CERN: Climate Models will need to be revised

[waldo]

You are clueless fool

no need to get testy and personalize this, hey Timmay! Just answer the question... the one you keep avoiding, over and over and over again:

why is that TimG? Why can't McIntyre, why can't his posse, put forward a formal challenge? Is there a problem? 3+ years and counting... is there a problem? In all your personal puffery, in your recently stated 100% certainty, why don't you take up the charge... personally? Bring him down... bring the Mann down! Is there a problem?

[TimG]

Just answer the question... the one you keep avoiding, over and over and over again:

Because it is an irrelevant distraction. I posted a technical argument. If you disagree then try to refute it by actually addressing the argument made. If you can't do that then simply admit your ignorance and shut up. Edited March 13, 2012 by TimG

[jbg]

You have failed to do so in 30 pages of posts.

He is great at multi-color, multi-level posting however. But he totally disrespects "eyeballing".

[waldo]

Because it is an irrelevant distraction. I posted a technical argument. If you disagree then try to refute it by actually addressing the argument made. If you can't do that then simply admit your ignorance and shut up.

an irrelevant distraction! The question isn't elaborate, isn't confounding - the question is quite 'matter of fact' and very straight-forward. You choose to repeatedly ignore the question because it, quite obviously, cuts to the heart of your BS.

your "technical argument"? Your "argument" was refuted the last time you pulled this, your obsession, out. In any case, your "argument" was soundly refuted - within this very thread. More pointedly, in line with the question you refuse to answer, your (McIntyre parroted) "argument" carries no formal weight, no formal substance, no formal bearing. But your "argument" sure plays well within the isolated confines of the denialsphere. Again, just answer the question... the question is not your declared 'irrelevant'. The question is not your declared 'distraction'. Just answer the question; again:

in your fluster to get Michael to answer your questions, how about you finally answer the one you keep avoiding - like the plague!

Much to your chagrin, prevailing science/papers isn't determined through denier blogs. Denier blog science does not rule! Clearly, legitimate scientists respond to formal challenge through the recognized scientific peer-review/response cycle

.

as I said just a few posts back, McIntyre has flogged this to the nth degree... he spawned a veritable cottage industry with his acolytes carrying his "decree" forward. As you're well aware, aside from McIntyre, there's also the cast of usual suspects who purport to have taken the Mann et al paper/supplementals open source apart. Yet... yet... over 3+ years not a single one them, including McIntyre, has seen fit to shift off the cozy confines of their sheltered and safe blogWorld to issue a formal challenge to the paper. Most remarkable in that, of course, Mann is the denier's symbolic attack figure head... so remarkable that, apparently, no one can bring themselves forward to attempt to formally/officially usurp the Mann et al paper. Go figure!

why is that TimG? Why can't McIntyre, why can't his posse, put forward a formal challenge? Is there a problem? 3+ years and counting... is there a problem? In all your personal puffery, in your recently stated 100% certainty, why don't you take up the charge... personally? Bring him down... bring the Mann down! Is there a problem?

just answer the question

[waldo]

You seem to have lots of problems with "eyeballs". At least I'm not alone.

The warmistas are never persuaded by actual temperature data.

He is great at multi-color, multi-level posting however. But he totally disrespects "eyeballing".

your last 3 posts are a testament to your drive-by nothingness... carry on!

[Michael Hardner]

no need to get testy and personalize this, hey Timmay!

Don't taunt, Waldo... it's not conducive to positive discussion...

[Michael Hardner]

You are clueless fool who has nothing

Insults...

[TimG]

Insults...

Sorry - it is frustrating dealing with someone who clearly does not understand the material or the arguments presented yet makes post after post insisting he does.

[Michael Hardner]

Still reading about this... Tijlander used varve thickness and Mann used the data but flipped the sign prior to transformation (?) and got a different coefficient as a result of the reconstruction.

That's two issues right ? Am I understanding that correctly ?

[TimG]

Still reading about this... Tijlander used varve thickness and Mann used the data but flipped the sign prior to transformation (?) and got a different coefficient as a result of the reconstruction.

The primary proxy in Tijlander is the X-ray density (more minerals/less organics mean higher densities and higher densities are claimed to mean colder temperatures). Varve thickness is also used but it does not have a relationship to temperatures which is as straight forward.

The X-ray density proxy is so contaminated that it suggests a dramatic cooling in the last 100 years that obviously did not occur. So when Mann used regression against temperatures over the last 100 years to determine which proxies should he should keep, he flipped the X-ray proxy upside down and found a spurious correlation with temps where there was none.

This correcting this error means one of the main conclusions of the paper is false (I don't want to get into this argument now - I would rather get agreement that the error was made first).

[Michael Hardner]

The primary proxy in Tijlander is the X-ray density (more minerals/less organics mean higher densities and higher densities are claimed to mean colder temperatures). Varve thickness is also used but it does not have a relationship to temperatures which is as straight forward.

Ok.

The X-ray density proxy is so contaminated that it suggests a dramatic cooling in the last 100 years that obviously did not occur. So when Mann used regression against temperatures over the last 100 years to determine which proxies should he should keep, he flipped the X-ray proxy upside down and found a spurious correlation with temps where there was none.

Flipped the proxy upside down means he fed the data into the model with a negative sign ?

Isn't that a separate issue from using contaminated data ?

This correcting this error means one of the main conclusions of the paper is false (I don't want to get into this argument now - I would rather get agreement that the error was made first).

I need to understand this one step at a time. Feel free to bail at any time, as this is as painful for me as for anyone... and I'm realizing that I will only arrive at the same impasse that lots of smart people have arrived at.

[TimG]

Flipped the proxy upside down means he fed the data into the model with a negative sign?

He used the data in a way opposite to what is required by the physics of the proxy. e.g. if he had mecury expansion data he used the data in a way that assumes that the volume of mecury increases as temperatures go down.

Isn't that a separate issue from using contaminated data ?

The contaminated data is why the error occurred. Without the contamination Mann's algorithm would have worked fine. i.e. I am not saying there is a flaw in Mann's algorithm - I am only saying that it was used incorrectly with this particular dataset.

I need to understand this one step at a time.

One step at a time is the only way to approach these discussions. Edited March 13, 2012 by TimG

[Michael Hardner]

He used the data in a way opposite to what is required by the physics of the proxy. e.g. if he had mecury expansion data he used the data in a way that assumes that the volume of mecury increases as temperatures go down.

Let's be specific here. The data was signed/unsigned in the original data set ? Did he change the sign before adding it to the model ?

Let's take this step by step.

One step at a time is the only way to approach these discussions.

[TimG]

Let's be specific here. The data was signed/unsigned in the original data set ? Did he change the sign before adding it to the model ?

Mann's algorithm makes no assumptions about the sign. It simply looks for a positive or negative correlation with temperatures and uses the sign that produces the best correlation. In this case, it found a what it thought was a positive correlation but that correlation was an accidental result of the contamination. The physics of this proxy says the data is negatively correlated with temperatures.

This means that when Mann used this proxy to create a reconstruction he used it 'upside down' (i.e. he used it assuming a positive correlation when the physics dictates a negative correlation).

Edited March 13, 2012 by TimG

[Michael Hardner]

Mann's algorithm makes no assumptions about the sign. It simply looks for a positive or negative correlation with temperatures and uses the sign that produces the best correlation. In this case, it found a what it thought was a positive correlation but that correlation was an accidental result of the contamination. The physics of this proxy says the data is negatively correlated with temperatures.

Did he change the sign (or remove/add a sign) of the core data before putting it into his model or not ?

I thought this was what was meant by 'upside down' ?

This means that when Mann used this proxy to create a reconstruction he used it 'upside down' (i.e. he used it assuming a positive correlation when the physics dictates a negative correlation).

To say somebody "used it upside down" to me means he "used" the data with the wrong sign, as in he fed it into the model with the wrong sign. If he fed it into the model as is, and the correlation came out differently then that's something else - is that what we're talking about ?

[TimG]

Did he change the sign (or remove/add a sign) of the core data before putting it into his model or not?

Yes - he assigned the wrong sign when he fed the the data into his model.

[Michael Hardner]

Yes - he assigned the wrong sign when he fed the the data into his model.

Ok. So that's the first thing. That would be a mistake, for sure. The result shouldn't have been affected by that, though since the regression would have ended up just giving a negative coefficient and -1 * -1 is 1.

Still, a mistake.

Are we together still ?

[TimG]

The result shouldn't have been affected by that, though since the regression would have ended up just giving a negative coefficient and -1 * -1 is 1.

I am saying that once you have factored in everything the sign is reversed - the data was used incorrectly. There is no -1 * -1 that cancels the error out.

The process is works like this:

1) Screen the proxy to determine the sign of correlation

2) Use the proxy data based on that sign of correlation.

Step 1) come up with the wrong sign for this dataset.

Edited March 13, 2012 by TimG

[Michael Hardner]

I am saying that once you have factored in everything the sign is reversed - the data was used incorrectly. There is no -1 * -1 that cancels the error out.

That's due to adding the extra years with dirty that though, right ?

The process is works like this:

1) Screen the proxy to determine the sign of correlation

2) Use the proxy data based on that sign of correlation.

Step 1) come up with the wrong sign for this dataset.

I'm not quite there yet...

[TimG]

That's due to adding the extra years with dirty that though, right ?

I am not sure what you mean. The Mann algorithm screens the proxies based on correlations with temperatures in the last 100 years. Therefore the algorithm had no choice but to use contaminated data when it did the screening.

[Michael Hardner]

I am not sure what you mean. The Mann algorithm screens the proxies based on correlations with temperatures in the last 100 years. Therefore the algorithm had no choice but to use contaminated data when it did the screening.

Sorry - very busy and trying to get on this thread when I can. This should have read:

That's due to adding the extra years with dirty data though, right ?

But... it seems that this isn't the case. Mann screened Tijlander's data and used the screened data, it seems. Is that right ?

[TimG]

That's due to adding the extra years with dirty data though, right ?

It is one data series where the final 150 years are contaminated. No one "added" dirty data - it is a part of the data and needs to be dealt with properly.

But... it seems that this isn't the case. Mann screened Tijlander's data and used the screened data, it seems. Is that right ?

Screening in this case means determining if the data is correlated with temperature and, if so, determining whether it is negatively or positively correlated. This data should have been screened out because there is no correlation with temps in the calibration period. The fact that it was included at all is the problem. Edited March 13, 2012 by TimG

[Michael Hardner]

It is one data series where the final 150 years are contaminated. No one "added" dirty data - it is a part of the data and needs to be dealt with properly.

Screening in this case means determining if the data is correlated with temperature and, if so, determining whether it is negatively or positively correlated. This data should have been screened out because there is no correlation with temps in the calibration period. The fact that it was included at all is the problem.

Reconstructions were performed based on both the “full” proxy data network and on a “screened” network (Table S1) consisting of only those proxies that pass a screening process for a local surface-temperature signal. The screening process requires a statistically significant (P < 0.10) correlation with local instrumental surface-temperature data during the calibration interval. Where the sign of the correlation could a priori be specified (positive for tree-ring data, ice-core oxygen isotopes, lake sediments, and historical documents, and negative for coral oxygen-isotope records), a one-sided significance criterion was used. Otherwise, a two-sided significance criterion was used. Further details of the screening procedure are provided in SI Text.

Ok... still learning and thanks for your patience...

Screening requires a statistically significant correlation with actual temperature provided by readings ? I would assume that this was done in later days. In other words if varve measurement is a significant predictor of temperature then it would be used ? Or is it individual datasets within varve measurement group ?

From scienceblogs.com, I read this:

" This shows the distribution of proxies in the data he used. It was over 1200 in total more than 400 passed the local temperature screen. "

Does this mean 1200 data sets or 1200 readings ?

Again, thanks.

[TimG]

Screening requires a statistically significant correlation with actual temperature provided by readings?

Yes.

I would assume that this was done in later days. In other words if varve measurement is a significant predictor of temperature then it would be used?

Yes

Or is it individual datasets within varve measurement group?

Each data series is assessed individually.

Does this mean 1200 data sets or 1200 readings?

1200 datasets, but a single paper like Tiljander has several datasets (4).

http://amac1.blogspot.com/2010/08/tiljander-data-series-data-and-graphs.html

But these datasets are all dervived from the Xray measurements:

http://amac1.blogspot.com/2011/08/lightsum-and-darksum-are-calculated-not.html

Note that Mann acknowledges that he should have used a "a one-sided significance criterion" where "the sign of the correlation could a priori be specified". But he did not. If he had he would not have used the data because it would have been screened out.

Edited March 14, 2012 by TimG

[waldo]

From scienceblogs.com, I read this:

ok, Michael... good on ya, you're reading one of the good guys blogs - an actual scientist! Of course TimG will have follow-up disparaging comment for 'the weasel'! From the date on that blog's quote reference, you're also about a full year behind on the "issues" more topical review. Blog science rules! At the end of your journey you will note, for the most part, other than within McIntyre's hobby-horse home, no one really gives this issue the time of day anymore... people have moved on... the opposing sides have firmed up... well, except McIntyre and his lappers, hey TimG?

and now TimG quotes directly from "Amac" - oh my! The guy who went from an acknowledged state of 'ground zero' knowledge on the issue, who floundered at length, forever asking questions of scientists (real scientists), to the point he became the self-proclaimed expert... while never actually performing any science/any reconstructions! Blog science rules!

Michael, not to take too many cycles from your rat-hole pursuit, further to my suggestion to you, might you be inclined to answer the very question TimG steadfastly refuses to answer - to even acknowledge it? Oh wait, he did call the question an, "irrelevant distraction". I know from your writings in the past you've held to a clear distinction between peer-reviewed and blog science. In that vein, why do you believe/interpret no one has formally challenged the prevailing MBH PNAS comment on this paper... over the period of 3+ years? That challenge, if successful, would most certainly cause MBH formal pause and to possibly have them reply in kind... that formal challenge, if successful, would most certainly (even if only momentarily) affect the full statement on "1100 versus 1300" years warming. That 200 years, of course, supposedly being the "real driver" behind McIntyre's obsession!

on an "oh snap" edit moment: Let me play devil's advocate here... let me, hypothetically, up front, state that everything and anything DenierBlogWorld asserts about Mann et al 2008/9 is correct. Why won't "those denizens" of DenierBlogWorld formalize their claimed "correctness" within the recognized & accepted vehicle of scientific advancement - Peer Review? Why, why not?

Edited March 14, 2012 by waldo