Chopping Up The Hockey Stick

Wanna know why ‘science’ gets no respect these days. Read this tale of manipulated data.

http://pjmedia.com/blog/the-death-of-the-hockey-stick/

 

The idiotic post above isn't worth investing much time responding to, so I'll just copy/paste some material I posted elsewhere that addresses just one particular attack on Mann's hockey-stick made in the hit-piece linked to in that post.

The above "hockey-stick from spectral noise" argument is breathtakingly incompetent, and on a number of levels. Here's a (hopefully) not-too-technical "plain English" explanation that shows why that is the case. There is a fair bit of technical jargon here, but reasonably intelligent folks should be able to grasp the "big picture" even if they don't understand all the technical details.

 

Name-calling is not going to change the fact that Mann got caught falsifying data to make a flawed model work.

Nobody believes Mann or his apologists.

Penn state fired Joe Paterno and they should fire Mann.

 

Intelligent folks look at all the data, not just that which already backs up their bias.

 

OMG dont tell me you actually believe that (*)(*)(*)(*). Dr. Mann has been lying his ass off for years. Most of what you said are blatant lies or strawmen designed to muddy the waters. Since Dr. Mann never gets questioned on his lies he gets away with them. You probably believe that McIntyre asked for an Excell file too since Dr. Mann said so.

Lets deal with your claims one by one.

Too bad no credible statistician has agreed with that conclusion. Its nothing more than a spitball thrown out by the team. Neither Wegman or the NAS panel agreed that the hosking.sim would leave a hidden temperature signal nor has anyone been able to show that it does. Further simulation with other red noise generators have shown the same thing. That is why Amman and others used words like "may have" and "could have" because they cant show it. Furthermore other analysis using different red noise generators have shown the same thing. To quote the team member Rob Wilson from the climategate 2 e-mails

You already know the answer to this. Dr. Manns hockey stick did contain a real signal the Gaspe CO2 signal causing a higher eigen value than purely random red noise. However the eigen value is relative to the rest of the series. You get passable eigen values with a real signal or random noise. Manns eigen value method doesn't mean that you won't get a hockey stick from purely random noise. You argument is a strawman.

Your argument just shows how bad Manns method is because your high eigen value comes for a signal that isn't temperature. No on ever said that Mann's hockey stick is only wrong in one way. His inclusion of a known CO2 proxy is another. Its wrong 6 ways from Sunday. One mistake doesn't disprove another. Your argument seems to be that because Mann (*)(*)(*)(*)ed up in so many ways that he didn't (*)(*)(*)(*) up at all.

Another "red haring" argument. Of course half of them turn down. You would expect that with random data. The problem is that after the PC method the down turning PCs are either discarded or turned right side up. The fact that half of them turn down is meaningless because the PC analysis is only the first step. The later steps will deal with the down turning data. Either discarding it or assuming an inverse relationship to temperature and turning it upside down. This is critical in later proxies because for example Mann 08 turns the Tiljander series upside down falsely creating a super hockey stick. This is just another way in which Dr. Mann (*)(*)(*)(*)s up. So once again you are arguing that because Dr. Mann (*)(*)(*)(*)ed up in so many ways he had to be right.

You are falsely arguing that we focus solely on PC analysis. The hockey stick shape has been reproduce with red noise using PC analysis and not using PC analysis. Any method will inherently mine for hockey sticks regardless. Its fundamental to the approach of using a purely statistical reconstruction because temperature has only been rising during the calibration period. This is why Loehle used only proxies that have a proven scientific correlation and didn't use statistical correlation.

As for using proper Eigen value you are again just proving the (*)(*)(*)(*) up. Manns strong Eigen values don't come from crucial data but bad data. Its just another mistake beyond his problem with random data. You are arguing that because he (*)(*)(*)(*)ing up in more ways than one that he is correct. You are again arguing that because his reconstruction included a spurious CO2 signal that it is correct. No that is just another way in which it is wrong.

They did realize that. The eigenvalue method only compensates if the eigenvalue is representing a real temperature signal. In his analysis its not. Its the product of a CO2 proxy. Were the CO2 proxy not in the series you would still get a hockey stick using Manns eigenvalue method which has been proven over and over again.

Actually all you have done is show that Dr. Mann (*)(*)(*)(*)ed up in more ways than one. Which we already knew. Mann's reliance on a CO2 proxy is well accounted for in the skeptics argument. You just choose to ignore it. And its not an error. He knew full well that

You haven't in any ways shown that Dr. Manns method doens't produce hockey sticks from random noise only that Dr. Mann's hockey stick had a stronger eigen value than hockey sticks produced by random noise. You neglect to menntion however that his reconstruction contains a CO2 proxy. And that is what produces the high eigen value. You are arguing that one mistake makes up for antoher which isn't the case.

You of course know that becasue its been told to you here.

http://www.desmogblog.com/little-kn...eptics-may-be-their-confidence#comment-719102

You chose to disregard this as "moving the goal posts". You would just chose to put your head in the sand and pretend that Dr. Mann didn't use a bad proxy and instead chose to argue that one mistake justifies another.

 

Actually on Desmog caerbannog was told that his high Eigen value came from a spurious proxy, which he didn't dispute. He simply chooses to ignore it. Like I said his entire argument is that one mistake justifies another. He is choosing to take a very myopic view of the hockey stick because he believes that he can discredit that random noise argument by the inclusion of another error. This is only an effective argument if your opponent doesn't already know that your argument is based on another error. That is why he didn't mention it in his post. He is lying by omission.

The desmog post went like this.

Carabong: Random data hockey sticks don't have the same eigenvalue as the Mann hockey stick so the Mann hockey stick is accurate.

Nullius In Verba: The high eigenvalue is solely the product of a spurious proxy.

Carabong: Quit moving the goal posts!

 

What you can see with someone like caerbannog is that he is very careful in his wording. He likes to sound like he is making a scientific argument but he is really playing word games.

As you can see in the la nina thread. He goes off on a litany of issues that are common to all temperature reconstructions surface or satellite because they are inherent regardless of method but carefully words his diatribe to make it sound like these issues only apply to the satellite record while also being careful not to slip up and say that the same problems don't also apply to surface stations. Its a giant lie by omission. Now I work a lot with kids and about once a week I have to explain to a child how lying by omission is still lying. They tend to get it. So even children understand the concept. That is why I really don't tolerate lying by omission by adults. Its seems that only warmmongers and politicians cant grasp it.

 

The more I look around the internet where carabannog has posted else where and I think I see where carabannog fracked up. He never really read MM2003. I think that caerbannog just took everyone elses interpretation of what MM2003 says. He is also conflating refutations of other hockey sticks. Caerbannog is arguing that purely random red noise wont produce a hockey stick. Congratulations cearabonnog you have proven that you never read MM2003 I guess I gave you too much credit. M&M have never once said that you get a hockey stick from pure red noise. They have always said there has to be a hockey stick within the series. Their analysis is of psudo-random noise. They have never said otherwise. If you ever bothered to read what they wrote you would know that. They have never once argued that and they routinely correct not so informed skeptics that make that mistake on their site climateaudit.org perhaps you should go there and learn something. As far as the Mannian PC method MM and every other major player in the skeptic community have always said that the MBH98 PC method mines for hockey sticks. The hockey stick has to exist in at least one of the series. Any series that contains a hockey stick no matter how many other series are in the reconstruction will produce a hockey stick. When you ran your analysis you didn't use the same pseudo-random data that MM used so your series didn't contain any hockey sticks to mine for.

Now where I think you are conflating things is that most of the other hockey sticks dont use PC analysis and they will in fact produce from purely random red noise which has been documented many many times. I believe you are confusing the analysis of these reconstructions with the analysis of MBH98 done by M&M. It probably is all one big hodgepodge in your mind.

 

Now Guys....Penn State did an internal closed investigation of Mann and his data and climate gate and determined he did nothing wrong..........

And we all know Penn State never covers up any wrongdoing........

Elmer.....PSU 1979, and having a hard time being proud of it these days between Sandusky, the Admin, the Board, and Mann......

In many ways Mann is the bigger criminal than Sandusky.....Mann is trying to screw the lifestyle of 300 million people to protect his academic integrity....

 

It was not just Penn State but several investigators

But like "bothers" denialists just will not let go

 

LOLOLOL.......oh my, denier cultists and their mindless myths are soooooo funny.... and sooooo full of it.....

Climate myths: The 'hockey stick' graph has been proven wrong
In fact, later studies support the key conclusion: the world is warmer now than it has been for at least 1000 years.
New Scientist
Michael Le Page04 September 2009
(excerpts)

The "hockey stick" graph was the result of the first comprehensive attempt to reconstruct the average northern hemisphere temperature over the past 1000 years, based on numerous indicators of past temperatures, such as tree rings. It shows temperatures holding fairly steady until the last part of the 20th century and then suddenly shooting up. It provided yet more evidence that the rise in greenhouse gases due human activity is causing warming, although the case for this was already very strong. The conclusion that we are making the world warmer certainly does not depend on reconstructions of temperature prior to direct records.

2006 report of the US National Academy of Science (pdf). The academy was asked by Congress to assess the validity of temperature reconstructions, including the hockey stick. The report states: "The basic conclusion of Mann et al. (1998, 1999) was that the late 20th century warmth in the Northern Hemisphere was unprecedented during at least the last 1000 years. This conclusion has subsequently been supported by an array of evidence that includes both additional large-scale surface temperature reconstructions and pronounced changes in a variety of local proxy indicators, such as melting on ice caps and the retreat of glaciers around the world".

Update: as suggested by the academy in its 2006 report, Michael Mann and his colleagues have reconstructed northern hemisphere temperatures for the past 2000 years using a broader set of proxies than was available for the original study and updated measurements from the recent past. The new reconstruction has been generated using two statistical methods, both different to that used in the original study. Like other temperature reconstructions done since 2001 (see graph), it shows greater variability than the original hockey stick. Yet again, though, the key conclusion is the same: it's hotter now than it has been for at least 1000 years. In fact, independent evidence, from ice cores and sea sediments for instance, suggest the last time the planet approached this degree of warmth was during the interglacial period preceding the last ice age over 100,000 years ago. It might even be hotter now than it has been for at least a million years.





(In accordance with Title 17 U.S.C. Section 107, this material is distributed without profit to those who have expressed a prior interest in receiving the included information for research and educational purposes.)

 

Appeal to numbers is a logical fallacy livefree. You have no reason to believe that they aren't making the same mistakes. And since they all work together and collaborate by e-mail we know that is a fact.

 

Nice to see that an open mind

 

Liar.

Not one datum in Mann was falsified. Unlike your false statement above. Which I flatly predict you don't have the honesty to retract.

 

Hmmmm. What's inside that ellipsis, one wonders? Did Windigo even check it out? Might it be important? Or did Windigo just lift that directly from WUWT, where the same quote contains the exact same ellipsis in the exact same place?

Reading what's inside that ellipsis proves enlightening. So how did Rob Wilson actually generate hockey sticks? Just randomly? Nope. Not even close!

Here's what Anthony Watts hid from his readers, and what Windigo hid from you:

In other words, out of 1000 random series, Wilson selected that less-than-5% that just happened to look like hockey sticks. And in fact, since his selection CL was only 90%, the fact that he only got 48 (rather than the expected 100) indicates that the actual NH signal is pretty darn non-random.

Nice job lying, Windigo.

 

You again don't understand what the hell you are talking about. It will be very hard to explain to you exactly how proxy reconstructions work because you are obviously clueless.

First your mistake is assuming that all the series included in a reconstruction actually make it into the reconstruction. This isn't true. Most of the series will be filtered out because they do not correlate well to temperature. Then either through PC eigen weighting or weighting on the correlation to temperature or both even more will be thrown out as their weights will be insignificant. Eventually you end up with only a handful of proxies accounting for the bulk of the final reconstruction. In the case of MBH98 its final shape was dependent on only one series and that series was not a temperature proxy. It was a CO2 proxy so all Dr. Mann showed was that CO2 has climbed in the 20th century.

 

You might want to check MBH98 Mann makes the fraudulent claim that his method had passed R and R2 verification. It didn't. He lied.

 

Repetition does not exonerate Mann.

 

Mann needs no exoneration. He was right and has been confirmed as such.

 

Is anyone denying the earth is getting warmer?

I mean really.
Why spend so much time trying to pick apart one persons data when the whole scientific world knows the earth is getting warmer?

For you people from countries other than the U.S, this is called "The bait and switch"

 

So I guess than means you now admit that Loehle 2007 (whose graphs you have posted and praised) is a pile of crap. Because he didn't do any of those things.

 

But that's not what you said. You said Mann falsifed his data. When I challenged you to find such a datum, you move the goalpost and say, well no, he falsified statistics.

We can have the statistical argument later, but only after we agree on this: Mann didn't falsify a single datum. And you lied when you said he did.

 

Your R values are data. They are formally called residuals. By falsely claiming that his model passed R verification he falsified his data.

 

Actually Loehle didn't use tree rings and did no statistical reconstruction. What separates Loehle from other reconstructions is that he used verified temperature proxies with a proven scientific relationship between to temperature rather than the pure correlation models that most temperature reconstructions are. Since it has already been proven that a pure correlative reconstruction will always produce a hockey stick.

 

And that is the key. The argument of whether the MWP was warmer or Mann's studies is accurate are just tactics used to distract us from the obvious. If just the sea ice were declining and the glaciers were expanding and the oceans were cooling and growing seasons were getting shorter, the skeptics might have a point of the data being corrupt. But the sea ice is declining, the glaciers are melting, the oceans are warming, the oceans are becoming more acidic, the growing seasons are getting longer, the nights are getting warmer. ALL lines of evidence from all over the globe by scientists of numerous countries and are pointing in the same direction.
Some acknowledge that the system is getting warmer but it is not CO2. Then the same people show up in a thread and argue that the data is corrupt. How consistent is it in one post to acknowledge the earth is getting warmer and in the next post to deny the data?