A Simple Test of Marcott et al., 2013

see Robert Rohde to Andy Revkin
I’m no scientist but as soon as I read over the SI it had me thinking that the study cannot treat intervals of less than a few centuries with high confidence at all. Thus, how can Mann & friends already be saying that the latest hockey stick shows something decisive about an unprecedented “rate” of temp. change in any part of the 20th century? We can’t know from this study what rates of temp. change might have existed before the 20th century…. (and it’s not clear how they got their hockey stick blade at all). The pre-20th centuries are much smeared/blurred with this method. One could not know if there were rapid increases (or decreases) in certain prior periods under 300-400 years.
fyi, Robert Rohde notes that the study’s methods lack resolution in 100-400 year intervals, hence one cannot directly compare recent temp. changes to past centennial events:

“… In essence, their reconstruction appears to tell us about past changes in climate with a resolution of about 400 years. That is more than adequate for gathering insights about millennial scale changes during the last 10,000 years, but it will completely obscure any rapid fluctuations having durations less than a few hundred years…. Because the analysis method and sparse data used in this study will tend to blur out most century-scale changes, we can’t use the analysis of Marcott et al. to draw any firm conclusions about how unique the rapid changes of the twentieth century are compared to the previous 10,000 years. The 20th century may have had uniquely rapid warming, but we would need higher resolution data to draw that conclusion with any certainty. Similarly, one should be careful in comparing recent decades to early parts of their reconstruction, as one can easily fall into the trap of comparing a single year or decade to what is essentially an average of centuries. To their credit Marcott et al. do recognize and address the issue of suppressed high frequency variability at a number of places in their paper.”

I think that the Suyts article referenced above by Theo Godwin is excellent for the layman to grasp these parlour tricks. What’s going on here is so dishonest for a subject so important that it would put many a Victorian seance charlatan to shame.
I would encourage Anthony to explore the possibility of carrying that post here at WUWT to give it some more exposure.

“When hockey players are choosing a hockey stick they tend to be pretty picky about what they like, and what they simply can not stand to play with. The decision mainly comes down to the “feel” of the hockey stick. “Feel” is the term used to describe the player’s ability to accurately sense and control the puck with the hockey stick that they are using. This includes both offensive and defensive maneuvers, which consist of but are not limited to puck-handling, shooting, passing, reach and poke-checking. As a player develops these skills, they also acquire a better sense of which stick will do the trick for them.”
http://www.hockeygiant.com/hockey-giant-buying-guide-how-to-buy-hockey-sticks.html

Dave, have a look at TN057-17. It ranges up to 1950 but swings wildly around the mean with variations > 4C between samples. It is sensitive to something but it doesn’t appear to be temperature and makes one of it’s wild excursions right at 1950 (BP=0). It has me scratching my head as to why it was included as one of the proxies. What are your thoughts?
Regards,
Hank Hancock

Skiphil says:
March 11, 2013 at 8:21 pm
…Few if any of the proxies offer any (visual) support for the blade of the hockey stick, so what is in the stats “food processor” that produces such an outcome?
Turn up the heat with Viadgraph, the new climate virility drug from Hyde D. Klein!

The paper talks about resolution in a very weird way (short excerpt provided under fair use):
“Because the relatively low resolution and time- uncertainty of our data sets should generally suppress higher-frequency temperature variability, an important question is whether the Holocene stack adequately represents centennial- or millennial- scale variability. We evaluated this question in two ways. First, [they mistakenly simulated with white noise instead of red noise], with essentially no variability preserved at periods shorter than 300 years…. Second, spectral analysis indicates that the variance of the Holocene proxy stack approaches that of the global CRU-EIV reconstruction of the past 1500 years (2) at millennial time scales and longer (figs. S20 and S23).”
In other words, they found no preservation of variability for short periods even with the wrong kind of simulation (white noise, not red, a red noise simulation would have shown more variability) but that’s ok because the spectral analysis of their reconstruction matches the spectral analysis of the CRU 1500 year reconstruction (like that means anything). Did they even notice that the lack of preservation of variability means they can go no further? But they go further:
“Our global temperature reconstruction for the past 1500 years is indistinguishable within uncertainty from the Mann et al. (2) reconstruction”
So apparently their methodology wins 2 goals to 1. Finally, they have the temerity to add some fake variability back in as red noise in figure 3 (after using white noise incorrectly). In figure 3 they show that the 2000-2009 period is well within the temperature “probability distribution” (I hate those things) of their proxy series plus some red noise.
This paper is pathetic. Was it worth $20? Under normal circumstances for entertainment, yes. But they are using this to maintain the drumbeat of propaganda in the IPCC report, not to add to greater understanding of climate. To answer some people above, Marcott is clearly the next Mann. The climaterati will circle the wagons around him and defend this crap to the death. He will bring in new grants and go on speaking tours.

So the time resolution for the paleo data is 140 years on the horizontal axis.
The entire data set on the horizontal axis is 16000 years. In other words the possible error for the paleo data is less than 1% of the horizontal axis.
You could shift the blue line to the left or right by this amount and you would never notice it.
This quiblble is meaningless in the interpreattion of the long term data compared compared with the very raid rise in temperature as measured by instruments for the last century and a half.

@- David Middleton – Re:- previous atmospheric CO2 levels
“Two words: Plant stomata.”
Two more: radiocarbon dating.
If there had been any significant changes in atmospheric CO2 levels during the Holocene it would show up in a very profound correction necessary in the radiocarbon dating curve. The only reason that radiocarbon dating works without a big correction is that solar activity and the carbon cycle have been extremely stable for the recent period.
Until anthropogenic effects intervened….

***
Nick Stokes says:
March 11, 2013 at 4:46 pm
However, the spike of the last century does have special status, apart from being well observed. It follows the insertion of 350 Gtons C into the atmosphere from fossil fuels. That hasn’t happened before.
****
“It follows the change from .00028 to .00039 CO2 of the atmosphere from fossil fuels.”
There, fixed it for ya.

Excuse my ignorance but what annual event off the coast of Oregon creates layers of sediment that accurately measure temperature?