Was There a "Mike's Nature Trick" to "Hide the Decline?" Part 1 - Misreading CRU Emails

In November 2009, someone likely stole[1] and publicly released over 1,000 emails from Climate Research Unit (CRU) at the University of East Anglia. Following this, people combed through all these emails  looking for evidence of frauds and coverups on the part of climate scientists. They found very little damaging information, but not everything they found was was good. There were, as best I can tell, four categories of content that were discovered:

1. The normal kinds of communication you'd expect to find in correspondence between scientists
2. Words and phrases that were twisted and contorted to support an imagined conspiracy that had been imposed on the emails.
3. Evidence of mean-spirited attacks on contrarian scientists and heavy-handed tactics intended to punish journals that published contrarian studies. 
4. A single email describing the process of creating a graph for a WMO statement that used the phrase "hide the decline." 

Here I plan to tackle the "hide the decline" email, since so many contrarians think there's evidence here of some sort of fraud and coverup. Most of these accusations depend on misquoting and then misunderstanding the email. It's actually on a life of its own, independent from the actual text and context of the email. What I would like to do is examine the email in the context of the graph that it references. We'll examine whether the email describes anything unethical that was done, and if so, whether this affects the integrity of the graph in any way.

The Email

The email was written on Tue, 16 Nov 1999 in a correspondence between Phil Jones of CRU and the authors of two proxy reconstruction papers. The primary recipients were Michael Mann, Ray Bradley and Malcolm Hughes with a CC to Keith Briffa and Timothy Osborn. In this email, Jones writes,

I’ve just completed Mike’s Nature trick of adding in the real temps to each series for the last 20 years (ie from 1981 onwards) and from 1961 for Keith’s to hide the decline. Mike’s series got the annual land and marine values while the other two got April–Sept for NH land N of 20N. The latter two are real for 1999, while the estimate for 1999 for NH combined is +0.44C wrt 61–90. The Global estimate for 1999 with data through Oct is +0.35C cf. 0.57 for 1998.

Jones is describing the process he used to construct a graph that was placed on the front cover of a WMO statement[2] on global climate during 1999. It's this graph here.

The description of this cover art was written as, "Front cover: Northern Hemisphere temperatures were reconstructed for the past 1000 years (up to 1999) using palaeoclimatic records (tree rings, corals, ice cores, lake sediments, etc.), along with historical and long instrumental records." There is also an explanation for this graph on page 4. Here's the relevant portion of the explanation. 

"Our knowledge of pre-20th century temperature variations in the Northern Hemisphere has increased dramatically in recent years. The availability of natural archives of past temperature such as tree rings, banded corals, ice cores and lake sediments, in addition to historical and long instrumental records, has enabled NH temperature variations to be reconstructed for the last 1000 years at an annual resolution. It is not yet possible to do the same for the Southern Hemisphere due to the lack of adequate palaeoclimatic records. Despite their different emphases on annual or extended summer seasonal temperatures and their different geographical biases, all the reconstructions (shown on the front cover as 50-year smoothed differences from the 1961–1990 normal) indicate that against the background of the millennium as a whole, the 20th century was unusually warm. Uncertainties increase earlier in the millennium due to the sparser and imperfect nature of proxy data (95% confidence errors of ± 0.3°C for 1000–1500, reducing to ± 0.1°C by the early 19th century on this 50-year time scale)."

Jones appears to have adapted the WMO graph from Briffa 1999[3]. Briffa's graph shows the Apr-Sep land instrumental record from 1871 to 1997, NH tree-ring density proxies for temperature from 1550 to 1960 from Briffa's previous work[4][5][6] and proxies from 1000 to 1980 from Mann's studies[7][8].

Jones' adaptation of this graph used his own reconstruction[9] through 1980 with Mann's through 1980 and Briffa's through 1960. He then added the appropriate instrumental record to each of those time series. Because each time series shows different things, his email describes precisely what he did to each time series.

1. To Jones' time series, he added land Apr to Sep NH temperatures following 1980.
2. To Mann's time series, he added land and ocean annual NH temperatures following 1980.
3. To Briffa's time series, he added land Apr to Sep NH temperatures following 1960.

So the WMO statement graph is very similar to that produced by Briffa in 1999, with the big difference here is that the appropriate instrumental record is seamlessly added to proxy record so you can't tell by looking at the graph where the proxy evidence ends and where the instrumental record begins. We know from the image description that there was a transition to the instrumental record, but we don't know when that was. The transition from proxy to instrumental is seamless, with no indication in the graph or its explanation when the change in data type took place in each time series.

Jones described that his reasoning for choosing 1961 for Briffa's time series was to "hide the decline." He's referring to the decline in tree ring density following 1960 that affected samples from boreal forests. This was to address what was (and is) a well-known problem called the "divergence problem," and there were multiple papers describing this problem in the peer-reviewed literature, including papers by Briffa. I'll probably share more about this in a future post, but for now, this is a problem affecting select trees, mostly from boreal forests, where tree ring densities fail to match instrumental temperatures following 1960 or so. The evidence we have shows that it's a recent problem, possibly caused by selection bias introduced by sampling forests with a higher proportion of younger trees. 

"Mike's Nature Trick"

In public discourse, there has been a great deal of discussion of this email that is based on a misquote of its contents. Some critics have shortened the quote to "Let's use Mike's trick to hide the decline." But that conflates "Mike's Nature trick" with "hide the decline." It also removes the word "Nature," which was the name of the journal Mann published his hockey stick reconstruction in 1998. As we'll see, it's important to keep these phrases separate. Let's deal first with "Mike's Nature trick."

Michael Mann published two papers with Bradley and Hughes (the other two recipients of the email) in 1998 and 1999. We'll refer to these as MBH98[8] and MBH99.[9] MBH98 was published in Nature, so the "trick" must have come from the 1998 paper. Here's the graph from that paper.

The graph quite clearly shows a global reconstruction back to a.d. 1400 with reconstructed temperatures and instrumental temperatures ("actual data") delineated separately, but both included together on the same graph for easy comparison. That was "Mike's Nature trick." It was simply "adding in the real temps" to the proxy data, and in MBH98 it's clearly stated that's what he was doing. The same is true for MBH99, which reconstructed Northern Hemisphere temperatures from a.d. 1000. Here's the graph from MBH99, published in GRL.

This is the data that was used for Mann's time series in the WMO statement and the IPCC Third Assessment Report (TAR). Notice that the reconstruction is clearly stated to cover a.d. 1000 to 1980 while the instrumental record from 1902 to 1998. And in case there is any doubt, here's the graph from TAR.

Each of the above reconstructions can be said to have used "Mike's Nature trick," but none of them could reasonably be accused of doing anything unethical for showing both the proxy reconstruction and the instrumental data on the same graph. There is absolutely nothing wrong with this "trick" as it was used by MBH98, MBH99 and the IPCC's TAR. Whatever we end up thinking about what Jones did to "hide the decline," that problem does not extend to the hockey stick in general but to its presentation in the WMO statement.

"Hide the Decline"

As I've already hinted at above, Jones did not use "Mike's Nature trick" to "hide the decline." Mike's Nature trick was simply to add both the proxy and instrumental data together on the same graph. What Jones did to "hide the decline" was to choose 1961 for Briffa's time series instead of 1981. He writes, "I’ve just completed Mike’s Nature trick of adding in the real temps to each series for the last 20 years (ie from 1981 onwards) and from 1961 for Keith’s to hide the decline." The "decline" that Jones wanted to hide was the decline in tree ring density following 1960, and that was a problem with Briffa's time series, not the other two. The proxy reconstructions of the other two time series were not limited to tree ring data, so they continued to 1980. What Jones was "hiding" was the unreliable data in Briffa's tree ring-only proxy reconstruction. In one popular criticism of the WMO graph, the proxy data from Briffa following 1961 is shown without the instrumental record.

As you can see, while Mann's and Jones' series track with the WMO graph through 1980, Briffa's series plummets after 1960 because of the quality of the tree ring data in Briffa's paper. Some have suggested that this means that Jones (and the others included in the email discussion) were involved in "hiding" this discrepancy between the instrumental record and the proxy data. In the graph above, it's even said that this is what the reconstruction would look like if "Mike's trick is not used" - a claim that is based on an explicitly dishonest reading of the email. However, the fact that unreliable proxy data wasn't included in the published WMO graph doesn't mean that anything was hidden. The graph claims and is intended to show NH temperatures from a.d. 1000 to 1999; it's not specifically a graph of proxy data. It turns out that 1) each of these time series was already published in the peer reviewed literature with the instrumental data distinct from the proxy reconstruction and 2) the divergence between tree ring proxies and instrumental data was already in the peer-reviewed literature when the WMO statement was published. 

The following comes from Briffa 2000,[10] You can clearly see the divergence between the instrumental record (thick black line) and tree ring proxies following 1960.

Scientists would also have been aware of the divergence problem in other treatments of the subject going back at least to 1995. In this paper from 1998,[6] Briffa also goes into a considerable amount of detail describing the problem. This graph shows the discrepancy between the Apr - Sep mean temperatures in the instrumental record with proxy data.

If Jones was intending to hide the divergence between tree ring proxies and the instrumental record, he did a pretty poor job, because that was already known and discussed in the scientific literature for about 5 years. Jones deleting Briffa's tree ring proxy data following 1960 could not have hidden the divergence problem. There was nothing hidden from the public and certainly not from scientists familiar with the divergence problem; it was already published in the scientific literature. 

Was Jones Ethical in the WMO Graph?

This particular email has taken on a life of its own with unending charges of fraud, data manipulation, and hiding data that contradicts global warming. I consider these kind of ridiculous. But the fact remains that Jones did say he wanted to "hide" something, and hiding is generally not consistent with transparency. I think we have evidence here of a lack of transparency, at least on the part of Jones. All Jones had to do is to indicate on the graph and/or explanation at what points of time the proxy and instrumental data were merged - 1980 for Jones and Mann and 1960 for Briffa. He could easily have changed the time series to a dotted line at 1960 or 1980 and indicated the transition to the instrumental record, and if he had, none of the critics would have anything legitimate to complain about. I think we all believe it's important for scientists both to do competent work and be truthful about it, but I think equally important is transparency. It's transparency that builds trust, and it's the lack of transparency that creates suspicion - that suspicion could have been avoided if Jones had disclosed exactly what he had done.

That said, the graph itself is essentially correct. It presents three time series that show the best evidence at the time for NH temperatures from a.d. 1000 to 1999. As I see it the authors did good, competent work. Even showing three time series instead of just one give readers a sense of the uncertainties in the various proxy reconstructions. If the graph simply noted where the instrumental record begins in each time series, I would have nothing to complain about.

So my conclusion is this. Jones erred in not disclosing when the transitions to the instrumental record began. The failure was a failure of transparency - he didn't fully disclose what his readers should have been told. However, the graph itself is a truthful presentation of NH temperatures. Someone reading that graph and taking it at face value would have an accurate understanding of the available evidence of NH temperature variability over the previous millennium. Now 22 years later, we have far more extensive proxies and a better understanding of global temperatures, but this graph is valid for the evidence its authors had available to them at the time.

References:

[1] There’s some dispute over whether the stolen emails were due to a hack or a leak. While I don’t know of any conclusive evidence, the servers for the RealClimate blog were also hacked, and the emails were uploaded to their website. This would suggest that it wasn’t a leak from someone at CRU, but rather a hacking job.

[2] WMO, WMO Statement on the Status of the Global Climate in 1999. WMO-No. 913. https://library.wmo.int/doc_num.php?explnum_id=3460

[3] Briffa, K. R. (1999). CLIMATE WARMING: Seeing the Wood from the Trees. Science, 284(5416), 926–927. doi:10.1126/science.284.5416.926
10.1126/science.284.5416.926

[4] Briffa, K., Schweingruber, F., Jones, P. et al. Reduced sensitivity of recent tree-growth to temperature at high northern latitudes. Nature 391, 678–682 (1998). https://doi.org/10.1038/35596

[5] Briffa, K. R., Jones, P. D., Schweingruber, F. H., & Osborn, T. J. (1998). Influence of volcanic eruptions on Northern Hemisphere summer temperature over the past 600 years. Nature, 393(6684), 450-455. https://doi.org/10.1038/30943

[6] Briffa, K. R., Schweingruber, F. H., Jones, P. D., Osborn, T. J., Harris, I. C., Shiyatov, S. G., … Grudd, H. (1998). Trees tell of past climates: but are they speaking less clearly today? Philosophical Transactions of the Royal Society B: Biological Sciences, 353(1365), 65–73. doi:10.1098/rstb.1998.0191

[7] Mann, M., Bradley, R. & Hughes, M. Global-scale temperature patterns and climate forcing over the past six centuries. Nature 392, 779–787 (1998). https://doi.org/10.1038/33859 and

[8] Mann, M. E., Bradley, R. S., & Hughes, M. K. (1999). Northern hemisphere temperatures during the past millennium: Inferences, uncertainties, and limitations. Geophysical Research Letters, 26(6), 759–762. doi:10.1029/1999gl900070

[9] Jones, P. D., Briffa, K. R., Barnett, T. P., & Tett, S. F. B. (1998). High-resolution palaeoclimatic records for the last millennium: interpretation, integration and comparison with General Circulation Model control-run temperatures. The Holocene, 8(4), 455–471. doi:10.1191/095968398667194956

[10] Keith R Briffa, Annual climate variability in the Holocene: interpreting the message of ancient trees, Quaternary Science Reviews, Volume 19, Issues 1–5, 2000, pp. 87-105,