Keith, the censorship at RealClimate is central to the theme of this thread. How is my comment out of line? I am glad to restate it in its most factual form:

I'd love to state that at his blog, DeNihilist, but I am censored there, so I cannot do so ...

How's that? No personal anything, just the facts.

Lucia, the "wo both" only contains a hockey stick if we include the instrumental data. I see no theoretical basis for that inclusion, nor is one presented in the paper. In general, proxy and instrumental data cannot be compared directly.

For example, note the almost 0.4°C difference between the "wo both" reconstruction and the instrumental data around 1960. Given those differences, we cannot simply add the instrumental data to the mix.

Gavin, my apologies. I had clicked on your link, saw it looked nothing like the Figure 8a in the Mann2008 paper and didn't say Figure 8a, saw it was not published until 2009, assumed your link was incorrect, and went back to the paper. You had said that:

This issue was clearly acknowledged in the M08 paper and both of these possibilities (with and without 'tiljander') were shown

However, both of the possibilities were not shown in the M08 paper as you claimed, so I fear your claim was far from clear. To avoid confusion you might have mentioned that this was a revision from a year later, not (as you claimed) what the paper originally said.

You are correct, they later published a figure which looks nothing like 8a. However, their figure (without both sets of proxies) totally substantiates my claim. Taking out both the bristlecones and the Tiljander proxies makes a very large difference in the early years of the reconstruction. With the bristlecones and the Tiljander proxies, the most recent proxy values were equal to or higher than historical values.

Without them, however, things are very different. The historical values dwarf the current values, and lead to a very different conclusion than the one drawn in the original paper.

Which was my point.

I'd love to state that at his blog, DeNihilist ... but I am ruthlessly censored there ... science at its finest.

Post, Part 2:

So that is the situation regarding Tiljander. Now, let me return to how this relates the subject of the thread. This discussion of Tiljander, and Gavin's willingness to answer, shows the great value of climate scientists engaging those of us outside of mainstream climate science who also study these issues. It allows everyone to see whose claims are supported by the data. In this case, we can all see that, despite his fervent belief that this issue has been decided and settled, Gavin's claims are wrong ... which is how science progresses.

I applaud Gavin for his willingness to discuss this question here, and I encourage other climate scientists to do the same. Yes, the blogs are a rough-and-tumble kind of place ... but the mainstream scientists are losing the battle for the public (and hence the political) support for their ideas. I believe that in large part this is because most of the mainstream scientists are unwilling to publicly discuss and debate the issues. Various reasons are given for this unwillingness, some of which are excellent reasons. Others of the reasons, such as Gavin's claim that the issues have already been settled, may or may not be valid.

But the net result is that we all lose when mainstream climate scientists are unwilling to publicly defend their results. The public loses because they don't get to hear the mainstream views. The skeptics lose because whatever erroneous claims we might have don't get corrected.

But the mainstream climate scientists lose triple. First, erroneous skeptical views go unchallenged. Second, erroneous mainstream views (such as the idea that Mann2008 looked at the effect of removing both bristlecones and Tiljander) do not get corrected. But the greatest loss is that the public looks at the blogs and says "Well, if the mainstream scientists won't show up to defend their views, maybe they're not really sure about them, maybe there's something they don't want us to find out."

I don't think that's true ... but it is certainly the impression that the public is left with.

So I am grateful to both Keith and Gavin for providing both a theoretical subject and an practical example to discuss these issues.

My regards to all,

w.

Keith, please allow me a short excursion through the Mann 2008 sensitivity analyses, because they have great implications for the subject of this thread. I will have to split it into two parts because your blog won't allow long posts, so this is part one. Gavin, you say:

Given the methodology used in that particular paper (Mann et al, 2008) (weighting based on a local calibration to temperature in the modern period), the 'tiljander' proxies can only be used one way. If there is a contamination in the modern period by non-climatic influences (which the originating authors suggested there might be), then they just can't be used. This issue was clearly acknowledged in the M08 paper and both of these possibilities (with and without 'tiljander') were shown (it made almost no difference to the final reconstruction).

Supp. Fig. S8a

(Note that figure also shows the difference dropping all the tree-ring data makes, and what happens when the neither the tree ring data nor the Tiljander proxies are not used).

Sadly, this is not true. Contrary to Gavin's claim, Figure 8a does not show the situation when neither the tree-ring nor the Tiljander proxies are used. The caption for Figure S8a says:

Fig. S8. Comparison of long-term CPS NH land (a) and EIV NH land plus ocean (b) reconstructions (full global proxy network) both with and without the seven potentially problematic series discussed.

The "seven potentially problematic series" are the four Tiljander series, along with three others. Here is the entirety of what the Mann2008 SOI says about the sensitivity analyses they conducted:

Sensitivity Analyses (NH Temperatures). Sensitivity to use of tree-ring data in early centuries. We examined the influence of use of tree-ring data on the resulting long-term CPS and EIV reconstructions in Fig. S7 below.

Potential data quality problems.

In addition to checking whether or not potential problems specific to tree-ring data have any significant impact on our reconstructions in earlier centuries (see Fig. S7), we also examined whether or not potential problems noted for several records (see Dataset S1 for details) might compromise the reconstructions. These records include the four Tijander et al. (12) series used (see Fig. S9) for which the original authors note that human effects over the past few centuries unrelated to climate might impact records (the original paper states ‘‘Natural variability in the sediment record was disrupted by increased human impact in the catchment area at A.D. 1720.’’ and later, ‘‘In the case of Lake Korttajarvi it is a demanding task to calibrate the physical varve data we have collected against meteorological data, because human impacts have distorted the natural signal to varying extents’’). These issues are particularly significant because there are few proxy records, particularly in the temperature-screened dataset (see Fig. S9), available back through the 9th century. The Tijander et al. series constitute 4 of the 15 available Northern Hemisphere records before that point.

In addition there are three other records in our database with potential data quality problems, as noted in the database notes: Benson et al. (13) (Mono Lake): ‘‘Data after 1940 no good— water exported to CA;’’ Isdale (14) (fluorescence): ‘‘anthropogenic influence after 1870;’’ and McCulloch (15) (Ba/Ca): ‘‘anthropogenic influence after 1870’’.

We therefore performed additional analyses as in Fig. S7, but instead compaired the reconstructions both with and without the above seven potentially problematic series, as shown in Fig. S8.

Note that, contrary to Gavin's claim, they first looked at the "full global proxy network" both with and without tree rings (Fig. S7). Then they separately looked at the "full global proxy network" with and without the Tiljander and other problematic proxies (Fig. S8). In no case did they look at the situation without both groups, much less without the equally problematic Tornetrask proxies.

Gavin, thanks for your reply. There is an oddity about the Mann2008 paper. It uses both the bristlecones and the upside down Tiljander proxies. As you point out, if you take out either one of these, you still get the same result.

But that is not a sufficient test. Neither set of proxies should be used. And if you take both of them out, you don't get the hockeystick.

See here for a discussion of the issues, and the comment here for the sources of the hockeystick. Out of all of the proxies, the majority do not show any hockeystick shape at all. It is entirely contained in the bristlecones, the Tiljander proxy, and (to a lesser extent) Briffa's "adjusted" Tornetrask proxies. All three of these are known to have problems. In addition, the Tornetrask proxies are unarchived. Problems with these proxies have been widely discussed in the literature, yet Mann2008 uses all of them.

You seem to be impressed that removing one of the three proxy groups known to have problems doesn't make much difference. All that shows is the weakness of the test. Take them all out, and it makes a huge difference.

Although these are a minority of the proxies, the others average out, and as a result, we get a hockeystick.

w.

PS - It is curious that you neglected to mention the bristlecones, after I specifically asked about them. Your comments on that, on the Tornetrask issue, and on the use of unarchived datasets would be welcome.

Gavin, many thanks for your comment. You say:

Most often the issue in question has been discussed dozens of times previously and is usually based either on an irrelevancy, or was acknowledged clearly in the original or subsequent paper or is based on some misperception of the science. [See Mann et al (2008) paper.]

Could you point out where the problems with the Mann et al (2008) paper are "based either on an irrelevancy, or was acknowledged clearly in the original or subsequent paper or is based on some misperception of the science." I don't find, for example, that the use in that paper of the bristlecones, or the use of upside-down Tiljander proxies, fit any of those criteria.

Thanks,

w.

First, my appreciation to both our host and the two participants for a most interesting discussion. I have been a commenter on the blogs of both participants, and found them both very open, inquisitive, and supportive.

I do have an issue with Bart's comment, viz:

But the thing is, right now, a lot of the discussion that is purportedly about climate science, is actually much more about the different ideas people have on how to respond to an issue like this: those who want to do something about the problem, and those who don’t want to do something. That’s what the disagreement is really about, I think.

In my experience, this is a totally false dichotomy between "do something" and "do nothing". There is a third way, which I call the "no-regrets option".

This is to take those actions which will be of value whether or not CO2 is an issue. All of the projected catastrophes of increased CO2 are with us today. We already have the floods, and the droughts, and the diseases, and the rising sea levels, and all of the myriad Biblical plagues foretold by AGW adherents. If an increase in those is the issue, and we are facing those same issues today, to me the reasonable path is to work on those problems today. If increased drought turns out to be a problem in 50 years, we should start working on solutions for drought today. That way, whether CO2 is an issue or not, we have a fifty year head start on the drought problem. I discuss these issues further at "Climate, Caution, and Precaution".

Keith, you have presented a very interesting interchange. My thanks to Lucia, Bart, and yourself.

w.