Question on whether climate is chaotic or not

[Astronuc]

Problem is gavin from RC, and in fact most climate modellers don't want to accept that the climate is a chaotic system because it would falsify the predictive potential of their climate models.

So they play semantic games by sidestepping the elephant in the room.

I think folks need to be careful and refrain from making statements about others' comments or others' motivations.

It's one thing to cite published statements, but it becomes hearsay to make an unsubstantiated claim about what another party said or wrote privately.


Certainly the public has a poor understanding of chaos and predictability. The climate models might do a good job in predicting trends - within uncertainty.

Obviously, if some data have been excluded such that the measurements are 'made' to agree with models, i.e., the data are intentionally and artificially biased, then there is a big problem with the data integrity and models.

 

[Coldcall]

I think folks need to be careful and refrain from making statements about others' comments or others' motivations.

It's one thing to cite published statements, but it becomes hearsay to make an unsubstantiated claim about what another party said or wrote privately.


Certainly the public has a poor understanding of chaos and predictability. The climate models might do a good job in predicting trends - within uncertainty.

Obviously, if some data have been excluded such that the measurements are 'made' to agree with models, i.e., the data are intentionally and artificially biased, then there is a big problem with the data integrity and models.

Okay well i have referenced multiple evidence showing these climate models are not credible from any scientific perspective. I can do no more than point people at the actual science.

I mean no disrespect to Gavin, but the fact he won't define the basic physics underlying his idealisation of the climate *system* is very disturbing. It has also become apparent that most of these climate modellers have no background in non-linnear dynamics as has been pointed out by many engineers who use complex models idealising something from the real-world.

In fact its is interesting how so many working engineers are sceptics. But not surprising since engineers actually have to work with real-world problems which require real results that can be proven or falsified.

 

[vanesch]

I have been playing a bit around, and just as a pedagogical illustration and nothing more, I've been doing the following.

I took the Lorenz attractor (see http://en.wikipedia.org/wiki/Lorenz_system) and took the source code there as a starting point, and adapted it to the free matlab-like environment scilab (see http://www.scilab.org/ where you can download it for different systems). I prefer scilab over octave because it has a better windows-implementation as far as I know.

The Lorenz attractor is a very well known chaotic system, as a function of 3 parameters, sigma, rho and beta (see wiki for an explanation).

If you calculate a solution to the Lorenz system starting out from arbitrary starting conditions, you get a solution, and given that the system is chaotic, each different initial condition will give you a wildly different solution.

But then I calculated, along a solution, an average (here the average of x(1) * x(2), but you could think up any quantity based upon the state of the lorenz system).

It turns out that that average is, even though calculated on the basis of a solution to a chaotic system, rather "stable".

Next I calculated that average for different Lorenz systems, where I varied one of the parameters, namely sigma, from 7 to 15 (in steps of 0.2) and calculated the average of my "interesting quantity" x(1) * x(2) for each of these. This gives me a (somewhat noisy) dependency of my "interesting quantity" on sigma.

I did that several times, and I find rather comparable evolutions.

So I can conclude that even though the underlying system is chaotic, I can say things about the dependency of my average "interesting quantity" as a function of a parameter (sigma).

Here's the code:

// Lorenz Attractor equations solved by ODE Solve
// x' = sigma*(y-x)
// y' = x*(rho - z) - y
// z' = x*y - beta*z
function dx = lorenzatt(T,X);
    global rho;
    global sigma;
    global b;
    dx = zeros(3,1);
    dx(1) = sigma*(X(2) - X(1));
    dx(2) = X(1)*(rho - X(3)) - X(2);
    dx(3) = X(1)*X(2) - b*X(3);
endfunction
// Using LSODE to solve the ODE system.
clear all
global rho;
global sigma;
global b;
k = 1;
ssig = 7:0.2:15;
for sigma = ssig
  rho = 28; 
  // sigma = 10; 
  b = 8/3;
  t = 0:0.01:100; X0 = rand(3,1);
  X=ode(X0,0,t,lorenzatt);
  // param3d(X(1,:),X(2,:),X(3,:))
  uu(k) = mean(X(1,:).*X(2,:));
  k = k + 1;
end;
plot(ssig,uu,'g')

 

[Coldcall]

Vanesch,

"It turns out that that average is, even though calculated on the basis of a solution to a chaotic system, rather "stable"."

Thats a good point though because chaotic systems are stable. Chaotic behaviour has nothing to do with instability. The term chaos refers to the unpredictability because of sensitive initial conditions.

Hence why we define some chaotic systems as self-organising because they appear able to maintain a status quo (within a limit cycle) creating long term stability through positive and negative feedbacks. What better example do we have than a climate system that has remained stable for some billions of years.

The agw proponents would have us believe this climate stability is now endangered by our moderate increase in Co2 emissions. Maybe they are right, but i suspect they are not.

 

[vanesch]

Vanesch,

"It turns out that that average is, even though calculated on the basis of a solution to a chaotic system, rather "stable"."

Thats a good point though because chaotic systems are stable. Chaotic behaviour has nothing to do with instability. The term chaos refers to the unpredictability because of sensitive initial conditions.

I think you are pretty confused about what "chaotic" actually means. It has a rather well-defined mathematical meaning, which has been discussed before in this thread.

You don't seem to have noted what this little example illustrates. It illustrates that you can have a chaotic system (the Lorenz system here, a very simplistic "weather forecaster"), of which certain average quantities (here taken to be < x1 x2 > ) which correspond to "climate" which is "average weather" are nevertheless rather well-defined and can be calculated, and moreover, of which these averages can be calculated as a function of a varying external parameter (here sigma). Nevertheless, all the "solutions" to the lorenz system that I calculated were "wrong" simply because it is a chaotic system and that my computer has finite numerical resolution. So all these numerical solutions have deviated from the "true" solution after small time intervals (although within these time intervals, they were each time rather good approximations).
But even with these "wrong" solutions, I could calculate a rather well-defined average, which was reproducible, which didn't vary wildly with initial conditions, and which had a smooth dependency on sigma.

So for this chaotic "weather" I could calculate rather well-defined "climate", and I could also calculate the influence of a changing boundary condition (sigma) on "climate" - even though the underlying "weather" is totally chaotic.
(again, I need to emphasise that this is a toy example for illustrative purposes only, and has nothing to do with a real climate system).

Hence why we define some chaotic systems as self-organising because they appear able to maintain a status quo (within a limit cycle) creating long term stability through positive and negative feedbacks. What better example do we have than a climate system that has remained stable for some billions of years.

Don't confuse issues please, so-called self-organisation has nothing to do with the definition of a chaotic system as we take it here.

The agw proponents would have us believe this climate stability is now endangered by our moderate increase in Co2 emissions. Maybe they are right, but i suspect they are not.

Absolutely not. You should maybe try to understand what the scientific AGW claim is. It is not about "rendering the climate chaotic", or "rendering the climate instable" or something of the kind. It is about changing the climate under the influence of human activities. And changing the climate means "changing average weather".

These changes might be relatively large and we might not like the new "average weather".

So the whole idea is to try to estimate as well as we can, what will be the "new average weather" as a function of certain projections of human activities in the near future, and how much it will be different from the current "average weather".

So for the near-term projections (next century or so), any "chaotic behaviour" of the climate system itself doesn't really matter - and these are the only projections that matter on the level of political and societal decision making.

There can be an academic question of whether on the very long run, the 30-year weather average itself (climate) has a chaotic dynamics - in other words, whether the chaotic behavior of the underlying weather dynamics has very slow components that exhibit themselves a chaotic dynamics. I guess that's an open question, but it has nothing to do with the social implications of AGW. The little bit of trajectory that we need to find out about, namely the next century or so, has nothing to do with this potential very-long-term behaviour of the weather/climate system.

 

[Coldcall]

Vanesch,

I'm not going to get into a slanging match about the defintion of chaos, but this wiki defintion should suffice:

"Chaos theory is an area of inquiry in mathematics, physics, and philosophy which studies the behavior of certain dynamical systems that are highly sensitive to initial conditions. This sensitivity is popularly referred to as the butterfly effect. Small differences in initial conditions (such as those due to rounding errors in numerical computation) yield widely diverging outcomes for chaotic systems, rendering long-term prediction impossible in general"

"So for this chaotic "weather" I could calculate rather well-defined "climate", and I could also calculate the influence of a changing boundary condition (sigma) on "climate" - even though the underlying "weather" is totally chaotic.
(again, I need to emphasise that this is a toy example for illustrative purposes only, and has nothing to do with a real climate system)."

Exactly, so what is your point? All you've shown is what we know already about chaotic systems. On the one had you say that using your simple example tells us we can make average projections/predictions which within a margin of error are correct-ish. We've always known that.

Then you go on to admit that your example is nothing like the climate system. Correct.

"Don't confuse issues please, so-called self-organisation has nothing to do with the definition of a chaotic system as we take it here."

Theres no confusion, but the climate is indeed self-organising. And it is relevant because we are discussing the stability factor which you brought into the discussion with the attractor, and the self organisation of the climate with negative and positive feedbacks is what helps it stay stable. Its the positive and negative feedbacks which applies self-organisation to such a system as the climate otherwise it would not have stayed so stable.

Do you not agree with that above statement? If not, why not?

"Absolutely not. You should maybe try to understand what the scientific AGW claim is. It is not about "rendering the climate chaotic", or "rendering the climate instable" or something of the kind. It is about changing the climate under the influence of human activities. And changing the climate means "changing average weather"."

Ahh but it should be about rendering the true physics of the (chaotic) climate into the climate models because otherwise they are not describing anything we know of in the physical world.

And yes it does appear to be rendering the climate as "unstable" if we add more human Co2 emissions to the mix. They talk of tipping points and run-away warming which would be events causing instability to the once stable climate. This is what we are supposed to be getting so worried about.

"So for the near-term projections (next century or so), any "chaotic behaviour" of the climate system itself doesn't really matter - and these are the only projections that matter on the level of political and societal decision making."

It seems to me you are saying that we can just add a huge caveat re margin of error and that will insulate from any major errors in policy making based on those projections. That depends on measures taken in view of such a projection.

How large is that margin of error? Look i could say that the temperature will not fall or increase within a + or - 2 degree spread in the next 50 years. I will likely be 100% correct, but from a scientific point of view its a worthless statement. I don't need any science to make that projection, just common sense.

"There can be an academic question of whether on the very long run, the 30-year weather average itself (climate) has a chaotic dynamics - in other words, whether the chaotic behavior of the underlying weather dynamics has very slow components that exhibit themselves a chaotic dynamics. I guess that's an open question, but it has nothing to do with the social implications of AGW."

Its not just an academic question. As I said earlier if they are building climate models which are supposed to (with some degree of accuracy) describe the physical climate then it is important to factor in the chaotic nature of the beast - so to speak.

It seems to me you are saying that we can ignore the physics of chaos and build an accurate climate model which does not take it into account.

And reading on how you treat "chaos" it seems to imply that you think the only chaotic factors which would be relevant would be dramatic ones which cause the system to spin out of control or become unstable.

For instance the butterfly effect shows how a small almost negligible disturbance can cause a much greater effect in another part of the system. That's chaos and it needs no dramatic tipping point to be classed as an important property of complex non-linear systems - such as the climate system.

What you appear to say above is that there is no such effect with the climate, or if there is we can account for it and its not important for projections. Since we don't even know what the causal sensitivity is in the climate i find it really diffiult to be satisifed with your hand waving away of the chaotic effects in climate systems.

So let's put it to the test, as science should be done. But so far the GCMs have failed every test and i have referenced on this htread the many independent papers which politely mock the models ability to project or predict anything significant.

PS: Just to add i sense a serious contradiction in your argument (if you are supporting the run-away warming). On the one hand you want to minimise the amplitude of chaotic effect, but that contradicts the idea that Co2 is going to cause the system to tip into some new dangerous configuration. Can you not see the contradiction?

 

[vanesch]

Vanesch,

I'm not going to get into a slanging match about the defintion of chaos, but this wiki defintion should suffice:

"Chaos theory is an area of inquiry in mathematics, physics, and philosophy which studies the behavior of certain dynamical systems that are highly sensitive to initial conditions. This sensitivity is popularly referred to as the butterfly effect. Small differences in initial conditions (such as those due to rounding errors in numerical computation) yield widely diverging outcomes for chaotic systems, rendering long-term prediction impossible in general"

Yes. Nothing to do with "self-organisation" or the like. Sensitive to initial conditions (and some other conditions such as mixing).

"So for this chaotic "weather" I could calculate rather well-defined "climate", and I could also calculate the influence of a changing boundary condition (sigma) on "climate" - even though the underlying "weather" is totally chaotic.
(again, I need to emphasise that this is a toy example for illustrative purposes only, and has nothing to do with a real climate system)."

Exactly, so what is your point? All you've shown is what we know already about chaotic systems. On the one had you say that using your simple example tells us we can make average projections/predictions which within a margin of error are correct-ish. We've always known that.

So why are you insisting that because weather is known to be chaotic, climate predictions are useless ?

Theres no confusion, but the climate is indeed self-organising.

What does this actually mean, and what does this have to do with the chaos in weather systems which might, or might not, make climate predictions for the next century or so impossible ?

And it is relevant because we are discussing the stability factor which you brought into the discussion with the attractor, and the self organisation of the climate with negative and positive feedbacks is what helps it stay stable. Its the positive and negative feedbacks which applies self-organisation to such a system as the climate otherwise it would not have stayed so stable.

You will have to need to be a bit more explicit here. Do you think that climate has been stable over time ??

Do you not agree with that above statement? If not, why not?

Climate has not been "stable" over time, if by that you mean hasn't changed. That's pretty sure.

Ahh but it should be about rendering the true physics of the (chaotic) climate into the climate models because otherwise they are not describing anything we know of in the physical world.

That's what people try to do. But again, nobody knows whether this dynamics is chaotic in the end, but for sure this is not relevant on time scales of a few centuries.

And yes it does appear to be rendering the climate as "unstable" if we add more human Co2 emissions to the mix. They talk of tipping points and run-away warming which would be events causing instability to the once stable climate. This is what we are supposed to be getting so worried about.

Not at all. Don't confuse popular accounts with scientific statements. We are supposed to get worried about changing average weather from what we have now. From having different patterns of precipitation, of average temperature, from different snow and ice cover from what we have now. From *change*. We are supposed to get worried because some things will change. Will not be identical to what we have now.

"So for the near-term projections (next century or so), any "chaotic behaviour" of the climate system itself doesn't really matter - and these are the only projections that matter on the level of political and societal decision making."

It seems to me you are saying that we can just add a huge caveat re margin of error and that will insulate from any major errors in policy making based on those projections. That depends on measures taken in view of such a projection.

No, what I'm saying is that the academic discussion of whether the climate system shows signs of chaos over, say, 10 000 years or so is of no relevance to eventual policies that only need to know what will be the prospective changes for within 50-100-200 years or so.

How large is that margin of error? Look i could say that the temperature will not fall or increase within a + or - 2 degree spread in the next 50 years. I will likely be 100% correct, but from a scientific point of view its a worthless statement. I don't need any science to make that projection, just common sense.

Well, I think that the current "consensus" is that a doubling of CO2 will lead, within something of about a century or so, to a temperature increase grossly between 2.1 and 4.5 degrees WORLD AVERAGE, but which can hide much wilder local changes. In polar regions, this could be 10 degrees more and oceans maybe just 1 or 2 degrees. Note that this is for a CO2 doubling. If we have a 4-fold increase of CO2, grossly these numbers have to be multiplied by 2, so a temperature change between 4.2 and 9 degrees world average.

"There can be an academic question of whether on the very long run, the 30-year weather average itself (climate) has a chaotic dynamics - in other words, whether the chaotic behavior of the underlying weather dynamics has very slow components that exhibit themselves a chaotic dynamics. I guess that's an open question, but it has nothing to do with the social implications of AGW."

Its not just an academic question. As I said earlier if they are building climate models which are supposed to (with some degree of accuracy) describe the physical climate then it is important to factor in the chaotic nature of the beast - so to speak.

It seems to me you are saying that we can ignore the physics of chaos and build an accurate climate model which does not take it into account.

You don't "build in" chaos in a model. You build physical hypotheses in a model, and you can observe it to show chaos or not. It's something you discover after the fact, when you run the model.

And reading on how you treat "chaos" it seems to imply that you think the only chaotic factors which would be relevant would be dramatic ones which cause the system to spin out of control or become unstable.

Nobody talks about "spin out of control and become unstable". You are repeating this over and over again, but that's not what climate change is about. It is about, well, climate change. About changing climate, about having a different climate 100 years from now than we have now.

So let's put it to the test, as science should be done. But so far the GCMs have failed every test and i have referenced on this htread the many independent papers which politely mock the models ability to project or predict anything significant.

In order to put it to the test, you should run a climate model for, say, 200 000 years, make observations for about 200 000 years, and tell us whether you find any agreement or not.
Can be done but it takes a loooooong time.

PS: Just to add i sense a serious contradiction in your argument (if you are supporting the run-away warming). On the one hand you want to minimise the amplitude of chaotic effect, but that contradicts the idea that Co2 is going to cause the system to tip into some new dangerous configuration. Can you not see the contradiction?

No serious scientist considers run-away climate. It is not impossible, but you need extreme conditions (I think ocean temperature must be something like 60 degrees in order for vapor feedback to give you a runaway condition). As I pointed out already, a run-away climate is in fact rather easy to calculate, and has no chaotic behaviour: the oceans boil away, we know exactly the amount of water vapor in the atmosphere (the entire ocean), there's no life, there's no ice, there's no precipitation. Pretty simple.

 

[Coldcall]

Vanesch,

"Yes. Nothing to do with "self-organisation" or the like. Sensitive to initial conditions (and some other conditions such as mixing)."

It has plenty to do with self-organising systems. In fact it was through chaos theory we started understanding how non-living systems self-organise as do living ones, through positive and negative feedbacks. Every major text on chaos and complexity defines self-organisation as a natural property of those types of systems. Please read Davies (Cosmic Blueprint), Gribbins (Simplicity), Gleicks (Chaos), Stewarts (Does God Play Dice?). Are they all wrong?

"So why are you insisting that because weather is known to be chaotic, climate predictions are useless ?"

I never said they were completely useless, and i am in favour of modelling chaotic systems in general. My point is that one must accept the limitations of the physics of the underlying system, which is not the case with Gavin and GCMs. a) They don't accept the climate is a chaotic system b) They make highly exaggerated claims about the certainty of projections.

"What does this actually mean, and what does this have to do with the chaos in weather systems which might, or might not, make climate predictions for the next century or so impossible?"

The self-organising feature has nothing directly to do with climate predictions; i never said it did.
The part about self-organisation was related to your comments previously about the stability of a chaotic system. Those are two different strands of the same conversation.

"You will have to need to be a bit more explicit here. Do you think that climate has been stable over time ??"

"Stable" is a relative term, but yes within certain boundaries the climate has remained stable even during quite large fluctuations in temperature and Co2, as the record shows. If you look back just over the last million years you will see a very stable pattern of peaks and troughs.

"Climate has not been "stable" over time, if by that you mean hasn't changed. That's pretty sure."

Thats a daft statement :-) Sort of like using the term "climate change" to imply human causality. Stable in a chaotic system does not mean it stays exactly the same all the time. It varies as we have natural variation in climate. To suggest because the climate changes it is not stable is a false statement ( at least in the perspective of chaotic systems like the weather and climate).

"Not at all. Don't confuse popular accounts with scientific statements. We are supposed to get worried about changing average weather from what we have now. From having different patterns of precipitation, of average temperature, from different snow and ice cover from what we have now. From *change*. We are supposed to get worried because some things will change. Will not be identical to what we have now."

Thats a ridiculous thing to get worried about because there will be climate change no matter what. Of course the climate will never be identical :-) But you and i both klnow that if you used that rather weak argument for spending trillions on Co2 taxes there would be no COP15 right now. So instead we need the alarmism to panic folks into acting. Thats what is going on right now - in my opinion.

"Well, I think that the current "consensus" is that a doubling of CO2 will lead, within something of about a century or so, to a temperature increase grossly between 2.1 and 4.5 degrees WORLD AVERAGE, but which can hide much wilder local changes. In polar regions, this could be 10 degrees more and oceans maybe just 1 or 2 degrees. Note that this is for a CO2 doubling. If we have a 4-fold increase of CO2, grossly these numbers have to be multiplied by 2, so a temperature change between 4.2 and 9 degrees world average."

The problem is averaging out our global temperatures because we live in a localised world. But If those projections are from the GCMs then obviously i am going to take them with a large sack of salt (for all the reasons I've already stated in this thread).

But hey that's pretty alarming and would consitute a run-away warming, totally anamolous from a historical perspective. In th previous paragraph you stated: "Will not be identical to what we have now". Which sounds a lot less dramatic an occurence than an increase in temperature of 4-9 degrees!

"You don't "build in" chaos in a model. You build physical hypotheses in a model, and you can observe it to show chaos or not. It's something you discover after the fact, when you run the model."

No, you first define what is the underlying physics of the system so it can be mapped accurately and according to known laws of nature involving extremely complex non-linear dynamical systems. They have failed to do this first step, and instead play semantics over the physical definition of the climate system.

"No serious scientist considers run-away climate. It is not impossible, but you need extreme conditions (I think ocean temperature must be something like 60 degrees in order for vapor feedback to give you a runaway condition). As I pointed out already, a run-away climate is in fact rather easy to calculate, and has no chaotic behaviour: the oceans boil away, we know exactly the amount of water vapor in the atmosphere (the entire ocean), there's no life, there's no ice, there's no precipitation. Pretty simple."

Thats funny becuase a temeprature increase anywhere near 9 degrees would be absolutely catastrophic for certain parts of the world. Sorry i can't square your two statements about what the consensus view is, as i mentioned before; just the climate changing or not staying identical is a far cry from the sort of temperature projections you state are the consensus.

Climate change is okay as its been happening since the birth of earth. But 4-9 degree increase is not normal climate change; on a graph -depending on scale - it will look like we hit a wall!

 

[vanesch]

"Yes. Nothing to do with "self-organisation" or the like. Sensitive to initial conditions (and some other conditions such as mixing)."

It has plenty to do with self-organising systems. In fact it was through chaos theory we started understanding how non-living systems self-organise as do living ones, through positive and negative feedbacks. Every major text on chaos and complexity defines self-organisation as a natural property of those types of systems. Please read Davies (Cosmic Blueprint), Gribbins (Simplicity), Gleicks (Chaos), Stewarts (Does God Play Dice?). Are they all wrong?

You are just calling in some vague associations: you hear the word "chaos" and you jump to "self-organization", you hear "feedback" and you think "chaos". But even "self-organizing" systems which have an underlying well-defined dynamics can be calculated, without needing to know this. But moreover, all this has nothing to do with climate or weather. Moreover, the works you cite are full of popular mumbo jumbo.

"So why are you insisting that because weather is known to be chaotic, climate predictions are useless ?"

I never said they were completely useless, and i am in favour of modelling chaotic systems in general. My point is that one must accept the limitations of the physics of the underlying system, which is not the case with Gavin and GCMs. a) They don't accept the climate is a chaotic system b) They make highly exaggerated claims about the certainty of projections.

And how, exactly, do you think one "models a chaotic system" differently from modeling a non-chaotic system ? There's no difference ! You just DISCOVER that a certain system's dynamics you've modeled, has a chaotic behavior.
Look at the Lorenz system: they're just a set of differential equations, that resulted from a symplified physical modelisation. Lorenz didn't realize it was a chaotic system, he discovered it. The modeling work is identical, he didn't say "hey, let's put in some chaotic description". That doesn't mean anything, btw.

The "limitations of the physics" can only be discovered by putting them into a model and looking at the behavior of that model, by running it on a computer... exactly what GCM do.

"What does this actually mean, and what does this have to do with the chaos in weather systems which might, or might not, make climate predictions for the next century or so impossible?"

The self-organising feature has nothing directly to do with climate predictions; i never said it did.

So then why did you bring it up in this context ?

The part about self-organisation was related to your comments previously about the stability of a chaotic system. Those are two different strands of the same conversation.

I'm addressing aspects the original question of this thread, namely, the alleged impossibility of calculating climate change because of chaotic behavior and at the same time the requirement of chaotic behavior in order to be able to cry wolf.

"You will have to need to be a bit more explicit here. Do you think that climate has been stable over time ??"

"Stable" is a relative term, but yes within certain boundaries the climate has remained stable even during quite large fluctuations in temperature and Co2, as the record shows. If you look back just over the last million years you will see a very stable pattern of peaks and troughs.

Oh, well, then be reassured that it will remain stable within those boundaries

I think this has been the misunderstanding in this thread from the beginning. There's nothing totally exceptional going on. Just maybe a faster climate change than usual, and at least partly induced by human activities.

"Climate has not been "stable" over time, if by that you mean hasn't changed. That's pretty sure."

Thats a daft statement :-) Sort of like using the term "climate change" to imply human causality. Stable in a chaotic system does not mean it stays exactly the same all the time. It varies as we have natural variation in climate. To suggest because the climate changes it is not stable is a false statement ( at least in the perspective of chaotic systems like the weather and climate).

Well, stability means that there is a value of the state around which small deviations will give rise to a counter reaction that brings back the system to its original state.

If you use words with another meaning than the one it usually has in the theory of dynamical systems, then you can understand maybe the strange reactions you can get.

If you tell me "climate is stable", it means it didn't change, but moreover, that IF something tried to change it, there will be a mechanism that brings it back to its original value. If you understand now that it is "limited to a certain bounded range of values", that's totally different. Well, we will probably remain within that bounded range of values. No worry.

"Not at all. Don't confuse popular accounts with scientific statements. We are supposed to get worried about changing average weather from what we have now. From having different patterns of precipitation, of average temperature, from different snow and ice cover from what we have now. From *change*. We are supposed to get worried because some things will change. Will not be identical to what we have now."

Thats a ridiculous thing to get worried about because there will be climate change no matter what. Of course the climate will never be identical :-) But you and i both klnow that if you used that rather weak argument for spending trillions on Co2 taxes there would be no COP15 right now.

Nevertheless, that's exactly what it is about. People don't want the climate to change - especially so much in so few time. There's no more to it.
People want to limit the temperature deviation over a century to less than 2 degrees. Now, you know as well as I do that climate has both been hotter and colder than 2 degrees in the history of the earth. But people consider that that's too much of a change to adapt to in a century's time.

"Well, I think that the current "consensus" is that a doubling of CO2 will lead, within something of about a century or so, to a temperature increase grossly between 2.1 and 4.5 degrees WORLD AVERAGE, but which can hide much wilder local changes. In polar regions, this could be 10 degrees more and oceans maybe just 1 or 2 degrees. Note that this is for a CO2 doubling. If we have a 4-fold increase of CO2, grossly these numbers have to be multiplied by 2, so a temperature change between 4.2 and 9 degrees world average."

The problem is averaging out our global temperatures because we live in a localised world. But If those projections are from the GCMs then obviously i am going to take them with a large sack of salt (for all the reasons I've already stated in this thread).

Why ? That's nevertheless the best possible way of doing things: to model it physically.

But hey that's pretty alarming and would consitute a run-away warming, totally anamolous from a historical perspective. In th previous paragraph you stated: "Will not be identical to what we have now". Which sounds a lot less dramatic an occurence than an increase in temperature of 4-9 degrees!

That's absolutely not a run-away climate, and it has been hotter in the (far) past.

"You don't "build in" chaos in a model. You build physical hypotheses in a model, and you can observe it to show chaos or not. It's something you discover after the fact, when you run the model."

No, you first define what is the underlying physics of the system so it can be mapped accurately and according to known laws of nature involving extremely complex non-linear dynamical systems. They have failed to do this first step, and instead play semantics over the physical definition of the climate system.

That's nevertheless exactly what people try to do when they build GCM's.

"No serious scientist considers run-away climate. It is not impossible, but you need extreme conditions (I think ocean temperature must be something like 60 degrees in order for vapor feedback to give you a runaway condition). As I pointed out already, a run-away climate is in fact rather easy to calculate, and has no chaotic behaviour: the oceans boil away, we know exactly the amount of water vapor in the atmosphere (the entire ocean), there's no life, there's no ice, there's no precipitation. Pretty simple."

Thats funny becuase a temeprature increase anywhere near 9 degrees would be absolutely catastrophic for certain parts of the world. Sorry i can't square your two statements about what the consensus view is, as i mentioned before; just the climate changing or not staying identical is a far cry from the sort of temperature projections you state are the consensus.

Nevertheless, that's still within those "boundaries of stable historical climate" you were talking about, and - as you say - it does worry people. And NOT because of run-away. And not because of "instability".

Climate change is okay as its been happening since the birth of earth. But 4-9 degree increase is not normal climate change; on a graph -depending on scale - it will look like we hit a wall!

That's because it is going faster now than we used to have it, and because we have had a very long time of rather small temperature variations. So now you are realizing maybe why some people think that some action should be undertaken. But, again, this is still "within normal climate boundaries".

Now, 4 - 9 degrees is the prospect of a 4-fold increase in CO2 levels, which is very high. People usually table on something like a factor of 2, and then the increase should be 2 - 4.5 degrees, but that depends on CO2 exhaust. If we burn all the coal this century, it will be more than twice the pre-industrial level.

Just for what it is worth, there's the wiki entry on global temperature reconstruction. Don't take it as a solid source, I just cite it here because I'm too lazy to look up publications about it. I hope it is not considered "controversial", and there's no detailed point to be made apart from the fact that I think that it is generally accepted that temperature has both been higher and lower than now in the distant past by several degrees.

http://en.wikipedia.org/wiki/File:All_palaeotemps.png

 

[Coldcall]

Vanesch,

"You are just calling in some vague associations: you hear the word "chaos" and you jump to "self-organization", you hear "feedback" and you think "chaos". But even "self-organizing" systems which have an underlying well-defined dynamics can be calculated, without needing to know this. But moreover, all this has nothing to do with climate or weather. Moreover, the works you cite are full of popular mumbo jumbo."

The works i cite are all excellent texts on chaos, complexity and self-organising systems. If you can find something erroneous in those books please do let me know. But so far you are just insulting the authors (all of them except Gleick are phsicists and mathematicians). I'm really dissapointed you have resorted to ad homs against highly respected men of science.

I'm going to cut this discussion short now because i have referred you to numerous texts, peer-reviewed studies backing up my argument, but you have provided no evidence for your assertions. Instead you insist that we can use these GCMs to make accurate projctions/predictions, even while I've shown you studies which ALL question the credibility not only of the predictive capacity of GCMs but the very physics on which they are based.

I can do no more :-)

 

[sylas]

The works i cite are all excellent texts on chaos, complexity and self-organising systems.

It is the application which is ... odd. The general texts may be fine, but the inferences you are drawing for specific cases are not in the texts and don't follow from the texts.

 

[sylas]

Maybe we need a separate thread on the effects of volcanic eruptions on climate.

One paper in which you might be interested is the following.

• Wigley, T. M. L., C. M. Ammann, B. D. Santer, and S. C. B. Raper (2005), http://www.agu.org/pubs/crossref/2005/2004JD005557.shtml, in J. Geophys. Res., Vol 110, D09107, doi:10.1029/2004JD005557.
  http://www.kore-net.com/documents/volc.doc

This paper describes the way in which climate sensitivity is estimated from the recovery times after a large volcanic eruption. The discussion helps explain how the inference works. This paper was also introduced and discussed in the thread [thread=307685]Estimating the impact of CO2 on global mean temperature[/thread].

The three eruptions considered are in the latter part of the twentieth century. They were large eruptions, but not super eruptions. Their effect was significant and easily noticeable in global climate records, and recovery afterwards occurs over several years.

Weather is basically driven by the need to shed energy from the Sun. Given a comparatively constant input of solar energy, you don't tend to get drastic shifts in climate (averaged weather) unless something else changes to alter energy balances: such as a shift in Earth's orbit, or the composition of the atmosphere, or the arrangement of continents. You can potentially get a shift in climate if weather patterns alter, which in turn affects cloud distributions and energy balance from their effects on radiant energy transfers; but this is not a cummulative effect. It means only that (in principle) we might expect climate to shift between different modes occasionally.

But generally, after a big eruption the climate soon comes back to basically what you had previously. Climate is not chaotic in the sense that a small shift can grow without limit to a large different future trajectory. In fact, it tends to be pretty stable for a given set of conditions, and only shifts when you get some significant shift in the energy flows for some reason. In that sense, the behaviour of "weather" and "climate" are actually quite different.

Cheers -- sylas

 

[Coldcall]

It is the application which is ... odd. The general texts may be fine, but the inferences you are drawing for specific cases are not in the texts and don't follow from the texts.

If you really think that then you clearly have either not read those books, or you have forgotten the content. I suggest you pick them up and have a good read and they will confirm my posts on the strong connection between chaotic systems, complexity and self-organisation.

 

[Coldcall]

Guys,

This is my last reference on this issue because the conversation is getting silly mwith people starting to deny some basic physcial laws we are well aware of.

Please read this paper, one of hundreds on self-organisation and chaotic systems.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC128561/

"The National Academy of Sciences convened an Arthur M. Sackler Colloquium on “Self-organized complexity in the physical, biological, and social sciences” at the NAS Beckman Center, Irvine, CA, on March 23–24, 2001. The organizers were D.L.T. (Cornell), J.B.R. (Colorado), and Hans Frauenfelder (Los Alamos National Laboratory, Los Alamos, NM). The organizers had no difficulty in finding many examples of complexity in subjects ranging from fluid turbulence to social networks. However, an acceptable definition for self-organizing complexity is much more elusive. Symptoms of systems that exhibit self-organizing complexity include fractal statistics and chaotic behavior. Some examples of such systems are completely deterministic (i.e., fluid turbulence), whereas others have a large stochastic component (i.e., exchange rates). The governing equations (if they exist) are generally nonlinear and may also have a stochastic driver. Many of the concepts that have evolved in statistical physics are applicable (i.e., renormalization group theory and self-organized criticality). As a brief introduction, we consider a few of the symptoms that are associated with self-organizing complexity."

 

[vanesch]

Guys,

This is my last reference on this issue because the conversation is getting silly mwith people starting to deny some basic physcial laws we are well aware of.

Please read this paper, one of hundreds on self-organisation and chaotic systems.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC128561/

"The National Academy of Sciences convened an Arthur M. Sackler Colloquium on “Self-organized complexity in the physical, biological, and social sciences” at the NAS Beckman Center, Irvine, CA, on March 23–24, 2001. The organizers were D.L.T. (Cornell), J.B.R. (Colorado), and Hans Frauenfelder (Los Alamos National Laboratory, Los Alamos, NM). The organizers had no difficulty in finding many examples of complexity in subjects ranging from fluid turbulence to social networks. However, an acceptable definition for self-organizing complexity is much more elusive. Symptoms of systems that exhibit self-organizing complexity include fractal statistics and chaotic behavior. Some examples of such systems are completely deterministic (i.e., fluid turbulence), whereas others have a large stochastic component (i.e., exchange rates). The governing equations (if they exist) are generally nonlinear and may also have a stochastic driver. Many of the concepts that have evolved in statistical physics are applicable (i.e., renormalization group theory and self-organized criticality). As a brief introduction, we consider a few of the symptoms that are associated with self-organizing complexity."

Self-organization is a vague concept. Some people call phase transitions "self-organizing". What is said here, is that systems that, according to some criteria are "self-organizing" what ever that may mean, exhibit also often chaotic dynamics.
That doesn't mean that any dynamical system that exhibits chaos will be "self-organizing", and moreover being "self-organizing" doesn't tell you much about the possibility of modelling it and using the dynamical model.

Now, in what way does that have anything to do with the potential of climate models to give us an idea of future climate on the relatively short term (say 100 years) ?

Because how do you think that people set up those so-called "self-organizing" system dynamics, in a different way from the usual way of setting up a dynamics ?

You still haven't addressed that. How are people supposed to handle "climate dynamics" differently if it would turn out to be chaotic (which is not certain at all), and whether it would show aspects of "self-organization" (whatever that might mean in climate dynamics: phase transitions in climate ??). In what way would the usual way of setting up the dynamics be any different, and why would that approach not be valid ?

 

[Coldcall]

Self-organization is a vague concept. Some people call phase transitions "self-organizing". What is said here, is that systems that, according to some criteria are "self-organizing" what ever that may mean, exhibit also often chaotic dynamics.
That doesn't mean that any dynamical system that exhibits chaos will be "self-organizing", and moreover being "self-organizing" doesn't tell you much about the possibility of modelling it and using the dynamical model.

Now, in what way does that have anything to do with the potential of climate models to give us an idea of future climate on the relatively short term (say 100 years) ?

Because how do you think that people set up those so-called "self-organizing" system dynamics, in a different way from the usual way of setting up a dynamics ?

You still haven't addressed that. How are people supposed to handle "climate dynamics" differently if it would turn out to be chaotic (which is not certain at all), and whether it would show aspects of "self-organization" (whatever that might mean in climate dynamics: phase transitions in climate ??). In what way would the usual way of setting up the dynamics be any different, and why would that approach not be valid ?

First of all you claimed that self organisation has nothing to do with chaotic systems. You've been proved wrong on that score. I've shown you a direct correlation as stated by that paper and many others, including all the books on this subject i mentioned previously.

Now you are backtracking by claiming its a vague concept. Its not vague at all and well understood in relation to chaotic systems which maintain that self organisation through feedback. From wiki:

"

Self-organization usually relies on four basic ingredients:

Positive feedback
Negative feedback
Balance of exploitation and exploration
Multiple interactions"

Are you claiming the climate system lacks any of those four properties? If so please provide such evidence.

 

[sylas]

If you really think that then you clearly have either not read those books, or you have forgotten the content. I suggest you pick them up and have a good read and they will confirm my posts on the strong connection between chaotic systems, complexity and self-organisation.

You continue to miss the point. The connections given in these general references are not what is in dispute, and have not been denied in the thread. The specific conclusions you draw with respect to climate are not in the books either. You need to actually understand the specific climate example a lot better before you can draw strong conclusions based on general texts for chaos theory.

Cheers -- sylas

 

[Coldcall]

You continue to miss the point. The connections given in these general references are not what is in dispute, and have not been denied in the thread. The specific conclusions you draw with respect to climate are not in the books either. You need to actually understand the specific climate example a lot better before you can draw strong conclusions based on general texts for chaos theory.

Cheers -- sylas

No what you are basically saying as is Gavin at RC and anyone who has invested emotionally or professionally in the GCMs is more or less: "we can't be sure climate system is chaotic, so we'll ignore the basic physics involved". Hence we are back to square one.

I think I have invested too much time already trying to convince you guys of some basic facts regarding the real-world implications and physics of climate systems.

You still insist they are not chaotic. Fine, I'm happy for you to wrap yourself up in a that false sense of security, but it aint science.

 

[sylas]

You still insist they are not chaotic. Fine, I'm happy for you to wrap yourself up in a that false sense of security, but it aint science.

Actually, I have not; though I have indicated some forms of chaotic behaviour that don't apply to climate. You understand me about as well as you understand Dr Schmidt, climate, science, and how to apply basic reference textbooks.

 

[vanesch]

First of all you claimed that self organisation has nothing to do with chaotic systems. You've been proved wrong on that score.

No, I disputed the fact that a chaotic dynamics by itself has anything to do with self-organisation, as if self-organisation (whatever that may mean !) is an automatic consequence of "chaotic dynamics", especially in the case of climate systems.

Chaotic dynamics is well-defined: there is a clear definition of what it is. Self-organisation is much vaguer. It comes down to "a system that increases its internal complexity", but what's that supposed to mean exactly ?

For instance, consider an example of "self-organisation" in magnetisation which is "spontaneous magnetisation". A typical theoretical system here is http://en.wikipedia.org/wiki/Ising_model"

This is basic statistical physics. The physical modelling is done "as usual", what is observed is the behavior of the output of the model. Nobody put "self organisation" in the model (by definition you can't put it in the model !).

You make of that concept of "self organisation" what you want. You don't win much by introducing the word: it doesn't change any of the actual work behind the modelling.

Self-organization usually relies on four basic ingredients:

Positive feedback
Negative feedback
Balance of exploitation and exploration
Multiple interactions"[/B]

What does that mean, a system that has a "balance of exploitation and exploration" ?

When does a differential equation have a "balance of exploration and exploitation" ?

Are you claiming the climate system lacks any of those four properties? If so please provide such evidence.

Apart from "balance of exploration and exploitation" of which I don't know much, most electronic amplifiers can be said to satisfy those criteria: there are usually positive and negative feedbacks in it, and as there are multiple components, you can say that there are multiple interactions. Are they self-organizing systems ? Are they chaotic then ? Can't we simulate them by putting together the models of their constituents (for instance in a SPICE simulator) ?

Do you understand why I say that it is a vague concept ?