backtop

Print E-mail del.icio.us 104 comment(s) - last by Shuxclams.. on Sep 4 at 6:35 PM

Satellite imagery of El Nino currents in 2006  (Source: NASA)
Carbon-dioxide out; "synchronized chaos" in

"Arctic Ocean Getting Warm; Seals Vanish and Icebergs Melt" --Washington Post headline, November 2, 1922.

If there was any doubt that fear-mongering has long been cherished by the media, the above headline should put the question to bed. But that 80-year old news story also illustrates two of the great problems for the global warming theory -- its inability to explain sudden climate shifts in the Earth's past, and to explain why the Northern and Southern Hemispheres are so unequally affected by warming. 

A team of mathematicians have come forth with a startling new theory that solves both these problems. Led by Dr. Anastasios Tsonis, their model says the known cycles of the Earth's oceans -- the Pacific Decadal Oscillation, the North Atlantic Oscillation, El Nino (Southern Oscillation) and the North Pacific Oscillation -- all tend to try to synchronize with each other. 

The theory is based on a branch of mathematics known as Sychronized Chaos.  The math predicts the degree of coupling to increase over time, causing the solution to "bifurcate," or split. Then, the synchronization vanishes.  The result is a climate shift.

Eventually the cycles begin to sync up again, causing a repeating pattern of warming and cooling,  along with sudden changes in the frequency and strength of El Nino events.

Better yet, their theory has predictive power. The model predicts past shifts in the year 1913 (explaining the strong warming of the 20s and 30s), 1942 (resolving the post-WW2 cooling trend) and 1978 (covering our current warming). The model predicts another shift to occur around the year 2033. Most shocking of all is their prediction for the year 2100 to be slightly cooler than present day, despite the assumption of a doubling of atmospheric CO2 levels.  Eye-popping indeed.

Is carbon-dioxide really so ineffective at warming? A new study by Belgium's Royal Meteorological Institute seems to think so. Its conclusion is that, while CO2 does have some effect, that "it can never play the decisive role attributed to it" in global warming, and that its effects have been grossly overstated.

Comments     Threshold


This article is over a month old, voting and posting comments is disabled
Predicting the past does not a theory make
By Kuroyama on 8/14/2007 8:34:37 PM , Rating: 2
History is littered with theories which accurately predicted the past but turned out to be useless for future predictions. In fact, it's not all that hard to make any model you like fit historic data. Based on a quick Google search it is clear that going back 10+ years Dr Tsonis has been a global warming skeptic, so he is certainly not an impartial observer. Of course this does not discredit his views, for instance the mathematician Serge Lang was a very vocal critic of what he considered to be shoddy research into the connection between HIV and AIDS up until he died recently, and no one really questions his intelligence (although he could be quite a jerk).

It'll be interesting to see what these papers have to say. Although my specialty is not dynamical systems, chaos, or atmospheric science, I do have a PhD in a closely related field so I figure I have as much a chance as most of making out the mathematical integrity of the many pages on his web page (will report back in the next day or so).




By Zurtex on 8/14/2007 10:27:07 PM , Rating: 2
I was skimming over this, the odd reference to topological terms and what seemed like topological terms. It's a bit new to me when dealing which such applied mathematics (I mainly keep my head in pure, but I'm only starting a Masters).

A particular phrase got me a little curious that he has on his front page of his web site:

"Total number of links (connections) at each geographic location."

When he says connections, is he speaking of the differential geometry term, some geo-physicist term or some more relaxed informal "English style" definition?

I leant about connections last semester, and them being generalizations of exterior differentiation to vector-valued differential forms, so I'm still getting a bit of a buzz off finding related mathematics to it that I can understand. So I'm not too bothered if I don't get an answer, probabily look it up myself at some point.


RE: Predicting the past does not a theory make
By Kuroyama on 8/14/2007 10:53:25 PM , Rating: 4
I have to lend my voice to those who say that Masher is mis-representing the Geophysical Research Article. There have been plenty of comments on specific points (graphs, etc), so I'll stick to an overall explanation of the paper instead.

Here's my understanding of it. The article considers changes in 4 variables (basically certain measurements of oceans). It considers what happens when they are synchronized (i.e. basically correlated) and/or coupled (i.e. movement of one explains movements of the others). These are not the same because as we all learn in intro to statistics, in a non-linear system things can be highly coupled but have zero correlation (the classic example being that fuel efficiency of your car depends in a clear way on its speed, but yet they can have zero correlation).

Anyways, the authors of this paper claim only that when the variables become synchronized, and then more and more strongly coupled, then this is followed by a sudden change in climate behavior. They find that without the increase in coupling then the climate change doesn't occur.

In layman's terms, look at the stock market. There is a general trend, that in the long term prices go up. Beyond this things are fairly random. However, if we look at local fluctuations (say over 7-15 years as in this paper), when we notice an increasingly strengthening trend of everyone saying "buy buy" or "sell sell" (i.e. synchronized and increasing coupling) then it is quite likely that a sudden change in the market is likely to occur. Happened at the end of the bubble in Japan, end of the dot-com boom in the US, and may be happening in the US right now with real estate.

Interesting stuff. But this does not mean that the overall trend is wrong. It may be that a market transition goes from increasing prices, then suddenly jumps to a period where past models are invalid and prices drop, then suddenly they are increasing, then suddenly stable, etc. Each time the synchronize + coupling effect preceded the market change, but it may still true that the overall trend is an increase in prices.

If we want to conclude anything from this it is that regardless of the underlying trend in global temperature changes (existent or non-existent), we should expect eras that buck the overall trend entirely. Many of you would suggest that the current "global warming" era is the one bucking the trend, but it might be the "global cooling" era. Realistically we need to look at data spanning the longest possible time period in order to try to discern the underlying trend around which these miscellaneous fluctuations are occurring. This seems to be a warming trend (whether anthropogenic or otherwise), and this paper seems to agree on that point at least.

Last but not least, the suggestion about 2100 is ludicrous. If we believe this paper then after the 2035 turning point the current paradigm will basically become broken, and at least as far as short-term predictions are concerned then anything after that would be speculation. To repeat a point made in previous paragraphs, long-term predictions are a different matter and this is primarily concerned with the behavior of the randomness / chaos in the system.

One last illustration of the idea. As long as you can look around and easily find "flaws" in the global warming idea (eg. more ice in Antartica bucking the trend in the Arctic) then things are likely to stay in the post-1970's paradigm (gradual temperature increases), but when it starts to feel like "ALL trends point towards warming" then it may be the time to bet your money against this. But climate change proponents haven't really said otherwise; witness all the claims that warming in Greenland might cause a flood of fresh water and a sudden cooling in Europe (or even in NYC as the silly movie "The Day After Tomorrow" suggested).


RE: Predicting the past does not a theory make
By masher2 (blog) on 8/15/2007 12:56:48 AM , Rating: 2
> "Interesting stuff. But this does not mean that the overall trend is wrong."

Obviously, the purpose of this research was not to disprove global warming by itself, but rather simply to prove that the past climate shifts (particularly the one beginning in the mid-1970s) are explainable without resorting to rising CO2.

That said, when one removes the forcing given in this model, the the justification for CO2 does become considerably weaker. If nothing else, the rate of the long-term warming is considerably less steep.

> "If we believe this paper then after the 2035 turning point the current paradigm will basically become broken"

I'm not sure why you say this; the paper specifically identifies bifurcations occurring at both 2033 and 2073. The model doesn't end at 2035.


RE: Predicting the past does not a theory make
By Zurtex on 8/15/2007 1:08:14 AM , Rating: 2
I've not done bifurcations properly since year 1 of my BSc in mathematics (did a quirky coarse on the fundamentals of chaos theory). But it's essentially when you have a graph split off in to multiple directions (namely 2 directions). So if a bifurcation of predicting weather climate occurs 2035, then the environment can take different paths at that point right? (I may be seeing this completely wrong)

So if this does occur in the environment, then any prediction after that would not really work, because there is a potential for your prediction relying on one thing in 2035 and it doing entirely the opposite.

Given that Dr. Anastasios Tsonis has one of his articles linked to on the front page entitled "Is global warming injecting randomness into the climate", he's probabily not fundamentally opposed to the idea of global warming.


RE: Predicting the past does not a theory make
By masher2 (blog) on 8/15/2007 1:17:52 AM , Rating: 3
> "But [a bifurcation is] essentially when you have a graph split off in to multiple directions...if a bifurcation of predicting weather climate occurs 2035, then the environment can take different paths at that point right?"

More generally in chaos theory, a bifurcation is a period doubling that indicates the onset of a chaotic response. It doesn't imply the climate model has two solutions, neither of which can be chosen.

> "Given that Dr. Anastasios Tsonis has one of his articles...entitled "Is global warming injecting randomness into the climate", he's probabily not fundamentally opposed to the idea of global warming"

To Dr. Tnonis, global warming means "a global warming trend" . It does't mean what it does to the general public, which is something more akin to "a warming trend caused by anthropogenic greenhouse gases, which will continue and eventually lead to catastrophe".

If you read the paper you mention above, you'll see it contains no references to CO2, anthropogenic or otherwise.


By Zurtex on 8/15/2007 1:32:51 AM , Rating: 2
> If you read the paper you mention above, you'll see it contains no references to CO2, anthropogenic or otherwise.

I did read it, and yes it was a look at "what if a warming trend continues" it didn't question the whys, as far as I skim read it anyway.

> More generally in chaos theory, a bifurcation is a period doubling that indicates the onset of a chaotic response. It doesn't imply the climate model has two solutions, neither of which can be chosen.

Indeed, my statement, was completely wrong. But never the less, the general idea of it being very difficult to predict anything about the climate after that becomes very difficult because of the bifurcation was sorta correct, no?

What was coming to mind was graphs like this:

http://en.wikipedia.org/wiki/Image:Chaosorderchaos...

The bits where the line splits apart is where the bifurcation occurs, which indicates, as you say, a period doubling. When tends to chaos (though not always and kind of depends how you define chaos).


RE: Predicting the past does not a theory make
By Zurtex on 8/15/2007 1:10:41 PM , Rating: 3
> If you read the paper you mention above, you'll see it contains no references to CO2, anthropogenic or otherwise.

I'd just like to comment on this again. This is a discussion in a mathematical sense of what happens to prediction of the global climate if there is a general trend of warming, why would there be any references to CO2? It's not a chemist or geographers papers as in trying to determine a cause of a global warming. It's a straight mathematical discussion, with an assumption clearly started at the start and then an attempt to try and see what that means, it's all about numbers and models not chemicals and natural/human influences.

Reading through this stuff has reminded me what I like applied mathematicians more than theoretical physicists, they're still all about the numbers, not the words. (Once the problems and assumptions have been clearly stated).

I find it worrying because you see a paper about global warming, and it doesn't mention CO2 or human influence anywhere, you then imply that it backs up your argument that any global warming has nothing to do with CO2 or human influence.


By onelittleindian on 8/15/2007 1:36:04 PM , Rating: 2
quote:
I find it worrying because you see a paper about global warming, and it doesn't mention CO2 or human influence anywhere, you then imply that it backs up your argument that any global warming has nothing to do with CO2 or human influence.
I can't believe you typed this with a straight face. He was implying nothing of the sort, it was an obvious rebuttal to your ridiculous post that Dr. Tnonis believes in CO2-based climate change simply because he wrote a paper with the phrase "global warming" in it.


RE: Predicting the past does not a theory make
By Zurtex on 8/15/2007 1:50:46 PM , Rating: 2
:D

This is what I said:

> "Given that Dr. Anastasios Tsonis has one of his articles...entitled "Is global warming injecting randomness into the climate", he's probabily not fundamentally opposed to the idea of global warming"

This is what he said:

> "If you read the paper you mention above, you'll see it contains no references to CO2, anthropogenic or otherwise."

Who was the first to mention CO2? It certainly wasn't me, get your facts straight rather than just plain mis-representing me and doing exactly what you claim I'm doing.


By onelittleindian on 8/15/2007 2:00:03 PM , Rating: 2
You forgot to quote the paragraph above that one, which clearly states his reasons for the CO2 reference:
quote:
To Dr. Tnonis, global warming means "a global warming trend" . It does't mean what it does to the general public, which is something more akin to "a warming trend caused by anthropogenic greenhouse gases, which will continue and eventually lead to catastrophe".
Obviously Tsonis isn't "fundamentally opposed" to the earth getting warmer (the whole point of his latest paper is to explain that by ocean currents, remember?).

But you were trying to make it sound like he believed in human-causeed warming from greenhouse gases. You were wrong, and you got called on it.


RE: Predicting the past does not a theory make
By Zurtex on 8/15/2007 2:10:29 PM , Rating: 2
No, I really wasn't trying to make it sound like that at all. If you misinterpreted what I said, fair enough, I apologize for not making it clear enough.

If I was trying to:

> make it sound like he believed in human-causeed warming from greenhouse gases.

I would not type things like:

> why would there be any references to CO2? It's not a chemist or geographers papers as in trying to determine a cause of a global warming. It's a straight mathematical discussion, with an assumption clearly started at the start and then an attempt to try and see what that means, it's all about numbers and models not chemicals and natural/human influences.


By onelittleindian on 8/15/2007 2:41:18 PM , Rating: 2
Ok, so what WERE you trying to say? If you really intended nothing more to than let us know Tsonis believes the earth did warm some, then you were simply making a redundant, blatantly obvious non sequitur that overlooks the entire point of this article, has no relevance to the post you replied to, and doesn't advance, support, or even counter any of the arguments you or anyone else in the thread made.


RE: Predicting the past does not a theory make
By masher2 (blog) on 8/15/2007 2:51:34 PM , Rating: 2
Let's put this argumentative thread to bed right now. I certainly didn't intend to suggest that, simply because one scientist failed to mention CO2 in one paper, that the entire theory of GHG-based warming was suddenly disproven. And I accept that Zurtex was not trying to suggest that Dr. Tsonis is a supporter of human-induced global warming.


By Zurtex on 8/15/2007 3:02:09 PM , Rating: 2
Yeah, I was just replying to onelittleindian but I think I'll let this die of death.

I really don't have much of an opinion on whether or not the theory of GHG-based warming is true or not. I just get annoyed when people skew things horribly out of proportion. Reading back on your other posts, I see you've not done this as much as I assumed. But here in England, it's pretty widely accepted by the government (whatever the political party) that GHGs are a problem and they need to be tackled, so I do find it odd to meet someone who is willing to so strongly argue the opposite.


RE: Predicting the past does not a theory make
By Kuroyama on 8/15/2007 2:03:28 AM , Rating: 5
Regarding post 2035 stuff, my reading of the paper is that they used the SRESA1B model to predict upcoming changes in the 4 variables they measure, and then checked whether an "event" would occur. However, without knowing what the paradigm shift will be then I do not believe they can say reliably say anything after 2035. I would conjecture that they plot the highest probability paradigm shift, but that due to the inherent uncertainty at this phase transition of sorts, I think any subsequent predictions should not be taken very seriously.

As for what the authors are doing in this paper, the concluding remarks (i.e. CO2 stuff) were not the reason this was published; the purpose was to demonstrate that the network approach to complex systems can be used for something as complex as climate systems.

Anyways, although the authors state in the conclusions that this suggests the 1970's climate change event need not have been "due to greenhouse gases overcoming shortwave reflection effects due to aerosols", it appears that they identify only a means of identifying when a sudden change will occur, and don't really delve into whether some external event such as greenhouse gases was not in fact the cause of the synchronization and coupling that occurs before the climate change.

To revert to the stock market analogy, the popping of a bubble is not caused by some arbitrary chaotic event, but is typically due to excess optimism leading to extreme valuations, which in turn causes high synchronization and coupling when humans follow a herd mentality. Can the methods of this paper be used to show that stock market "events" are explainable without resorting to using say the "herd mentality" explanation of human behavior? Not really, this is only identifying a sign of when the "events" will occur and says little to nothing about the causes.

Well, I find this all interesting enough that I'll try reading through the longer papers tomorrow. The 1910 vs 2035 line in the conclusions seem to suggest the authors believe they have said more than I am giving them credit for, although I don't yet see how their implicit suggestion is possible.