Bishop Hill Mann, straw man and SciAm
Jan 10, 2012 Climate: Mann Climate: Models Michael Mann is in Scientific American today, with a podcast discussing computer models and the hockey stick man has a pop at Freeman Dyson:
I have to wonder if Freeman Dyson will get on an airplane or if he’ll drive a car because a lot of the modern day conveniences of life and a lot of our technological innovations of modern life are based on phenomena so complicated that we need to be able to construct models of them before we deploy that technology.
This is a bit of straw man, and Mann in fact goes on to discuss the fact that
we can’t do experiments with multiple Earths and formulate the science of climate change as if it’s an entirely observationally based, controlled experiment.
However, after a lot of words about how models are used, he returns to his strawman.
And again, does Freeman Dyson, assuming he is willing to get on an airplane even though models have been used to test the performance of the airplane, assuming he does and he knows he’s going somewhere where they’ve predicted, where weather models have predicted rainfall for the next seven days, does he not pack his umbrella because he doesn’t believe the models? It's just in that case the worst that will happen is somebody gets wet when they wouldn’t otherwise have. In this case, the worst that can happen is that we ruin the planet.
It's not desperately edifying, is it?
Reader Comments (125)
I wonder if Prof Mann is aware that they also test fly planes?
"we can’t do experiments with multiple Earths and formulate the science of climate change as if it’s an entirely observationally based, controlled experiment"
But we CAN do experiments with aircraft before we let passengers anywhere near them.
Mann is an idiot, and any comparisons of his pathetic utterings with the thoughts of Freeman Dyson is a travesty.
The Mann's an idiot and clearly doesn't understand the difference between models based on proven science, tested over and over again both in the real world and in the laboratory versus Mickey Mouse junk science models based on a science in its infancy whose predictions have failed time and time again.
More to the point is the question as to whether Mann should ever travel on a plane or in a car given that MANN's models predict it will destroy civilisation!
If ever I wished to see the difference between a genius and an imbecile the ScIAm article reveals it.
It's pathetic.
I like the way he indirectly compares himself to Peter Higgs.
If aeronautical models were as incapable of making verified predictions as are climate models none of us would be flying in aircraft.
We've got a huge functioning analog computer perfectly modelling the earth's climate. Let's examine it.
============================
He may be an idiot, but he's purposely obfuscating the meaning of models. His hybris is in believing he can blow enough smoke ahead of him to banish the nemeses.
====================
If this chap had any credibility left with technically educated people he just lost it.
"It's not desperately edifying, is it?"
Desperate but not edifying.
Engineering computer modelling is based on well known and validated scientific theories and principles where as climate science is based on unvalided guess work.
I disagree with the above.
Mann is not stupid, just evil.
Can anybody point to a single passenger carrying vehicle (land, sea or air) that has gone from a computer model to production without any intervening physical prototypes to test the model?
Mann does himself little credit by referring to Freeman Dyson in this way.
As for weather forecasts for the next seven days based on models, my experience shows they are tolerably accurate for the next 24 hours but for anything more they are no better than guesswork.
Can Mann actually be that stupid?
He does realise that plane builders do stuff called 'testing' as well, doesn't he?
Maybe the word is unfamiliar.
Michael Mann: a dwarf standing in the shadow of a giant.
Surely, one of the main problems with actually doing experiments is that it requires getting out from behind your computer terminal...
Of course you can't experiment on the entire Earth, but there is no reason at all why controlled experiments can't be carried out to provide insights into something like the CO2 induced CAGW theory. Also I think there is still a lot of scope for climate scientists to be out collecting good quality data.
Not particularly constructive but, first reaction is "the Cheeky So and So" (polite version).
;)
Mann's example of packing a brolly - the precautionary principle alarmists love so much would require Freeman Dyson to take one with him, just in case.
Some things really make me mad. The aircraft models Dr Mann refers to are designed by engineers and 'hard science' scientists. The models are subject to stringent test regimes to verify and validate that they meet their design aims and are fit for purpose. As part of the gestation process of a new aircraft every part of the aircraft and its systems will be subject to testing to software and hardware standards designed and agreed by engineers. When everyone is happy with the software and hardware simulation results the aircraft is handed over to yet more engineers for building to formally laid down, internationally agreed standards that are applied to everything from wings to light bulbs. Construction is followed by months of ground tests and then months of flight tests. Not until all this has been completed satisfactorily will anyone who is not an engineer get anywhere near the aircaft.
How dare he imply that there is a comparison between aerosystem models used in the design of aircraft and those that generate straight line predictions of future climate calamity.
Quote, Freeman Dyson, "my objections to the global warming propaganda are not so much over the technical facts, about which I do not know much, but it’s rather against the way those people behave and the kind of intolerance to criticism that a lot of them have."
"As for weather forecasts for the next seven days based on models, my experience shows they are tolerably accurate for the next 24 hours but for anything more they are no better than guesswork."
Slightly unfair as in when you have one of those large blocking high pressures that sit over Denmark in summer that tend to provide consistently good weather for a week or 2 then the forecast tends be consistently good for the next 7 days. Even then though they don't seem very good at predicting exactly when the blocking system will break down. Also using the well known forecast of tomorrow being the same as today probably does pretty well too.
Gareth
I think Mann's idea of the precautionary principle would require you to take a wet suit, thermal underwear, sun-blocker and umbrella for an afternoon in your own back garden.
The people I know who do computer modeling for aerospace, especially those doing the fluid modeling that use the same underlying principles as the climate modeling, think it is ludicrous that today's climate models are regarded as having any predictive value at all by many people. Because their own models have so often been wrong, and their successful models are only so after many iterations from comparisons to test results, they are deeply, deeply skeptical of the claims of climate modelers.
But then, they are not climate scientists, so their opinions don't count...
Bishop Hill's precis isn't desperately edifying no. But (in the SciAm version)Dr Mann asks a very pertinent question "The models are simply at some level a formulation of our conceptual understanding and when someone says they don't like models then I’m wondering what alternative they have in mind."
I like models , but I gather you don't so please tell me what alternative does Bishop Hill have in mind?
Mann also has an article in the BBC "science" magazine Focus entitled "Can we solve climate change with geoengineering?" It is another article in which Mann reveals a complete lack of understanding of real world engineering.
Example: "better yet, burying it [CO2] beofre it even gets there [the air] (carbon capture and sequestration) - seems relatively harmless and worthy of consideration".
He says we will need to cut carbon , it will not be easy and we might have to make difficult lifestyle changes. "We might have to consider the acerbic taste of nuclear energy".
That's the first time I've heard nuclear energy described as "acerbic"!
Hengist
You don't use models in the real world of engineering until you have validated them and done lots of testing with real data. Understand? It is fairly obvious ro anyone who gets out in the real world and does some engineering.
Recently Mann compared climate scientists to research physicians as a means of special pleading for climate scientists. Now he is comparing climate scientists to engineers who create and test airplane designs and, once again, he is making the comparison as a means of special pleading for climate scientists. Where will he go next? Both comparisons are idiotic. Medical research exists to serve the purpose of relieving suffering not the purpose of seeking truth about the universe to satisfy curiosity. Aerospace engineering exists to improve the quality of aircraft not to discover the fundamental truths about flight. If Mann had any commonsense or humility he would argue that climate science exists to improve meteorology. But he cannot make that argument because no one would turn over management of world energy supplies to meteorologists, excepting GreenPeace and friends.
Has Mann (or Hengist) heard of crash testing? I'm sure new dummies are always required...
@Phillip Bratby
That doesn't answer the question posed by Professor Mann does it ? What alternative to climate models would you prefer ?
@James P
Thank you, yes I have heard of crash testing. So that's your answer is it ? You would prefer to crash test Planet Earth rather than check computer models. OK I reckon that sums up quite well the attitude of the Bishop Hill blog .
Nothing is preferable to a model with zero predictive capability.
The fact that Dr Mann seems to know nothing about aircraft test and certification, which is ultimately performed by flying real aircraft in the real atmosphere, tells us much about both his knowledge of engineering ability to do simple fact checking before making statements on-record.
I'll also note that although CFD code shares many structural similarities to GCMs (e.g. finite element models, Navier-Stokes equations, etc.) it is always validated against real data (flight and/or wind tunnel) whenever it's applied to safety critical systems... yet another basic fact that Dr Mann seems to be ignorant of or, more likely, simply chooses to ignore for the sake of a cheap sound-bite.
Hengist -- I don't, and I don't think most commenters here -- object to the principle of climate modeling to try to figure out what is going on with the climate. What many of us object to is using these fundamentally unverified models as justification for turning the world's economy upside down.
To turn Mann's analogy on its head, what Mann and his cohorts are asking us all to do is to get onto an aircraft that has modeled but not tested in any meaningful way for its maiden flight.
Hengist - which bit of the way those people behave and the kind of intolerance to criticism that a lot of them have don't you understand?
And I hope Josh won't get upset if I link to a post of mine on the topic of models, featuring a well-known competitor of his (plus a wonderfully hilarious vintage Fred Pearce on the power of computer models).
@ Phillip Bratby could you point me to a piece of peer reviewed science that supports your implication that climate models have zero predictive capability ?
Hengist - you are, of course, missing the point.
If you software-modeled a new airplane's behavior...then you looked at the results and compared them to the known behavior of aircraft already flying and found serious questions about your results, do you believe the model or do you believe empirical testing?
Further, let's say you model the behavior of the wings and the nose, but you don't really understand how the tailplane works well enough to incorporate it in your model so you leave it out but you don't tell anyone. Would you say your model's results are valid?
In reading Mann's crap, if one didn't (unfortunately) know better one would be inclined to wonder what brand of cereal box he got his Ph.D. out of.
Also what prevented David Biello from seeing the straw man? Isn't there anybody left at SciAm with a knowledge of the real world?
This is one of the reasons I cancelled my subscription to Scientific American. They have dumbed it down to populist pap and not a lot of real science. In some cases they appear to have become an arm of Fenton Communications. Their articles have become predictable.
Hengist, no one is asking for alternatives to climate models... all we're asking for is some way to test/validate their predictive powers by checking them against real-world data (e.g. the existence of an equatorial tropospheric 'hot spot').
I wonder what history will think of the "Brilliant" Mikey Mann. Probably in the same category as Lysenko or perhaps Bernie Madoff.
Hengist -- would you invest in Bernie Madoff if he got out of jail and opened a new investment firm because he has a new investment model?
I think that Mann's views provide a good test of what engineering is about.
That is to say: any true engineer (regardless of formal qualifications) can immediately spot what's wrong with Mann's views. Those who don't, should never try to do anything related to engineering.
Maybe I should use that as a test if I have to interview anyone for an engineering job.
@Hengist you are out of your depth a little here. Not an argument from authority but I was an engineer in the Design Computing Department of a major aircraft factory. I had to support the whole process.
Even after all the models, all the test flights, real planes are fatigue tested for years and years in water tanks and rigs. Extremely boring work, documented to the nth degree. It costs a lot of money. Because despite the all the work with such a machine until first flight, they still cannot have modelled/test proven everything.
Despite having some of the most mature models in IT, despite being historically the biggest investors in IT, despite having the best scientists and engineers on the planet, they know their models are not good enough. And these models are created against KNOWN benchmarks.
Fatigue stress is the equivalent of Climate, just climate is many degrees more complex.
Re: Hengist
How about experimentation?
Manns speciality is extracting temperature signals from tree rings.
Where is his experimentation for this? Is he growing trees in a controlled environment whereby he can analyse the rings in future years?
Does he have a section of forest full of instrumentation to measure the environment? Is he taking regular soil samples and analysing nutrients?
There are any number of experiments that Mann could have designed and run over the past dozen years that might improve the field of dendroclimatology but he has chosen not to.
In fact the only climate based experiment I am aware of is the one at CERN testing for possible links between cosmic rays and clouds.
@Peter B... spot on lad
For Hengist - start here:
http://climateaudit.org/2006/05/15/gerry-browning-numerical-climate-models/
http://itia.ntua.gr/en/docinfo/978/
Ah, but no doubt you are ahead of these guys.
You all may be a little too hard on Prof. Mann. Haven't any of you been to the Smithsonian or the Wright-Patterson Museums and seen the famous Wright brothers flight test computer or the reems of notebooks filled with code (painstakingly written by hand)?
Me neither.
Perhaps Mann should have used space vehicle launches. There, all the billion dollar budgets count for nothing.
The proving is in the actual usage. No other way of testing.
How many mission failures do we have now? Even after 60 years?
Mission failures are a fact of life despite the modelling. Despite the trillions spent.
Peter B - exactly right. There may be many technically skilled people who have been following the climate controversy at a distance, who are reserving judgement on the big picture because the detailed knowledge, calcs and physics are not their personal expertise. However, as you say, anybody with technical understanding reading Mann's ill informed and poorly disguised ad hom will now know he is a fool. Maybe the full transcript will show a kinder interpretation on his words. If not, given the readership Scientific American reaches, I reckon this should signal the end of MM's career at his current level.
Hengist McStone said:
Forgive the appeal to authority but would the word of Kevin Trenberth be enough?
More knowledge, less certainty
The IPCC are only now getting round to examining the possibilities of making predictions out to about 30 years (which will require observing for a sufficient period to validate those predictions*) and continuing with their projections (not predictions) to 2100 and beyond.
* As evidenced by Trenberth writing "Although important progress has been made in this area, the techniques are not yet fully established 5. In part because it takes at least a decade to verify a ten-year forecast, evaluating and optimizing the models 6 will be a time-consuming process."