A friend of mine was working on a lens calculating program. HE went to the standard Optical reference, a published collection of papers in one volume.

He was working through it to find the relevant math. He came across one article that was particularly helpful, and he found it very easy to understand. So he went looking for the authors so that he could find more material.

Well then it made sense why it was so understandable. He had written it years previously with a colleague.

When I re-read this post, I must confess that I could barely remember writing it. I had a similar experienc when I was googling and saw an interesting statement on the internet about the very different weights for two Colorado PIFL series in MBH a few miles apart, one with HS and one without – I’d forgotten all about it, but it was an interesting observation and worth incorporating in one of the zillions of papers that I’m supposed to write. (But don’t blame the blog; if I’d not written this down for the blog, I’d have still made the calculation and then totally forgotten about it.)

This result suggests that the large-scale loss of climate sensitivity documented by Briffa et al. (1998a, b) is unique to the 20th century, which argues for an anthropogenic cause ….

Now wait a minute. The proxies are no longer sensitive to temperature? And we are using them as temperature proxies? But they are wonderful proxies for temperature 1,000 years ago? And we know this how?

“But the 20th century ramp here is largely created by two high-altitude low-latitude foxtail sites, where CO2 and other fertilization (phosphate, nitrate) issues have been specifically identified by specialists.”

Interesting. The Bristlecone site are about 70 miles away from one of the most intensely farmed and fertilized agricultual regions, the Central Valley. The famous Fresno Fog increases distant tree growth? Might check trees on the western side of the Sierra Nevada, the effect there would arguably be more prominent.

Comment by Follow the money “¢’‚¬? 14 March 2006 @ 5:46 pm

…(and played in a squash tournament)…

And haven’t told us the result. Is that because you are feeling grumpy, or feeling modest ?

Steve: The bad news was reported here

Comment by fFreddy “¢’‚¬? 14 March 2006 @ 6:00 pm

Thinking more about the Owen’s Valley and fog…there’s the famous Owens Lake dust storms of certainly anthropogenic origin.

Should have known you discussed it before!

Owens Lake Water Diversion for L.A. and Bristlecones
Could account for increased phosphate, etc. And doesn’t need to hike over the Sierra.

Comment by Follow the money “¢’‚¬? 14 March 2006 @ 6:00 pm

What would be the result if instead of using latitude for North-South, the clmatological growing classification was used?
(The one used for gardeners)

Comment by Gerald Machnee “¢’‚¬? 14 March 2006 @ 7:25 pm

4. The presence/absence of Jasper and Quebec don’t look like they make any difference. The results are driven by foxtails (and Jacoby’s Sol Dav, Mongolia.) How many times have we seen these series turn up like bad pennies? So how seriously should we take it if these series are now said to be magic bullets that prove that there exists some still unknown anthropogenic cause for the Northern sites? It doesn’t rise above cargo cult.

Comment by Steve McIntyre “¢’‚¬? 14 March 2006 @ 7:28 pm

And another point about fertilizers and desert lakes without outlets. How about potash? Owens lake had a history of extraction attempts for potash. I know they had a soda plant there in the past.

In the immediate area around the lake the vegetation is brushy or high steppe. Would be interesting to see if the east side Sierra trees upslope record increased tree ring growth. Could correlate with the White mountain bristlecones and dust/smoke from Lake Owens.

Comment by Follow the money “¢’‚¬? 14 March 2006 @ 7:38 pm

But, how did you do in the squash tournament?

Steve: Lost in the finals of the Canadian over-55 doubles, which was very good for me. Beat last year’s champions in the quarters and the top seeds in the semis and lost to the #2 seeds. Close but no cigar. Was bone tired after playing 4 matches in less than 48 hours, which is a lot at my age.

Comment by jae “¢’‚¬? 14 March 2006 @ 9:15 pm

“In short, Cook et al 2004 resolves nothing. They provide no explanation of the Divergence Problem for North Sites that rises above cargo cult science.”
This is not the issue – the issue is not whether the divergence is explained. The issue is that the same series that show divergence in 20th century do not show this in MWP, thus the argument that there might be a divergence in MWP is not supported by the data. The late 20th century divergence remains unexplained, but one should make it clear that the issue is not a general issue for tree rings, and that the issue is over-hyped by many here.

Comment by bart s “¢’‚¬? 15 March 2006 @ 2:10 am

But that’s just it, if nobody understands WHY divergence is happening, how can you be sure it wasn’t happening in the past, and we just can’t detect it? Maybe if the proxies were divided up into different bins, rather than say “North” & “Not so North” like they do now, divergence in the past WOULD be evident? Just because the selection of proxies studied here have divergence in the 20th century and not the 14th does not mean a different set of selections, maybe even from the same pool, would not show similar properties in a different time period.

Or put it this way, you get some proxies from location X that tell you temperature went up in 1492. You get proxies from location Y that tell you temperature went down in 1492. Now, you can claim that this means there was no average change in temperature in 1492 because they cancel out. But how do you know this isn’t just a case of divergence? Just because divergence is more pronounced now does NOT rule out its happening at other times.

Statistical methods will only serve to give meaning to these kinds of errors. That’s the last thing you want in a statistical analysis. Especially if finding nothing of significance (i.e. no trend in the middle ages) is in itself significant.

Comment by Nicholas “¢’‚¬? 15 March 2006 @ 2:37 am

It seems that the phrase “the Divergence Problem” is a euphemism for “not a temperature proxy”, or “misleading if used as a temperature proxy”.

Comment by JerryB “¢’‚¬? 15 March 2006 @ 3:20 am

Re 2,7 – Not bad at all. Take “em next year.

Comment by fFreddy “¢’‚¬? 15 March 2006 @ 5:26 am

“The issue is that the same series that show divergence in 20th century do not show this in MWP, thus the argument that there might be a divergence in MWP is not supported by the data. ”

How exactly do you know there was no divergence in the MWP?

The point of the “divergence problem” is the divergence between the tree ring record and the thermometer-based “instrumental” temperature record – if there are no thermometers…

Comment by John S “¢’‚¬? 15 March 2006 @ 5:26 am

Bart, I think that John S has pretty well nailed this point.

Comment by James Lane “¢’‚¬? 15 March 2006 @ 5:43 am

re 9&13. Nicholas, go and read Cook´s paper. He has “binned” the data as you suggest, and the results are quite convincing. The 20th century pattern is not seen in earlier times, thus tree-rings seem to work in the same manner as in the 18th and 20th century prior to the klast decades. It is the last 50 years that are odd, and only at high lat, not elswhere. People are therefore making too much out of this and the issue is over-hyped. For what purpose one can only guess…The divergence in the 20th century has a geographical pattern. There is no such pattern in the preceeding centuries of the last millennium. If one is concerned about the issue of the divergence in some areas in the last decades, one should be equally aware of the times and areas when there is consistency.

Comment by bart s “¢’‚¬? 15 March 2006 @ 6:53 am

Why do they even go back 500 or 1000 years into the past, when they only have accurate temperature data for the past 100 or so years? Aren’t they over-complicating their task?

They are attempting to not only predict the future, based on the past 100 years, but also to predict (or “post-dict”) the past. They don’t have data for the future, nor do they have accurate data for the past.

Half a degree, 1000 years ago? When temperatures fluctuate by 10 degrees, or more, day-to-night/day-to-day/month-to-month/season-to-season?

Do they think they’re gods?

Comment by BradH “¢’‚¬? 15 March 2006 @ 7:16 am

I’m sorry but I still don’t see how it’s valid to assume that tree rings are good temperature proxies if, for the only period where we have both instrumental and proxy data, much of the proxy data fails to track the temperature for a large portion of that period.

What makes the assumption particularly startling to me is this: we only have instrumental data from the last 150 years or so. Also, if we’re going to have any major effect on the climate, it will also be during the last 150 years or so. Therefore, deviation of proxies from instrumental data can *always* be excused as some kind of anthropogenic forcing. However, where is the evidence that is the case? Just because historical proxy data correlates with other proxies does NOT prove that ANY of the proxies are actually accurately reflecting your chosen piece of data.

I’ve read some of Cook’s paper and it didn’t address the quandry above. It’s the same quandry that we get out of the “bristlecones” and other proxy series again and again, which is that they don’t seem to be accurate proxies now, so what makes us think they ever were accurate proxies? In the absense of strong evidence to suggest they are good proxies, I think the prudent course of action for anyone is to assume they are not.

I grow sick of this. The only reason many proxies seem to be considered temperature proxies is because they correlate with temperature. Yet many of them actually do not! In the absense of a very, very good reason for that I can’t understand defending them. “Unknown anthropogenic forcings” (i.e. fudge factor) is not a sufficiently good excuse.

Comment by Nicholas “¢’‚¬? 15 March 2006 @ 7:22 am

“Half a degree, 1000 years ago? When temperatures fluctuate by 10 degrees, or more, day-to-night/day-to-day/month-to-month/season-to-season?”

Wow, now that IS a disingenuous comment. Dave, I KNOW you know the difference between weather and climate and that you KNOW variation between years, or decades, is way less than 10C on a global or hemispheric basis.

“Do they think they’re gods?”

No, do you think Steve is one?

Steve: Peter, please stop ragging.

Comment by Peter Hearnden “¢’‚¬? 15 March 2006 @ 7:27 am

Sorry, I mean Brad not Dave – doh!

Comment by Peter Hearnden “¢’‚¬? 15 March 2006 @ 7:28 am

bart s, the “divergence” problem is so ubiquitous it even occurs on basically the same site among the same species, let alone over large geographical areas:

http://www.climateaudit.org/?p=478

Comment by beng “¢’‚¬? 15 March 2006 @ 7:44 am

#14 ROFLMAO… if there are no records to diverge from then there isn’t a problem then Bart? Look even Hughes admitted at the NAS that there is no way of knowing whether there has or hasn’t been divergence prior to the establishment of an instrumental temperature record in the 19th Century. In that case, how does anyone actually trust the proxy record as being remotely accurate?

Comment by John Lish “¢’‚¬? 15 March 2006 @ 7:55 am

16. I agree. There is now a tremendous amount of evidence that tree rings are simply not generally suitable as temperature proxies. It’s time to “move on.” There are other proxies that are much more reliable, such as the Sargasso Sea proxies. The problem is that if these good proxies are used, they clearly show a MWP and LIA, thereby spoiling the hockey stick and clearly showing that we very well may be experiencing part of one of the normal warming and cooling cycles.

By the way, it is relatively easy to demonstrate statistically that the averaging of proxies with a lot of red noise (caused primarily by large dating errors in tree ring data) will “wash out” any long-term non-linear trends, leading to a “hockey stick shaft.” The greater the number of such proxies included in a “reconstruction,” the straighter the shaft gets. Individual proxies make much more sense than “averages.”

Moreover, there are some good non-linear models that correlate very well with the better proxies, such as the Sargasso Sea proxies.

Comment by jae “¢’‚¬? 15 March 2006 @ 11:05 am

Things like the Sargasso Sea proxies are not necessarily magic bullets for millenial reconstructions. Their accumulation is slow. When you examine relative dating, there’s lots of wiggle-matching which is hard to prove.

Comment by Steve McIntyre “¢’‚¬? 15 March 2006 @ 11:32 am

Steve: I need to get a paper to you. How can I do it?

Comment by jae “¢’‚¬? 15 March 2006 @ 11:36 am

First off I apologize in advance if this has already been discussed. Traveling and so I don’t get to read everything as thoroughly as normal.

But Steve has started to use the term “Cargo cult” I like the sound of it, but didn’t know the real meaning, So I searched, and since this is the first result I’m sure others have done the same and found it, I’m sure Steve knows about it and all, but I so enjoyed reading it, and it is so directly applicable to all of the discussions here I figured I’d post it for those that may not have seen it.

http://www.physics.brocku.ca/etc/cargo_cult_science.html

Feynman I’m sure could have used direct examples from climate research if he’d done the talk now instead of 1974.

_____________________________________________________
“It’s a kind of scientific integrity, a principle of scientific thought that corresponds to a kind of utter honesty–a kind of leaning over backwards. For example, if you’re doing an experiment, you should report everything that you think might make it invalid–not only what you think is right about it: other causes that could possibly explain your results; and things you thought of that you’ve eliminated by some other experiment, and how they worked–to make sure the other fellow can tell they have been eliminated.”

R Feynman

Comment by ET SidViscous “¢’‚¬? 15 March 2006 @ 11:52 am

I’ve updated this post to show my emulation of the North and South Site averages, which do not show the same amount of reported coherence as Cook et al. The correlation over the full period is very low, but is relatively high in a selected subperiod. Also I’ve plotted up the autocorrelation functions of the two series.

Cook et al indirectly acknowledge a lack of coherence in the MWP, but do not report the results, which are worse in this emulation than people would expect from reading Cook et al. Whether anything rises to statistical significance relative to similar autocorrelated series is not discussed in Cook et al. – hey, this is the Hockey Team. I’ll bet that David Stockwell could answer this pretty quickly. If not, I’ll return to it on a later occasion.

Comment by Steve McIntyre “¢’‚¬? 15 March 2006 @ 12:19 pm