At least a part of the IPCC 2007 report has been put in question. We cannot blame it on a few irresponsible individuals. The entire esteemed climate research community has to take responsibility.
Laboratory Fellow, Remote Sensing Team Leader, ISR-2 MS-B244
Los Alamos National Laboratory
http://www.thegwpf.org/opinion-pros-a-cons/218-petr-chylek-open-letter-to-the-climate-research-community.html
Roger Pielke senior tweeted a link to a talk he gave recently.
Up to slide 30 or so there's a very good overview of the failings of climate models.
(And at the end of the talk there are some Josh cartoons!)
Those of us whose education failed to include Wittgenstein's "Philosophische Untersuchungen" may be wondering why the referenced post talks about some people only seeing rabbits, while others only see ducks. The reference below may help.
Rabbit-Duck illusion
Martin A - I think you'll find that is a red herring, "Climate Science" has moved on from the tired old debate about whether climate models matter....
Like Shine, Rob Varley of the Met Office also described climate models as "state-of-the-art". [Audio clip at the BBC article on the new £97m MO supercomputer.]
There are also state-of-the-art fusion energy facilities. Someday their successors may be productive, too.
Assumptions about future CO2 levels?? I would tend to think that they cannot predict the future exactly since they depend on a hundred other assumptions by the scientists who code them.
mcraig
The wordcrafting skill of the Met Office editors is wondrous to behold in their ability to convey the impression they wish to convey but with always a get-out clause.
They convey, to anybody but the most cautious reader, that their models could tell the future exactly were it not for the assumptions they had to make about future greenhouse gas emissions and possibly (encoded in the 'for example') on other assumptions.
^^^^ spot on @mcraig
For those interested in Facts and Fiction:
http://www.metoffice.gov.uk/media/pdf/9/2/factfiction.pdf
Martin A
Consider the possible linkage from the Met Office through Phil Jones to the University of East Anglia and thence to its well-regarded Creative Writing Department.
Climate models are good for keeping climate boffins employed. That's about it.
nby - interesting pamphlet to add to the collection.
I thought you'd like it - I wonder if Christy and Zorita have read it?
Related, Nick Stokes has a post where he claims the models don't match reality because weather is chaotic. In chaotic systems, models initially coincide with reality but slowly diverge. The IPCC models diverge right at the beginning.
Nov 10, 2014 at 12:04 PM | shub
===================================================================
"because weather is chaotic".
Well, whooda thunk it?...
Models follow the laws of physics
Why plumbers always site radiators under windows the coldest parts of a room
The law of thermal dynamics heat travels from where its cold to where it's colder not from where it's already hot to where it's hotter
Worth reposting Barry Woods post from unthreaded yesterday about what the Met office guy said on the BBC about the £97m new super computer
"Comments borrowed from James Anaan's blog: (There is an audio clip of an interview with Rob Varley of the Met Office as a sidebar to this BBC article).
Excerpt:
"... RV: We can tell you that the global average temperature is going to increase by 3 or 4 degrees if we carry on as we are, but the critical question is what is that going to mean for London? What is it going to mean for Scotland? What is it going to mean for my back garden? At the moment the general looks that we can produce really don't answer those kinds of questions.
Q: So you're saying you're going to be able to predict climate changes in temperature for individual British cities into the future?
RV: For regions of the UK into the future, in a way that is helpful for people who need to plan, in particular in terms of investment on infrastructure, in large-scale long-term projects.
Q: But you're still going to be dependent, aren't you, on the modeling, the computer modeling, which some people say is pretty inaccurate and can't really give us a proper picture of what is necessarily going to happen.
RV: Well, the modeling is state of the art for now, but with a more powerful computer, it will become...
Q: The models are going to be better...
RV: Yeah. ..."
"Climate models consist of complex equations which represent the process that occur in our weather and climate"
>But many simplifications are made to get these these equations to run and a lot is missing because we don't yet understand many natural processes nor can we replicate many of those we observe.
By running these models through a supercomputer we get forecasts of what will happen in the future. Although making climate models is a major part of our work, we'd be the first to admit they aren't perfect.
>Far from being just imperfect they are not even adequate.
But we do check models for accuracy by running them from points in the past and checking their predictions against what actually happened (called hindcasts). From this we know they represent the past climate accurately so we can have confidence in them for the future climate (the forecasts).
>Only a fool would pretend that a good hindcast begets a good forecast. Hindcasting is like taking an exam with the answers in front of you. Only forecasts can test a model and every medium-to-long-term forecast ever made by the Met has failed. A 100% failure rate is nothing to be smug about!
"The fact that lots of different models, put together by scientific organisations from all over the word, agree on the general forecast of a changing climate increases our confidence in the predictions."
>The fact that all of these models fail to represent reality means that they are all bad, not just some of them, and it is doubly foolish to assert otherwise.
"state-of-the-art climate models"
So it really is an art, not a science? Glad that's sorted.
but rest assured that it a bad case of climate denial to suggest that GCMs are not suitable tools for policymakers.
This wouldn't have been Betts by any chance, would it ?
The fact that lots of different models, put together by scientific organisations from all over the word, agree on the general forecast of a changing climate
So, their VV&T is that the models show that the clipmate changes, WOW £ billions for that master stroke of genius. Chuck another £billion their way for crying out loud. That's real value for money !
I don't think
What I see as the most telling exchange is:
What I see here is humility in the face of a difficult unsolved scientific problem. As Hoskins points out here and in other places in the interview this is exactly which climate science needs and what climate science currently lacks. it is why climate science is less and less scientific and more and more political. Both AGW consensus advocates and skeptics can learn from this
A lot of the problems the GCMs have can be traced back to the mindset that both spawned and unquestioningly accepted Mann's hockey stick. AGW was said to be "real" because they were unable to replicate 20th century temperatures without sensitivity to CO2 forcing. But if prior centuries' temperatures couldn't be replicated either, then this approach would have fallen flat on its face. And that was the local difficulty with the pre-Team era position which acknowledged past natural variability - the RWP, MWP, and LIA.
Then came Mann to the rescue. No past variability of any note, just a tenth or two of a degree of variation against a gently declining trend. Drinks all round.
The unfortunate but richly deserved consequence for all those climate scientists who acquiesed in Mann's junk statistics is that they have hugely underestimated natural variability. Now, with the pause, the screw is tightening: ignominy, ridicule and irrelevance are looming larger by the day.
Funny how all these models are just applying the "laws of physics" and they all produce different results.
The issue is not how bad models are (someone dishonestly mentioned economic models). The issue is that climate models have been presented as infallible and have been used to politically extort society.
"...[T]hey are like patients in intensive care who need to be monitored, have their fluxes adjusted, and generally be pampered and cajoled to keep going and stay within tolerable, or even desirable bounds...." --John Shade
"They're dead, Jim."
"...state-of-the-art climate models, which are our embodiment of the laws of physics..."
Translated: "...the very latest climate models we could cobble together, our Frankenstein's monster version of the laws of physics."
"It may be dawning on them that there's something fundamental they don't understand and they're not just getting it wrong, they're not including it at all."
Or to cite Wolfgang Pauli: "Das ist nicht nur nicht richtig, es ist nicht einmal falsch!" (that is not only not right, it isn't even wong)
What a telling unguarded comment from Hoskins. There is no intention to 'solve the problem'.
Julia said we have to trust the models: http://www.metoffice.gov.uk/media/pdf/q/e/julia-slingo.pdf
'So we have to trust these models and understand the scientific basis behind them, and accept that they are our best way forward for looking at what the future could look like for this planet in terms of our climate.'
i.e. whether the models produce a single testable prediction is immaterial.
I know I'm quoting a fictional character, but I recently got reminded of one of the good Sherlock Holmes quotes at a conference. It seemed relevant.
'It is a capital mistake to theorise before one has data.
Insensibly one begins to twist facts to suit theories, instead of
theories to suit facts.'
An open blog with very light moderation, 43 comments up after 1 day and still nothing from anybody offering a counter proposition that GCMs offer anything of value to policy makers on the most pressing issue of all time?
Why isn't this apparent lack of substance reflected in this headline hitting report - the "most important document" on climate change? What is the basis for Topic 2?
http://www.ipcc.ch/pdf/assessment-report/ar5/syr/SYR_AR5_LONGERREPORT.pdf
Where are the journalists?
http://www.bbc.co.uk/news/science-environment-29803811
http://www.bbc.co.uk/news/science-environment-29861942
http://www.bbc.co.uk/news/science-environment-29898173
http://www.theguardian.com/environment/climate-consensus-97-per-cent/2014/nov/10/we-must-manage-global-warming-risks-by-cutting-carbon-pollution-ipcc
And note; Christy and Emanuel's comments that Zorita references predate the SYR report by 6months:
Russ Roberts
PERMANENT LINK | MARCH 24, 2014
John Christy and Kerry Emanuel on Climate Change
EconTalk Episode with John Christy and Kerry Emanuel
http://www.econtalk.org/archives/2014/03/john_christy_an.html
There are similarities to seismic interpretation. Geophysicists from a physics background often make an interpretation which geologists say is unrealistic or wrong. Modelling groundwater once one moves from a simple uniform aquifer , especially the movement of a contaminant plume, often ends so over simplifying reality that the result is of little use.
One can model airflow around an object, vibration or stresses in an object but once of the natural world is too complex and involves too many non-linear relations. I think the real breakthrough will come when we become aware of what we can model and what we cannot, and that can only be perceived through extensive monitoring, over prolonged periods.
The problem is that there are vast numbers of egos and salaries dependent upon computer modelling.
Once again the skeptics are proven correct: Skeptics have been pointing out for years that climate models are not useful.
Charlie, my fellow Aussie Geoff Sherrington, who comments here, is a retired mining geologist with decades of experience who used models of various kinds over his career. The takeaway message (and he can correct me if I am wrong) is that you never know until you actually dig a hole. You can improve the odds of success somewhat, sometimes, by using models, but definitely not guarantee it.
Johanna,
As a full-time modeler, I'd agree with that.
It's kinda obvious if you think about it: most of the models that we work on, still need work. No one is working on the Ideal Gas Law or Poisson's equation any more - these are settled models, and while they're reliable, most of the value has been already sucked out of them. So we move on to harder, more complex models, particularly in industry.
So there are settled models, tried-and-true, but these aren't the kind of thing that active researchers are going to work on. 'Cause they're settled.
Climate models reproduce many major features of the climate (and a helluva lot more than they used to), but then, they also screw up some big features, too. These'll get worked out in time, and climate modelers will move on to more and more arcane or subtle aspects of the climate.
So two days on and still not a word explaining how GCM's can be relied on for policy decisions?
Just what is this based on?:
http://www.theguardian.com/environment/2014/nov/12/china-and-us-make-carbon-pledge
Anybody?
Yeah, "our embodiment" and "state-of-the-art" are weasel words, carefully avoiding making any actual assertion about accuracy or usefulness. But "not fit for purpose" is the compatible truth.