The WUWT Hot Sheet for Saturday August 31st, 2013

Good to know that peaple make Michael E. Mann a fool LOL
https://twitter.com/MichaelEMann/status/373133788005400576

Quote “the entire report and its Summary for Policymakers are already invalidated even before publication.”
If this is true, and I have no doubt it is true, then where is the outrage from the scientific community? Surely, the IPCC reports have been the center of the evidence on which the international policy of “decarbonization” was based, and these reports were supposedly based on science.
All we are hearing form the likes of the Royal Society, the American Physical Society, etc.etc. are the “Sounds of Silence” (Simon and Garfunkel).
Who will lead the charge from the scientific community to say “Enough is enough!”? I nominate Prof. Judith Curry.

From DJ on August 31, 2013 at 6:03 am:

Good to know that peaple make Michael E. Mann a fool LOL

That and other tweets in the sequence here:
http://wattsupwiththat.com/2013/08/28/requesting-some-assistance/#comment-1403068
Watts’ reply, several comments later:
http://wattsupwiththat.com/2013/08/28/requesting-some-assistance/#comment-1403189

Breaking News from UAH (eh, close enough)
http://blog.al.com/breaking/2013/08/summer_of_2013_among_coolest_o.html

Summer of 2013 among coolest on record in Alabama, continues cooling trend, climatologist says
By Paul Gattis | pgattis@al.com
…
on August 29, 2013 at 4:39 PM, updated August 29, 2013 at 4:40 PM
HUNTSVILLE, Alabama – From Huntsville to Mobile, summer temperatures have been cooler than normal this year.
In fact, the summer of 2013 will be listed as one of the coolest summers in the past 131 years, according to state climatologist John Christy, director of the Earth Systems Science Center at the University of Alabama in Huntsville.
Christy announced today that this summer is the fifth-coolest on record in Alabama with average temperatures that are almost 2 degrees below normal.
…
According to Christy, 1967 shows up in most Alabama weather stations as having the coolest summer since reliable weather record keeping started in 1883. Other summers on the cool end of the record have all been since 1960 and include 1997, 1992, 1961, 1989, 1994, 2003, 2004 and 2013.

I would like to see an explanation of this part:

In calculating the average temperatures for each region, the announcement said, the climatologist’s office uses only daytime high temperature readings for stations within a 50-mile radius of each city center.
Only daytime highs are used because nighttime low temperatures at some stations have slowly been contaminated by the growth of urban areas around those stations. The data were combined in a process that takes into account station moves and instrument changes.
…

UHI contamination of nighttime lows, heard of it. But to throw them out completely while using only the highs which also have problems? It’d be nice if we could get a short write-up of that by Dr. Christy.
See article for more info and stats for statewide and select large cities.

“Climate campaigners seem to think they have a winner with this takedown of elected officials who reject global warming science,” This approach again illustrates how desperate the warmists are with their failed climate science . Unable to support their claims by valid science they are constantly looking for new ways to discredit those who diagree with their questionable science . If this approach should ever materialize, and I don’t think it ever will , there are countless climate science failures that could be tagged with the certain climate scientists that will soon even the playing field in the eyes of the public

J. Philip Peterson on August 31, 2013 at 7:25 am:

I’m now a tweet follower of Michael Mann. Does that now mean that I am a Mann stalker from WUWT?

Nah, but you are a Mann ego-stroker.
I was warned long ago to avoid stupid people who do stupid things, as the stupid rubs off and you can get infected by the stupid. I’ve seen enough tweets from Mann’s clique of twits to have initial confirmation.
You better buy some surgical gloves so you can practice safe stroking.

“A model that predicts is susceptible to falsification. A model that projects is insusceptible to falsification. The climate models project. They do not predict.”
Businesses use that word for future growth, profit, etc
http://www.forbes.com/sites/matthewgarrett/2013/08/22/traditional-forecasting-leads-to-traditional-results-failure/
Some of those businesses known as banks used computer models to make those projections!

I was going to respond to Terry’s comment about the two words and thought it was misleading when I came to your reply Richard while I was mulling over what to wrote,now I will just say good reply!
In 1990 IPCC report they originally used the word PREDICT and got burned on it thus the use of the word Projection came along which gives them some wiggly room to their definition on what is probable in their future forecasts.

Terry Oldberg says:
August 31, 2013 at 8:03 am
In claiming that the climate models have been falsified, this post at the Hockey Schtick errs. A model that predicts is susceptible to falsification. A model that projects is insusceptible to falsification. The climate models project. They do not predict.
—–
A distinction without a difference. Let’s ask: (a) what is it to report the output of a climate model? and (b) what is the purpose of a speech utterance or narrative that reports the output of a climate model?
Surely (a) is an assertiion that the global average temperature will rise x degrees in y years (hedged probabilistically, if it is honest)? If it’s not a statement about the future, what else could it be? And if the purpose of reporting the model output is not to communicate an expectation of future temperature, then there seem few alternatives other than to treating it as dissembling with the purpose of deceiving the listener/reader.
It seems to me that the only reason for drawing a distinction between prediction and projection in this manner and in this context is to be able then to assert that the models reflect reality despite the mounting discrepancies between them and the world. Mere scholasticism.

Tweeting with Mann reminds me of an old saying : “Never wrestle with a pig—you get dirty and the pig likes it”

If Michael Mann is so eager to sue everybody – maybe he’d better consider whether it’s smart to libel Anthony with his comments about Koch,
REPLY: Yes, I’m actually considering some options there. I’ve never r=taken a dime from Koch, and Mann doesn’t know what he’s talking about. For somebody who is in the libel lawsuit business, he’s certainly a loose cannon. – Anthony

rogerknights says:
August 31, 2013 at 12:39 pm
Terry Oldberg says:
August 31, 2013 at 8:03 am
In claiming that the climate models have been falsified, this post at the Hockey Schtick errs. A model that predicts is susceptible to falsification. A model that projects is insusceptible to falsification. The climate models project. They do not predict.
It seems to me that they make conditional predictions–e.g., if GHGs rise at rate A, then surface temperatures will rise by rate Y, plus or minus amount Z. If the condition is satisfied (CO2 rises at rate A), then the prediction is triggered.
——–
You are correct, but only because it is tautological (albeit not self-evidently so) that all predications that involve causation are necessarily conditional statements. In contrast, the statement “1 + 1 = 2” is not a prediction – not even of a child learning its arithmentic – because the result is not an effect of the arithmetic operation.

can we start naming climate prediction failures after scientists and activists?
The Hansen Hiatus
Not a bad idea, but it needs reworking.
“The Hansen Hiatus ” makes is sound like he predicted the Hiatus.
“Hansen Hallucination” is more on target.
“Hansen Hype-a-Temp” covers the GISS manipulation of the historical records.
“The Hansen Hoax” need to be reserved to describe the whole past quarter century of CAGW.

richardscourtney:
Thank you for taking the time to respond.
In reference to the issue that you raise, “anonymous” responds that ( http://hockeyschtick.blogspot.com/2013/08/why-forthcoming-un-ipcc-report-is.html#comment-form ) “OK, not falsified. How about “shown not to provide information useful for evaluating human impact on the global climate”? Anonymous is spot on but his response diametrically differs from yours. How can this be so?
In the investigation of a suspected example of an equivocation fallacy, one translates the language of the suspected equivocation into an unambiguous language for the purpose of establishing whether or not a term changes meaning in the midst of the argument. Rather than disambiguating the language of your response, you respond in an ambiguous language that supports your fallacious conclusion.

.Luther Bl’t:
Thank you for raising the penetrating question. To paraphrase this question, is it not true that a climate model makes a conditional prediction (e.g., if GHGs rise at rate A, then surface temperatures will rise by rate Y, plus or minus amount Z)? The answer is NO.
In other terms, a “conditional prediction” is a “predictive inference.” Logic contains what we know about the rules by which correct inferences may be distinguished from incorrect ones. Thus, your question has logical significance.
In response to this question, I’ll point out that a “predictive inference” is an extrapolation from an observed state of a system to an unobserved but observable state of the same system. By convention, the observed state is called the “condition” while the unobserved but observable state is called the “outcome.” An example of a condition is “cloudy.” An example of an outcome is “rain in the next 24 hours.”
A pairing of a condition with an outcome is a description of an event. The complete set of these events is an example of a statistical population. No statistical population underlies the IPCC climate models. Hence, the predictive inference that is made by such a model is undefined.
In a model that possesses an underlying statistical population, events of a particular description possess relative frequencies. The numerical values of these relative frequencies relate to the numerical values of the associated probabilities. The various probability values map to the information that is conveyed to a policy maker about the outcomes from his or her policy decisions in advance of these outcomes.
For the IPCC climate models, there is no underlying statistical population. Hence there are no relative frequency or probability values. Thus, these models convey no information to a policy maker about the outcomes from his/her policy decision in advance of these outcomes (http://wmbriggs.com/blog/?p=7923 ) . .
It follows that these models are worthless for the purpose of making policy on CO2 emissions. The appearance that they are worthwhile for this purpose is a product of repeated instances of the equivocation fallacy on the part of global warming climatologists. Among these climatologists are skeptics as well as their opponents.

“@terry Oldberg”
If the ipcc is only making projections and not predictions, then 1] it is not doing real science, and 2] there is therefore no basis upon which to argue that we should modify anything we humans are doing.

Bruce Cobb above misread what I wrote. I didn’t write “statists.” I wrote “stasists,” my term for folks eager to maintain existing energy norms in America. More examples: http://dotearth.blogs.nytimes.com/?s=stasists

richardscourtney:
In your post of Aug. 31, 2013 at 10:15 am you contend that I have yet again posted untrue nonsense and deplore the injustice of this infraction. However, your argument fails to support your contention.
My contention is stated in unambiguous and logical terms in the peer-reviewed article at http://wmbriggs.com/blog/?p=7923 . The referee for this paper, William Briggs, holds a master’s degree in atmospheric science and a PhD in mathematical statistics; he is a professor of statistics at Cornell University; formerly, he sat on the standing committee of the American Meteorological Society on probability theory and statistics. In a previously published peer reviewed paper http://judithcurry.com/2011/02/15/the-principles-of-reasoning-part-iii-logic-and-climatology/ I go over similar ground. The referee for the latter paper, Dr. Judith Curry, is a professional climatologist and is the chair of Earth Sciences at Georgia Tech University. In articles that are cited in the latter article, Gray, Green and Armstrong and Trenberth reach conclusions that are similar to mine. Gray holds a PhD in physical chemistry from Cambridge University and is a longtime IPCC expert reviewer. Armstrong is a professor at the Wharton School of Business of the University of Pennsylvania; he edits the Journal of Forecasting. Green holds a PhD degree and is a university professor. Trenberth holds a PhD degree and is a high-level professional climatologist.
As I reach my conclusion via a syllogism, this conclusion is true unless it can be proved that one or more of the premises to my argument is false or unproved. Your post of Aug. 31, 2013 at 10:15 am is devoid of such a proof.

If you haven’t seen Dr. Murry Salby’s utter destruction of the AGW hypothesis through a stunning mathematical analysis of the data, it is worth your time. You will need some physics and Math background. Even if you don’t have that, you will be able to follow the logic. I don’t think the IPCC will be inviting him for a presentation any time soon. Here’s the link:

part of the introduction in German but the rest in English.

Terry Oldenberg writes: For details on the equivocation fallacy in global warming arguments, see the peer-reviewed article at http://wmbriggs.com/blog/?p=7923 .
Rarely have I ever had the dubious pleasure of reading a piece of sophist crap like this, “peer-reviewed” (whatever that means in the context of a blog post) or not.
In addition to openly inventing terminology to create an absolutely artificial distinction — model vs “modele” — the only thing this article does is carefully delineate the following circumstances:
a) If the climate programs, built on top of what are loudly trumpeted to be physical laws and a variety of stated and unstated assumptions, each of which can be treated as a bayesian prior and weighted with a prior probability of being true, are models, they can be falsified by ordinary hypothesis testing and validated to the extent that they prove successful in predicting the future. They are science. They aren’t even “good” or “bad” science — science doesn’t use value labels — they are merely computational models of a general theory that should be given weight according to their predictive success, in their own way no different from abstruse model computations in condensed matter physics or predictive models built with empirical methodology that predict the stock market, customer preferences, or tomorrow’s weather. Even models built without an underlying theory — neural networks, for example — are without exception called “models” by every expert in predictive modeling in the world, where yes, I am such an expert. Show me a single textbook that supports your utterly artificial, bullshit “disambiguation” and we’ll talk. I’ll cheerfully show you a small stack of very current textbooks on pattern recognition, predictive modeling, scientific modeling, and Bayesian statistics where no such distinction is ever made because it does not matter.
b) If the climate programs (specifically, of course, the General Circulation Models or GCMs), built on top of physical laws incorporated in code and a variety of stated and unstated assumptions (bayesian priors) are modeles, you assert that by definition, they cannot be falsified by ordinary hypothesis testing or validated. This does not make the slightest bit of difference in how sensible people should treat their “projections”. They are computational “modeles” of a general theory that should be given weight according to their predictive success because mere common sense suggests that it is stupid to give much weight to non-falsifiable models that don’t work, whether or not they are technically “falsified” (technically because of course they are falsified — all one requires to be able to apply a hypothesis test is a null hypothesis, and model or modele, there is always a null hypothesis. One can, and people do, apply hypothesis testing to far more abstruse “modeles” than GCMs — for example a “modele” that claims that prayer affects outcomes in disease, claims of pure prophecy, claims of astrology. Note well the “modele” company you are putting GCMs into — religious claims are all modeles, are they not? Otherwise, what is it, precisely, that differentiates the two beyond a mere statement of purpose, a desire to protect a modele from falsification (as do the purveyors of any religious belief)? Not a damn thing.
Not only is your distinction, then, between modele and model absolute nonsense — we already have plenty of adequate terminology for predictions made on a religious basis, even those that “claim” to be a science such as astrology or other numerological nonsense and we do not hesitate to reject the use of those predictions in common political discourse on the grounds of mere common sense when the predictions thus made fail to come to pass — it is incorrect in the specific case of climate science where the GCMs are without any possible doubt science based predictive models that are without question subject to falsification, one at a time. If this were not the case, they would not be science at all, you twit.
Let me be specific. In both AR4 and (leaked) AR5, there are graphs of the output of certain specific GCMs when a Monte Carlo method is used to add noise to the inputs and various possible future trajectories are computed. The spread of these trajectories is then used to determine both a mean future climate and the envelope of statistical uncertainty around that mean, with the implicit null hypothesis of “this model is correct”. Let us state this in clear statistical terms. If this model is correct (as a prior assumption) and if we perturb the initial conditions enough to obtain a reasonable statistical spread of future outcomes, then the future trajectory should lie withing that spread with a probability directly related to the spread. In other words, we expect the actual future climate to be in the densest part of the distribution of possible futures, and to spend relatively little time out on the wings of the distribution of possible futures.
For most of the models that contributed to the predictions of AR4 and AR5 — where AR4’s summary for policy makers did not hesitate to make statistical claims containing percentages, even though those claims are utterly indefensible using any form of statistical theory I’ve ever heard of and hence the IPCC is treating them like models, not “modeles”, they are called General Circulation Models, not “Modeles”, and where their predictions are being used to make statistical assertions about the future as we expect of scientific models, not religious “modeles” — the actual future climate trajectory, relative to the time of the model’s construction and application, is at the outside envelope of a spread consisting of hundreds of runs. It is even worse than this, as none of the runs have either the right noise or variation, and even the runs that generate the envelope to not spend all of their time down where the actual climate has been, but rather individually drop down for short periods to help collectively define the range. The p-value that can absolutely be estimated per model, for most of these models is well under 0.01. The models can, for the most part, be rejected at the 99% or better confidence level even over the last 15 years alone, one at a time.
Collectively, they aren’t even a model or modele, they are just a bad mistake, an utterly indefensible application of the terminology of statistics in circumstances that do not satisfy the axioms of statistics as prior conditions for its relevance.
Note well that there are in the literature genuine efforts, underway at last, to test the GCMs. A recent paper, for example, compared four GCMs applied to the same toy problem — a uniformly heated, untilted “water world” — and got four future climates that bore not the slightest resemblance to one another. This too, is an elementary hypothesis test and we can thus say that each model is at most 25% likely to be correct because they cannot all be correct and a toy problem like this probably does, in fact, have a unique solution (we just don’t know what it is). Or perhaps it does not — which is even worse news for climate modelers, because if we cannot even get a single consistent answer for a toy problem, then what is the point of modeling, or “modele-ing”, any sort of future with GCMs and pretending that it has some predictive, projecting, or prophetic force?
To conclude, the issue is not sophistry like an artificial and arbitrary distinction between “models” and “modeles”, it is a much simpler distinction between science and religion. Science is by definition strongly falsifiable and weakly verifiable — any claim, whether “predictive”, “projective”, or “factual” made by a numerical computation implementing a scientific theory or hypothesis, can be disproven easily and rapidly by contrary observational evidence, and while it can never be fully proven by observational evidence, to the extent that it is in continual agreement we accord the theory a greater degree of belief. Good science is not dogmatic, it is flexible and it changes to accommodate contrary evidence so that it incrementally approaches a consistent theory that works to cover all or nearly all observations (where it is “nearly” all we consider the science to be “unsettled” — that’s where “research” comes in).
Personally, I think that GCMs are science. Science that has, in many specific cases (there are many different, distinct GCMs) been shown to be in very serious conflict with observation, conflict so profound that we should correctly doubt both the implementation of known physics and the stated and unstated hypotheses and the code itself (there are many places a complex computationall theory can fail). That is, they are clearly models as their name asserts (unsurprisingly, since nobody ever heard of “modeles” before you just made the term up as a religious apologia). If they are not science, they are religion (which might offend the climate scientists who think otherwise) but as far as politician, scientist, or lay public alike are concerned it does not matter. Call them anything you like, they are still wrong, still fail elementary hypothesis tests at the level of 99% on up, they are still things that mere common sense suggests should not be taken particularly seriously as they almost certainly contain serious errors in fact and implementation and are not even sufficiently consistent to give a single answer in implementation to a problem that is far simpler than the Earth’s actually noisy, tilted, highly elliptical orbit and its crazy array of oceans, solid surface, and biosphere.
As a professional predictive modeler, one that uses modeles such as neural networks to obtain results that cannot be connected in any way outside black (bayesian neuromathematical) magic to the predictions, this really is the bottom line. A model, or modele, either works or it doesn’t. If it doesn’t it is pretty useless no matter what term you invent to try to excuse it from the necessity of actually working. If it doesn’t work, you don’t get paid.
rgb

RACookPE1978:
Your argument is illogical. If this topic is of interest, I’d be happy to expend upon it.

This is because he makes irrational assertions based on
(a) his own definition of terms
and
(b) refusal to provide explanation of the terms he uses.
He often claims he can “expand” or “explain” what he says but – when pressed – he never does.
Time honored ways of trying to win an argument, at least in your own mind, of course. I have nothing against the idea of disambiguating terms as this can indeed be a major problem in any philosophical or logical debate, and it is closely tied to e.g. shifting sands fallacies and so on that often underlie it (inherent in the refusal to provide explanation and example of the new “definitions” inserted). Hence my request for a concrete example, ideally in context, for what he has in mind. They’d better not be GCMs, of course, because those are painfully obviously ordinary predictive models and a perfect match for NON-ambiguous definition 1 from the wordnet list of definitions of models. They embody many hypotheses (so that the hypothesis itself is quite complicated) all of them ultimately based on observation and data because all of science including the laws of physics are based on observation and data in a consistent Bayesian network or empirically supported reasoning.
GCMs in particular are further fueled by the use of empirically set, tunable parameters and are more or less without exception individually validated on a training set of climate data used to tune these parameters. When released into the wild, they promptly have diverged from the future, and were never ever able to hindcast the past. Their use has openly been to predict the dire climate future, and has invariably been accompanied by spurious, erroneous, and deliberately misleading claims of the “probability” of that future in quite quantitative terms, made in precisely the context where they would have the maximum political impact and be safely shielded from any sort of scientific or statistical criticism.
If there is something else that he wants to consider a “modele”, all he has to do is — without obscenity, ad hominem, or shrillness — present one or more clear examples and indicate how they cannot fit into the list of accepted definitions of the term model and how they are relevant to the discussion of the general circulation models used as the fundamental statistical predictive elements of the politicized side of climate science. Or, perhaps, his point is that they should not be so used because one cannot make a specific Bayesian argument with full knowledge of their priors, but that is a common feature of nearly all predictive models and doesn’t stop them from being models built on top of a collective hypothesis, it just means that when they fail a hypothesis test it is reasonable to reject that hypothesis and try try again. More generally, one should weaken the conclusion of a complex argument to the extent that the prior probabilities are unknown or poorly known, which is why multivariate predictive modeling is a goddamn hard game to play — it doesn’t take a lot to weaken a complex argument to damn near a coin flip, and it means that a wise person will want a lot of weak verification before he trusts his life or fortune to its predictions. Taleb’s The Black Swan presents a lovely argument for why in his polemic against the abuse of the normal, against ceteris paribus, against egregious assertion of Bayesian priors that are more properly unknown and sometimes unknowable.
rgb

The only reply I have time to make is that the metaphor (rather than reply) given about concerning disambiguation of the word “plane”, in context, would have us reasoning about the kind of clothes GCMs wear as they parade down runways.
In other words, only a complete idiot could look at the content, context, and purpose of general circulation computational predictive models and then confuse them with any meaning of the term “model” but meaning 1 in my list above, which does not under any circumstances require disambiguation either logically or semantically. There is a name for this particular logical fallacy — could it be “straw man”? Inventing an issue out of whole cloth, writing entire articles on the subject intended to provide apologia to something that isn’t a model/modele issue but is just poorly done science, improperly used by scientists to influence political issues under the guise of settled science when in fact it not only isn’t settled, it fails basic hypothesis tests applied to the GCMs one at a time as hypothesis based predictive models.
There is no need to kick the GCMs by calling them modeles (a term that no one will understand because you just made it up, Terry) — their failure as mere models is quite straightforward and unambiguous. And no, I’m not calling you an idiot — I’m saying that both your example and mine illustrate that there isn’t the slightest shred of contextual ambiguity here. If you want to work on disambiguating various terms used in religious discussion, such as whether or not atheism is a theistic belief, or the term Universe (meaning everything that has objectively real existence) does or does not (by definition) include God and hence logically precludes any possibility of the Universe being created by a dualistic God from a manifestly contradictory complement to the set of everything that actually exists, I’m all over that — endless, harmless fun in an issue that has no real empirical basis and hence is fundamentally undecidable at least with the data in hand. Trying to play this game in predictive modeling is simply unnecessary, confusing, and a bit silly.
Beyond that, Richard really doesn’t need my help, but we’d all appreciate you making your point. Are general circulation models models in the commonly accepted sense of the term in this context or not? If not, be well-prepared to defend your assertion otherwise, on the basis not of your made-up term “modele” which we have not yet agreed is logically or semantically necessary. If it is not necessary, adding it is unacceptable and confusing and violates several precepts of efficient argumentation. Note that there isn’t any doubt what the creators of those models think that they are, what AR1-4 assert that they are, what AR5 is ready to assert what they are, what the scientists who wrote the summaries for policy makers claim that they are, and what they are used for, without exception, worldwide.
They are failed predictive models, but I would argue that there is no doubt that they are predictive models in design, intent, purpose, application, and all subsequent use. Where is the ambiguity?
rgb

rgbatduke:
A phrase that you use in your latest post, “predictive model,” is notable for employing two polysemic terms. They are “predictive” and “model.” If you wished, you could assist the cause of rationality in debates over global warming by switching to the use of monosemic terms. For you to serve this cause, it is not necessary for you to use the specific words (including modele) that I use in my papers. You could, for example, use made up words. The essential aspect of a disambiguation is not the words that are chosen but rather that the chosen words are monosemic in their description of the methodology of the research. I’m flabbergasted at your resistance to joining me in this endeavor. Does this mean that you prefer irrationality?

rgbatduke:
In your latest post, you set up a strawman and knock it down. I have nowhere stated that adjectives cannot be used in disamguating the terms of an argument. The fact is that you steadfastly refuse to use adjectives or any other means of disambiguation in making arguments about global warming. A consequence is for you to repeatedly draw conclusions from equivocations in violation of a logical principle.