In today’s post, I’m going to discuss a previously undiscussed example of “Hide the Decline”, one that precedes Briffa and Osborn (Science 1999), the earliest example discussed so far. CRU did not report it in their submission to Muir Russell.
Jones et al 1999 (Rev Geophys) was published in May 1999, the same month as Briffa and Osborn (Science 1999). However, the latter was a comment with a shorter peer review process, placing Jones et al 1999 at least several months earlier in the queue. Its section 5.6 discussed proxy reconstructions, with its Figure 6 being a spaghetti graph, consisting of the same elements as the later IPCC TAR spaghetti graph: CRU plus the Mann, Jones and Briffa reconstructions. In this case, they used the Briffa et al 1998 (Nature) reconstruction – IPCC TAR used the then unpublished version from Briffa et al 2001. The Briffa series has been truncated in the mid-20th century.
Original Caption: Figure 6. Northern Hemisphere temperature reconstructions from paleoclimatic sources. The three series are Mann et al. [1998, 1999] (thick), Briffa et al.  (medium) and Jones et al.  (thin). All three annually resolved reconstructions have been smoothed with a 50-year Gaussian filter. The fourth (thickest) line is the short annual instrumental record also smoothed in a similar manner. All series are plotted as departures from the 1961–1990 average.
The following graph is an emulation of Jones et al 1999 Figure 6, applying the adjustments (re-scaling and re-centering) of Tim Osborn’s March 1999 program osborn-tree5/pl_mannbriffajones.prom showing the Briffa version in orange. The versions plotted here clearly match the ones in the Jones et al 1999 figure – except for the deletion of Briffa values from 1960 or so on.
CRU Submission to Muir Russell
In their submission to Muir Russell (page 38), CRU described the “hide the decline” methodology used in IPCC and most other publications (as opposed to the WMO splice) as follows:
(b) Depict the reconstruction over the period for which it can be considered to be a reliable estimate of past temperature variations (typically we have curtailed it in 1960, considering the post-1960 data to be a poor representation of temperature) and overlay separately the appropriate instrumental temperature observations.
They purported to justify the deletion of adverse data (“hide the decline”) as follows:
Presentation type (b) has typically been selected for publications whose purpose is to convey the available range of evidence for past temperature variations. In this situation, it is arguably reasonable to show only data that are considered to provide some information about temperature variations. In the same way that the early parts of a reconstruction might not be shown if they were considered to be unreliable (due, for example, to unacceptably low replication), the post-1960 values are also excluded. This avoids the presentation of values that are known to be unrepresentative of the real temperatures. Of course, the recent divergence in these data will be less clear if post-1960s values are excluded and that represents a potential disadvantage of this exclusion if this divergence is important for assessing confidence in the earlier reconstructed values. Section 1.2 outlines various situations where this is or is not a concern. If this is a concern, then to avoid the exclusion of the recent period resulting in an overly confident impression being given of the accuracy to which past temperatures can be reconstructed, we include appropriate caveats and references to the articles where the limitations are explored in greater detail. Examples of this type of presentation include Plate 3 of Briffa et al. (2001); Figure 2A of Jones et al. (2001); Figure 8 of Briffa et al. (2004); Figure 6.10b of Jansen et al. (2007); and Figure 5b of Hegerl et al. (2007).
The Muir Russell “inquiry”, with their typical negligence, failed to address this “explanation”, which, in Graham Stringer’s words, is “difficult to parody”. Obviously the divergence is “less clear” if the adverse data is deleted.
CRU stated that “If this is a concern, then to avoid the exclusion of the recent period resulting in an overly confident impression being given of the accuracy to which past temperatures can be reconstructed, we include appropriate caveats and references to the articles where the limitations are explored in greater detail.”
This assertion is a fabrication. In virtually all cases in “presentation (b)” style, they included no caveats whatever. To date, I have not identified a single publication in which they explicitly state that they have deleted post-1960 data and why. In some cases, the caption says that the Briffa data is from 1402-1960, but in such cases, there is no explicit statement that data was deleted and why. In other cases, there is not even a hint that the data has been chopped back to 1960.
Jones et al 1999
Jones et al 1999 is a pernicious example where there is not even a hint that data has been deleted. Worse, the running commentary all too often depends on the deletion of the adverse data.
Over the last few years, a number of compilations of proxy evidence have been assembled following the pioneering work of Bradley and Jones . In Figure 6 we show three recent reconstructions of Northern Hemisphere temperature for part of the last millennium. The reconstructions are all of different seasons (annual [Mann et al., 1998, 1999] and two definitions of summer [Briffa et al., 1998; Jones et al., 1998]). The short instrumental record on an annual basis is superimposed.
Agreement with the annual instrumental record is poorest during the nineteenth century, partly because of the different seasons (summer in two of the series) used. The instrumental record also rises considerably in the last 2 decades, and this cannot be seen in the multiproxy series because they end before the early 1980s, as some of the proxy records were collected during these years. The most striking feature of the multiproxy averages is the warming over the twentieth century, for both its magnitude and duration. The twentieth century is the warmest of the millennium and the warming during it is unprecedented (see also discussion by Mann et al. [1998, 1999] and Jones et al. ). The four recent years 1990, 1995, 1997, and 1998, the warmest in the instrumental
series, are the warmest since 1400 and probably since 1000. The end of the recent El Nin˜o event (such events tend to warmer temperatures globally) and the greater likelihood of La Nin˜a (which tends to lead to cooler temperatures) as opposed to El Nin˜o conditions during 1999 and 2000 means that 1998 will likely be the warmest year of the millennium. The coolest year of the last 1000 years, based on these proxy records, was 1601.
Neither the caption nor the running text gives any hint that the reconstruction of Briffa et al 1998 (Nature) has been deleted after 1960. Yes, the article cites Briffa et al 1998 – which reports the decline – but Jones et al 1999 contains no “caveats” about the decline or about the deletion of data.
Important comments in the running text on Figure 6 also depend on the deletion of the decline. For example, Jones et al state: “Agreement with the annual instrumental record is poorest during the nineteenth century.” Actually, the worst agreement is between the Briffa decline and the late 20th century temperature increase – but that’s been hidden.
Jones et al say: “The instrumental record also rises considerably in the last 2 decades, and this cannot be seen in the multiproxy series because they end before the early 1980s, as some of the proxy records were collected during these years.” Again, this is completely misleading. The Briffa multiproxy series do not end before the early 1980s because of they were collected during those years, but because the post-1960 was deleted.
Jones et al say: “The most striking feature of the multiproxy averages is the warming over the twentieth century, for both its magnitude and duration.” Again, this statement depends entirely on the deletion of the decline. If the Briffa series were shown, the inconsistency of the Briffa reconstruction with the 20th century temperature history becomes the most striking feature of the graphic – calling the validity of the reconstructions into play.
In my opinion, locating the earliest known example of “hide the decline’ in Jones et al 1999 (Rev Geophys) places hide the decline in a remarkable new light. I think that it’s fair to say that most of us have assumed that “hide the decline” originated with Mann or Briffa. However, it seems to me that this new evidence suggests that the lead author of Jones et al 1999, Phil Jones himself, may have been responsible for CRU’s decision to hide the decline in the spaghetti graph comparisons – initially Jones et al 1999 Figure 6, later, as we all know, IPCC TAR Fig 2.20.
To my knowledge, “hide the decline”, as a technique, was used in every subsequent spaghetti graph using the Briffa reconstruction except one – the Zero Order Draft of IPCC TAR presented to the Lead Authors in Arusha in September 1999. This spaghetti graph – which didn’t hide the decline – went over like a lead balloon with IPCC – thus the busy emails of September 1999, which I’ve previously discussed. The IPCC spaghetti graph in its First Order Draft (October 1999) adopted a hide-the-decline strategy modeled on the techniques “pioneered” in Jones et al 1999 and Briffa and Osborn 1999 (though Mann varied the method somewhat.)
In preparing my emulation of Jones et al 1999, I used the adjustment parameters of Osborn’s program pl_mannbriffajones of March 1999, almost contemporary with the preparation of Jones et al 1999. When one does a cross-chronology of the Climategate programs, there is some further interesting information on the genesis of “hide the decline” that I’ll discuss in another post.
Update Mar 23:
Here is an updated emulation of the Jones et al 1999 spaghetti graph, showing that the hide-the-decline is effective in 1950 (rather than 1960) and that the source information is the 4th column (NHD2) in the Briffa et al 1998 archive.
Figure 2. Excerpt from Jones et al 1999 with annotation.