How do they get away with it? That’s the inevitable question as I examine one of the most specious – and despicable – “climategate” allegations from contrarians, namely the oft-repeated claim that climate scientists at East Anglia University’s Climate Research Unit inappropriately adjusted certain tree-ring temperature proxy data to provide a better match with instrumental temperature record.
In an extraordinary interview with FoxNews in December of last year, economist Ross McKitrick alleged that scientists were “faking the match” between proxy and temperature data. McKitrick averred this was “not being honest with the reader” and even invoked a comparison to falsified experimental medical research. Stephen McIntyre of ClimateAudit recently claimed that the “artificial correction” seemed “to have entered the CRU toolkit” (in his submission to the U.K. parliamentary committee examining allegations arising from stolen CRU emails, no less).
Yet a close examination of the computer code and ensuing research bears out the key contention of the firm rebuttals from East Anglia University, as well as CRU scientist Tim Osborn, namely that the specific adjustments in question were for private exploration only and never incorporated into CRU scientists’ subsequent published research temperature reconstructions. So, once again, McIntyre and McKitrick have made odious and unfounded accusations, based on non-existent evidence. They should withdraw these specious allegations immediately.
Most readers will recall that among the “climategate” documents stolen last November from East Anglia’s Climate Research Unit, were a number of research computer programs, with code and comments that appeared to show the use of a “fudge factor” to “artificially correct” tree ring-data after 1960. A number of bloggers, led by Anthony Watts, fell over themselves to bring forth this supposed “smoking gun” proof of fraudulence perpetrated by scientists.
As the furor over the stolen emails and documents mounted, Fox News prepared a one-hour documentary, provocatively entitled Global Warming … or a Lot of HotAir? The program was based largely on interviews with McKitrick and McIntyre (as well as Patrick Michaels), and aired a few days before Christmas. The program was subsequently archived at WattsUpWithThat in six convenient segments (unfortunately, the links are no longer operative due to flagrant copyright violation).
The remote split-screen interview with McKitrick is in the third segment and appears to have been conducted with McKitrick in his book-lined office at the University of Buckingham (where he was visiting professor until January). At around 3:30 of that segment, the documentary discusses the “fudge factor”.
FOXNews narrator: Then there is the suggestion that climate scientists were improperly massaging their data. It ‘s found within the computer codes that they used in their temperature research. One program contained information referred to as a fudge factor.
FOXNews interviewer: Can you explain what the fudge factor is?
Ross McKitrick: The fudge factor is actually a line written into a computer program at the Climate Resaerch Unit, where they’ve run into a problem with their data. They want to fit some tree-ring proxy data to their temperature data, but the tree-ring proxy data goes in one direction; it goes down, and the temperature data is going up.
Narrator: In other words, at times the tree-ring data and the temperature didn’t match. The thermometers said warmer, and the tree-ring said cooler.
Interviewer: It almost sounds like it’s faking it.
McKitrick: It’s faking the match. The problem is, if they haven’t explained that they use this fudge factor to force it to match, then it’s not being honest with the reader.
Narrator: Another programmer note reads, “These will be artificially adjusted to look closer to the real temperatures”.
Interviewer: Do all scientists do this?
McKitrick: I sure hope not. I mean where would we be in medical science, for instance, if they took all the people that died from the experimental treatment and just recoded so it looks like they got better.
A few minutes later, a denial is desultorily displayed:
The University of East Anglia added that the “fudge factor” computer codes were never used for published articles or data and defended the Climatic Research Unit’s work, as its director, Phil Jones, has temporarily stepped aside for a university investigation, and the world has learned that the Unit had discarded much of its temperature data.
Aside from the obvious implication that the University denial is untrustworthy, since the head of CRU is “under investigation”, note that Fox introduces the canard that CRU had discarded “much of its temperature data”, data which was not actually CRU’s, and not relevant to the research discussed in any event.
Naturally, the debate about the “fudge factor” has become a key part of the U.K. parliamentary Science and Technology Committee investigation of the stolen CRU documents. In a submission to that committee, CRU paleoclimatologist Tim Osborn gave a detailed explanation of the matter.
8. CRU’s Computer Code is Fit For Purpose and Does Not Reveal Secret Manipulation of Data
A small sample of my computer programming code was included in the disclosed files. It has been argued that comments within the code such as “Fudge factor” and “shouldn’t usually plot past 1960 because these will be artificially adjusted to look closer to the real temperatures” demonstrate that data have been manipulated in an inappropriate and undisclosed manner. My programs that were highlighted on BBC Newsnight that contained comments such as these were not the basis for any published article or dataset, and thus are not a valid indication of inappropriate data manipulation. If we do need to make adjustments to data that are scientifically justified, then we state clearly both the justification and the adjustment when we publish the article or dataset. In relation to the second of the two highlighted comments, it was simply a note that should have read “…because these will have been artificially adjusted…” to remind myself that I had applied an adjustment to this particular set of data (for the purposes of exploring the consequences of recent trends for the calibration of tree-ring temperature proxies) and that if I did plot them it would give a false impression of the agreement between tree-rings and temperature because of the adjustment. Thus, rather than indicating that an undisclosed adjustment would be made, it was a warning to avoid using adjusted data without realising it. To re-iterate: I have made no adjustments to data except those that are scientifically justified and stated in published papers. [Emphasis added].
Yet despite these convincing and forthright explanations, McIntyre forged ahead with his own reckless accusation in his submission to the parliamentary committee.
7. One of the underlying problems in trying to use tree ring width/density chronologies for temperature reconstructions is a decline in 20th century values at many sites – Briffa’s 1992 density (MXD) chronology for the influential Tornetrask site is shown at left below. The MXD chronology had a very high correlation to temperature, but went down in the 20th century relative to what it was “expected” to do and relative to the ring width (RW) chronology (which had a lower correlation to temperature.) So Briffa “adjusted” the MXD chronology, by a linear increase to the latter values (middle), thereby reducing the medieval-modern differential. This adjustment was described in private as the “Briffa bodge” …
“Bodging” then seems to entered [sic] into the CRU toolkit to get reconstructions to “look” right, as evidenced by the Climategate documents containing annotations that the method contains “fudge factors” or “very artificial corrections for decline” (e.g. …
;****** APPLIES A VERY ARTIFICIAL CORRECTION FOR DECLINE*********
Although neither McKitrick and McIntyre gave any specific instances where this undisclosed “method” to get “reconstructions to look right” had actually been used in subsequently published research, that inference was very clear. And McIntyre even implies that this was a regular occurrence, as the manipulation was supposedly part of the “CRU toolkit”.
So there is a clear conflict between the fanciful accusations from McIntyre and McKitrick on the one hand, and the actual CRU explanations on the other. However, for the benefit of anyone still in doubt that Osborn is more credible than McKitrick or McIntyre, I’ll now look a little more deeply at some of the actual code.
The computer program in question is the very one from which McIntyre quoted (briffa_sep98_e.pro, apparently written by Ian Harris of CRU). John Graham-Cumming discussed the “artificial correction” as implemented in that particular program back in November. He explains how the correction table is used to create a linear interpolation and adjust MXD (maximum latewood density) according to year.
Up to 1924, proxy values are unaffected, while post-1960 values are boosted considerably.
Clearly, as Graham-Cumming surmised, the point was to study an aspect of the well-known “divergence problem“, whereby the MXD proxy at high-latitude locations show marked divergence from local temperatures in the late twentieth century.
But was this program actually used in published research? Or rather was it used merely for the exploration of calibration issues, as Osborn stated?
Looking at the code, one can see that two methods for standardizing tree-ring series were compared (standardization is a necessary step in combing individual tree-ring samples into a site or regional series).
printf,1,'Correlations between timeseries' printf,1,'Age-banded vs. Hugershoff-standardised' printf,1,' Region Full <10 >10 >30 >50 >100'
As we have seen before, multi-centennial reconstructions using tree-ring proxies need to account for age-related growth characteristics (usually younger trees grow faster than older ones), but in a way that preserves “low frequency” information as much as possible. (Hugershoff standardization on the other hand, preserves detailed interannual variation, suitable for, say, studying the relationship between temperature and volcanic eruptions).
Keith Briffa’s “regional curve standardization” (RCS) is the best known technique for accomplishing this low-frequency retention, but requires large amounts of tree-ring data from a single species in a realatively prescribed region.
So Briffa and Osborn also proposed an alternative “age-band decomposition” (ABD) as a compromise technique for situations where RCS would lead to unacceptable uncertainty levels. The technique was first described in a 2001 JGR paper, “Low-frequency temperature variations from a northern tree ring density network”, co-authored with Osborne, Harris, Phil Jones and three other researchers.
In a useful summary of this and other papers, available at the CRU site, Osborn states:
The MXD measurement series from individual tree cores were processed using the Age-Band Decomposition method introduced and described in this paper. The rationale for this new method is that if data are initially analysed and combined within relatively narrow ranges of tree age, then it is no longer necessary to detrend individual tree-core time series to remove any dependence on tree age. This detrending, which is part of traditional standardisation techniques, results in the loss of variance on long time scales that might be caused by changing climate. Because the Age-Band Decomposition method does not require this detrending step, the reconstructions obtained with it have stronger variance on longer time scales. The method may introduce different types of uncertainty compared with traditional standardisation methods, especially when the sample of tree-ring measurements is small, so we consider that the regional and hemispheric reconstructions produced using this method are more appropriate for the study of climate variations on multi-decadal time scales and longer.
It seems that the Briifa_sep98_e program is early exploratory work using ABD with these same chronologies. The program lists the chronology data sets to be opened and plotted:
The 2001 paper gives nine MXD regional northern hemisphere chronologies. By the time of the paper, the acronyms identifying the chronologies have changed, although the regions they represent are recognizably the same ones. Here’s the caption from the polar map of the regions studied:
Figure 1. Locations (circles) of the 387 tree ring density chronologiesto, getherw ith the boundaries of the nine arbitrary regions: NEUR, northernE urope; SEUR, southernE urope; NSIB, northern Siberia; ESIB, eastern Siberia; CAS, central Asia; TIBP, Tibetan Plateau; WNA, western North America; NWNA, northwestern North America; ECCA, eastern and central Canada.
But the actual paper, far from using manipulation (undisclosed or otherwise) to force a match between post-1960 MXD data and temperature, examines the effect of different combining techniques on the full time range of the series, up to 1994, clearly showing overall “divergence” between the instrumental temperature record and proxies in the late 20th century.
And the paper also states the proxy data for many chronolgies after 1960 was unreliable and should not be used for calibration; this period was also excluded for the various cross-comparisons between regional chronologies and with other reconstructions.
The period after 1960 was not used to avoid bias in the regression coefficients that could be generated by an anomalous decline in tree density measurements over recent decades that is not forced by temperature[B riffa et al., 1998b].
So at no point, was the “manipulation” in the above exploratory computer program used in this particular published research. And, since this paper appears to be the only one that discusses these nine regional chronologies and ABD in any detail, the inevitable conclusion is that there exists no evidence of any published research that may have used the output of the Briffa_Sep98_E program specifically referred to by McIntyre. None whatsoever.
If either McIntyre or McKitrick had any evidence at all to back up their accusations, surely they would have released it by now.
Now there is only one course of action open to them. McIntyre and McKitrick must withdraw, immediately and unequivocally, their false accusations. And they must also publicly apologize to Keith Briffa, Tim Osborn, Ian Harris and Phil Jones, as well as to East Anglia University and its unjustly maligned Climate Research Unit.