Bishop Hill GCMs and public policy
Aug 24, 2014 Climate: Models Climate: Parliament In the thread beneath the posting about the Chen and Tung paper, Richard Betts left a comment that I thought was interesting and worthy of further thought.
Bish, as always I am slightly bemused over why you think GCMs are so central to climate policy.
Everyone* agrees that the greenhouse effect is real, and that CO2 is a greenhouse gas.
Everyone* agrees that CO2 rise is anthropogenic
Everyone** agrees that we can't predict the long-term response of the climate to ongoing CO2 rise with great accuracy. It could be large, it could be small. We don't know. The old-style energy balance models got us this far. We can't be certain of large changes in future, but can't rule them out either.
So climate mitigation policy is a political judgement based on what policymakers think carries the greater risk in the future - decarbonising or not decarbonising.A primary aim of developing GCMs these days is to improve forecasts of regional climate on nearer-term timescales (seasons, year and a couple of decades) in order to inform contingency planning and adaptation (and also simply to increase understanding of the climate system by seeing how well forecasts based on current understanding stack up against observations, and then futher refining the models). Clearly, contingency planning and adaptation need to be done in the face of large uncertainty.
*OK so not quite everyone, but everyone who has thought about it to any reasonable extent
**Apart from a few who think that observations of a decade or three of small forcing can be extrapolated to indicate the response to long-term larger forcing with confidence.
So, let me try to explain why I think GCMs are so important to the policy debate.
Let us start by considering climate sensitivity. As readers here know, the official IPCC position on climate sensitivity is largely based on the GCMs. This time round we have had some minor concessions to observational estimates, but a significant proportion of the probability density of the observational studies remains outwith the IPCC's likely range of 1.5-4.5°C. Proponents of GCMs might counter that the upper end of the GCMs are ignored too, but I would suggest that one should conclude that an ECS of 5-6°C in the light of temperature history.
Estimates of climate sensitivity - and therefore in practice GCM estimates of climate sensitivity - directly inform estimates of the social cost of carbon. So when people like Chris Hope are arguing for a carbon tax of $100/tCO2, this is a function of GCMs. I recall, I hope correctly, that Chris suggested a figure of $18/tCO2 if one used an ECS of 1.6, in line with observational estimates. This matters of course, because the policy response, if any, to an $18 problem is significantly different to that for a $100 problem.
Wherever we look in the interactions between scientists and politicians on climate questions, we see an emphasis on catastrophe. We see no confessions of ignorance, but only occasional reference to uncertainties. Here's some notes of Tim Palmer addressing the All-Party Climate Change Group:
With the amount of carbon dioxide already in the atmosphere, future emissions will need to be reduced by half to that of historical emissions to limit global average temperature rise to 2°C. However, if emissions are not curbed (under the business as usual scenario), the amount of carbon dioxide in the atmosphere will be three times the historical emissions and the temperatures might rise up to 4°C.
And on the other hand they might not. This idea does not, however, seem to have been put forward for consideration.
Readers might also wonder what explanations were given to our political masters on the credibility of the GCMs. Here's what Palmer said:
Climate models are only flawed only if the basic principles of physics are, but they can be improved. Many components of the climate system could be better quantified and therefore allow for greater parameterisation in the models to make the models more accurate. Additionally increasing the resolution of models to allow them to model processes at a finer scale, again increasing the accuracy of the results. However, advances in computing technologies would be needed to perform all the necessary calculations. However, although the accuracy of predictions could be improved, the underlying processes of the models are accurate.
Apart from the transport of heat to the deep ocean, if Friday's paper from Chen and Tung is to be believed.
You can see that policymakers are getting a thoroughly biased picture of what GCMs can do and whether they are reliable or not. They are also getting a thoroughly biased picture of the cost of climate change based on the output of those GCMs. They are simply not being asked to consider the possibility that warming might be negligible or non-existent or that the models could be complete and utter junk. They are not told about the aerosol fudging or the GCMs' ongoing failures.
And this is just scratching the surface.
[BTW: Could commenters who like to amuse themselves by baiting Richard please refrain from so doing!]
Reader Comments (306)
Aug 27, 2014 at 7:02 AM . . . and Then There's Physics
You should possible look up the concept of energy conservation. Even chaotic systems have to obey that.
This is a perfect illustration, and I do mean perfect, of why it is so difficult to carry out rational discussions with aTTP.
aTTP's habit of throwing out these completely useless comments in such a condescending manner has led me to think that ATTP does not have sufficient understanding of the subject matter to offer useful comments.
Energy is always conserved under all conditions; always has been always will be. Everyone knows that. All GCMS have representations of the energy equation. The Lorenz 1963 system contains a representation of the energy equation.
And then aTTP suddenly discovers that Earth's climate systems are open relative to energy, that interactions between subsystems, and intra-actions within subsystems affect the radiative energy balance at the ToA and that climate change is not a BVP. There's a lively discussion over there.
The latter has been noted above in this thread several times and recently has been noted even at aTTP's blog and also here which is a summary of the flawed BVP argument presented in this post
So much for the bounds and constraints offered by conservation of energy, I guess.
I must commend AndPhysics - willing to stick around and discuss. Greater credit goes to the venue-and-participants.
Dan,
I'll grant you that I can be condescending if you accept that you can be obnoxious. Give me a reason not to be condescending, and I'll try harder. Also, if you read my past comments, you may discover that I wasn't actually saying what you appear to think I've been saying. A suggestion : try interpreting what people say as they intend it to be interpreted, not as you'd like to interpret it. Just to be clear, I'm not denying that there are aspects of our climate that are chaotic, simply that there appears to be little evidence to suggest that its chaotic nature is likely to play a major role in setting the overall state of the climate on timescales that are currently of interest for us. Simply stating that something is possible, does not mean that it's likely. Of course, if you want to continue thinking that I have too little understanding of this topic to continue this discussion, that's also fine. My tolerance for infantile exchanges is getting smaller and smaller.
Shub,
Thanks, I think.
I agree however Dan, Attp's 'arguments' consist mainly of trivial truisms and the commenting style consists mainly of condescending remarks to sceptics who are stoopid enough to reject said truisms.
ATTP now says "...GCMs provide an additional tool and allow us to study processes on a smaller scale "
Isn't that what Richard Betts said? Isn't that what's contested? Examine the AR4 attribution case, if you are unfamiliar: there is a well-known graph with pink and blue error bands that plots model output with and without human-added CO2. GCMs sit at the centre of the attribution argument, not just as an 'additional tool' like a screwdriver.
aTTP
I understand that people may have good reasons for posting anonymously on this blog, but there is a danger involved in doing that. It's a recipe for a generally decent person to reveal unattractive aspects of his or her personality. A good rule, during composition of a post, is to ask yourself whether you are prepared to be known publicly for the comment you are making, and indeed whether you could make that comment face to face with the person you are addressing it to.
So I think that in making these rather snide sign-off remarks you may be letting yourself down somewhat. I think that is what some others here are trying to tell you.
The discussion you are having here about chaos is way over my head, but I get the impression you are well qualified to hold your own. It's a pity to spoil that with the snide stuff.
M.C. Tucker,
You appear to mean well, so thank you.
FWIW, I don't think I've said anything that I wouldn't have said if I were known. I find the online climate science debate a constant battle between decency (trying to be polite) and honesty (saying what I actually think). Sometimes I err towards the latter more than the former (you should have seen the comments I deleted :-) ).
Possibly, but - I would argue - that the manner in which people says things does matter. Of course, maybe they're trying to illustrate what I'm supposedly doing by doing the same themselves, but it might be more effective if they set a better example, rather than a worse one. Alternatively, maybe people think that the rules only apply to others and - to be fair - that does appear to be a common theme in the online climate debate. If all the people who complained about the tone of the debate actually tried to improve their own, there would - IMO - be a marked improvement.
> Nothing you've written has any impact on the discussion with ATTP.
This implies that Spence's
has no impact on his discussion with AT. On the other hand, we now know that Spence's "specific technical meaning" is "clear" but has no "agreed mathematical formalisation, even though he seems able to claim things like "if a process f(x) is chaotic, then an average (or locally integrated) f(x) is also chaotic".
As the Stanford entry shows elsewhere, even the "specific technical meaning," whatever which one Spence's referring to, has practical limitations.
"Chaos" is quite chaotic.
***
> if you think it is wrong, then all you have to do is publish the refutation [say of K's chaotic theory of climate] of this body of work.
How to reverse the burden of proof in one single step.
Spence's the one who claims that K's pet theory's the best. And that's an understatement: he's not far from implying that anyone who does not dig K's chaotic theory's an idiot. That there's a body of work competing with K's pet theory suffices to show that there might exist people in climate science for whom handwaving "kolmogorov" cuts little ice.
And to show that K's pet theory's the best, Spence arm waves at at K's pet theories. Check kiting, anyone? In any case, that ought to put his "I'm not relying on appeal to authority" into perspective.
As far as I am concerned, there's no need to publish that going all-in on chaos ain't seem compatible with falsificationism. If reverse stunts are allowed on this ClimateBall (tm) field, I'll simply say it should be obvious to anyone who has a grasp on the technical meaning of "falsifiability."
***
Speaking of body of work, all we have on this thread are the usual suspects: "Cohn, Lins, Koutsoyiannis." Are there other who published on this since the early 2000s, and why do we always keep hearing the same names over and over again?
The only truly exogenous variable in the climate system is insolation. And even that is modulated by internal factors which are part of the climate state vector (e.g. clouds).
Again, you use the word chaotically here as if moving to a different state is chaos. Chaos in terms of climate is not about state, but about the dynamics of the system. This once again stems from not having a sufficient depth of understanding of what we are talking about here.
I gave you a link to a presentation which clearly divides modelling into two distinct categories (and subcategories for each of these), with scientific justification. This is not a question of viewpoint, these are scientific positions. Although from your reply I am guessing you did not even look at the presentation at the link. You can also see a more informal discussion at the climatedialogue link willard gave earlier (on true vs. false dichotomies).
It is no good insisting that the position is a viewpoint when I have already given you a scientific justification for that position.
People misusing technical terms has no impact on a discussion? How ridiculous! Misuse of technical terms hinders scientific discussion - anyone should understand this. But at least ATTP is trying to engage in coherent debate. A random number generator would make more sense than your posts. For example:
Well, it is clear to someone who has advanced beyond high school mathematics, but perhaps not clear to you. I will explain a little further, although I suspect we will hit a roadblock since your diploma from google U does not extend to understanding.
From the link that you provided, we acknowledge no formal mathematical definition, but rather a worded definition (but still words that have technical meaning). Chaotic systems conform to a number of criteria, and one of those is exponential error growth from initial conditions, which ensures any arbitrarily small error will overwhelm the solution in linear time.
So how can I go from such an informal definition to my claim? Well, because I understand basic maths. If we assume weather variability is simply a form of random variation and that averaging this can reduce the error, I will note that the best reduction of error this can achieve is the square root of the timescale over which we average. So we have two competing parts of the error term: an exponential increase in error as a function of time, compensated by a square root reduction of error in timescale.
So, in willard's strange world, there may exist a rigorous mathematical definition in which a square root reduction in error swamps an exponential increase in error at increasing time. But back in the real world, I know that under no conditions can an exponential increase in error be compensated by a linear reduction in error. At some time scale, the exponential error term will dominate.
So a basic knowledge of mathematics allows us to know that exponential error term will always swamp the linear reduction in error, and that the averaging process makes negligible difference to our time horizon of predictability - which for weather is typically of the order of 5 to 10 days.
If you believe weather has exponential divergence from initial conditions, so does climate. Unless you believe linear terms in an equation can swamp exponential terms, in which case you may as well believe in pixies at the bottom of the garden.
The same reasoning can be applied to the other criteria, which I do not have time to do here but I believe Dan Hughes has set out detailed explanation on this topic at his blog in the past. Anyone who tries to make the claim that weather is chaotic and climate isn't doesn't understand chaos. They are the same system and however you define it, they are either both chaotic, or neither chaotic, take your pick.
The real, interesting debate is that once you realise that the systems are constrained by the same limits of predictability, moving on to determining what those limits of predictability are, which are governed by the time horizon (Lyapunov exponents) and the fractal dynamics. But instead of having the useful interesting debate, instead willard tries to defend the indefensible by claiming exponential terms in an equation are swamped by linear terms. Fantastic.
Spence_UK,
No, I didn't. Clearly I'm in capable or writing a sentence that you can understand and you're incapable of interpreting what I say in the manner in which it was intended. Not uncommon on this debate, as far as I can tell. If there is the smallest chance that someone can interpret what another person has said in such a way as to invalidate what they've said, they'll grasp it with glee. In fact, that appears to be the main theme of this blog. If Andrew wants suggestions for a tagline, mine would be "misinterpreting climate science and climate scientists since (insert year)".
Yes, but we have little evidence to suggest that internal variability plays a big role in setting these boundary conditions. I realise that they appear to be able to influence them on short-timescales (decades) and I realise that the Dansgaard–Oeschger events are probably internally forced. But, just because something is possible doesn't make it likely, especially on the timescales that are of interest to us today.
In fact our discussion made me go and look up a bit more about this and I found the following on Steve Easterbrooks page,
You may think this is some kind of appeal to authority, but it's not. Maybe Steve is wrong, but it appears that this general view is that the long-term stability of our climate is largely set by the boundary conditions, which is largely what I've been getting at. Maybe, everyone I've ever heard talk about this is wrong, but a discussion on a blog in which one party thinks that a horde of experts is simply wrong, just seems remarkably silly (especially when they can't even properly interpret fairly simple sentences).
While these boundary conditions / constraints seem to be all-important relative to the future trajectory of Earth's climate, they are missing in action. So far as I am aware, no one has written down the differential equation system to which these apply and the actual boundary conditions have never been displayed.
This discussion potentially could be more useful if the concepts of focus have been explicitly presented and the effects on the subject equation system demonstrated.
Otherwise, this is known as hand- and arm-waving. Waving in very thin air. Apparently that's where these all important boundary conditions and constraints exist.
Until these critically important conditions have been explicitly displayed and discussed, those presenting the argument are doing so solely from authority.
ps
The difficulties of trying to carry out rational, objective discussions in blog-world are clearly illustrated
near the end of this thread.
Evidence that some people simply cannot avoid presumptive labeling and condescending comments is on display in abundance. And it's interesting to observe the feeble attempts to justify that approach. It's all a mis-understanding, don't you know.
The entire thread has dissolved into rationalizations for un-productive behavior.
Dan,
I don't get what you mean by this. Are you really suggesting that the models don't consider variations in Solar flux (which has to be an input) changes to the atmospheric composition (which they have to as modelling feedbacks is crucial), or changes to albedo (which - as I understand it - they do).
Sorry, is poor old you feeling put upon. Here you are trying to be so reasonable and thoughtful and others are just behaving so badly. It must be tough being you, but you just keep on setting this fantastic example of engaging in good faith and maybe the rest of will follow suite. (hint : in case it isn't obvious, this is sarcasm. If you don't like people being condescending, maybe stop being obnoxious and full of self-pity. People who bemoan the behavour of others without at least trying to acknowledge their own are, in my opinion, just irritating. I'm more than happy to acknowledge that my behavour isn't always ideal, but given how others behave it's sometimes hard not to respond in kind. Maybe also look up the meaning of the word "condescending".).
Dan,
I've just followed your link in which you bemoan the difficulties of carrying out a rationale, objective discussion, which turns out to be the end of a comment stream in which I apologise for what I've said and for not doing a better job of moderating. I've also just deleted my first attempt at this comment because it was sufficiently extreme that I decided not to post it. Use your imagination. I'm guessing you can make a fairly good stab at what my opinion of you is at this moment in time. I also wouldn't hold back if I was you (in terms of letting your imagination run wild).
Are you really suggesting that the models don't consider variations in Solar flux (which has to be an input) changes to the atmospheric composition (which they have to as modelling feedbacks is crucial), or changes to albedo (which - as I understand it - they do).
Nope, never said that.
I asked what are the specific constraints and boundary conditions to which you and Easterbrook refer. You have so far listed two. Changes in albedo in the atmosphere are not specified; they are instead calculated based on the model equations.
The two so far listed are standard for GCMs. Importantly they do nothing what so ever, absolutely nothing, to alleviate the lack of predicability associated with the spatial-temporal chaotic response of Earth’s climate as reflected in the GCM calculations.
The chaotic response from GCMs is the very basis for the use of ensembles of calculated results. If the calculated response was not chaotic, ensembles would not be necessary. Thus it is clear that none of the standard boundary conditions can do anything to alleviate the lack of predicability.
I suggest a Google Scholar search that requires all of the words; chaotic, chaos, GCM, predicability. Here’s a quote from one of the hits from the introduction to a chapter in a book published in 2012:
The nonlinearity of the primitive equations is also the main source of chaos and lack of predictability for long times in the weather and climate prediction. As has first been recognized by Lorenz (1963), even a simple three-variable nonlinear dynamical system (the so-called Lorenz attractor), based on the idealized convection cell with cooling at the top and heating at the bottom, exhibits extreme sensitivity to initial conditions. Nowadays, the Lorenz attractor is considered a canonical textbook example of chaos in a nonlinear dynamical system, with many illustrations depicting two nearly identical initial conditions evolving into two unrelated trajectories after a short period of time. In more complex dynamical systems with advection terms, nonlinear chaos develops in much more sophisticated fashion, making long-term forecasts difficult and uncertain. [ edh bold ]
Dan,
That would seem to be a good thing, unless I'm missing your point. If they were specified then they would be unable to respond to changes. Although I thought I'd said this a number of times before, I'll try again. It's my understanding that it is unlikely that internal variability can significantly influence factors like albedo, and atmospheric concentration. Hence, once these are specified, the system is likely to settle into a broadly stable state (in an overall sense) that is largely determined by these conditions. I'm not suggesting that it's impossible or that it has never happened (D-O events for example), simply that it seems unlikely given the current state of our climate. As I understand it, this is also the view of many people who work in this field. I may be wrong. They may all be wrong. We're unlikely, however, to resolve this on a blog whose motivation appears to be to mock any mainstream climate science.
I don't disagree with that. Seems like quite a reasonable comment. Maybe the next thing to consider is whether or not climate models are attempting to make forecasts. It's my understanding that they're not.
AndPhysics, the exact same thing that was done with SoD was done with me on your Salby thread. I requested people not to dismiss out-of-hand, but to be clear which of his claims were false and why. The thread degenerated into a browbeating charade. I wish I had stood up against the nonsense some commenters carried on in that discussion but it is hard when the host is up against you - well, that is my excuse anyway. You do not summarily ban SoD or Richard Tol because there is a price to be paid for getting rid of or going against prominent bloggers who may be aligned with your non-sceptical cause. On the other hand, it takes little to prevent the same against everyday sceptical bloggers and there may even be prizes behind-the-scenes.
Think about it - the moment SoD wanted to discuss natural variability there was the questioning of motives and scientific integrity. It does not matter if you are SoD, or Judith Curry or Roger Pielke Jr or Roger Pielke Sr - the result is the same.
Shub,
What I do now is very different to what I did a year ago. I've learned a lot and don't dispute that I've got a lot wrong (or would have done things differently if I knew then what I know now). If you could maybe show some appreciation of how difficult it is to moderate a blog, maybe we'd actually get somewhere. Tol I would happily ban as his contributions are ridiculous and his behaviour is infantile. I'm constantly amazed that he is actually a Professor of Economics. However, since I actually write posts about him, that doesn't seem fair. SoD, on the other hand, is extremely well-informed and knows a great deal about this subject. That he and SB clashed yesterday is unfortunate and I'd rather that they hadn't. You on the other hand were arguing on a post about Salby's ideas. Salby's ideas are so obviously wrong that I really can't be wasting my time moderating a thread with someone who thinks they might have merit. They really, really, really don't.
You say things like this and you expect me to possibly go "oh, maybe I made a mistake banning Shub. Such a thoughtful, insightful pleasant person, who could really make a positive contribution"? Seriously? I've banned a small number of people all of whom I have reason to dislike and all of whom have shown little understanding of climate science. I have no problem with thinking about whether or not I should relax this, but I'm certainly not going to do so if the same people make silly comments about "the cause" and "prizes behind the scenes". Makes me suspect that the Recursive Fury paper may be the most credible paper ever to have been retracted.
ATTP, I'm going to split my response into two parts, because I have a feeling they are going to get long. The first is the communication barrier caused by your misuse of technical terms, the second I will address Easterbrook's comments.
The problem is I am trying to interpret your statements as if they were scientific. But the words you use have scientific meaning which is simply nonsensical in the way you are using them. Let's look in detail.
First problem is the use of "state". What does this mean? Weather / climate is not quantised, and I am not aware of any formal definition of distinct states.
Some (sceptics and consensus) have argued to me that there are states - the most obvious of these being the change from interglacial to glacial period in the current ice age. So I looked to see if I could detect any evidence of distinct states. I plotted the Dome C ice core (after converting to uniform sampling by interpolation - otherwise the modern era is overrepresented). If there were distinct states, this should be visible in a bimodality in the distribution.
I saw no such thing. The distribution was unimodal, with no sign of preferred positions. To this date I have seen no convincing evidence of distinct states for climate - just one continuous space that the is traversed by the chaotic attractor. (I do use the term "state vector" to describe the state of the climate system, but that is not a "state" in the sense that you use it - it is a numerical modelling term to describe the vector of parameters, e.g. temperature, pressure, etc, that describe the system at a given point in time)
So we are moving through phase space in some way, and you define this motion as shifting "chaotically". I really have no idea what you mean. I think you might intend to mean randomly or unexpectedly. But with chaotic systems, the motion is always deterministic, and it is predictable over short time horizons, but not long time horizons. Sudden or slow changes are unlikely to have a big difference in randomness or predictability. So a big change or a small change are equally "chaotic", to misuse the term myself in an attempt to try and understand what you are getting at.
So there are two absolutely fundamental issues with what you have said that make no sense to me whatsoever scientifically. Both the concept of "state" and "shifting chaotically" make absolutely no sense to me at all, from a scientific perspective.
This isn't a show attempt to misunderstand you. I genuinely think what you say makes no sense in the context of the dynamics of complex systems.
Spence_UK,
I really wouldn't bother with the second part. The first part is enough to convince me that attempting to discuss anything with you is entirely a waste of my time. I know I've been accused of being condescending, but you really do take the title. I'll stop now before I say anything that I'm regret.
Tell you what. Imagine an incredibly rude response to this. Then make it ruder. And again. Now you're getting close.
Moving on to Steve Easterbrook's comment. As a side note to this discussion:
No, I don't consider quoting someone (even to the extent of endorsing them) is an appeal to authority if you are providing the complete train of thought they are giving. An appeal to authority takes the form "X is correct because Y says so", with the appeal being who says it. If you give a complete set of reasoning from another person as an example or because you agree with it, and that line of reasoning is available to debate, then that is most certainly not an appeal to authority - and I often reference Dr. Koutsoyiannis' work in the same way.
You also say:
Well, yes, Steve is absolutely wrong on the fundamentals of what he describes. The problem comes right at the beginning.
This is a big problem. To maintain stable boundaries, a chaotic system must exhibit Markovian dynamics, in which the attractor does not move out of one particular part of phase space and completes regular orbits filling most of the region it remains in. The Lorenz attractor is often used as an example, because it has Markovian dynamics. Unfortunately, the Lorenz attractor is an exception, rather than the norm, particularly for complex systems.
Complex chaotic systems are more likely to exhibit fractal dynamics, which means they do not have stable long term averages.
So, how can we tell which type weather is? Well, one option, as Dan Hughes points out, is we can write out the solution to the complete set of equations governing climate and examine the behaviour. Unfortunately, as Dan points out, nobody has ever written down the complete set of equations of climate, we only have a mixture of crude approximations and parameterisations which are most certainly not the full set of equations.
But that isn't the only option. We can perform statistical analysis on climate observations, to look for properties that will give us a clue to the nature of the dynamics. One is to test for embedding dimensions; a low number of embedding dimensions makes it more credible that Markovian dynamics could dominate. Many people have looked into this, and low dimensionality is consistently rejected for weather / climate. There are some claims near the limit of detectability (around 7-8 dimensions) but this has to be taken with a rather big pinch of salt, since near the limit of detectability false positives become more likely. So by and large the evidence for Markovian dynamics here is weak and unconvincing.
There is a second way, which is to directly test the climate indices for the presence of fractal dynamics. This can be tested for in a number of ways, for example by looking for evidence of long term persistence (suggesting that there are no "stable averages"). As far as I know, around a dozen or so climate researchers have investigated to see if long term persistence is present in the climate system. Every single one of them that I am aware of concludes that yes, long term persistence is present, pointing strongly towards fractal dynamics.
This provides evidence against Markovian dynamics (through the rejection of low number of embedding dimensions) and support for fractal dynamics (through the presence of long term persistence in observations). Both of these show that Easterbrook's claim that climate has boundaries that "keep long term averages stable" is almost certainly incorrect, and the conclusions Easterbrook draws from this are simply wrong.
Well, I posted my response to Steve Easterbrook's comments before seeing your reply. I can see you are very angry, which is ridiculous as I am merely presenting a scientific perspective on complex systems.
Technical language is important in communicating technical ideas. For example, what Dan Hughes says makes perfect sense to me because he is using the technical language of complex systems accurately. I absolutely confess that my use of technical language is not as accurate or careful as Dan's, I am sometimes sloppy, and am always happy to be corrected to help clarify discussion.
If you continue to abuse technical terms, and refuse to clarify what you intended to mean so communication becomes impossible, then meaningful discussion is impossible.
Spence_UK,
I'm guessing you missed my intermediate comment. Pity, could have saved you wasting your time writing the second part of your comment. You've written an entire comment about chaotic systems without mentioning the boundary conditions at all. I'd respond further but I really, really, really think that discussing things with you is a complete and utter waste of my time. You carry on believing that you're right and everyone else is wrong (apart from some people who published papers about chaos in the 1980s) and I'll try to remember not to bother discussing things with you in future.
Spence_UK
Wow, aren't you perceptive.
Indeed. On the other hand if you are entirely incapable of understanding a fairly basic sentence, there's not much I can do. FWIW, by "state" I meant the state set by the boundary conditions. For example, the equilibrium temperature. Now I imagine that you will still find some reason to regard this as abusing some technical term, but that would be par for the course, as far as I can see. (Again, you could imagine a much ruder response to your comment if you wish).
Nonsense. In my first main response to Steve Easterbrook's quote:
I recognise boundary conditions (which are present in systems with either Markovian or Fractal dynamics), and note that they are only stable in the presence of Markovian dynamics.
From that point onwards, the only interesting thing is the type of dynamics in the system. If we have fractal dynamics, we do not have a stable measure of boundary conditions (even though they do exist in the system, they are virtually impossible to measure with any accuracy).
That's why my discussion was to do with the dynamics of the system, because it is the elephant in the room. The boundaries don't tell you the limits of predictability, the dynamics of the system do.
And how can you tell the difference between a shift due to a change in boundary conditions, vs. a shift due to traversal of the chaotic attractor, other than making it up as you go along?
Sorry I should have added this:
Your use here is not conventional, and the word "state" usually has a different meaning, but at least now I know what you were trying to say. Thank you for that.
Spence_UK,
I was going to write response, but it truly is a waste of my time. You're just spouting details about chaotic systems without showing any understand of basic physics, the role of external forcings, or any of the other major factors that influence our climate. You're obeying one of my rules of the online climate debate : anyone who is willing to state that what someone else has said is nonsense, is probably talking nonsense themselves.
Spence_UK,
Because if the boundary conditions don't shift, the system can't change state. It's not that complicated. It's to do with energy conservation. Now I really am wasting my time.
That doesn't tell me how you tell the difference between a change in boundary conditions and the normal motion of the attractor. In fact your response, using your definition of "state", is just a tautology which says if the boundary conditions don't shift, the boundary conditions don't shift.
You are avoiding the elephant in the room. How do you tell the difference between a change in boundary conditions and the motion of the attractor?
This is not an answer. Energy conservation is a very weak constraint on a dynamic system. Neither Markovian nor fractal dynamics are excluded by energy conservation, and energy conservation has exactly zero bearing on the predictability of the system. It tells us nothing. I really don't understand your strange obsession with it.
That quote is an absolute keeper. I haven't laughed so much or so hard in a long, long time. Thanks.
You don't even know what is external forcing and what isn't until you can understand internal variability. You've put the cart before the horse.
Now there's a challenge I can't pass up. First hit from google on your site:
http://andthentheresphysics.wordpress.com/2014/08/14/a-challenge/
Just two weeks ago, quoted directly from your own blog. What are you saying about yourself, ATTP?
Spence_UK,
Seriously, this is absolutely ludicrous.
If the boundary conditions do not change, then the system cannot change into a completely different state (by which I mean one with - for example - a different equilibrium temperature). The system being chaotic does not change this. The normal motion of the attractor cannot change the overall state of the system if the BOUNDARY CONDITIONS DO NOT CHANGE! If, on the other hand, internal variability could change one, or more, of the boundary conditions, then the system could change into a different state, but then THE BOUNDARY CONDITIONS WOULD HAVE CHANGED. However, there is little evidence to suggest - given our current state - that this is possible (again, not impossible, just unlikely). Therefore, since there is little evidence to suggest that internal variability can significantly influence our boundary conditions, the overall state of our climate cannot change unless THE BOUNDARY CONDITIONS CHANGE. You continuing to tell me that chaotic systems and the motion of the attractor does not change that the overall state of the system cannot change unless the BOUNDARY CONDITIONS CHANGE!
Because it is a f**king fundamental law of physics.
Spence_UK,
Fair point, but anyone who thinks that Salby's ideas have merit is truly lost in the wilderness. How anyone can understand chaotic dynamics and not understand that Salby's ideas are complete nonsense is completely beyond me. I would regard that as a position that is virtually impossible to hold.
Right. So far we have ascertained that:
You have redefined the word "state" to mean "boundary conditions"
You have repeated, pointlessly, over and over again, the tautology that a state change (by your definition) is a boundary condition change.
You arbitrarily, and with no understanding of nonlinear dynamics, declare certain things (which are unclear) to be likely and other things (which are equally unclear) to be unlikely
And yet you still have not answered my simple question - how can you tell, in the presence of fractal dynamics, whether a change is a boundary condition change or part of the normal trajectory of attractor.
I can tell you why you have not answered the question: because to do so would require at least some understanding of complex non-linear dynamics, which it is becoming increasingly clear you do not possess.
I can list a dozen "fundamantal laws of f**king physics" which are equally irrelevant to the point at hand. You don't seem to have the same obsession with those. Perhaps I should start talking about Maxwell's equation every post and swearing at you in the hope that it means something?
FWIW I have never read any of Salby's work and my quote from your blog was absolutely not intended to imply any endorsement or otherwise of any claims he has made. There are limits to how many different climate theories it is possible to take an interest in and I genuinely have no idea what his claims are.
Spence_UK,
Because if the presence of fractal dynamics could, for example, result in our climate getting warmer (an increase in overall energy) without changing the boundary conditions, then either it has violated a fundamental law of physics (energy conservation) or that law is wrong (which seems unlikely).
No, it appears that you do not understand the role our boundary conditions play in setting the overall state of our climate.
Well, apologies for being rude, but this has been one of the most frustrating discussions I've had for a long time. Also, being told I don't understand the evolution of dynamical systems by someone who appears not to understand the relevance of energy conservation is remarkably irritating.
I think is a complete waster of time, so you can carry on appearing to suggest that chaotic systems can violate energy conservation. I'll continue to think that they can't.
By this token, you seem to believe that weather violates a the law of energy conservation. It doesn't.
Fractal dynamics are quite stationary, and therefore do not violate the conservation of energy. However, the boundary conditions are external to the fractal dynamics of the system.
Fractal dynamics can result in, for example, trends of arbitrary length (millions of years if need be). This is achieved purely by variations of internal parameters of the system (in particular, the hydrological cycle - water vapour, clouds etc)
This has all been explained to you, yet you persist in believing that the conservation of energy somehow prevents this from happening. There is no scientific reason or basis for it.
I actually don't think you know what boundary conditions really are. Do you think cloud cover is a boundary condition? Do you think conservation of energy limits the amount of cloud cover the earth has? If the entire earth became covered in cloud, would this affect the temperature at the surface, and would it violate the law of conservation of energy?
I am genuinely scratching my head at the belief that conservation of energy somehow disallows these things. There is absolutely no basis for it.
Spence_UK,
No, that's the point I've been trying to make. The weather is chaotic. Our climate is not. I'm not arguing that the system isn't chaotic. I'm arguing that internal processes (chaotic or not) are unlikely to change the boundary conditions by themselves. Hence the overall state (the total energy) of our climate is set by the boundary conditions which are typically not influenced by the chaotic nature of the system.
You've explained lots of things, none of which really seem relevant. Whether you believe it or not, I do understand the concept of deterministic chaos.
I'm disallowing the possibility that the total energy in our climate system can change without a change in one of the boundary conditions (albedo, atmosphere, Sun). That's really all. As it stands, there is no evidence that internal processes can significantly influence these conditions. Telling me that the system is chaotic does not change this.
But, as I've said before, this is truly an amazingly frustrating discussion, especially as you know appear to be arguing against your own strawmen. Tell you what, why don't you imagine what my next response will be (because your interpretation of what I've said so far is essentially the same as simply making it up) and then argue against it. It will save me a lot of time actually responding.
*Rollseyes*
Wow. You really have learned nothing. This claim is simply false, as I have clearly explained over and over again. The solution to the system of equations that governs both weather and climate is chaotic, or not. They are the same system, and timescale has no impact on whether the solution is chaotic, or not. So they cannot have different states.
"Albedo" and "atmosphere" are not boundary conditions in the conventional sense of the word, they are parameters which are part of internal variability of the system. So once again you abuse technical terminology.
But let us, just for a moment, accept your botched use of the term.
So cloud cover, which is clearly an internal parameter and not a boundary condition, cannot change albedo? Cannot change radiative imbalance? There is no evidence that internal processes affect cloud cover, which in turn changes these parameters which you incorrectly call "boundary conditions"?
But wait, I've already pointed this out to you, at least twice (possibly more) and every time you refuse to acknowledge this obvious point. You just cling to your belief that changes in cloud cover must result in a violation of conservation of energy, despite the fact that something as trivial cloud cover so obviously shows this claim to be false.
This implies that Spence's
Once again, you abuse the word "chaotically". It has a specific technical meaning.
has no impact on his discussion with AT.
> People misusing technical terms has no impact on a discussion? How ridiculous!
Spence claimed that nothing I've written "has any impact" in the discussion. All my comments so far undermine Spence's vapid claim that "chaos" has a specific technical meaning. The only valid conclusion is one Spence himself finds ridiculous.
Fancy that.
***
> Chaotic systems conform to a number of criteria, and one of those is exponential error growth from initial conditions, which ensures any arbitrarily small error will overwhelm the solution in linear time.
This seems to refer to the definition (SD, which stands for Sensitivity Dependence):
http://plato.stanford.edu/entries/chaos/#ChaDetQuaMec
See section 1.2.5. Defining chaos that way excludes a weaker definition that does not specify any rate of divergence; see (WSD) in the same section. The entry observes that Poincaré's examples satisfied WSD.
Anyway. These definitions are incomplete. But even if we patch them, SD is still problematic:
It is only by turning this into a food fight that Spence managed to prolong that silly thread. He almost succeeded.
***
> The real, interesting debate is that once you realise that the systems are constrained by the same limits of predictability, moving on to determining what those limits of predictability are, which are governed by the time horizon (Lyapunov exponents) and the fractal dynamics.
The Stanford entry, again:
This is the second time we quote that paragraph. This time, we quoted it in full. With our emphasis.
***
No wonder Spence clings to his intuitions. The problem here is that Spence's intuitions leads him to claim things like "Anyone who tries to make the claim that weather is chaotic and climate isn't doesn't understand chaos." Now there's a challenge I can't pass up:
http://www.realclimate.org/index.php/archives/2005/11/chaos-and-climate/
Spence might need to have a word with James Annan, who wrote this paragraph.
***
As an aside, I noticed Dan Hughes' claim:
> If the calculated response was not chaotic, ensembles would not be necessary.
Any travelling salesman would agree, no doubt. But I'm not a travelling salesman. So an explanation might be nice.
This kind of claim is too "complex" for my little head, if you catch my (post-modern?) drift.
Hmm, odd, I answered this last night but it hasn't shown up. Probably an issue my end.
Wow. Another quote keeper. It is, it isn't, I'm not arguing it isn't. What?
There is more than one concept to grasp in deterministic chaos. It is like someone who understands the speed of light is fixed then claims to understand the concept of general relativity - but again there are different levels of "understanding", ranging from Lorentz invariance through to Riemannian geometries. I guess in your own terminology, understanding is not a black/white thing.
Ugh. Albedo is not a boundary condition. Albedo changes due to internal dynamics of the system, and internal dynamics of the system changes albedo. It is a part of the non-linear dynamics, and has strong ties to the hydrological cycle (sea ice, clouds).
A true boundary condition for climate is that albedo cannot be less than zero, and cannot be more than one. Those are actual boundary conditions. "Albedo" as a whole, is not.
Atmospheric composition and greenhouse gases are also not boundary conditions. The quantity of water vapour in the atmosphere, which dominates the greenhouse effect, is a complex function of many variables (evaporation, condensation, precipitation, evapotranspiration). In turn, it affects other aspects of climate (ranging from temperature, to competition between life forms, to geography). So we have two-way non-linear dynamics which means the composition of the atmosphere is absolutely not a boundary condition, but a part of the attractor.
The sun is not a boundary either, it is an external input, and not a constant one. It is not a part of the attractor (since the sun affects climate, but the climate does not affect the sun in any obvious way I can think of).
So what we conclude is that you have arbitrarily re-labeled parts of the phase space of the chaotic system as "boundary conditions", and insisted they are different, even though there is absolutely no qualitative difference between these things and other parts of phase space. Hmm.
But I can go one stage further (and in fact already have). Let's adopt your terminology and show it is logically inconsistent (reductio ad absurdum).
In climate science parlance, the hydrological cycle is not an external forcing, but exists only as internal variability and feedback. You claim albedo is a boundary condition, and that natural variability cannot affect it. Yet cloud cover clearly affects albedo. Cloud cover is changed by all manner of internal variables within the climate system, from the hydrological cycle above (with things like evapotranspiration affected by competition between lifeforms), geography, nucleation, water vapour concentrations which are affected by pressure, turbulence, all manner of internal dynamics which are absolutely not "stable long term averages" (who on earth would describe competition between life as "stable long term averages"!). All of these things continuously change in non-linear and unpredictable ways, not only on a decadal scale, but examples from above vary on centennial scales, thousands, millions of years and more.
Clearly, this shows that cloud cover can change, which changes albedo, which by your misuse of technical terms is a "boundary condition". So we have absolutely unambiguously, internal variability can change the state of the system. You just don't know how much by, but unless you can quantify it, all we can say for now is unambiguously, yes, internal variability can change what you mistakenly call "boundary conditions" by some unknown amount, on pretty much all scales.
It is interesting that I have at least twice raised cloud cover, and you have ignored it twice. This doesn't surprise me. The claim that the hydrological cycle is somehow a feedback (and therefore neither a "forcing" or "boundary condition" as misused by climate activists) is completely at odds with the claim that albedo cannot be changed by internal variability. It is the most glaring inconsistency in the model of climate used by activists and it doesn't surprise me that you've avoided addressing this one because there is no answer in the activist toolbox of arguments.
Wotty is a fool and a bluffer who has no idea what he is talking about. This was illustrated earlier in the thread, when he claimed the climate was not chaotic on longer time scales, and when asked how he knew that by Nullius he warbled about next year's winter! Now he is waffling incoherently about boundary conditions, and it's clear that he does not even understand what these words mean. When this is pointed out, he resorts to shouting and swearing.
Statements such as "if the boundary conditions don't shift, the system can't change state" are complete gibberish. Systems of nonlinear ordinary differential equations, which don't have boundary conditions, just initial conditions, can 'change state' either spontaneously or as a result of a small change in a parameter. I teach this stuff.
Similarly "internal processes are unlikely to change the boundary conditions by themselves" is complete nonsense.
Spence, I admire your patience!
Paul,
Yes, because they DON'T HAVE BOUNDARY CONDITIONS! FFS, this is not a difficult concept! Are you really a mathematician, or is the Paul Matthews at Nottingham someone else and you're just impersonating them?
Quite possibly, but at least I recognise this as a possibility!
No it isn't, because we don't have evidence to suggest that this is possible given the current state of our climate. Again, this is not a difficult concept.
Of course, I'm absolutely wasting my time trying to even engage with you as you have clearly shown yourself to be a remarkably unpleasant individual whose main goal is to undermine anyone with whom you disagree. The tagline on your Twitter account should really be "Science should be open and honest, unless I disagree with it or it challenges my rather narrow world-view". The idea that you can pontificate about scientific integrity and honesty is absolutely laughable.
Spence
I don't know if you are still visiting this thread but I note this comment
I was doing a thought experiment: what would happen to surface temperatures if the earth was completely covered in clouds? would it make any difference if they were high clouds or low clouds? My conclusion was that the increased albedo would mean surface cooling. And of course there would be no SW heating of the oceans so the world as a whole would cool. So understanding clouds must in my view have a profound influence on climate.
I don't know if you have seen this graph posted by lapogus on another thread
http://s1114.photobucket.com/user/Chief_Hydrologist/media/cloud_palleandLaken2013_zps73c516f9.png.html?sort=3&o=16
the correlation with surface temperature appears from eyeballing to be quite good but the interesting thing is the relatively sharp rise in cloudiness just after the year 2000. Is this an example of fractal dynamics in action?
And as Paul says I admire your patience in dealing with ATTP
With Google Scholar requiring all the words; GCM chaotic chaos predictability, and then choosing Since 2014, the following was the first hit displayed.
D. T. Mihailović, G. Mimić, and I. Arsenić, Climate Predictions: The Chaos and Complexity in Climate Models, Advances in Meteorology, Volume 2014 (2014), Article ID 878249, http://dx.doi.org/10.1155/2014/878249
Abstract
Some issues which are relevant for the recent state in climate modeling have been considered. A detailed overview of literature related to this subject is given. The concept in modeling of climate, as a complex system, seen through Gödel’s theorem and Rosen’s definition of complexity and predictability is discussed. Occurrence of chaos in computing the environmental interface temperature from the energy balance equation given in a difference form is pointed out. A coupled system of equations, often used in climate models, was analyzed. It is shown that the Lyapunov exponent mostly has positive values allowing presence of chaos in this system. The horizontal energy exchange between environmental interfaces, which is described by the dynamics of driven coupled oscillators, was analyzed. Their behavior and synchronization, when a perturbation is introduced in the system, as a function of the coupling parameter, the logistic parameter, and the parameter of exchange, were studied calculating the Lyapunov exponent under simulations with the closed contour of environmental interfaces. Finally, we have explored possible differences in complexities of two global and two regional climate models using their air temperature and precipitation output time series. The complexities were obtained with the algorithm for calculating the Kolmogorov complexity.
It’s a free open access article.
This is not your 1980s chaos literature :-)
? You started the thread!
There is a difference between "Salby's ideas might have merit" and "Show me where he is wrong". I have clarified countless times saying the latter does not imply the former. Have never seen such a pathetic lack of understanding.
You were much more reasonable back then - you understood and adopted the line of argument yourself - 'Assuming that Salby is correct, what is the implication on ocean dissolved CO2?'
Did asking this question make you someone 'who thinks [Salby] might have merit'? If the answer is 'No', the same applies to me.
You say: "Salby's ideas are so obviously wrong ..."
If Salby's ideas are so obviously wrong, they must be the easiest to demonstrate as being wrong.
But the caveat is you should understand exactly he is saying and rebut it, instead of your own caricatural simplification of it.
"It's so obvious" is good rhetoric for teenage girls, not scientists.
My impression is that people who say that Salby's ideas are obviously wrong have not always understood exactly what he has said.
While readers have commended and encouraged AndPhysics, this is him talking to himself, even as he commented here:
I don't give up easily as Mr Physics can himself attest but for some reason, I find it hard to take people who use the word 'denier' seriously.
I didn't post in this thread for a while because my posts weren't getting through - I assumed my posting activity on this thread, which is quite different from my normal posting levels, may have triggered a spam filter. But it seems from Andrew's latest post that there were issues with the blog. So I guess it is okay to continue.
Paul, thanks for the comments. I am sure you would do a better job than me (I'm familiar with the issues at hand, but this is not really a specialist subject for me), but I also recognise that the discussion appears to be largely futile since ATTP really does not have a clue about the subject, so I can certainly understand why engaging is not worthwhile. But I find it interesting to see the extent of the understanding of the standard activist lines on this topic.
Willard's latest comment is a good example of this - apparently, willard is continuing to try to defend the idea that a linear reduction in error can outstrip exponential error growth, and that perhaps we are inferring too much from positive Lyapunov exponents for chaotic systems. In which case I look forward to willard's perfect prediction of weather six months from the date of prediction.
Interestingly, willard also links to that article from RealClimate - which I think is an example of exactly the type of argument that willard and ATTP are trying to present. Unfortunately this argument - by analogy of a simple, low dimensional system with Markovian dynamics - is of an exceptional, rather than typical, example of a chaotic system, and real (high dimensional, fractal dynamics) systems, which are more common in nature, behave nothing like this. But I think it is quite informative in the sense, this is where willard and ATTP get their claims from. They are impressed by the presentations like the ones at RealClimate, because it has technical terms they can sort-of relate to, but they do not have sufficient knowledge to see the arguments are full of holes.
When they present the claims to someone more knowledgeable, who asks the questions they had insufficient knowledge to ask, they both get angry because they can't answer them, and resort to the limited options available to them. Shout, swear and search google for an article to quote which doesn't have any bearing on the current discussion. That has to be better than learning something, right?
H2O, yes that is a good example of fractal dynamics (strong presence of low frequency variation). These low frequency fluctuations are every bit as unpredictable as the high frequency variations.
I always think clouds are a good example to give because the activists are trained to tell us the hydrological cycle is not forcing, but feedback / internal variability. So giving them clouds as an example really puts them in an impossible position. Everyone knows that clouds affect albedo, and therefore change energy balance / radiative balance. You know, on a summers day in an air-conditioned building, I want to know how warm it is, I can't use the temperature around me as a guide, so to guess I look out the window to see if it is cloudy. Cloudy will mean cooler. Clear sky will be warmer. The opposite is true at night (when the clouds insulate). This is usually accurate enough to decide if I need my jacket or not. Oddly, I never feel the need to check the current CO2 level.
The fact that clouds have such a profound effect on albedo, but activists cannot declare them to be forcings because it breaks their strange way of labelling things, means they cannot address it head on. I've made this point to ATTP at least three times during this conversation, and every time the point is ignored and ATTP tries to deflect the discussion away from that point.
There is a further interesting point about clouds - as I mentioned above, their impact depends on where they occur within the diurnal cycle. So you can have zero change in average cloud cover, yet a redistribution of cloud cover about the diurnal cycle, which could cause a much bigger change than a 10% increase or decrease in average cloud cover distributed evenly across the cycle.
Spence
thanks for the response
perhaps also worth mentioning that the effect of cloud on temperature throws up one of the weaknesses of using a mean temperature anomaly. As you mention, more cloud at night will have an insulating effect which would mean higher lows producing higher means but unless I miss something it is hard to see how higher lows are going to have us all going to thermageddon. Given that weather is driven by energy gradient, if we were to have a narrowing of the diurnal temperature range then one would assume that more benign weather would follow.
another point if I may. There is much talk about climate sensitivity - if the climate is governed by fractal dynamics is it fair to say that to the extent that climate sensitivity makes any sense it is unlikely to be a constant?