Role for Eurasian Arctic shelf sea ice in a secularly varying hemispheric climate signal during the 20th century?

This is Wyatt / Curry's Stadium Wave (Marcia Glaze Wyatt, Judith A. Curry, Clim. Dyn., Sept. 2013; henceforth W+C), but you don't get a title like that past a staid journal like Climate Dynamics.

Note: the copy of W+C I started writing this from which I found at Curry's site offers graphics of truely outstanding industrial-strength awfulness. Really: if you don't believe me, go look. Most of them are completely unreadable (you need to go about half way down the comments at Curry's before you find anyone who notices this, strongly suggesting what the comments there also suggest: that few have troubled themselves with reading the paper). I subsequently found a better version elsewhere, so I can't claim ignorance as I was planning to.

Note: what this most reminds me of, off the bat, is the "Antarctic Circumpolar Wave" (White and Peterson 1996. I even made a small contribution to the literature). Its not the same - the ACW has a clearish mechanism and dynamics, and a very different though still "slow" period. But if I were W+C I'd definitely have mentioned it. Odd that they don't.

Introductory

Behavior of numerous and diverse geophysical indices – from fish populations to cosmic nuclides – fluctuate at a quasi-periodic 50-to-80 year tempo (e.g. Ogurtsov et al. 2002; Patterson et al. 2004; Klyashtorin and Lyubushin 2007).

But Ogurtsov et al. say "It is seen from Figure 2 that two significant century-scale oscillations really are in the spectrum of sunspot variability, a 90–100-year cycle (since the second part of the 18th century) and 50–60-year cycle (until the first part of the 19th century)" (my bold). The bolded bit is important: it means the 50-60 years stuff has gone by 1900. So its useless to W+C. Interestingly, O et al. only say this in the body, not the abstract - perhaps some carelessness or confirmation bias from W+C there.

Patterson et al. covers ~1440–4485 years before present, so isn't much use, especially if you actually believe Ogurtsov et al. which tells you that the periodicity comes and goes; and Patterson et al. also say the periodicity comes and goes (and the 50-to-80 stuff isn't terribly convincing in their spectral plots either).

K+L doesn't look terribly valuable; bizarrely, it was posted by J Marohasy in 2008 but without comment. And note that the copy I've pointed you at is hosted by "klimarealistene.com/" who are An organization for those who do not agree with the UN climate panel, the IPCC.

None of this matters a great deal. People often lard their papers with some boilerplate. But its odd that W+C can't find better refs to support their desire to see a 60 year cycle.

Onwards

WKT invoked numerous observational and model-based studies to support suggested physical dynamics potentially conveying connectivity within the stadium-wave network. Candidate mechanisms include...

This very definitely reminds me of numerous poor met/climate papers I've read that string together chains of "possible mechanisms" but end up offering nothing solid.

...This paper extends WKT by investigating the underlying physical mechanisms associated with the stadium wave through analysis of an expanded network of geophysical indices... The expanded index collection provides insight and perspective on attribution and potential predictive capacity of the stadium wave...

Um, so, still no fundamental dynamics then? Just more looking at indices? While I'm here, I'll also note that

WP hypothesized that failure to find the signal in any of the CMIP-model simulations might reflect the absence or poor representation of network dynamics fundamental to signal propagation

feels like the kind of model-bashing you'll see as a staple at WUWT.

Methods

Prior to analysis, all raw indices were linearly detrended (least squares method)... The cornerstone of our analysis is the Multichannel Singular Spectrum Analysis (M-SSA...). M-SSA is used to extract and characterize dominant spatio-temporal patterns of variability shared by indices within a network. The technique is particularly skillful...

This may be a sign of trouble ahead, because it reads very much like "we took a pile of data and stuffed it into a stats package we found but didn't understand"; Curry has form in this area.

Other evidence that the stats are dodgy: in methods, W+C say:

One-hundred-year time series for each index were auto-correlated. The auto-correlation plots showed that the maximum autocorrelation after one year among the eight indices was r~0.65. Using this auto-correlation value in the Bretherton formula (Bretherton et al. 1999), the effective number of degrees of freedom was estimated to be 40. From this, the projected decorrelation time is N/N* = 100/40 = 2.5 years

which is plausible. However in the caption to figure 3 they say Error bars are based on North et al. (1982) criterion, with the number of degrees-of-freedom set to 40, based on decorrelation time of ~2.5 years which has reversed the causality between d.o.f. and decorrelation time. By itself its a small point; but it doesn't inspire confidence in their stats.

Hiatus?

Curry's PR is pushing this paper as "‘Stadium Waves’ Could Explain Lull In Global Warming". But you should never trust PR. The denialists are all over this, in a rather short-sighted way: they love the "lull" or "hiatus" in the PR, because in their odd minds that means "global warming has stopped / was always a fake". But the only possible interpretation of this paper, were it correct, would be that GW is valid and this is merely "variability". However, very few people are actually reading the paper, which Curry is relying on.

Because the claims of explanatory power, and even more the claims of predictive power, that Curry's PR is pushing just aren't backed up by the paper:

We suggest that the stadium-wave hypothesis holds promise in putting in perspective the numerous observations of climate behavior; offers potential attribution and predictive capacity; and that through use of its associated proxies, may facilitate investigation of past behavior that may better inform our view of future behavior.

And the future? Again, we've got people sounding off:

“The stadium wave signal predicts that the current pause in global warming could extend into the 2030s,” Wyatt said, the paper’s lead author

but those are Wyatt's words, not those of the paper. The paper itself offers little in the way of prediction: While evidence strongly supports our hypothesis of a secularly varying climate signal propagating through a hemispheric network of synchronized ocean, atmosphere, and ice indices during the 20th century, we cannot know if this variability, tempo, and sequential chronology will continue into the future is about it.

Predictions

Yeah, we love predictions, they're really good. W+C are bold enough to make at least one:

Rebound in WIE, followed by ArcSib should occur after the estimated 2006 minimum of WIE and maximum of AMO...

An odd sort of prediction, you might say, because 2006 is well in the past. But their figure 3 (which I take to be an important figure; its one of the few that can actually be read) only runs up to 2000. I'm rather unclear what period their study covers; I'm going with the labelling on their figures, because I can at least see that; the "data" section is tolerably vague about time periods; at one point it says " These indices include 20th-century instrumental values of well-known indices" which is at least consistent with stopping in 2000. Its an odd end-point though; you're throwing a way a decade and more of good-quality data.

Its also an odd sort of "prediction", because as we all know Arctic sea ice has been on a steep decline over the last few decades, and doesn't seem to have shown any signs of recovery post 2006. The quote is from the summary+discussion section, and you'd have expected an alert referee to pick up on this. WIE includes, according to W+C, Greenland, Barents and Kara. And a recovery or "rebound" is not obvious.

To be continued

You'll notice that amongst all this, I haven't actually offered any critique of the paper's main thesis. Well spotted. That will have to wait until I have another spare evening. In the meantime, I'd be grateful for links to other people's analysis of the paper (and I mean a detailed analysis, not just an "oh, look, there's a badger" sort) either pro or con or neutral. I had a look, but didn't find any. I did trog through the endless comments on Curry's site, and found Wyatt saying we wanted to really understand dynamics propagating and sustaining the wave. That is the essence of the paper which sounds good, but does it live up to itssss promise, Baggins?