Climate Audit

At the close of Boulton’s April 9 interview with CRU, the only such interview relevant to the proxy reconstruction controversies that constitute 99% of the Climategate emails, Boulton asked CRU to comment on Ross McKitrick’s National Post op ed last October during Yamal. The response was given to Muir Russell on or after June 16 and the “report” doesn’t refer to it. But it contains some interesting answers pertaining to long-standing questions about the Polar Urals chronology that were not addressed in the “report”. (Whether the answers make sense is a different question.)

Remarkably, CRU’s explanation for never reporting the Polar Urals chronology of Briffa et al 1995 with the incorporation of additional measurements is that they never bothered calculating the impact of the additional measurements. And check out their explanation for why they didn’t do a regional Yamal-Polar Urals-Schweingruber chronology.

Briffa et al 1995 (Nature) had published an influential 1000-year chronology from this site used in Jones et al 1998 and several other IPCC reconstructions. Additional measurements for the site (more or less doubling the number of measurements in the sparsely covered early portion) were available by at least 2000. The figure below shows the core counts in Briffa et al 1995 and with the additional measurements. With the additional cores, there are at least seven cores in all years back to the 10th century, much improved over the sparse early coverage in Briffa et al 1995.

Figure 1. Polar Urals Core Couns – before and after additional measurements

In a mineral exploration business, if you have new drill holes, you have to report them. It’s an ethical requirement that is also a legal requirement. A mining promoter would have had to report the impact of the new measurements and intuitively it’s always seemed to me that so should CRU. Particularly when, as in this case, the additional results affect the previously reported chronology. In my submission, I compared the Briffa et al 1995 series to a chronology calculated by Esper including the additional measurements.

Although I did not know it for sure, it seemed inconceivable to me that reasonably diligent scientists would not have calculated the effect of the additional measurements. Even Mann had calculated the impact of excluding (censoring) the Graybill bristlecones and inadvertently left the results in his CENSORED directory ( he and Phil Jones later commiserating about the inadvertent leaving of files on his FTP site.)

Briffa and Osborn seem like thorough sorts, but here is what they told the Muir Russell panel – with some indignation at the implication:

We had never undertaken any reanalysis of the Polar Urals temperature reconstruction subsequent to its publication in 1995.

“Never undertaken any reanalysis”. Amazing.

Nobody from the panel bothered asking them about this admission, which hardly speaks well of thoroughness at CRU. Particularly when dendrochronologists regularly complain about the scarcity of long (1000-year) chronologies. Here was a substantial increase in relevant measurements for a long chronology that CRU scientists had worked on. You’d think that they’d at least be a little bit interested in the effect of the additional measurements. But apparently not.

In their comments on the minutes of the April 9 meeting, Briffa and Osborn re-iterated that they had never “examined” any chronologies incorporating additional Polar Urals measurements (which they called “McIntyre’s version”) and asserted that the idea that they had calculated and “hid” such a chronology was “ridiculous”:

CRU had not examined the “updated Polar Urals” series or assessed the quality of what was McIntyre’s version of these data until his submission and his implied allegation that we effectively hid it is ridiculous.

They dismiss the update of the more highly replicated version of their own chronology (Briffa et al 1995) as follows:

The chronology is poorly replicated in parts (more so than the Yamal series), and a calibration against local climate data has never been published as far as we know.

Funny – I thought that Briffa et al 1995 had reported the calibration of Polar Urals larch against climate.

I had previously raised the discrepancy between Polar Urals and Yamal in a number of IPCC Review Comments, which had been answered by Briffa. So the issue wasn’t entirely new. As an IPCC reviewer, I stated:

You need to state clearly that proxy series from nearby sites may give very different results e.g. Yamal and the Polar Urals update. [Stephen McIntyre (Reviewer’s comment ID #: 309-100)]

Rejected – this would imply a greater instability than current evidence supports.

Or again here:

There is an updated version of the Polar Urals series, usedin Esper et al 2002, which has elevated MWP values and which has better correlations to gridcell temperature than the Yamal series. since very different results are obtained from the Yamal and Polar Urals Updated, again the relationship of the Yamal series to local temperature is “ambiguous” [Stephen McIntyre (Reviewer’s comment ID #: 309-41)]

See response to comment 6-1143 and note that the Polar Urals and Yamal series do exhibit a significant relationship with local summer temperature.

or here

Problems can be observed elsewhere e.g. the Yamal series and the Polar Urals Update have very different properties with the Yamal series being a big contributor to HS-ness while the Polar URals series has a strong MWP. The Polar Urals Update correlates better to gridcell temperature than the Yamal series and one cannot help but suspect that the decision to use the Yamal series in all studies except Esper has been done with one eye on the MWP-modern relationship. [Stephen McIntyre (Reviewer’s comment ID #: 309-62)]

Accepted – text revised to stress overlap in early centuries of the last millennium. However, please note that the reviewer’s “suspicion” is unfounded

Although these comments deal very directly with differences between Polar Urals and Yamal and despite his obligations as an IPCC assessment author, Briffa apparently couldn’t be bothered even examining the additional measurement data for the Polar Urals site that he had reported on in Nature.

A Regional Yamal-Urals Composite
Their response to the National Post article also contains an interesting response to a then rhetorical question by Ross in his op ed.

If you think back to the Yamal controversy – and this issue got completely lost because of disinformation from CRU and realclimate – the form of my re-examination of Yamal was triggered by the new data from Taimyr, where the measurement data that had just become available for the first time. My analysis showed that the Taimyr regional composite of Briffa et al 2008 had incorporated a Schweingruber series about 600 km away from the Taimyr sites of Briffa 2000 – something that was not mentioned in the text or supplementary information to the original publication. (Nor was the connection to Schweingruber mentioned in the evidence to Muir Russell.)

The original replication at Taimyr had been much greater than at Yamal. So why didn’t Briffa add in nearby Schweingruber sites to improve Yamal replication. I discussed one such Schweingruber site in September 2009 – a site that was derided by realclimate but which actually appears to be co-located with some Yamal subfossil samples.) For that matter, why wouldn’t they have made a regional composite along the lines of the Taimyr-Avam composite and Tornetrask-Finland composites of Briffa et al 2008. Yamal and Polar Urals are much closer than the sites combined in the Briffa et al 2008 Taimyr-Avam combination. Ross posed the rhetorical question as follows:

Combining data from different samples would not have been an unusual step. Briffa added data from another Schweingruber site to a different composite, from the Taimyr Peninsula. The additional data were gathered more than 400 km away from the primary site. And in that case the primary site had three or four times as many cores to begin with as the Yamal site. Why did he not fill out the Yamal data with the readily-available data from his own coauthor? Why did Briffa seek out additional data for the already well-represented Taimyr site and not for the inadequate Yamal site?

In his initial response to Yamal in October 2009, Briffa said that they simply “did not consider” inclusion of additional Schweingruber Khadyta River data (though it met their criteria):

Our current practice when selecting data to incorporate in a regional chronology, is to include data exhibiting high levels of common high-frequency variability (i.e. on the basis of high inter-site correlations, where these are calculated using high-pass filtered data). Judged according to this criterion it is entirely appropriate to include the data from the KHAD site (used in McIntyre’s sensitivity test) when constructing a regional chronology for the area. However, we simply did not consider these data at the time, focussing only on the data used in the companion study by Hantemirov and Shiyatov and supplied to us by them.

Their answer to Muir Russell is a bit different.This time they say that they did consider doing a regional chronology for the Yamal area as part of Briffa et al 2008 – as they had done for the other two sites – but it could not be “completed in time”.

Some historical context for our 2008 paper might shed some light on this issue. Some time ago we began work on a multi-institution paper intended to describe the sensitivities in producing tree-ring-based climate reconstructions to the methods of chronology construction and subsequent climate calibration, illustrated using the examples of various tree-ring chronologies across northern Eurasia. When we later received a request to submit a paper to a planned themed issue of the Philosophical Transactions of the Royal Society about ‘The boreal forest and global change’, Briffa and colleagues decided to use some of the material to hand in preparing a draft. It was intended that this should describe 3 continuous 2000-year ring-width series, each originally planned to represent the integration of a large-regional data set of subfossil and living tree data. The focus was to be on representing large-regional growth signals and initial comparisons with equivalent regional temperature data. The western, ‘Fennoscandia’, series would incorporate near tree-line pine data from northern Sweden and Finland; the Avam-Taimyr series would integrate larch data from near the Taimyr peninsula tree-line region. Between these we had intended to explore an integrated Polar Urals/Yamal larch series but it was felt that this work could not be completed in time and Briffa made the decision to reprocess the Yamal ring-width data to hand, using improved standardization techniques, and include this series in the submitted paper.

Amazing. They’d been working on this sort of stuff for years. Yamal had been specifically placed at issue in AR4 Review Comments and yet they couldn’t complete “in time”. Given that they said that they hadn’t examined the additional Polar Urals measurements as at the time of my submission (in Feb 2010), it doesn’t seem like they did so in the following two years either. Any responsible “inquiry” should surely inquired as to how long it would have taken to do the corresponding calculations for a combined Polar Urals-Yamal network. With a program that had already done the calculations on the other two networks, why would it take all that long to do the same calculations for Yamal-Polar Urals?

In this connection, there’s an intriguing Tim Osborn email (684. 1146252894.txt) of Apr 28, 2006, which refers to the three regional groups of Briffa et al 2008 (Phil Trans B), including a regional group combining Yamal, Polar Urals and shorter sites as follows:

Date: Fri, 28 Apr 2006 15:08:05 +0100
To: philip.brohan
From: Tim Osborn
Subject: Re: Standardisation uncertainty for tree-ring series
Cc: Keith Briffa,simon.tett
Hi Philip,

we have three “groups” of trees:
“SCAND” (which includes the Tornetrask and Finland multi-millennial chronologies, but also some shorter chronologies from the same region). These trees fall mainly within the 3 boxes centred at:
17.5E, 67.5N
22.5E, 67.5N
27.5E, 67.5N

“URALS” (which includes the Yamal and Polar Urals long chronologies, plus other shorter ones). These fall mainly within these 3 boxes:
52.5E, 67.5N
62.5E, 62.5N (note this is the only one not at 67.5N)
67.5E, 67.5N

“TAIMYR” (which includes the Taimyr long chronology, plus other shorter ones). These fall mainly within these 4 boxes:
87.5E, 67.5N
102.5E, 67.5N
112.5E, 67.5N
122.5E, 67.5N

We do some analysis at the group scale, and for this we take the JJA temperatures from each box and average to the group scale to obtain a single series from each of SCAND, URALS and TAIMY.

We do some analysis at the overall scale, and for this we take these three group temperature series and average them to get an overall NW Eurasia temperature for boxes with tree chronologies in them.

We did also try using a wider average for the region, including all LAND
Temperatures from grid boxes within a rectangular region from 12.5E to 127.5E and from 57.5N to 72.5N, but I don’t think it correlated so well against the tree-ring width data (I can’t remember the exact correlations), so we didn’t pursue that.

Does that give you enough information to be going on with? I’d recommend using CRUTEM3 rather than HadCRUT3, because the correlations seem to deteriorate with the inclusion of SST data in some cases — though of course you can look into this yourself.
Cheers
Tim

I, for one, would like to know more about the regional URALS composite described in Osborn’s email of April 2006, what it looked like and why they didn’t use it. But wasn’t that what they “inquiry” was supposed to do?

In their evidence to Muir Russell, CRU spend a great deal of energy condemning the updated Polar Urals chronology relative to the Yamal chronology – something that I’ll discuss on another occasion. (There’s moving of the thimble here as their argument mostly compares Yamal replication against Polar Urals without the additional measurements. As illustrated in the graphic below, the original Yamal is definitely not uniformly more replicated than updated Polar Urals over the relevant record. Moreover, it is evident that the Polar Urals data is a substantial expansion of the data set in the earlier period. In our October 2009 discussions, I observed that there were statistical issues in the combining of sites to make regional chronologies – however, these issues already existed in the Yamal chronology where the homogeneity of the POR and YAD sites to the subfossil sites is something that, in my opinion, needs to be demonstrated rather than asserted.

The following image shows the total available cores from Polar Urals (original and additional in shades of purple), original Yamal (in cyan), Schweingruber Khadyta (beige) and the “new-found” Yamal data in other colors. (The POR and YAD data give Yamal its hockey stick, as opposed to JAH.)