Click images for more details



Recent comments
Recent posts
Currently discussing

A few sites I've stumbled across recently....

Powered by Squarespace
« Happy Christmas! | Main | US usurps EU's role of climate fool »

Seitz is no guarantee

Readers are no doubt familiar with Harvard physicist Russell Seitz, a frequent commenter in these parts. If so you may be interested in an email I received today:

Take a look at this 1990 article by Russell Seitz, placed online recently here. It's colourfully written, but ironically it sets out a sceptic position rather well. Does this sound like something that you might have written?

A disturbing reality confronts us:  the deliberate creation of a double standard, with one set of facts for internal scientific discourse and another for public consumption.

On whether CO2 is a "big" problem:

Clearly, a sharp-toothed carnivore is on the prowl. But we've yet to see a full-grown specimen.  Are we dealing with Snoopy or Cerberus? It's hard to tell- it's only just a foundling pup, and the question of its diet remains to he wrestled with-it might grow into either. But grow it will-slowly, and for a long while undetectably. One of these centuries, we're going to have a real dog in our front yard. But what kind?  And when?  An interdisciplinary consensus on the magnitude of the "greenhouse effect" and its impact on sea levels in the next century won't come cheap-or soon.

On activists and scientists:

On CO2, some [scientists] have cast objectivity aside and openly made common cause with the eco-politicians. The salvation of the world affords an enchanting pretext for those predisposed to societal intervention. They have already raised the abolitionist banner, pointing to the prospect of Bangladesh awash and water  skiing down the Mall to the Capitol-a prospect no more likely in my lifetime than nothing happening.

On nuclear:

Rather than embarking down the soft energy path that leads back beyond the Industrial Revolution's roots into a future dark age, the Greens should pause to consider the effect on the environment of renewing and perfecting our mastery of the atom's pale fire.  The prospect of nuclear power's second coming presents environmental millenarians with a real source of cognitive dissonance: it is they who are the problem. It is their delaying tactics that wasted years and squandered billions at Seabrook and elsewhere. And it is their past indifference to the environmental consequences of the fossil fuel that the reactor might have saved that makes a mockery of their present rhetoric.

PrintView Printer Friendly Version

Reader Comments (233)

You mean kind of like Meteor strike insurance just in case one hits ones car? One can't ever rule out the event taking place with confidence therefore its better safe than sorry right?



Dec 31, 2015 at 1:01 PM | Unregistered CommenterMailman

Phil Clarke

I see you posted Gavin Schmidt's blog post response to Demetris Koutsoyiannis' peer-reviewed paper. Nothing wrong in principle with a blog post response, but one should always check to see if errors are made (whether a paper is peer-reviewed or a blog post). Of course, because Gavin failed to document his criticisms in a peer-reviewed paper, it means there is no peer-reviewed response to his article, but there is a blog response by professor Koutsoyiannis over at climate audit here:

Prof Koutsoyiannis responds to Gavin Schmidt

The first thing to note is that Gavin makes a number of basic errors in understanding Hurst-Kolmogorov processes. You need to understand the basics before responding, and Gavin clearly has not sufficiently understood the Hurst phenomenon yet. As such, his criticisms are without merit. (See Demetris' response for more details)

I have, for some time, been trying to bring the Hurst phenomenon to the attention of climate scientists with relatively little success, however Demetris did make some progress in IPCC AR5 which acknowledges the presence of the Hurst phenomenon in climate, and acknowledges that climate scientists have no idea what impact Hurst-Kolmogorov processes will have on our ability to predict climate. To be fair, I think this is an accurate characterisation of Gavin's ignorance on the topic. It is not an accurate characterisation of the papers by Prof Koutsoyiannis, which show conclusively that climate models fail to capture these dynamics, and therefore fail to capture one of the most fundamental aspects of climate correctly.

Getting climate scientists to understand this will doubtless take a considerable time. In my experience they are very reluctant to accept new ideas outside their narrow experience, even when evidence for these things is as clear as it is for Hurst-Kolmogorov dynamics.

Dec 31, 2015 at 1:02 PM | Unregistered CommenterSpence_UK

Phil Clarke - please clarify: can statistics inform or not on the agw theroy? If so, how precisely? If not, then do you agree that it is a theory that cannot be falisfied?

Extracts from up thread:
1) lack of any definition from the proponent side for the falsification criteria for agw

Statistically significant failure of the planet to warm in response to the radiative imbalance would do it. Not happening so far.

Dec 30, 2015 at 5:41 PM | Unregistered CommenterPhil Clarke


I suggest you take your arguments up with these guys rather than me:

Why would I? The first is a discussion of statistical models of temperature time series used by the Met Office, therefore it has 'no bearing on our understanding of the climate system or of its response to human influences such as greenhouse gas emissions', to quote the paper itself.

Dec 31, 2015 at 3:03 AM | Unregistered CommenterPhil Clarke

Dec 31, 2015 at 2:15 PM | Unregistered Commenternot banned yet

Phil C,

As I said, you don't understand what you link said and are merely repeating talkning points. And trying more of those when called out ...

You first introduced the concept of 'forcing' as a possibility to 'falsify' your pet theory. However, since this is merely a concept (a hypothesized unmeasurable metric, that the the AGW-hypothesis hopes is an accurate description of causation) it too must be established first. And for it to even be a viable scientific hypothesis it too must be falsifiable ..

But hey, I guess this is way over your head now. Better just to retort 'because Gavin said so on his blog ... '

As I said, this pretty much is the level I expect. And the responsens too after just one pointed question.

As I expected,you hadn't even read your own link. But your are right, rehashing memoried

Dec 31, 2015 at 4:38 PM | Unregistered CommenterJonas N

Russell has gone dark? Perhaps he has seen the errors of his ways?


Dec 31, 2015 at 5:03 PM | Unregistered Commentermailman

I see you posted Gavin Schmidt's blog post response to Demetris Koutsoyiannis' peer-reviewed paper. Nothing wrong in principle with a blog post response, but one should always check to see if errors are made (whether a paper is peer-reviewed or a blog post). Of course, because Gavin failed to document his criticisms in a peer-reviewed paper, it means there is no peer-reviewed response to his article […]The first thing to note is that Gavin makes a number of basic errors in understanding Hurst-Kolmogorov processes.

The first thing to note? Maybe but given the startling conclusions of the paper - Climate models have no predictive skill - it would have perhaps attracted more formal comment and reaction from Gavin, and indeed the climate science community if it had appeared in a climate science publication rather than a journal of hydrology. The H-K stuff is interesting, but it seems to me an example of a McIntyresque focussing in on an insignificant aspect of the criticism and amplifying it, while ignoring the meat, that really requires addressing, such as the 'apples and oranges' issue of using a single model run rather than an ensemble …..

Dec 31, 2015 at 11:15 PM | Unregistered CommenterPhil Clarke

Oh...wait...spoke too soon. Sock puppet is here.


Jan 1, 2016 at 12:27 AM | Unregistered CommenterMailman

Wow - I've heard some odd arguments in the climate debate before, but this is a new one even on me. I've heard sceptic arguments dismissed because they're in blogs and not in the peer-reviewed literature before, but not had someone complain that it isn't in a peer-reviewed journal that is sufficiently climate-y.

This point is, on the face of it, ridiculous. The idea that hydrology has nothing to do with climate is delusional. Not even Gavin made such a claim. But lets just dig a little deeper to confirm; in the Journal of Hydrology guide to authors (here, my emphasis):

The Journal of Hydrology publishes original research papers and comprehensive reviews in all the subfields of the hydrological sciences including water based management and policy issues that impact on economics and society. These comprise, but are not limited to the physical, chemical, biogeochemical, stochastic and systems aspects of surface and groundwater hydrology, hydrometeorology and hydrogeology. Relevant topics incorporating the insights and methodologies of disciplines such as climatology, water resource systems, hydraulics, agrohydrology, geomorphology, soil science, instrumentation and remote sensing, civil and environmental engineering are included.

Indeed, climate papers are commonplace in the Journal of Hydrology. I looked at the most recent complete issue, no. 513 (Dec 2015), and a quick scan through the abstracts showed around 5 or 6 papers discussing climate. I also looked at it from the other side of the fence - I picked a big-ish chapter from AR5, WG1 chapter 2 "Observations: Atmosphere and Surface", and found six references to the Journal of Hydrology. I'm sure there are many more references to this journal throughout the IPCC reports.

So if the Journal of Hydrology has nothing to do with climate, someone needs to inform the IPCC and the journal editors, both of which think it is an entirely valid outlet for climate related articles. On the other hand, it seems rather more likely that Phil Clarke is just plain wrong, and that the journal of hydrology is an entirely suitable outlet for climate papers.

As for amplifying one insignificant aspect, you are absolutely wrong. Hurst-Kolmogorov dynamics defines the fundamental limits of uncertainty associated with climate change. It's like saying the Heisenberg uncertainty principle is just some little insignificant detail of quantum mechanics, and the real meat is in measuring position and momentum of electrons. Without properly understanding the former, you are doomed never to fully understand the latter.

Jan 1, 2016 at 1:49 AM | Unregistered CommenterSpence_UK

Oops, looks like a broken closing tag on the link. Apologies for that. While I remember, happy new year to all as well.

Jan 1, 2016 at 1:53 AM | Unregistered CommenterSpence_UK

Not many physicists can write as brilliantly as Seitz. He is a man of towering intellect yet like so many academics he lacks common sense.

It is always fun to cross swords with him even though it is easy to get him on the defensive. When he can't explain simple things like "The Pause" he resorts to ad hominem attacks.

Jan 1, 2016 at 7:03 AM | Unregistered Commentergallopingcamel

Happy New Year to all. Does anybody have a model for how much time needs to elapse before lack of response to direct questions becomes statistically significant? Just wondering ...

Jan 1, 2016 at 11:12 AM | Unregistered Commenternot banned yet

Spence: More than half of Gavin's criticisms were about the flawed methods in the 2008 paper viz.

They are using single realisations of model runs, and so they are not testing the forced component of the response (which can only be determined using ensembles or very long simulations). By correlating at the annual and other short term periods they are effectively comparing the weather in the real world with that in a model. Even without looking at their results, it is obvious that this is not going to match (since weather is uncorrelated in one realisation to another, let alone in the real world). Furthermore, by using only one to four grid boxes for their comparisons, even the longer term (30 year) forced trends are not going to come out of the noise.

Problems that were not addressed effectively in the 2010 paper. Nor is it true to say there has been no peer-reviewed response. See DOI:10.1080/02626667.2011.610758

Discussion of “A comparison of local and aggregated climate model outputs with observed data” A black eye for the Hydrological Sciences Journal

Abstract: A paper published by Anagnostopoulos et al. in volume 55 of the Hydrological Sciences Journal (HSJ) concludes that climate models are poor based on temporal correlation between observations and individual simulations. This interpretation hinges on a common misconception, that climate models predict natural climate variability. This discussion underlines fundamental differences between hydrological and climatological models, and hopes to clear misunderstandings regarding the proper use of climate simulations.

And in the discussion notes that 2 out of 3 reveiwers recommended rejection based on these flaws and says The idea of evaluating climate simulations initialized in the 19th century based on temporal correlation with observations at the yearly time scale is incongruous for anyone familiar with climate simulations and I could not understand how the paper had been accepted for publication. 

If the conclusions of the paper about the predictive skill of the models had merit, one might expect more of a response from the climate modelling discipline in the more than 5 years that have elapsed since, in reality they seem to have shrugged and moved on, while the response in the hydrology community is to describe it as a 'black eye' and puzzle over how such a flawed study could ever have appeared in print.

Jan 1, 2016 at 12:01 PM | Unregistered CommenterPhil Clarke

Phil Clarke, if you read the response by Prof Koutsoyiannis you will see he has already addressed the point made by Gavin Schmidt on methodology (his point 4a). So you are raising a question that has already been addressed. I will add a bit more detail to try to help those who have not invested the time to fully understand what has been done.

In order to understand Prof Koutsoyiannis' response, it is important to realise that the study did more than just look at the local temperature. Koutsoyiannis' study looked at the point samples, the temporal means, the standard deviation and the autocorrelation function (the latter measure characterised as the Hurst coefficient). Armed with these pieces of information, you can fully characterise the statistical properties of the time series at any scale, which means you can, from theory, exactly characterise the properties of the distribution after averaging (e.g. via ensembles, or spatial scales, etc).

What Koutsoyiannis et al found is that none of these properties matches reality. The idea that you can take a statistical distribution with the wrong mean, the wrong standard deviation, the wrong autocorrelation function, then average this to get the "right" answer is complete nonsense. You may be able to "get" a match at one scale, and one scale only, but it will be for the wrong reasons, and all other scales will be necessarily wrong. In other words, there is no recovery in averaging, due to the other analysis conducted in the study - which, oddly, neither Huard or Gavin discuss at any point.

As for Huard, you have neglected to provide the formal response by Prof Koutsoyiannis. It can be found here (hopefully with proper tags this time!):

Opening Closed Eyes

A couple of selected quotes from the response:

Huard writes: “The idea of evaluating climate simulations initialized in the 19th century based on temporal correlation with observations at the yearly time scale is incongruous for anyone familiar with climate simulations.” A reply to this argument can already be found in section “Justification of the methodology”, subsection “Scale of comparison”, of our paper. Instead of pointing out what we possibly got wrong there, which could start an interesting discussion, Huard merely repeats the already-replied-to argument in its original form. In repeating it, he focuses on the yearly time scale, thus missing that our comparisons are also made at climatic (30-year) time scale, and that, apart from simulated series alone, we also compared statistical characteristics of climate model outputs versus those of the real climate.

On the topic of the decision to publish:
Whether the HSJ got “a black eye” is for the reader to judge, as is whether “reviewers A and C rejected the paper on technical and methodological grounds, not philosophy”, since the entire review file is now public. The reader may also assess whether we made “factual errors obvious to anyone familiar with climate science” and Z.W. Kundzewicz, the Editor of the HSJ (as well as of the International Association of Hydrological Sciences, for 14 years) failed to see them. It is also possible that Kundzewicz and Stakhiv (2010) share “the same misguided assumptions about climate simulations”. However, if Huard thinks that two Coordinating/Principal/Chairing Lead Authors of the IPCC freshwater chapters in IPCC Assessment Reports (Kundzewicz in the Fourth—Kundzewicz et al. 2007, 2008, Stakhiv in the First, Second and Third—Lins et al. 1990, Stakhiv et al. 1992, Kaczmarek et al. 1995, Arnell et al. 2001) are not familiar with climate science, then he should be more concerned with the IPCC than with the HSJ; but this is also for the reader to judge.

Prof Koutsoyiannis would never be this unkind, but to be blunt (and IMO), Huard's article is exactly the kind of gatekeeping which holds climatology back as a scientific discipline. However, it is important to note that being published in 2010, these papers predate climategate, where some of the gatekeeping was exposed. I believe things have improved a bit since then. But only "a bit".

Jan 1, 2016 at 2:21 PM | Unregistered CommenterSpence_UK

Phil C

Don't you find it a bit peculiar that some David Huard (a lobbyist for hire? (*)) felt compelled to try to object a Koutsoyiannis-publication, with a 'Discussion' and trying to give a black eye to the journal?

Particularly since Huard is neither mentioned (or his work chriticized) in the orignal article or brings any work of his own on the matter to the table in his discussion?

Did you even read the article? I understand that you picked up on the above snippets, and that you've been searching for 'dismissals' as eg by Gavin Schmidt, and want to pretend that it (and you?) speaks for "climate modelling discipline" and "the hydrology community" (two more red flags imho). But did you actually read and understand what Huard wrote and objected to? Not only the phrases and the offhand claims you clearly are searching for.

I did, and can't say that I was impressed. Or rather: It immediately was clear that this was (mostly a hit-piece) from the sidelines rooting for the home team. Often poorly argued, even contradictory, but pretending to speak for the 'we' of the purported 'climate consensus'.

It doesn't mean that there cannot be any valid points or just valid other perspectives vs the original article. But If there were, those are diluted and obscured by the amount of handwaving. It looks to me though as if you took this as gospel too (as you did Gavins blogpost, and others beliefs of yours as eg the IPCC - forcing) and that you regurlarly don't understand the details or get them right.

As I said, I don't expect much more.

In case you are intrested (and I am very sure you aren't) there was a response to D Huard's 'Back Eye' by Koutsoyiannis et al, you can read it here. I picks up on many of the 'objections' in Huard's 'Black Eye' and shows them mostly to be strawmen and red herrings.

In stark contrast to Huard's 'Black Eye' it is concise and to the point, and easily picks apart the (mostly) semantic-style 'arguments' by Huard.

So what will your response be now? Looking for an even more recent blog comment by Gavin, or a 'Discussion' by a lobbyist that once more contradicts them and proclaims that 'we' are 'climate science' and only 'our' terminology, 'physics', 'interpretations' and 'metods' are valdid in 'our field'?

As I said, I don't expect much more. But it is fascinating to watch your side's methods to defend the orthodoxy ... your Huard-link was a good example of that.

(*) BTW Huard presents 'research' on Climate Denialism via his home-page, which puts him more on par with the barking Lewandowsky-types on the sidlines, where he also uses the typical 'Science Denial' canards: "Apollo moon landings were staged, Vaccines cause autism, HIV does not cause AIDS, Creationism" etc. It seems he really is your kinda guy ... ;-)

Jan 1, 2016 at 3:06 PM | Unregistered CommenterJonas N

Rse itz is a bow tied fraud with his double V blog. In English & on a UK site, it's double U, you bilabial fricative twerp.

He did nothing for years.

Jan 1, 2016 at 3:23 PM | Registered Commenterperry

I noticed that Spence_UK had linked to the same Koutsoyiannis-response while I wrote about it. He also made some additional comments and attempts to explain 'the meat' of it.

I am not sure Phil Clarke understands or wants to understand. It is certainly a more technical and detailed argument than various versions of 'But Gavind says it doesn't matter ... ' or 'Not published in a proper Climate journal .. ' or even 'But the climate community has ignored it ...' and even the 'hydrological community describes it as a black eye' ..

These are exactly the kind of responses I would expect from somebody who desperately wants to avoid the contents and find out about its substance and what it means, but who wants to continue to believe that the side he is rooting for is the righteous one.

As I just said, I read the Huard piece, found it peculiar in many ways, and tried to find out who he is and why he wrote it. And couldn't figure that out. But he has a homepage where he shows off his 'accomplishments' (including the 'Black Eye' discussion) and it gave a pretty clear picture of where this guy is coming from.

Phil Clarke apparently wanted to elevate him to the ultimate 'authority' on the matter. Which in my view is typical, and what I would expect from those who believe that (alleged) 'consensus' is a scientific argument.

Jan 1, 2016 at 3:38 PM | Unregistered CommenterJonas N

Hang on a moment, upthread it was stated that there exists no peer-reviewed response to his article, 

Now we see that there not only was there was a response in the same journal, there was a reply to the reply. Very odd.

My opinion on the Huard response shall we say, differs. As peer review was claimed as a quality gate, I note that his response was presumably peer-reviewed, and so I will rise above the attempt to denigrate his response with feeble ad hominem mudslinging.

His main criticism (and that of the majority of reviewers) is simply stated:

The main issue with the AKCEM paper is that it is based on a false premise, namely that the selected climate simulations predict (forecast) climate in a deterministic sense. Climate models may indeed be used in a weather forecasting mode, and this is one way of evaluating their sub-grid scale parameterization at the sub-daily time scale (see for instance section 8.4.11 in Randall et al. 2007). Some are even used to “forecast” climate at the decadal scale (Keenlyside et al. 2008, Smith et al. 2007) using observed oceanic and atmospheric initial conditions, the oceanic inertia constraining the atmospheric model. However, these experiments are still considered highly experimental (Keenlyside and Ba 2010) and never claim to correlate with the inter-annual variability. Climate simulations included in IPCC's TAR and AR4 also make no pretence of predicting/forecasting weather or climate. As Smith et al. (2007) put it:
Previous climate model projections of climate change accounted for external forcing from natural and anthropogenic sources but did not attempt to predict internally generated natural variability.
Evaluating climate models based on temporal correlations with observations is meaningful only if those models claim to forecast the year-to-year climate variations due to natural variability. The climate simulations analysed by AKCEM make no such claim, and the paper's main conclusion, that models are poor, is irrelevant.
The reply to the reply completely fails to address this point, and indeed throws in some more
fundamental misunderstandings about climate models for good measure.
As to 'gatekeeping', none was exposed in the stolen correspondence. The 'money quote' most frequently rolled out is

'I can’t see either of these papers being in the next IPCC report. Kevin and I will keep them out somehow — even if we have to redefine what the peer-review literature is!'

But this was just hyperbole, as the correspondents would have known well. Indeed both the papers in question,  McKitrick and Michaels (2004) and Kalnay and Cai (2003) were cited and discussed in AR4, and both have pretty much sunk without trace. An unkind observer might conclude that Koutsoyiannis et al 2010, mentioned briefly in AR5, has suffered the same fate.

Jan 1, 2016 at 3:45 PM | Unregistered CommenterPhil Clarke

Hang on a moment, upthread it was stated that there exists no peer-reviewed response to his article,

I see you quote-mine like a creationist. No peer-review response was with regard specifically to Gavin's comments, which would have been clear if you'd quoted the full sentence, rather than no peer review response at all. Let me know when you have a link to Gavin's peer reviewed article.

As peer review was claimed as a quality gate

Peer review is indeed a gate, which can be both used and abused, and is far from a guarantee of quality. See John Ioannidis' work on this topic amongst others.

Huard's criticism quoted by you is fully addressed in Koutsoyiannis' reply, which you clearly have not read or understood. I will explain one more time. Averaging cannot fix a broken model, in the sense that if the standard deviation and autocorrelation function are wrong, averaging cannot "fix" them. Unfortunately a level of numeracy is required to follow this argument, which we have already ascertained in previous threads that you do not possess. I have explained this repeatedly and you have no answer to it than to repeat the original false claim ad infinitum. So try growing a bit as a scientist, and explaining to me how averaging can fix a broken autocorrelation function, which is a necessary condition to support the claim you quote and the flawed signal-to-noise argument.

As for gatekeeping, that is far from the only example in the climategate e-mails, and certainly not the most significant. From "going to town" on reviews of critical articles, arranging for friendly scientists to perform reviews, and pressurising editors all contribute to gatekeeping. (Pressurising editors does not just come from climategate either, I know of at least one example in an open review online). The effect is cumulative and toxic to a scientific discipline.

Oh and it is not ad hominem to point out these failures in the peer review process, so you are quite wrong about that too. Ad hominem is an explicit argument of the form "person X is a fool, therefore he is wrong". My claim about peer review does not fall into this category at all.

Jan 1, 2016 at 4:15 PM | Unregistered CommenterSpence_UK

Spence_UK, Jonas N: I am really thankful for your kind comments regarding my works.

Phil Clarke: I am thankful to you too for your criticism, even though I do not see any element not already replied to. I guess you have no reason to become an "unkind observer" as my papers are already satisfactorily recognized and cited, not only by IPCC per se... (see e.g. ).

All: Happy New Year with health, peace, love and good luck,


Jan 1, 2016 at 4:32 PM | Unregistered CommenterDemetris Koutsoyiannis

The narrative flows
Sinking slow and sedulous
To the very truth.

Jan 1, 2016 at 4:46 PM | Unregistered Commenterkim

Phil C

The reply to the reply completely fails to address this point, and indeed throws in some more fundamental misunderstandings about climate models for good measure.

Well, it looks like you are now pretending to speak for some 'fundamental understandings about climate models'.

(And previously for both the "climate modelling discipline" and "the hydrology community")

But of course you don't, and the term 'fundamental' wrt to science or 'basic physics' when defending model simulations are a dead give-away (as is any incarnation of 'denialist' etc).

So waht does you post boil down to?

You copy-paste a part of the same 'Discussion' we've already read once more!? While you make an obviously false statement wrt to it!? Namely:

As peer review was claimed as a quality gate ..

Whereafter you want to change topics once again?

Let go back to what you actually say ('the meat of it' so to speak). It is indeed strange that you claim that they " completely fai[ed]s to address this point", as they explicitly do:

We feel that Huard addresses these statements to the wrong audience. We generally agree with his diagnosis, and our results confirm it

Now, you may believe that the Huard-discussion lands some hard hits, and that the reply 'fails to adress' them. But I can't see you even attemptimg to make them. You just re-paste a paragraph saying it is damning criticism. But as far as I can see, only based on a similar belief in Huard's (assumed) authority on the matter (and possibly Gavin's too).

So sorry Phil, you did not add once tiny piece of substance here.

Precisely as I would have expected.

BTW What do you consider ad hominem and mudslinging? Being contradicted? Not accepting Huard as an authority based on his 'say so'? Could you please elaborate?

Or is poor grasp of the meaning of words and phrases once more the 'explanation'?

Jan 1, 2016 at 5:03 PM | Unregistered CommenterJonas N

So try growing a bit as a scientist, and explaining to me how averaging can fix a broken autocorrelation function, which is a necessary condition to support the claim you quote and the flawed signal-to-noise argument.

Irrelevant, if the quantities you are averaging are not expected to correlate! This is the point made by Gavin, the majority of the reviewers, and Huard, and which, sorry simply has not been addressed. Pointless to repeat it a third time.

Jan 1, 2016 at 5:04 PM | Unregistered CommenterPhil Clarke

Good to hear from you Demetris, and of course good to see your work still being actively discussed online :-)

Your arrival reminds me that I should not get so frustrated in blog comments - I do aspire to a higher level of collegiality that I never quite seem to achieve! I still haven't learnt when to stop...

Wishing you (and all at NTUA) a fractally happy 2016.

Jan 1, 2016 at 5:18 PM | Unregistered CommenterSpence_UK

Irrelevant, if the quantities you are averaging are not expected to correlate!

You don't expect the autocorrelation function... to correlate? I don't think these words mean what you think they mean.

As I have already explained, there is a clear theoretical relationship between the autocorrelation function and the averaging behaviour of a system, which allows you to determine what will happen under averaging. This is H-K dynamics 101, but if you do not understand this step your reasoning will remain forever broken. To be fair, I don't expect the lay climate activist to understand this, but Schmidt and Huard really have no such excuse.

Sadly, we have reached an impasse. I am using the language of mathematics, and you are speaking an entirely different language. There seems little point in continuing.

Jan 1, 2016 at 5:33 PM | Unregistered CommenterSpence_UK

There seems little point in continuing.

None. You've consistently and persistently ignored the elephant in the room, simply expressed by Schmidt

By correlating at the annual and other short term periods they are effectively comparing the weather in the real world with that in a model. Even without looking at their results, it is obvious that this is not going to match (since weather is uncorrelated in one realisation to another, let alone in the real world)

Holed below the water line before leaving port.

Jan 1, 2016 at 6:07 PM | Unregistered CommenterPhil Clarke

Phil C, you now seem to claim (possibly as a 'proxy' for Gavin) that climate models are not supposed to be attempts to model and then simulate the actual cliamte, with its (averaged) response to GHGs, variability and autocorrelation?

Come again?

I can surely agree that climate models fail at that, that they are mutually inconsistent, even contradictory. That they fundamentally(!) disagree on central key metrics ... But are you really saying that they aren't even attempting> to get these correct, to provide a realistic representation of what the climate might behave like!?


I would be very suprised if Gavin would claim that (and afik he hasn't either). To me it looks like he is attempting to 'do away' with the ctricicisms by saying 'we didn't expect anything else' ... which is correct, but doesn't change it one bit.

And as Koutsoyiannis et al detail in their response is the main point. That, and that the models by very many are used exactly in the way which Gavin (now) says they never were intended to.

And I'd add that this is more than just a bit disingenuous, as those simulations all the time are uses as if their (ensemble) outcomes have skill in predicting a range of key metrics. Gavin's response now seems to say: 'Yes, we know they aren't that good for that purpose'. His blog post once more is mostly about clouding the issue, talking about other things, saying this is 'not useful' (and in one way, that's correct from his perspective)

The more interesting thing here is that Phil Clarke believes that Gavin or Huard are the ultimate arbiters one something he doesn't understand.

Jan 1, 2016 at 6:17 PM | Unregistered CommenterJonas N

"Holed below the water line before leaving port"(*)

Yes, you would like to believe in that, wouldn't you Phil. More than anything else. But it is still only belief in a strawman by somebody else (and even that this strawman is correctly errected)

It's also funny to see that you think you are the one who gets to decide what is 'the elephant in the room' ...

What 'authority' do you think autorized you to make such calls, Phil?

(*) I'd say this looks like one more attempt to replace substance and argument by sematics, prhases, poor use of lagunage and just proclamation.

Jan 1, 2016 at 6:26 PM | Unregistered CommenterJonas N

I thought his use of "lagunage" very effective. It's a good word.

Jan 1, 2016 at 8:48 PM | Unregistered Commenterjferguson

Jonas, thanks for your comments and continued responses. There comes a point where continued debate adds no further value, and when Phil comes up with mathematical nonsense such as insisting the autocorrelation function is not expected to correlate, it is probably time for me to give up. A five year old may talk gibberish and then declare themselves the winner of something; it adds no value.

The link Demetris gives to citations is quite informative; according to google scholar, the original study (credibility of climate predictions) has 119 cites, the follow up study 78 cites. By comparison, Huard's response gets a whopping 7 citations - even fewer than Demetris' response (which has 11, even though it is one of the 7 citations to Huards paper, which could not be reciprocated!)

It is clear that the Huard response has faded into obscurity. Citations alone are not the full story, but near 200 cites between the two papers shows that some climatologists are at least considering the Hurst phenomenon and its consequence for predictions. This is a definite step in the right direction, no matter what activists on blogs might say.

Jan 1, 2016 at 8:51 PM | Unregistered CommenterSpence_UK

Citations alone are not the full story, but near 200 cites between the two papers shows that some climatologists are at least considering the Hurst phenomenon and its consequence for predictions.

Have you reviewed those cites? From the 78, I personally would discount those from WattsUpWithThat and of course Huard is responsible for at least three. That legitimate citations are from the world of hydrology not from climatologists, by and large.

It really is this simple : The weather (natural variability) in the grid cell of a single model run is not expected to correlate with the actual weather in the stations contained therein. How could it?

As for Hurst, I tend to side with Clegg 2005

… measuring the Hurst parameter, even in artificial data, is very hit and miss. In the artificial data with no corrupting noise, some estimators performed very poorly indeed. Confidence intervals given should certainly not be taken at face value (indeed should be considered as next to worthless).

When considering real data, researchers are advised to use extreme caution. A researcher relying on the results of any single estimator for the Hurst parameter is likely to be drawing false conclusions, no matter how sound the theoretical backing for the estimator in question. While simple filtering techniques are suggested in the literature for improving the performance of Hurst parameter estimation, they had little or no effect on the data analysed in this paper.

Jan 1, 2016 at 11:06 PM | Unregistered CommenterPhil Clarke

4) Phil Clarke

Jan 1, 2016 at 11:35 PM | Unregistered Commenternot banned yet

Have you reviewed those cites?

I reviewed the cites of the 119 (on the credibility of climate predictions). One was hosted at WattsUpWithThat (Pat Frank's review), and two were by Huard. Well, that brings us down to 116 if you don't count those. The later paper, perhaps 3 or 4 are from WattsUpWithThat. Let's even throw out the ones Ross McKitrick wrote, I'm sure you won't like those either. So we have around 190 vs. 7 instead of 200 vs. 7. Doesn't really change my point one little bit, does it?

It really is this simple : The weather (natural variability) in the grid cell of a single model run is not expected to correlate with the actual weather in the stations contained therein. How could it?

It really is this simple: by your bad definition above, the 30 year mean, standard deviation and autocorrelation function are weather. Please explain to us, oh guru of climate, how standard deviation counts as "weather". (I would ask autocorrelation function, but I'm reasonably satisfied you don't actually know what an autocorrelation function is - you should at least have some idea of what a standard deviation is)

As for Clegg, yes, the Hurst parameter is difficult to estimate accurately. Which is why Demetris devised the climacogram, including bias corrections, gives a consistent and reliable estimate of the Hurst exponent. This doesn't help you though. Like it or not, Hurst dominates natural variability. The fact that it is hard to estimate doesn't change that - you still have to deal with it.

Jan 2, 2016 at 12:28 AM | Unregistered CommenterSpence_UK

Huh, looks like Phil Clarke is not going to come up with an explanation as to how standard deviation of temperature and precipitation is "weather". What a surprise!

I also realised my wrong approach with Phil. I've been explaining *why* averaging won't help, which requires a certain level of mathematics to follow. But the paper has a much simple demonstration of why averaging won't help, supported by evidence.

Let's imagine for one moment, as a thought experiment, that this problem was governed by signal-to-noise ratio, and that averaging improved that ratio. Signal-to-noise is not an on/off thing - you don't end up with zero correlation until you reach some magic threshold, then suddenly - blam - you get a strong positive correlation. In the real world, you get a gradual transition from zero correlation to strong correlation as the signal to noise ratio improves.

This possibility was checked in Demetris' paper. He studied correlations both at the 1-year scale and the 30-year scale. This is somewhere between a 5x to 6x improvement in signal-to-noise ratio. There was no improvement in correlation performance between the two; in fact the 30-year scale actually performed worse than random guessing in the table presented (I think for Albany).

Of course, this result is exactly what we expect based on theory, but here we have substantiated evidence that the theory is right; that the "signal-to-noise" argument presented by Huard, Schmidt and Clarke is wrong.

Of course, having shown them to be wrong, Demetris gave an opportunity for them to show they are right: repeat the experiment, only apply the averaging before the process and see what happens. As far as I am aware, none of the above three people have gone ahead and taken this simple step to prove their signal-to-noise argument, which they simply continue to assert without any supporting evidence.

Jan 3, 2016 at 5:50 PM | Unregistered CommenterSpence_UK

PostPost a New Comment

Enter your information below to add a new comment.

My response is on my own website »
Author Email (optional):
Author URL (optional):
Some HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>