Friday, October 18, 2013

Drop That Paper And Back Slowly Away

What, bunnies ask, do zombies wanting to eat your brains have to do with the Stadium Wave?




Well consider, the Weasel's take on Wyatt and Curry's Role for Eurasian Arctic shelf sea ice in a secularly varying hemispheric climate signal during the 20th century.  Eli's friend actually reads the paper and notices several things.  First, the waves, that W&C claim to be displaced from one another in several different climate indicators (see the figure below for a quick look, read the link to Stoat for more details) are the result of massive filtering of the AMO, PDO, etc.  To get an idea of how much smoothing, compare the two figures below).  Second, all of the signals were normalized in the figure displayed by W&C, even though most of them had little power in the frequency pass band.  Stoat goes on to show that

---------------------------------------------------
Removing this, along with the other three that have essentially no in-band variance, leaves me with this crudely retouched version of their figure 2 a. Its now much less obviously a wave; its just three (four really, but two essentially overlay) different lines filtered to within an inch of their lives into a 60-year-ish band.

I think that’s about it, really. All the stuff about exploding sardines is just fluff and can be ignored. The “mechanisms” is an extended exercise in self-delusion.


-----------------------------------------------------

Again, what does this have to do with zombies.  Well, there is a long, as these things go, history of amateurs and pros, trying to use the PDO alone or with the AMO or the AMO alone as a climate forcing to brew statistical moonshine

One can see exactly how much W&C are loosing by comparing the five year smoothed PDO shown in red with the figure above, let alone the yearly data.

Wyatt and Curry calculated correlations amongst time series
Correlated indicies, considered along with M-SSA results potentially add further insight into dynamics associated with the signal's propagation.  Prior to computing correlations between pairs of linearly detrended normalized raw time series, values were smoothed with a 13-year running mean filter to sort out shorter-term fluctuations in order to highlight longer term behavior of indicies.  We also experimented with a variety of filter sizes from five years to 20; results were virtually unhanged  values
That, as they say is a big no-no.  Obviously even a five year smoothing eliminates much of the variability, which if left would decrease the correlations.  Detrending is also problematical.  In a quoting the devil (well the citizen-devil) mode, Eli will send you to Climate Audit and a post by Willis Eschenbach, Data Smoothing and Spurious Correlation,  and here is one from Willard Tony hisself.

Calculating correlation and assigning significance to it after you have filtered out everything except a narrow part of the variation is, shall we say, a bit like Lizzie Bordon pleading for mercy as an orphan.  Some facts are missing.

8 comments:

Cynthia said...

He looks a bit like Edward Snowden, but he's not nearly as important as him.

http://www.youtube.com/watch?v=68xkSpUaUeQ

In a more sane and just world, Edward Snowden would be hailed as a national and international hero while Obama, Cheney/Bush, Bush, Clinton, and the remnants of the Reagan and Nixon administrations would go on trial for war crimes and crimes against humanity.

Nuremberg, where are you when the world really needs you?

Albatross said...

Eli,

Sorry to be a tease, but look who was on Wyatt's supervisor committee ;) They also worked with Scafetta at some point.

Knowing that, the mathurbation in Wyatt et al. is not at all surprising. I'm sure the (cough) "auditor" will be all over this very soon, casting aspersions on Curry's credibility and honesty ;)

Fake skeptics seem to find mathurbation addictive, a bit like... oh never mind.

PS: Old Revkin seems to have taken the bait.

@whut said...

Well they did mention LOD as a proxy measure of variability. It does seem to supply the right inflection point contributions to explain multidecadal changes. Since it has no bias it is perhaps a good candidate as a time series detrender and defluctuator.

test it out here:
http://entroplet.com/context_salt_model/navigate

EliRabett said...

All Eli could think about is that the Great Mentioner has taken a Climate MOOC.

Anonymous said...

For years deniers have refused to use mathematics to separate short term noise from long terms trends, so they can claim a cooling trend that isn't there. Now deniers go to the other extreme with mathematical fantasies like this.

The one analysis the deniers will not perform is the one that matches reality: its the CO2 dears.

Regards, Millicent

Anonymous said...

"they did mention LOD as a proxy measure of variability. It does seem to supply the right inflection point contributions to explain multidecadal changes"


Inflection points (where second derivative changes sign), are highly dependent on filtering.

The problem is that second derivative is very susceptible to noise, even more so than first derivative which is more susceptible than no derivative.

THE CLIMATE WARS said...

read down on the citizen debbil, link and you'll find Watts' criticizing Spencer, which never fails to amuse:


"Thanks Dr. Spencer for spelling out your position, and for noting my contributions.

Part of the problem that citizen science has is that we are often outside of the orthodoxy/literature path. What is an easy immediate lookup of a paper from most any .edu domain can take us days of emailing, begging (do you have a copy of Smith et al I can read?), and sometimes just flat-out refusal of authors to share papers with skeptics.

Sometimes, all we are left with is the “science by press release” residual to work with.

If only “big oil” would fund our subscriptions to journals.

In any event, I hear you about the time sink. My personal and business life suffers much like yours from emails asking for comments, I’ve finally just resigned myself that I can’t answer them all and focus on those tasks which do the most constructive good.

In any event, I see this little tiff as a bump. No worries from my end."

Susan Anderson said...

I was struck by this from Spencer: I didn't expect this from him!

(italics is all quote:
C’mon, folks! Do you really think that of the billions of dollars spent on designing, launching, and keeping these satellite instruments going, that no one thought to analyze the data? Really? That’s why hundreds of scientists and engineers collaborated on such projects in the first place!

Just because you can’t find some technical issue described in blogs doesn’t mean it hasn’t been addressed. It’s in the scientific literature, and in workshop reports, conference proceedings, etc.

In retrospect, it’s now clear that public interest in climate change has led to citizen-scientists like Willis taking matters into his/her own hands, since so little information is available in a form that is easily digested by the public. Career scientists like myself have not done enough public outreach to describe what they have done. And when we do such outreach, it is usually too technical to understand. We are too busy publishing-or-perishing.

As a result, just about every time someone posts an amateur analysis of data that becomes popular, I’m asked to read it, critique it, and respond. Well, I simply don’t have the time. But these things sometimes get legs, and when they do, I get even more e-mails.

For example, I still get the occasional e-mail because the Sky Dragon Slayers took a NASA report about CO2 cooling of the upper atmosphere (which we have known for at least 50 years) and spun it into ‘proof’ that CO2 can’t warm the lower atmosphere. Well, greenhouse gases cool the upper layers, and warm the lower layers, of planetary atmospheres. Nothing new there…except maybe to misguided public perceptions of the science, which usually only involve the warming effects of greenhouse gases.

Anyway, I applaud Willis, who is a sharp guy, for trying. But now I am asking him (and others): read up on what has been done first, then add to it. Or, show why what was done previously came to the wrong conclusion, or analyzed the data wrong.

But don’t assume you have anything new unless you first do some searching of the literature on the subject. True, some of the literature is paywalled. Sorry, I didn’t make the rules. And I agree, if research was public-funded, it should also be made publicly available.


There's some nonsense in there, but it's not all nonsense!

http://www.drroyspencer.com/2013/10/citizen-scientist-willis-and-the-cloud-radiative-effect/

link from Stoat's "Woy vs. Willis" (I *did* want to put Villis ...)