[podcast]https://www.wmbriggs.com/audio/wmbriggs_com_14oct2009_0004.mp3[/podcast]
On today’s episode:
Science and skeptical bloggers A small article in last week’s Science magazine frets that skeptical bloggers are teasing climate scientists over their failed predictions. Bloggers are pointing out that actual temperatures have not been friendly to climatologists, and have failed to rise as predicted. Some climate scientists respond by effectively saying: have no fear, warming is on its way—and this time we mean it!
Climate forecast failures? Since 1999, most climate model predictions have been too high by a factor of about 3. Plus, actual temperatures have been decreasing, or at least not increasing. Yet the belief in the accuracy of future forecasts—all of which predict yet more warming—have not abated. Why is this? Why, that is, do scientists believe the opposite of the evidence and is it rational to do so?
Polka! Grab a beer and listen to Yosh and Stan Schmenge sing “Cabbage Rolls and Coffee”. Yum; or, rather, Mmm, Mmm, good. This is the rare live version! My first meal out was cabbage rolls—at Sanders (pronounced saw-n-ders) on Michigan Avenue in Dearborn (Sanders disappeared for a while, but are back, incidentally).
Evidence, Faith, and Belief To be useful, all models—climate, physics, statistics, whatever—must explain, or fit, previously observed data. Fitting that old data is always the first goal of the model-building process, but it is, or should be, far from the last. Climate models do fit, in a statistical sense, old temperature data.
But that old temperature data is sparse before about thirty years ago, and from heterogeneous sources before that, and some of it is even guessed at. There should be, therefore, but is not, tremendous uncertainty that the climate models have reproduced that old data faithfully. No climate model actually predicts past temperatures exactly; they only do so statistically, by simulating climates that “look like” the old data.
Explaining old data is a necessary but not sufficient condition for a model to be valid. To meet that standard, they must also predict new data accurately. So far, climate models have failed in this. Yet the powerful belief that is induced by a model happening to fit old data is almost overwhelming. The model fits, its owners say to themselves, so therefore it is valid. It also helps to know that nearly all statistical procedures—like the kind that are used to verify climate models’ performance—are designed to give measures of how well models fit old data. Overconfidence is an all too common result.
Climate models are built with the assumption that carbon dioxide is important, but only when it operates with a positive feedback mechanism. The truth of this is asserted and the models are built and tweaked, twisted, and tuned so that they fit old data well. Forecasts are then made, which invariably pronounce warming is on its way. And this forecasted warming is—incorrectly!—taken as evidence that the carbon-positive-feedback process is true. Again, this overconfidence stems from enjoying too much the co-incidence of the models fitting (statistically) the old data.
What needs to be done is this: a fully-funded, fully-dedicated team, skeptical of the carbon-positive-feedback process builds, tweaks and tunes, a climate model that does not contain this process. Forecasts from this model are made, and then compared with actual new data and with the forecasts from the standard climate models. (I realize this description is too brief, and even partly unfair—listen to the podcast for more.)
Glenn Beck-John Coleman call global warming a scam Coleman believes that climate scientists are forced to conform to the consensus and are engaged in a “scam.” I do not agree; “scam” is far too strong a word. It might apply to some, but it does not apply to the bulk of climate scientists. I know many of these people and I can assure you that they are not actively engaged in trying to fool the world into believing something that they themselves do not also believe. Using such an inflammatory word makes your enemies shut up their ears and rightly dismiss you as cranky, if not worse.
Nothing makes a scientist’s career faster than proving something that others either did not know or believed to be false. If a young scientist could prove that the carbon-positive-feedback process is false, he would be sitting pretty. But understand: climate models are enormous undertakings, dozens to hundreds of people working on them, building them through multiple years of effort. Vast sums are efforts are involved and no one person can play more than a small part in the process. Therefore, it is almost impossible for one person to go his own way; further, because of the sheer complexity, it is likely that everybody involved will tend to believe the same things. This is why there is a rough consensus of climate model workers.
And this explains why independent people, like your author, are the skeptics.
Polka again I found the outro song on YouTube, but I am unable to rediscover its source.
Matt:
Nicely done. I am going to link it to Andrew Revkin’s NYT thread that is discussing whether Steve McIntyre should publish on Briffa et al before he says anything – NYT link. There is nothing new, except for the fairly consistent differences in style of argument between the pro-CAGW commentatators and the skeptics.
Your last song is about as melancholy a lament as I can remember – alas, it is way to early for me to reach for a bottle of Bass. It is an excellent reason for endeavoring to live as long as possible.
Didn’t you have a comment on favorite beers sometime ago? Perhaps October is a good time for a reprise.
Bernie!
Too early? Define this curious “early” concept. I’m reminded of the Rumpole story in which he advises his friend George not to marry: “Don’t do it, George! While unmarried men are reaching for their glass of breakfast Chablis and have a slow read of the Times obits…” Now that is civilized. I can recommend Spaaten’s Oktoberfest, a dark lager. That’s what I was drinking when I heard the opening song.
Briggs, you say that
“Nothing makes a scientist’s career faster than proving something that others either did not know or believed to be false. If a young scientist could prove that the carbon-positive-feedback process is false, he would be sitting pretty.”
Not quite. The innovator in science is a lot like the innovator in business–they most often end up “broke.” It is very difficult to buck trends by one’s self. Perhaps, with luck and a lot of talent, the innovator will get credit for having been right, but this will occur only after a large group of others (not necessarily a majority) come to agree, and they are just as likely to give themselves credit first.
Matt:
Drinks with Mortimer/Rumpole would be an experience! I will look for Spaaten and take tasting notes.
Kevin:
I tend to agree and that is probably one of the reasons why a theoretically competitive model has not emerged – though I would bet that some are being worked on.
Matt:
If there was a competitve model do you have a candidate for the forcing that CO2 provides currently?
I suppose I ought to offer an example or two. John Tukey is often credited with having invented the FFT. Yet, Vern Herbert of Chevron, Calgary was using it much earlier to transform siesmic data–in the 1950s. This is not to say that Tukey didn’t perhaps invent it much earlier–Tukey was a brilliant guy, after all, but Herbert’s use seems to me to predate the crediting to Tukey by many years.
L.W. Morley was a Canadian science administrator who, in his spare time, had worked out the large details of plate tectonics, and tried to get his idea published in 1961-1962. Various editors returned his manuscript with comments like “this is too speculative.” In 1963 Matthews and Vine put forth the theory, on essentially the same data and arguments, and are granted the credit. While most earth scientists readily recall Fred Matthews name and work, I doubt anyone knows of L.W. Morley.
The granting of credit in science is quite tenuous. One is much more likely to gain credit if there is a well prepared audience available already. Credit may go to an early adopter but not always to the innovator.
Kevin,
Sounds like yet another instance of Stigler’s Law of Eponymy: “No scientific discovery is named after its original discoverer.” Stigler was a statistician, incidentally.
Bernie,
There’s plenty of forcing mechanisms. Concentrate more on water. In fact, there’s many ways to look at it.
Briggs:
I agree – I guess I was more taken by how most of the publicized models seem to have settled on CO2 as THE forcing that we should be paying attention to.
Guys,
I’m a recently retired physician and what I’ve read here is probably the most reasoned approach to the global warming hypothesis (of which I’m quite skeptical for a variety of reasons). Kevin, you are absolutely right when you say that the scientist who disagrees with the prevailing wisdom is almost routinely discounted (I’ve had the experience in my medical career – I was roundly ridiculed as “dangerous”, a “cowboy”, “experimenting on patients”, etc, etc but I turned out to be right and the medical world now routinely practices my techniques but I suffered a decade of slings and arrows). Woe to the well-meaning investigator who has the temerity to challenge conventional wisdom……..
Steve
As I understand it, these models have many adjustable parameters. If you have enough adjustments, you can fit a curve to anything.
As an example of this, many years ago I measured radio propagation data at various locations. The data points were all over the graph paper. However, I had a least square curve fitting program that allowed me to weight each individual data point. With that I was able to produce a beautiful smooth curve that looked like the theory said it should look. luckily, when I submitted my report, nobody asked any questions about that nice curve.
One of the worst deficiencies of these models is that the grid is quite coarse, and in order to model processes at scales too small for the grid to model with differential equations, the model uses inputs at each time step from other models with adjustable parameters. The topic goes under the heading “parameterization” of subgrid processes. Subgrid processes include evaporation, thunderstorms, precipitation, radiation, etc. The subgrid models are not likely to represent processes on the Earth very well, and these then affect the global model in a manner that produces unrealistic results.
One indication of this being so is that the global climate models do not exhibit the correct statistics with regard to persistence of atmospheric state. The global climate models have a tendency toward too much persistence of state in the short term (two years or less) and not enough persistence in the long term. Long term persistence can look just like a secular trend. Edward Lorenz, in fact, found models of atmopheric state driven by abrupt changes in temperature that wandered for centuries before settling to a steady value.
This seems like a good way to produce “hockey sticks” of all sorts using the computerized global models. Take a model with insufficient long-term persistence, so that the long-term output looks (statistically) quite flat, then drive it with a secular influence like increasing CO2.
As far as competing models go, there is a fellow who posts quite often on Anthony Watt’s blog by the name of Bob Tisdale. Tisdale suggests that El Nino causes the observed warming, but I haven’t found his explanations very clear as yet. I need to read his posts a bit more closely, so that I can learn exactly what he proposes, and it is possible his model yet needs some work; but if one combines a forcing such as a large El Nino (1998 for instance), with a cascade of feedback mechanisms, then one could end up with the situation that Lorenz described. A mean temperature that wanders up and down for decades or even centuries before settling back to equilibrium, and one that people confuse with “climate change”. So perhaps Tisdale is on a fruitful path.
Agree completely re. the model verification.
Can I get a shameless plug in here, as I addressed this point a litle while ago:
“…forecasts are important and this is why. Testing the ability of any model to predict future outcomes is the only real test of its efficacy. Every time someone tells you how well climate models track past temperature trends or variations as “proof” of the robustness of the hypothesis, call bullshit on them. For a surprisingly large number of reasons this is simply not true and any trained statistician or econometrician will verify that fact. ”
http://geckkosworld.blogspot.com/2009/06/question-about-dangerous-climate-change.html
“Well I wish there could be podcasts every day!â€
Are you quoting somebody, Joy?
Geckko, thanks!
Ray, absolutely. Interesting project, incidentally.
Steve Parker, thank you.
Sorry, no, I was singing aloud, but it would be good. A radio show perhaps, but they’d have to vet your music!
There are 2 ways to be wrong .
.
Imagine William that you are part of the large thriving scientific comunity constructing epicycle models and explaining Universe .
Like everybody you know that the models reproduce well the past but have an infortunate trend to randomly fail in forcasts .
And now you dedicated yourself to a special problem which will make you rich and famous .
You have discovered why the Neptune orbit doesn’t follow the model forcasts . WOW !
Indeed despite the huge complexity of the models and of the underlying reality you have discovered with sophisticated statistical methods that the problem is with the circle N° 137 .
Indeed if you remove it and replace it by 2 new circles whose diameters you carefully computed , not only will you reproduce the past but also explain why the apparently wrong but unfortunately observed Neptune orbit was just a fluctuation and Neptune will soon rejoin its previously forcasted orbit because everybody knows that Neptune is on the way to leave the Solar system .
There was an error in the model , you fixed it , the models were right and everybody was right to believe that Neptune was leaving us .
.
Now imagine that you are not part of that comunity and come on the crazy idea that there is a gravitational field that makes the bodies move . Fortunately your idea reproduces rather well the past . But unfortunately it predicts that Neptune will stay with us for billions of years .
Your idea suggests that there is no special problem with the circle N° 137 , there is a problem with the whole circularity paradigm .
So you will be strongly attacked by the scientific consensus because somebody who opposes the geometry of circles established during more than 2000 years must be a crackpot .
But you will also be attacked as an ennemy of mankind by the politicians who have no clue about geometry of circles or gravitational fields and don’t care anyway .
Indeed it has been proven beyond any reasonable doubt that WHEN Neptune leaves us , the frequency of earthquakes will be multiplied by 10 .
Following the recommendations of the IPEC (International Panel for the Earthquake Change) all countries engaged in a multi trillion € program of rebuilding all towns on the planet .
The voters didn’t like much the idea of Earthquake taxes but the famous movie “Horror in San Francisco” and the influential and adequately violent SA (Squads Anti-earthquake) finished by persuading everybody about the errors of their ways .
It is surely not some isolated crackpot financed by Wall Street lobbies who will be allowed to put the whole planet in danger and the careers of the politicians at risk .
.
I contend that the error with the AGW models is of the second kind .
Indeed the most fundamental assumption that all models without exception have in common is that the Earth system is in equilibrium and can be adequately described by deterministic equilibrium physics with (small) perturbations .
This assumption is equivalent to the circularity paradigm of our honorable epicycle scientific comunity above .
It is also specially strange because everybody agrees that the system is NOT in deterministic equilibrium at short time scales (days-months-years) and on the large scales (thousands and hundred thousands of years) .
Yet there is a magical window in the time scales exactly equal to 30 years where the system goes in equilibrium and can be deterministically predicted .
Amusingly it can be noted that the system still stays non predictable at short and long time scales . You must not use the climate “models” at short or long time scales because they would horribly fail .
So what would be the right paradigm ?
Well physics can also deal with systems out of equilbrium through non linear dynamics whose subset is the chaos theory .
However there is a price to pay – the evolution becomes unpredictable at least not in a deterministic way .
It is a red herring to say that this would contradict “established physics” because it doesn’t . The QM laws , fluid mechanics and thermodynamics still stay perfectly valid . GHG still have the right properties .
What changes dramatically is that what matters is how strongly coupled to the evolution of the system are the different dynamical parameters .
There are some scientists who take the fact that the Earth is not an equilibrium system seriously – f.ex Tsonis .
What they say is that the dynamics of the system is given by coupled interacting subsystems (PDO , AMO , el Nino etc) . The result is deterministic chaos and quasi periodical oscillations at all time scales .
The CO2 is still doing its thing but it’s no more the explaining variable .
I think this is the right way to go .
Tom,
Excellent summary, thanks very much. I think, too, that climatologists have been persuaded of their beliefs by the obvious success in meteorological models, a success which they believe has rubbed off on to the climate models, which are, of course, very different creatures. But not so different that they belong to another genre.
There is some evidence that climate models (not GCMS) have very modest skill out to three months, and negligible skill out to one year. But there is no evidence anywhere that they have skill beyond this. And no GCM has yet, to my knowledge, demonstrated skill.
Joy, you don’t like Polka? Strange.
I presume Stigler’s Law of Eponymy was first discovered by somebody else.
George,
That’s what everybody says.
Briggs,
I did listen twice to the recent stats podcast just so as you know I’m still with the programme.
Pokher? It’s cool if it’s heard from a skilift but not up close.
Not so keen if it comes with a roudy bunch of Austrians with elbows and big feet. They like to stamp to that sort of music. (come to think of it so do the English lads.) Sounds like the place you speak of was quite pretty. Any bear stories? For another time maybe.
I’ve worked in the defense industry for years, including think tanks such as the Rand corp. What you find is that the people working in these fields are attracted to it because they happen to believe it’s necessary = because they have a slightly paranoid world view. Which is one reason why you see the military intelligence agencies are usually more hawkish in their projections than the CIA, which goes out of it’s way to hire liberal analyists.
It’s been my experience that the climate modelers are the children of the green culture that have co-opted the old anti-malthusian “yankee know how” culture and have the same inbuilt green biases. I think it’s also fair, given that any skeptic is branded a sellout if they’ve ever picked up a check from an oil company, to point out that if tomorrow a young scientist proved that positive feedback was not the case that the entire AGW industry, upon which billions of governent dollars are invested yearsly, would collapse. For that reason alone I think that said young scientist would be attacked and dismissed.
“Science” has undergone some ugly changes in the last several decades – mainly associated with it’s being funded on the whims of pols and operating in a culture that doesn’t value truth nearly as much as it once did.
Perhaps, you should correspond with Professor John Brignell , a gentleman from across the sea. His country is, at present in a more advanced state of decline than ours at the moment and he might provide some inside to you.
http://www.numberwatch.co.uk/number%20watch.htm
Tom Vonk, thank you very much for that illuminating (and amusing!) insight into the climate models.
Regarding “If a young scientist could prove that the carbon-positive-feedback process is false, he would be sitting pretty” and the near impossibility of this happening, it it still in the realm of possibility. Maybe there is another way to prove the global warmists wrong.
ACORN is a large, well funded organization that was built by a lot of people over a long time. Hannah Giles and James O’Keefe, a couple of inexperienced youngsers, found a fatal flaw in the organization, exploited this flaw and blew a large hole in ACORN without much experience, training or funding.
I have some hope that someone, not necessarily an inexperienced youngster, can find a fatal flaw in the AGW hoax that can be expoited to change the public’s opinion which would then influence our congress.
One of the things missing from this debate is that climatology is not an experimental science. How do you prove or disprove any of it? I’ve also heard it critized as not being a ‘
big Boy’ science. This refers to the fact that in climate science you can gather data that you don’t publish, manipulate it in ways that you don’t publish, then publish the results and not only aren’t laughed at, but get additional grant money. Try that in Physics or Chemistry. In the world of real science, you don’t need a Freedom of Information Act request to get the data.
Here’s an interesting quote from the late Michael Crichton,
“This fascination with computer models is something I understand very well. Richard Feynmann called it a disease. I fear he is right. Because only if you spend a lot of time looking at a computer screen can you arrive at the complex point where the global warming debate now stands.
Nobody believes a weather prediction twelve hours ahead. Now we’re asked to believe a prediction that goes out 100 years into the future? And make financial investments based on that prediction? Has everybody lost their minds? “