Nathan Rive (archive here) is the newest entrant; he has cross-checked one of the graphics in Swindle against the original graphic in Lassen and Friis-Christensen and found a discrepancy between the graphic and the original article. BTW I obviously endorse such cross-checking. I focus on the studies cited by IPCC simply because I have limited time and resources, but it’s a good idea to cross-check Swindle or Inconvenient Truth (which I’ve spent relatively little time on).
Rive observes a divergence problem in the solar cycle correlations. I agree that divergence in these sorts of studies is a serious problem as do most CA readers. So I hope that Rive will spend some time with the more serious divergence problem in proxy studies.
Here is the comparison between two graphics showing the problem identified by Rive (endorsed by Friis-Christensen). As you see, the original graphic on the right has a discontinuity in the solar cycle series during the 17th century – the Maunder Minimum period, while the Swindle graphic doesn’t. Rive made some over-the-top allegations of “fabricated” data – something that I haven’t noticed him saying about Mann’s extension of the Gaspé series. In this case, I presume that someone along the way didn’t notice the discontinuity in the solar series and assumed that the two series overlaid one another (as you sometimes get with computer generated graphics).
Thus the graphic on the left. It’s an embarrassing error to be sure. However, given that there’s a plausible explanation for the error, I would counsel some caution by Rive and Friis-Christensen about alleging that the producers of Swindle “fabricated” data. That’s the type of over-reaching that could easily backfire.
On Durkin’s side, the error must be rather frustrating since the section with the solar cycle discontinuity corresponds to the Maunder Minimum and there were lots of ways that they could have talked through that. But this is climate science so errors and ironies abound.
I don’t know Rive but he put me on his circulation list for a mass email today publicizing his statement. After confirming to my own satisfaction that there was an error, I wrote Durkin urging them to set a good example to climate scientists by forthwith acknowledging any error and amending the release accordingly. Perhaps he could turn an embarrassing situation to his advantage by showing a willingness to admit and correct mistakes – a practice that would surely contrast favorably with the behavior of the Team.
BTW I’m not particularly impressed with this particular correlation anyway, now that I look at it. I don’t think that the Groveman and Landsberg 1979 reconstruction should have been used in 1995.
Update: May 1, 2007: In response to my email, Martin Durkin replied, acknowledging that an error had occurred and that a correction would be made as follows:
You can confirm, if anyone’s interested, that we’re changing the graphic for all future transmissions and the DVD version.
The Divergence Problem
The Rive statement wanders into silly speculations about the producers rummaging through Be-10 records. I cannot imagine that this happened. It then criticizes Swindle for inadequate disclosure of a Divergence Problem – in this case between solar cycles and temperature after 1975. They say:
The sunspot cycle length data in the L+FC graph shown on the screen stops in 1975. To demonstrate that both solar activity and temperature increased simultaneously from 1975 to 1985, Friis-Christensen and Lassen in their original 1991 paper included partially filtered and even unfiltered data for the last points in the graph. Updated calculations by Lassen and Friis-Christensen (2000) [JGR, 105(A12): 27493-27495] confirmed this trend between 1975 and 1985. However, they also explicitly concluded that after 1985 the temperature continued to rise while the sunspot cycle length flattened out, and thus no longer correlated with surface temperature. This point was not included in the narrator”¬’¢s statement.
It’s pretty interesting to see a climate scientist treat insufficient disclosure of a Divergence Problem as a serious failing. Personally I agree 100% that such problems need to be fully and completely disclosed, whether the author is Durkin or Mann. I’m not sufficiently familiar with the data here to opine on this particular divergence, but, on the basis that there is a divergence between solar cycles and temperature post-1975, I agree that this should be fully disclosed, just as Mann should have disclosed the verification r2 of 0.01.
This reminded me of the cynical IPCC TAR truncation of the Inconvenient Divergence in the Briffa et al 2001 reconstruction. By now, the Divergence between ring widths, densities and temperature in the last half of the 20th century is notorious. I’ve done what I can to publicize the problem. Briffa et al 1998 mentioned the problem in passing as follows:
However, although temperatures rose again after the mid-1960s and reached unprecedentedly high recorded levels by the late 1980s, hemispheric tree growth fell consistently after 1940, and in the late 1970s and 1980s reached levels as low as those attained in the cool 1880s. Over the hemisphere, the divergence between tree growth and mean summer temperatures began perhaps as early as the 1930s; became clearly recognisable, particularly in the north, after 1960; and has continued to increase up until the end of the common record at around 1990. The reason for this increasingly apparent and widespread phenomenon is not known
Here’s a nice graphic showing the Inconvenient Divergence (there’s much else that can be said about this figure, but for now let’s just observe the Inconvenient Divergence).
What happened to the Inconvenient Divergence in IPCC TAR? It was deleted. Now this was hard to see because of the coloring in the spaghetti graph. I discussed this previously here showing the following blowup of the IPCC graphic:
In AR4, there are no fewer than 3 reconstructions in the IPCC spaghetti graph with 1960 truncations: the original Briffa et al 2001 truncated series; Rutherford et al 2005 which used the Briffa network and also truncates in 1960; and Hegerl et al 2006, which uses a different network but also truncates in 1960, presumably also due to an Inconvenient Divergence. IPCC was unable to completely ignore the Divergence Problem and says:
Others, however, argue for a breakdown in the assumed linear tree growth response to continued warming, invoking a possible threshold exceedance beyond which moisture stress now limits further growth (D”¬’¢Arrigo et al., 2004). If true, this would imply a similar limit on the potential to reconstruct possible warm periods in earlier times at such sites. At this time there is no consensus on these issues (for further references see NRC, 2006) and the possibility of investigating them further is restricted by the lack of recent tree ring data at most of the sites from which tree ring data discussed in this chapter were acquired
and, on one side of its mouth, acknowledges a problem, but elsewhere disregards these caveats in its estimation of confidence intervals.
However, this is one side of their mouth and elsewhere they proceed to disregard this caveat by, for example, publishing confidence intervals that pay no attention to this pretty serious problem.
Also if you read closely, they show that they are aware of the post-1960 truncation (they should be; as a reviewer, I specifically asked that post-1960 values be shown on the spaghetti graphs), but they dance around admitting it as follows:
In their large-scale reconstructions based on tree ring density data, Briffa et al. (2001) specifically excluded the post-1960 data in their calibration against instrumental records, to avoid biasing the estimation of the earlier reconstructions (hence they are not shown in Figure 6.10), implicitly assuming that the ”¬Ëdivergence”¬’¢ was a uniquely recent phenomenon, as has also been argued by Cook et al. (2004a). p. 473
OK, let’s grudgingly grant Briffa the 1880-1960 period for calibration. That does not imply that IPCC TAR or AR4 are entitled to delete the Inconvenient Divergence. If Briffa’s calibration on 1880-1960 has yielded unbiased coefficients, then IPCC should show the post-1960 reconstructions – Inconvenient or not. In this case, their failure to do so is not inadvertent, but intentional as the matter was specifically and clearly brought to their attention with a reviewer request that post-1960 values be shown.