Global warming discussion III

[DC]

We measure the temperature, or the heat (of the oceans), or the atmospheric concentrations. We do not measure the cause. We infer the cause, but we do so making assumptions about natural variability that many scientists have shown may not be correct. These scientists have published in the peer-reviewed literature, and you are unable to show fault in them. That's quite a serious problem for your claim; furthermore, the pause does rather strongly support the claim of long term persistence in natural variability.

And once again, no specific testable claims, no proper cites or quotes, just "the IPCC says so". Yet the IPCC contains no original research, and provides no answers at all to the problems raised by Cohn and Lins (amongst others).


Please tell me which survey that explicitly asked if there was "a problem", which is a claim both you and greybeard have made. I am aware of a number of surveys, but none of them asked if there was "a problem". But I may have missed one. So please tell me which survey this was and how you determined this opinion.

For example, I am aware of the Doran and Zimmerman survey, that included the question: "How convinced are you that climate change, whether natural or anthropogenic, is occurring now?". There was strong agreement with this, but it does not tell us there is "a problem". The Harris interactive poll of AMS asked about consequences, which could be argued to indicate "a problem", but 44% said they expected only moderate effects, and 13% saw relatively little danger - so hardly a resounding consensus on the question of whether there is "a problem".

So specifically - which survey exactly tells us there is a "consensus" that there is *a problem*, which is what you originally claimed.

no they didn't.

41% near catastrophic and 44% moderately dangerous sounds like a pretty stong consensus on a problem.

http://www.usnews.com/news/national...vey-tracks-scientists-growing-climate-concern

 

[Reality Check]

Nowhere in Naturally Trendy are random walks used, invoked, referenced, anywhere.

You may want to read the first paragraph, CoolSceptic:

From time to time, there is discussion about whether the recent warming trend is due just to chance. We have heard arguments that so-called ‘random walk‘ can produce similar hikes in temperature (any reason why the global mean temperature should behave like the displacement of a molecule in Brownian motion?). The latest in this category of discussions was provided by Cohn and Lins (2005), who in essence pitch statistics against physics. They observe that tests for trends are sensitive to the expectations, or the choice of the null-hypothesis.

and later

One difficulty with the notion that the global mean temperature behaves like a random walk...
And, to re-iterate on the issues I began with: It’s natural for molecules under Brownian motion to go on a hike through their random walks (this is known as diffusion), however, it’s quite a different matter if such behaviour was found for the global planetary temperature, as this would have profound physical implications.

The main point of the article is still that climate is a physics driven environment that is better described by physical models than statistical models. Thus Cohn and Lins's argument is moot.

 

[Reality Check]

Obviously the claim that the University rooftop station was recording 113 days is actually insane, but by looking at the data...

CoolSceptic, you might take this claim with a grain of salt since r-j has not been able to produce this so far imaginary data after a couple of weeks:
r-j (8 September 2014): Please cite the source you used to check the temperature records of from 1930 to about 1950 (1930's and 1940's) for the weather station mentioned in the [URL="http://www.realclimate.org/index.php/archives/2014/03/new-daily-temperature-dataset-from-berkeley/"]RealClimate article.[/URL]

r-j also cherry-picked the beginning of his data in a graph in order to get the trend that he wanted (and left the missed out data in the graph to make this obvious!):
r-j (19 September 2014): Please justify what seems to be an arbitrary selection of 1995 as the starting date of the trend line when there is pre-1995 data in the graph.

 

[Reality Check]

Moving a weather station from a cool park-like area, the University which is next to a huge lake, to a huge busy modern airfield, changed not only the highs, but the lows.

Which is what did not happen according to real climate scientists - the O'Hare weather station came from the top of a black roofed building:
New daily temperature dataset from Berkeley

For example, if you were studying the history of extreme heat in Chicago, you would find a slew of days in the late 1930s and early 1940s where the station currently at the Chicago O’Hare airport reported daily max temperatures above 45 degrees C (113 F). It turns out that, prior to the airport’s construction, the station now associated with the airport was on the top of a black roofed building closer to the city.

And really bad links:
The official all time hottest day for Chicago is 105, from July 24 1934 is the Tuesday, July 24, 2012 of the weather station. The URL has KORD in it and this is Chicago O'Hare, IL. On July 24 1934, O'Hare did not exist .
P.S. The data on the site for KORD starts at 1946.
KMDW is Bedford Park, IL, i.e. Chicago Midway, IL. Again for Tuesday, July 24, 2012.

The O'Hare station uses the O'Hare data!
The Midway station uses the Midway data!
Which station if any is the "official" station used for current temperatures does not matter. We are talking about temperatures of a specific weather station before and after it was moved.

ETA: It looks like there an actual modern University weather station:
U Chicago Hyde Park, Chicago, IL weather station located at Latitude: 41° 47.40' N Longitude: 087° 35.96' W Altitude: 690 feet
Chicago-O'Hare International, IL weather station located at Elev 205 m 41.99 °N, 87.93 °W
So one wonders what this "University" station that r-j goes on about actually is?

 

[CoolSceptic]

You may want to read the first paragraph, CoolSceptic:

All you've done is highlight another mistake by realclimate.

Nowhere does the actual article by Cohn and Lins (which I linked to, so you have no excuse for not reading it) mention random walks, or use the statistical properties of random walks.

The fact that realclimate incorrectly believe Cohn and Lins is about random walks, merely highlights the fact that realclimate didn't read the paper properly, either.

So, once again. Please show me where Cohn and Lins use or refer to a random walk. Please do not show me third party straw men which claim there is a random walk. Show me *in the actual publication*.

The main point of the article is still that climate is a physics driven environment that is better described by physical models than statistical models. Thus Cohn and Lins's argument is moot.

Internal natural variability is sensitive to initial conditions, cannot be captured by deterministic models, so is necessarily a statistical model. However, the model is firmly grounded in physics. Once again, that realclimate errs on this topic does not make it true.

The underpinning physics is explained in more detail in this article:
http://www.sciencedirect.com/science/article/pii/S0378437111000148

 

[ideogram]

So specifically - which survey exactly tells us there is a "consensus" that there is *a problem*, which is what you originally claimed.

41% near catastrophic and 44% moderately dangerous sounds like a pretty stong consensus on a problem.

http://www.usnews.com/news/national...vey-tracks-scientists-growing-climate-concern

Is it asking too much to expect you to acknowledge error?

 

[CoolSceptic]

no they didn't.

41% near catastrophic and 44% moderately dangerous sounds like a pretty stong consensus on a problem.
[/url]

Hardly - especially as 13% said "little danger". But it is virtually impossible to properly assess the survey without the actual question asked. The paper indicates that scientists were asked to assess the risk from 1 (trivial) to 10 (catastrophic). But how were the key scores (4-7) defined? What do they subjectively mean?

I presume this information is in the supplementary information, but it requires a log in (tut! most journals do NOT require a log in for SI data).

The term "moderate danger" (which was used in the paper, not "moderately dangerous" as incorrectly reported in the press) is conflicted, and I think without further information is almost impossible to assess. The other thing that is not explained is why a scale of 1-10 was grouped the way it was. Why ask for a response of 1-10 and then reduce to three groups?

If you could furnish us with the exact question asked, and the actual distributions from 1-10 (which I would hope would be in the SI) then that would allow a reasonable understanding of the results. Without that, we have too little information to understand what this broad category (4-7) of conflicted "moderate danger" is actually supposed to mean.

 

[ideogram]

44% said they expected only moderate effects

This is wrong.

 

[Reality Check]

All you've done is highlight another mistake by realclimate.

No - I am highlighting that you did not read the words in that RealClimate article, CoolSceptic. It does contains mentions of random walks.

There is no statement in the Naturally trendy? article that Cohn and Lins mention random walks in their paper. What they do is use statistical models when physical models are better.

 

[CoolSceptic]

Is it asking too much to expect you to acknowledge error?

DCs original source stated "moderate effects" - the press release he quoted stated "moderate danger". So we need to understand which of the two sources is more accurate.

The original paper and SI are paywalled, so it makes getting to the truth more difficult. I found a no-pay copy from google but this just confuses the matter further; it doesn't state the question asked, it doesn't define the values 4-7, but it does classify these as "moderate danger".

So at the moment there is conflict between the different sources of information. The best way to resolve this is with two pieces of information: (1) the original question asked, and whether the survey respondents would have unambiguously classified 4-7 as "moderate danger", and (2) the exact breakdown of survey results into 1-10 rather than the three-way split.

 

[CoolSceptic]

No - I am highlighting that you did not read the words in that RealClimate article, CoolSceptic. It does contains mentions of random walks.

Which has absolutely no bearing on the discussion at hand. Other than usefully showing errors on the part of the realclimate authors.

There is no statement in the Naturally trendy? article that Cohn and Lins mention random walks in their paper.

Exactly. Which is why your comment about random walks is a straw man.

What they do is use statistical models when physical models are better.

Luckily I have already provided you with a peer-reviewed paper showing that the model is indeed a physical model. So we understand you and realclimate are incorrect about this, too.

 

[ideogram]

Why yes, more information about the survey would be very helpful.

Do you acknowledge your error?

 

[Reality Check]

So specifically - which survey exactly tells us there is a "consensus" that there is *a problem*, which is what you originally claimed.

As you know 97% of climate scientists agree that AGW is happening.
The science is that this is a problem for low estimates of the future warming and a catastrophe for high estimates(Positives and negatives of global warming) but how many of those scientists agree with this is unknown. But since most of them can read, it is probably a high % !

 

[Reality Check]

Which has absolutely no bearing on the discussion at hand.

Once again - it was a comment about your ability to read, CoolSceptic.
You stated that the article contains no random walks when it did and still does.

You seem mistaken about the Hurst–Kolmogorov dynamics as a result of extremal entropy production paper and its physical connection to climate.
The physical source of entropy production for climate is the Sun. The Sun is not a physical cause of the current global warming (output has been constant for about 35 years).
As far as I can see, the entropy in the paper (to be more exact this preprint) is not the physical entropy - it is the entropy of the data being analyzed.

 

[CoolSceptic]

Once again - it was a comment about your ability to read, CoolSceptic.
You stated that the article contains no random walks when it did and still does.

My exact words were: "Nowhere in Naturally Trendy are random walks used, invoked, referenced, anywhere.". By "Naturally Trendy", I was referring to the article by Cohn and Lins, which is what I was discussing. So, my statement was quite correct. It is unfortunate that the realclimate article you were talking about had the same name, you incorrectly assumed I was just talking about the article you had linked. So we have a simple miscommunication where I stated one thing correctly, and you (reasonably) misinterpreted it to mean something else.

However, once we put the miscommunication to one side, we find that the situation remains the same: the Cohn and Lins article "Naturally Trendy" does not use, invoke or refer to random walks. Realclimate's critique, also called "Naturally Trendy" is therefore incorrect to refer to it in the context of that paper. The realclimate article contains an error, and is a strawman.

Unfortunately, you seem to be obsessively focussed on the miscommunication, rather than the substance of the debate. Which reminds me of why I have so little patience with people on this thread.

You seem mistaken about the Hurst–Kolmogorov dynamics as a result of extremal entropy production paper and its physical connection to climate.
The physical source of entropy production for climate is the Sun.

Wow, no. This is in the "not even wrong" category. Every process in nature has some calculable rate of entropy production. This is such a fundamental misunderstanding I'm not sure where to go from here.

 

[Skeptic Ginger]

I do believe this is significant: China's per capita carbon emissions overtake EU's

The biggest emitters were China, which produced 29% of the total, followed by the US at 15%, the EU at 10% and India at 7.1%

But in an interesting development, China's emissions per head of population exceeding those of the European Union for the first time.

While the per capita average for the world as a whole is 5 tonnes of carbon dioxide, China is now producing 7.2 tonnes per person, to the EU's 6.8 tonnes. The US is still far ahead on 16.5 tonnes per person.

 

[Reality Check]

My exact words were: "Nowhere in Naturally Trendy are random walks used, invoked, referenced, anywhere.".

That explains it: You did not read the title of the paper which was Nature's style: Naturally trendy !
Or see that I pointed to the RealClimate article Naturally trendy? (with a question mark).

The RealClimate article Naturally trendy? is never mentions random walks in the context of the paper contents. They refer random walks as general usage in statistics.

Oh wow, CoolSceptic - still have the reading comprehension turned off ?
There is physical entropy produced in physical activities and informational entropy contained in information. Thus

You seem mistaken about the Hurst–Kolmogorov dynamics as a result of extremal entropy production paper and its physical connection to climate.
The physical source of entropy production for climate is the Sun. The Sun is not a physical cause of the current global warming (output has been constant for about 35 years).
As far as I can see, the entropy in the paper (to be more exact this preprint) is not the physical entropy - it is the entropy of the data being analyzed.

The big clue to those in the know is the name Kolmogorov.

 

[CoolSceptic]

Why yes, more information about the survey would be very helpful.

Do you acknowledge your error?

What error?

I said I was unaware of any surveys that supported the claim about "a problem", from DC's original source. I asked DC to provide me with some. Note: I did not claim none existed, I stated I was unaware of any. These are quite different claims and I'm not sure you understand the difference. DC has provided a different source for the same survey which used slightly different language. It is possible that this does support his original claim, but I would like one or two further clarifications first.

On the topic of consensus vs. no consensus, it appears that the question was a simple request from 1-10. I need clarification of whether the scale had better definition.

The split was into 1-3 / 4-7 / 8-10, or a 3 / 4 / 3 split. So if no guidance was given (not I do not know if this is the case - but just speculating if it was), we can see that a random number generator would yield a split 30 / 40 / 30.

A true consensus would be 100 / 0 / 0, or 0 / 100 / 0, or 0 / 0 / 100.

The actual results were 13 / 44 / 41. That is actually rather much closer to the non-consensus 30 / 40 / 30 than it is to any consensus.

So it is not surprising that press release headlines stated things like "Few now doubt global warming, though they disagree on the severity of the danger." Because the results really don't support consensus at all.

http://www.usnews.com/news/national/articles/2008/04/23/survey-tracks-scientists-growing-climate-concern

Also, this is an actual quote from the paper:

In short, the demonstrated consensus over anthropogenic warming does not preclude continued debate among scientists over the dangers that this development portends for the future of the planet.

It maybe provides incredibly weak support for consensus that there is "a problem", but it has many caveats that undermine that position as well.

 

[W.D.Clinger]

You may have missed the main point of the Naturally trendy? article, CoolSceptic.
It is that it is quite wrong to assume that the statistics of random walks apply to a system with physical forcing.

Nowhere in Naturally Trendy are random walks used, invoked, referenced, anywhere.

LTP is not a random walk. If you don't understand something so basic about stochastics, you have no hope whatsoever of understanding the work of Cohn or Koutsoyiannis.

Nowhere does the actual article by Cohn and Lins (which I linked to, so you have no excuse for not reading it) mention random walks, or use the statistical properties of random walks.

The fact that realclimate incorrectly believe Cohn and Lins is about random walks, merely highlights the fact that realclimate didn't read the paper properly, either.

I think CoolSkeptic is right about this, to approximately the same degree that the RealClimate blog author cited by RealityCheck was right to say "We have heard arguments that so-called 'random walk' [sic] can produce similar hikes in temperature" and to place similar claims about AR, ARMA, ARIMA, and FARIMA models in the historical context of claims about random walks.

Here's the main thing to understand about what's going on here: To reject random fluctuations as an explanation of apparent trends, you have to select some statistical model of the kind of random fluctuations you're taking as your null hypothesis. The more complicated the class of statistical models you select for your null hypothesis, the more difficult it becomes to reject the null hypothesis at some standard threshold of significance.

As Cohn and Lins say in their paper:

Choice of trend test, however, does matter when computing trend significance. The simplest test, T_β,{0,0,0} (which assumes no LTP), finds strong evidence of trend, a p-value of 1.8 * 10⁻²⁷. T_β,{ϕ,0,0} (which allows for short-term persistence) yields a p-value of 5.2 * 10⁻¹¹, 16 orders of magnitude larger and still highly significant. The p-value corresponding to either T_β,{0,d,θ} or T_β,{ϕ,d,0}, an unadjusted LRT trend test that considers both short-term and long-term persistence, is about 7%, which is not significant under the null hypothesis. In changing from one test to another, 25 orders of magnitude of significance vanished.

In other words, those who wish to deny the statistical significance of climate trends will always be able to postulate a statistical model that's capable of such complex behavior that a null hypothesis based upon that statistical model cannot be rejected at customary levels of confidence.

Cohn and Lins are troubled by this:

This result is somewhat troubling given the uncertainty about the stochastic process and consequently about which test to rely on.

That doesn't bother me, because I have enough knowledge of physical sciences to realize that many physical phenomena can be simulated using sufficiently complex mathematics.

That is not to say all mathematical models of physical phenomena are equivalent. Some are more accurate than others, while some may be accurate enough but lack predictive power because their free parameters are data-dependent rather than universal.

I believe the RealClimate blogger is trying to say the more complicated classes of statistical models that are being used to deny the statistical significance of apparent climate trends acquire their requisite variety of behavior by adding data-dependent parameters, which is unsatisfactory on scientific grounds. I don't think the RealClimate blogger said it very well, but I think that's roughly what he was trying to say.

I think Cohn and Lins said it better in the last sentence of their abstract:

From a practical standpoint, however, it may be preferable to acknowledge that the concept of statistical significance is meaningless when discussing poorly understood systems.

I also think Cohn and Lins would have made that point more effectively had they not devoted their discussion section to an attempt to spin that valid, practical, and (statistically) fairly obvious point into a claim that "These findings have implications for both science and public policy."

 

[ideogram]

What error?

44% said they expected only moderate effects

This is wrong.