From Brown University:
Statistical physics offers an approach to studying climate change that could dramatically reduce the time and brute-force computing that current simulation techniques require. The new approach focuses on fundamental forces that drive climate rather than on “following every little swirl” of water or air. And yes, there’s an app for that.
PROVIDENCE, R.I. [Brown University] — Scientists are using ever more complex models running on ever more powerful computers to simulate the earth’s climate. But new research suggests that basic physics could offer a simpler and more meaningful way to model key elements of climate.
The research, published in the journal Physical Review Letters, shows that a technique called direct statistical simulation does a good job of modeling fluid jets, fast-moving flows that form naturally in oceans and in the atmosphere. Brad Marston, professor of physics at Brown University and one of the authors of the paper, says the findings are a key step toward bringing powerful statistical models rooted in basic physics to bear on climate science.
In addition to the Physical Review Letters paper, Marston will report on the work at a meeting of the American Physical Society to be held in Baltimore this later month.
The method of simulation used in climate science now is useful but cumbersome, Marston said. The method, known as direct numerical simulation, amounts to taking a modified weather model and running it through long periods of time. Moment-to-moment weather — rainfall, temperatures, wind speeds at a given moment, and other variables — is averaged over time to arrive at the climate statistics of interest. Because the simulations need to account for every weather event along the way, they are mind-bogglingly complex, take a long time run, and require the world’s most powerful computers.
One practical advantage of the new approach: the ability to model climate conditions from millions of years ago without having to reconstruct the world’s entire weather history.Direct statistical simulation, on the other hand, is a new way of looking at climate. “The approach we’re investigating,” Marston said, “is the idea that one can directly find the statistics without having to do these lengthy time integrations.”
It’s a bit like the approach physicists use to describe the behavior of gases.
“Say you wanted to describe the air in a room,” Marston said. “One way to do it would be to run a giant supercomputer simulation of all the positions of all of the molecules bouncing off of each other. But another way would be to develop statistical mechanics and find that the gas actually obeys simple laws you can write down on a piece of paper: PV=nRT, the gas equation. That’s a much more useful description, and that’s the approach we’re trying to take with the climate.”
Conceptually, the technique focuses attention on fundamental forces driving climate, instead of “following every little swirl,” Marston said. A practical advantage would be the ability to model climate conditions from millions of years ago without having to reconstruct the world’s entire weather history in the process.
The theoretical basis for direct statistical simulation has been around for nearly 50 years. The problem, however, is that the mathematical and computational tools to apply the idea to climate systems aren’t fully developed. That is what Marston and his collaborators have been working on for the last few years, and the results in this new paper show their techniques have good potential.
The paper, which Marston wrote with University of Leeds mathematician Steve Tobias, investigates whether direct statistical simulation is useful in describing the formation and characteristics of fluid jets, narrow bands of fast-moving fluid that move in one direction. Jets form naturally in all kinds of moving fluids, including atmospheres and oceans. On Earth, atmospheric jet streams are major drivers of storm tracks.
For their study, Marston and Tobias simulated the jets that form as a fluid moves on a hypothetical spinning sphere. They modeled the fluid using both the traditional numerical technique and their statistical technique, and then compared the output of the two models. They found that the models generally arrived at similar values for the number of jets that would form and the strength of the airflow, demonstrating that statistical simulation can indeed be used to model jets.
There were limits, however, to what the statistical model could do. The study found that as pace of adding and removing energy to the fluid system increased, the statistical model started to break down. Marston and Tobias are currently working on an expansion of their technique to deal with that problem.
Despite the limitation, Marston is upbeat about the potential for the technique. “We’re very pleased that it works as well as it did here,” he said.
Since completing the study, Marston has integrated the method into a computer program called “GCM” that he has made easily available via Apple’s Mac App Store for other researchers to download. The program allows users to build their own simulations, comparing numerical and statistical models. Marston expects that researchers who are interested in this field will download it and play with the technique on their own, providing new insights along the way. “I’m hoping that citizen-scientists will also explore climate modeling with it as well, and perhaps make a discovery or two,” he said.
There’s much more work to be done on this, Marston stresses, both in solving the energy problem and in scaling the technique to model more realistic climate systems. At this point, the simulations have only been applied to hypothetical atmospheres with one or two layers. The Earth’s atmosphere is a bit more complex than that.
“The research is at a very early stage,” Marston said, “but it’s picking up steam.”
###
Direct Statistical Simulation of Out-of-Equilibrium Jets
S.M. Tobias and J.B. Marston
We present direct statistical simulation of jet formation on a β plane, solving for the statistics of a fluid flow via an expansion in cumulants. Here we compare an expansion truncated at second order (CE2) to statistics accumulated by direct numerical simulations. We show that, for jets near equilibrium, CE2 is capable of reproducing the jet structure (although some differences remain in the second cumulant). However, as the degree of departure from equilibrium is increased (as measured by the zonostrophy parameter), the jets meander more and CE2 becomes less accurate. We discuss a possible remedy by inclusion of higher cumulants.
Phys. Rev. Lett. 110, 104502 (2013) The paper is available for download at Arxiv.org PDF
Dr. Judith Curry also has a discussion of this at Climate Etc.
Physics is exactly what’s required to put climate science on a firm footing. The relevant quote is from NZ’s own Ernest Rutherford: “all science is either physics or stamp collecting”, although at that time (1902) there was no climate science — else he might have added a less complimentary third adjunct. All physicists are fully qualified to evaluate climate models — mind you, the climateers will hotly dispute that because physicists are by and large a very skeptical group!
Sure statistical modeling is a much simpler way to go. It takes a whole lot less CPU cycles that direct simulation, it’s simpler to develop & code, it is easier to debug, it is much less subject to programming errors, etc., etc. However, you can’t model a system you don’t thoroughly understand, either with statistical modeling or direct simulation. Computers don’t think for them selves… they’re nothing but big fast adding machines.
I don’t think anyone currently knows enough about the oceans, atmosphere and climate to make accurate models using either of the above approaches work.
A climate model based on physics?
Words fail me.
About ****ing time.
hoffer. Its been happening for a long time
spend some time watching. best hour you will spend
http://www.newton.ac.uk/programmes/CLP/seminars/2010082310001.html
Replacing horsepower with mind power? This looks promising from the basic premesis.
…Brad Marston, professor of physics at Brown University.
Hmmm, gotta admire this man:
Speaking of Climate Models. If you can’t model ENSO with any precision you cannot predict the future state of our atmosphere. The feeling I get from this article is that we have a great understanding of Climate; it’s only a lack of a decent computer system that holds us back. If it was only that easy. Just when you think you have ENSO figured out, she pulls a fast one.
There are certain things that can be calculated easily using physics principles instead of doing really complicated things. For example, if we have a bag of sand on a frictionless rail and fire a bullet into it, we can find the resultant velocity of the bag if we know the mass of the bullet and its speed before entry as well as the mass of the bag of sand. We do not need to analyze billions of collisions between the bullet and a huge number of sand particles. Of course weather is much more complex! It will be interesting to see if things can be simplified and made more accurate in the process.
Sure. And the physics of carbon dioxide is “well known” and can be calculated into statistical modeling (and we have heard that before). There are many examples of garbage physics in our history. This is no different. I have no positive things to say about this research. It appears to be same old same old on the inside, just new wrapping.
Call me a cynic but the basic problem will still exist – garbage in, garbage out.
Applying physics to climate ‘science’ eh?
A novel idea.
I doubt it will prove very popular amongst people who, for example, did biology to avoid doing physics. Or people who struggle with maths and don’t know how to get Excel to make a graph…
they are mind-bogglingly complex, take a long time run, and require the world’s most powerful computers
42 (sorry, it’s Friday night) :0
This all seems something of a waste of time at least until we actually understand all of the cycles at work in the process — including the external factors like the solar cycles.
physics…..
We have chemical biology….so why not biological physics….
earth’s climate….climate change…..climate statistics……climate systems
They’ve used so many different words…what are they really doing here?
…are they modeling the weather….then they don’t need biology
if they are modeling man made climate change…then they do
GIGO is far less of a problem with physicists because they are trained to watch the signal-to-noise, and to know the difference. It’s an integral part of being a physicist.
davidmhoffer says:
All climate models are based on physics. The point here is that they are applying some techniques used by physicists to deal with the system on a less-detailed level, thus allowing for the possibility of more rapid calculations that still get the statistical properties of the climate right without getting bogged down in the weather “details”..
NZ Willy says:
(1) As a physicist, I don’t think that I would say that “all physicists are fully qualified to evaluate climate models”. We certainly have a good background on which to understand climate modeling and become knowledgeable about it…but there is still a lot to learn to actually become conversant in the subject. Also, in regards to the first part of your statement, I think that the notion of reductionism, i.e., that the only interesting science is physics (or, even more specifically, particle physics, is pretty well dead: Interesting phenomena emerge as the result of the collective macroscopic behavior that occurs as one looks at larger scales.
(2) Physicists may be a “skeptical” bunch in the true sense of the word, but if you mean that most physicists are generally not worried about AGW, I don’t think that is correct. The APS has issued a statement on the issue ( http://www.aps.org/policy/statements/07_1.cfm ), major introductory physics textbooks (such as the ones that we use for both algebra- and calculus-based courses at RIT) talk briefly about it, …
By the way, I should mention that I know Brad Marston, having overlapped with him when I was a grad student and he was a postdoc…and I have the utmost respect for him. He has also played a major role in the formation of the new Topical Group on the Physics of Climate within the American Physical Society.
I hope there is something useful here .
Time will tell, at least the man is open source and challenging others to come play with this idea.
I keep hoping a real science will arise from the ruin that is climatology, but will remain cynical until the ology turns to science.
Mosher, Shore,
I did not know that.
Do I need to add sarc tags?
GIGOQ&E
GIGO quicker & easier.
“The new approach focuses on fundamental forces that drive climate…”
===================
Don’t be coy, just tell us exactly what they are.
I am puzzled as to why so much discussion goes on about the effect of CO2 on radiated heat energy from the Earth’s surface and middle atmosphere. Reference is also made to the Glass house effect.
In a glass house, the air that has been warmed by incoming radiation from the sun convects upward, this warm air is trapped mechanically by the glass roof.
The glass structure prevents winds from sweeping the heated air out of the building, replacing it with cool air.
The glass isn’t there to prevent heat radiating out of the building, but is there to keep the air in the building trapped. Any heat loss from the glass house is by conduction through the glass and some radiation.
To my knowledge a large proportion of the heat transferred to the upper atmosphere is by convection.
If you live in the tropical regions you can observe the clear skies in the morning and watch the huge clouds forming during the day followed by massive thunder storms and heavy rain late in the day.
CO2 and its overall effect on heat transfer to space is minimal and I think can be almost ignored
Water vapour is transferred to high altitude by convection, where it condenses loosing its latent heat of which a good percentage gets radiated out into space. Water in this case is a natural refrigerant, and in my opinion forms an important component of the Earth’s temperature control system. I think CO2 plays a very small role in the Earth’s temperature control.
I know that some climate scientist rely on “models”. The programs in the computers can only reflect the opinion of the programmer/s, as computers cannot think, but merely perform very rapid calculations using algorithms programmed into them by people with opinions. I feel that very often these opinions are wrong as born out by the divergence of model predicted and observed temperatures.
At a minimum, this could be useful as a double-check.
No doubt this ‘model’ is logically equivalent to the standard Climate Schmience algorithm:
def run_climate_model(required_answer=”It’s worse than we thought”):
do_complicated_looking_stuff()
return required_answer
Theoretically, it’s a good idea. Practically, a lot of work is required. The approach is widely used for finite element modeling for solid mechanics and circuit simulation for electrical engineering, and no doubt many other areas that I have no knowledge of. However, it’s successful in those areas because the physics of the simulated bodies is well understood and predictable. From what I have seen, climate physics is very much not in that club. IMHO, statistical simulation is not likely to make much progress until the underlying physics of the climate is a whole lot better understood than it is now. And, that’s just not a simulation problem.
Strictly as a layman on this, but how well does statistical physics deal with boundary conditions and external inputs? That is things like transition form ocean to land, mountain ranges, or changes in solar output?
I am really surprised at all the negative comments, I think this approach has merit.
Yes the climate models are based on physics, but they have two problems IMO. The first is that they are trying to model the physics at a level of detail that is impossible to model. There isn’t enough compute horsepower on earth to model every molecule of the earth surface, oceans and atmosphere, and there never will be. So the models fill in the gaps with statistics and trends which is their second problem. That’s just a fancy name for curve fitting and curve fitting fails over and over when there is an underlying flaw. That’s exactly the problem we see with the models today. No matter how well they match known observations, they fail to match new observations. Which is why they were “right” in the early days but now have been wrong for…depending on the data set… as much as 23 years. As time goes on they will get more off track as the fundamental errors overtake the curve fitting.
This model takes a different approach. It is in part curve fitting too. But instead of being mostly curve fitting on top of physics full of gaps, it is mostly known laws of physics with a manageable amount of curve fitting. Like Werner said, one doesn’t have to know the interaction between a bullet and every grain of sand in the bag (traditional climate models) to figure out how fast the bag will be going after it absorbs the bullet (what these guys are doing).
The fact that they modeled something and got the same result as the climate models is very interesting. When you get the same answer two completely different ways, you are probably on the right track. They still need to verify against observations (and I wish they had done that) but the fact is they got the same answer, and here is the important part of that:
They did it with a fraction of the horsepower of the climate models, in fact, they did it with an app you can download and run yourself. Let’s consider the merits of that should their technique be proven out against observations. What else could they model with the same approach? I’m betting a lot of things. And the more things they can model with this technique, the more multiple models can be tied together to become a larger model. Sure, that’s what the traditional models do too, but you could do what they are doing with a couple of hundred desk tops instead of thousands upon thousands of compute cores. Isn’t that worth exploring? They’ve made their code public for gosh sakes, let’s wait until a few other people have had a look and let’s see what comes of it. Either it will be debunked or it will be proven of value, but I think there’s enough merit to give it a shot rather than just throwing it under the bus as yet another failed computer model.
Would like to see rgbatduke weigh in on this…
Regardless of whether the method has merit or not, applications in climate & solar science will have only one physical driver, only one statistical driver, and more generally one and only one driver: Politics.
Cool, in the future we will be able to get the incorrect answer even faster than today.
This is an interesting approach, but why bother with the “higher cumulant” (yet) instead of perturbing the simulation thermodynamically i.e add incident energy on an incline hemisphere (simulate energy flux on a tilted spinning “Earth”). Could you just use the Gibbs free energy differentials based on the ideal (or “real” even) gas equation or statistically…then add complexity incrementally like two fluids (water and air) then phase changes and see if ice forms at the poles for instance.
my two cents.
John Kaye (March 8, 2013 at 8:18 pm) wrote:
=
def run_climate_model(required_answer=”It’s worse than we thought”):
do_complicated_looking_stuff()
return required_answer
=
You sir have been awarded academic life on funding easy street.
All science is simplification. John of Okham, etc.
I worked on big complex software for 30 years, and it’s my observation that people who work on such systems are chronically addicted to complexity. The proposal that simple is better is an anathema to them. I am certain climate modelers have fallen into this trap.
This idea fascinates me. We have a climate modeling tool released into the wild where we can all work with it. It may not be much, after all it’s just an app and I didn’t see any mention of the source code, but it could really be the start of something. The approach could be very fruitful. I don’t know if my programing skills or knowledge of fluid flow physics would be enough to do very much, but surely some of us have enough of both to do a lot better. The possibility of developing some true open source models of how the Earth works is just tantalizing.
A few years back I read some Isaac Asimov articles on how ice ages worked and how the configuration of the continents might determine when they could happen. I tried to do my own crude little simulation to check out some of the ideas. I took a big chicken cooker, put a little water in it and set it on a small burner. Then I set some aluminum scraps in it to simulate mountains and such, sprinkled in a little sawdust, and turned on the heat. Sometimes the sawdust would move around in patterns that looked a lot like weather maps. I couldn’t make it work well enough to see what I was really after, but I still had the feeling it was showing me some important things about how weather worked.
This might be a whole lot better. I’ll have to see how much hardware it needs. I’m more of a penguin breeder than apple maggot or window shade, but I might have a mac in my cyber junk pike good enough to run it. then it’ll be time to dig out my old physics books and see what they tell me. After all it seems likely that I already know more physics and stats then the average “climate scientist” 😉 😉 😉
This still does not get over the reality that climate is more complex that just atmospheric physics. We will not get any sense from climate studies until for a start the true geological outputs of CO2 gases and the biological creation and usage of these gases are fully understood.
Why is is just accepted that man’s emissions are compared to the net of natural used and created gases i.e. the residual atmospheric CO2 rather than nature’s emissions. Where else outside climate studies is the comparison of net and gross quantities considered the norm?
Let’s be honest the so called science has always been just tenth rate statistical analysis or we could have accurate forecasts for any place or period in the future up to the claimed climate forecast date.
What climate statistics generated by such statistical models are useful?
Global average temperature isn’t experienced by anyone, so that’s out… Average number of hurricanes per year? Well, maybe the ones that actually hit land… Average rainfall per year? Well, that could be a single day flood, or a nice evenly wet year…
I mean, the real problem here is that what *matters* is the weather at a time/space resolution that is very small – I want to know what the weather is going to be for a given 100 square miles on a given day if I’m going to make some sort of grand plan. I’m not sure if this proposed exercise is going to give anyone any sort of actionable intelligence.
I guess as a first pass at useful info, they should build a model that accurately predicts just one statistical portion of climate – ENSO. Predict ENSO out say, 10 years the way you can do with tide charts, and maybe they’re onto something.
This approach will not work in any useful way. The time evolution of a chaotic system is deterministic but unpredictable. Apart from being able to put some upper and lower limits on average global average temperature based on energy flux from the sun, Boltzman and heat produced in the earth, which can be done anyway, there is not a snowball’s chance in hell of predicting a regional climate change natural or otherwise with any useful degree of confidence.
Try it with the 3-body problem. Apart from the constraints of total Gravitation potential energy and kinetic energy being constant and limiting how far the bodies can move apart there is no chance of predicting a behaviour after a specific time.
Only an APPLE app. Need an ANDROID version.
Replace Mann with an app. , sounds good to me !
davidmhoffer says:
March 8, 2013 at 8:47 pm
“I am really surprised at all the negative comments, I think this approach has merit.”
I would be surprised by anyone who claims that this approach is in any way better or of a higher quality than the existing broken models.
“The fact that they modeled something and got the same result as the climate models is very interesting.”
It is not surprising at all. They arrive at the same bogus conclusion faster. So in the past we saw these predictions for the year 2100. Now as computers become cheaper and faster they let their grad students write papers that predict the climate in the year 3000 (they have no other use for the CPU horsepower).
In the future we will see, with this great new technique of outputting nonsense faster, simulations to the year 100,000 C.E. (or whatever atheists use instead of A.D.).
And I forecast that these papers will tell you that you should be worried. Very worried.
DirkH;
I would be surprised by anyone who claims that this approach is in any way better or of a higher quality than the existing broken models.
>>>>>>>>>>>>>>>>>>
Your logic is that existing models don’t work, so it is impossible to build models that do. Why bother to even try.
Mandelbrot already looked into this. He said it could not be done and spawned chaos theory in trying to explain why.
davidmhoffer says:
March 9, 2013 at 8:04 am
“Your logic is that existing models don’t work, so it is impossible to build models that do. Why bother to even try.”
As long as the prevailing theory is that the Earth’s climate is NOT controlled by solar cycles but is a freely oscillating chaotic system it does make indeed not one wit of sense to try to model it a hundred years into the future and expect any predictive skill.
See also the definition of insanity according to I think Einstein.
DirkH, Dinostratus, son of mulder: You guys are confused about what chaos theory says. It does not say that any sort of prediction is impossible. It says that predicting details that are sensitive to the initial conditions is impossible.
Take Brad Marston’s example of the ideal gas equation: The motions of the individual molecules are subject to chaos theory (besides being simply way too numerous to follow!) and yet we can still make conclusions about the macroscopic behavior of the gas.
Likewise, it is impossible to predict the weather several weeks in advance and hard to predict the climate relative to the mean (e.g., whether this summer will be warmer or cooler than average)…but it is relatively easy to predict that the temperature here in Rochester will be some 20 C warmer come July than it was in January.
DirkH
As long as the prevailing theory is that the Earth’s climate is NOT controlled by solar cycles but is a freely oscillating chaotic system
>>>>>>>>>>>>
That’s not what they are doing.
So long as it is not necessary to spend $$$ as a result of its predictions without thoroughly verifying its output, then it sounds reasonable.
“joeldshore says:
March 9, 2013 at 5:46 pm
You guys are confused about what chaos theory says. It does not say that any sort of prediction is impossible. It says that predicting details that are sensitive to the initial conditions is impossible.”
I think you’ll be able to read in the example of the 3-body problem that I appreciate the point you are making but no amount of statistical methods will predict how the jetstream will react say over Europe and hence whether in 10 years or 20 years or 50 years or 100 years time Europe will be wetter or drier than now or warmer or colder than now, or windier or less windy than now ie something useful that we don’t know.
Impress me with a type of prediction that will be useful from this method.
joeldshore says:
March 9, 2013 at 5:46 pm
“DirkH, Dinostratus, son of mulder: You guys are confused about what chaos theory says. It does not say that any sort of prediction is impossible. It says that predicting details that are sensitive to the initial conditions is impossible.”
So you run your model 20 times, average it, call it a multimodel ensemble mean and print it in the IPCC report as if that would improve anything. Excuse me while I laugh; tell me again how big the state space is. What you are doing is pseudoscience, plain and simple.
“Take Brad Marston’s example of the ideal gas equation: The motions of the individual molecules are subject to chaos theory (besides being simply way too numerous to follow!) and yet we can still make conclusions about the macroscopic behavior of the gas.”
I have yet to see any such result from climate science. Meanwhile climate science has the audacity, the chuzpe to tell us how to rebuild our energy sector. Climate scientists should be held liable for the damage it has already caused.
“Likewise, it is impossible to predict the weather several weeks in advance and hard to predict the climate relative to the mean (e.g., whether this summer will be warmer or cooler than average)…but it is relatively easy to predict that the temperature here in Rochester will be some 20 C warmer come July than it was in January.”
Oh noes! The Horrors! I hear that any increase above 2 deg C is violating Schellnhuber’s planetary guidelines! Thermal runaway awaits us all!
Joel Shore, how is the global average temperature time series that a GCM computes (assuming for the moment that average had any statistical meaning or were defined) not affected by the chaos in the system? What separates the low frequency part of the power spectrum (where we find the “climate” signal ) from the high frequency part (where we find unpredictable and chaotic weather)?
How much dampening of high frequency fluctuation between model runs do you achieve in decibel when you average 32 model runs (assuming for the moment that the temperatures were normally distributed which they aren’t)?
‘“Say you wanted to describe the air in a room,” Marston said. “One way to do it would be to run a giant supercomputer simulation of all the positions of all of the molecules bouncing off of each other. But another way would be to develop statistical mechanics and find that the gas actually obeys simple laws you can write down on a piece of paper: PV=nRT, the gas equation. That’s a much more useful description, and that’s the approach we’re trying to take with the climate.”’
This so-called “direct statistical simulation” is what I have been asking climate scientists to undertake for years now. I fervently hope that these scientists succeed. If they do, they will establish beyond the shadow of a doubt that our understanding of climate is inadequate for prediction or even for reasonable forecasts of climate. That is because climate scientists do not have reasonably well confirmed physical hypotheses, something analogous to the gas equation, for most forcings or feedbacks. We simply have not done the empirical research necessary to learn what water vapor does or what clouds do under conditions of rising CO2 concentrations. We need research on water vapor and cloud formation employing technologies along the lines of the Argo project but far more extensive than Argo.
“Conceptually, the technique focuses attention on fundamental forces driving climate, instead of “following every little swirl,” Marston said. A practical advantage would be the ability to model climate conditions from millions of years ago without having to reconstruct the world’s entire weather history in the process.”
Finally, there would be physics in the models. At present, apologists for CAGW tell us that the models are based on the best physics but fail to tell us that the physics does not actually appear in the models as rigorously formulated hypotheses along the lines of the gas equation. One great advantage of using such hypotheses in the models is that on occasion such hypotheses can be falsified. Introducing falsification into models runs would be a step toward bringing them into the arena of science.
“The theoretical basis for direct statistical simulation has been around for nearly 50 years. The problem, however, is that the mathematical and computational tools to apply the idea to climate systems aren’t fully developed. That is what Marston and his collaborators have been working on for the last few years, and the results in this new paper show their techniques have good potential.”
I am encouraged that Marston is sophisticated enough to address the mathematics and the computational problems. To talk with climate scientists, you could come away believing that the model is transparent to the scientist. That is far from the truth. Any model worth serious attention is based on heuristics that are not scientific principles and are not akin to them. A serious discussion of computer heuristics in climate models is long overdue.
DirkH says:
No…What climate scientists are doing is doing science. And, then it is up to us as a society to decide whether we want our public policy to be based on the best available science or whether we want it instead to be based on the opinions of people with ideological blinders on.
What chaos theory tells you is that the exact trajectory that you follow is very sensitive to initial conditions. So, for example, if you run a numerical weather prediction model for 4 weeks and see what weather pattern you get, it will bear little resemblance to the actual weather pattern that occurs that day. (And, in fact, it will bear little resemblance to the weather pattern predicted by the very same model with slightly perturbed initial conditions.) However, the weather pattern it predicts will still be a reasonable one.
Likewise, when a GCM is run over 100 years of rising greenhouse gases with different sets of initial conditions, the various ups-and-downs of the global average temperature look different in the different simulations but they all predict roughly the same climate 100 years hence.
It is really not that complicated: If I do many different trials of flipping a fair coin 1 million times, then the specific pattern of heads and tails that I get will be different in the different trials, but the statistical behavior of the patterns will be basically the same. And, if I repeat the experiment with a coin that is now biased so that it lands heads more than tails, again the specific pattern of heads and tails that I get will be different in different trials, but the changes in the statistical behavior of the patterns from what I saw with the fair coin will be similar.
“joeldshore says:
March 10, 2013 at 4:40 pm
………What chaos theory tells you is that the exact trajectory that you follow is very sensitive to initial conditions.”
It tells more than that, every iteration in a computer model contains numerical roundings so with chaotic systems at every point in say a 100 year iteration, the model’s “system state” does not reflect the initial conditions. By the time you’ve done 100 years it will reflect nothing like the initial state that was set. There is no way of knowing if the “initial state” the model reflects at any point is realistic let alone after 100 years. Hence, although the result may look “reasonble”, I contend you can draw no conclusion that will be useful.
son of mulder: You may contend that, but you are wrong. We are not trying to predict the weather on some particular day 100 years from now (or even one particular year in the sense of whether one particular year will be a relatively warm El Nino year or a relatively cool La Nina year). We are trying to predict how the climate changes in response to changing greenhouse gases. The fact that the exact trajectory won’t be correct is irrelevant just like the exact trajectory is irrelevant to the question of how the climate differs between winter and summer here in Rochester.