Saturday, 23 April 2011

Jaeah’s Investigation


The struggles of Mother Jones fact checker, Jaeah Lee, (see video here) to locate the “decline” in the IPCC assessment report is an object lesson in the effectiveness of the IPCC’s strategy to hide the decline. That she ultimately succumbs to bafflegab from Keith Briffa, declaring “case closed”, is not the main lesson of her video. The main lesson is the success of IPCC’s efforts to hide the decline, even from a fact checker looking for the decline.

The efforts of a young and inexperienced fact checker contrast with the negligent failure of the “seven inquiries” to do likewise. Consider for a moment that none of these reports contained an analysis of “hide the decline” that answered the reasonable questions of the most junior fact checker. None of them even contained a graphic showing hide the decline.

While Jaeah made a few mis-steps along the way and unwisely relied on Mann and Briffa as sources, her struggles to figure out the meaning of Hide the Decline are an object lesson that should embarrass the climate science community.

Jaeah begins her story about trying to find the decline:

We thought it would be easy. Boy, were we wrong.

She then gives a short description of proxies and says:

We put on our magnifying glasses and set out in search of the data….

Like me some years ago (or anyone else), she began her quest with the 2001 IPCC Report:

Figure 1. IPCC Figure shown in MJ presentation.

She couldn’t find the decline in this figure and asked “where was the so-called decline”?

We started out with the 2001 IPCC report, the leading authority… This graph illustrated historical temperatures…… All of these lines told more or less the same story. That starting in the 20th century, the earth warmed at an unprecedented rate. But where was the so-called decline?

Six years ago, I wondered the same thing – see CA post here, leading me to do a blowup of the graphic, only then discovering that there had been an unreported deletion of post-1960 in the IPCC graphic.

But back to Jaeah’s story. She asked scientists for data and “the more we dug for the answers , the more we grew confused.” She reports:

Several scientists sent us data sets. One started to showed a decline, but the data ended in 1960. We emailed Mann, the originator of the graph. … Mann said that we were looking at the wrong tree ring sets and pointed us to an article published in 2002 in Science. When we plotted this data set, we saw a clear decline around 1960 but still had no idea what happened after 1960.

Here Jaeah is experiencing one part of the frustration in trying to deal with the Team, but she didn’t get the Full Monty treatment. The Team sent her data without her having to resort to a Freedom of Information request. They also tried to answer her questions. However, even with full Team cooperation, she couldn’t get data showing the decline.

The next figure is a screenshot of her plot of the data that she was sent. It can be conclusively identified as the (truncated) Briffa et al 2001 reconstruction as archived (see here) up to and including the deletion of data after 1960. In their graphic, they’ve added 14 deg C to the archived version (which is in anomaly deg C). The graphic below plots the archived version plus 13 deg C (the overlay at 14 deg C is exact.) For her smooth, Jaeah used a 40-year rolling average (rather than the gaussian smooth that Briffa normally used). My quick experiments indicate that this was done using forward values (rather than centered), without padding, but all available values – thus the smooth extends into the 1990s, although the underlying Briffa version ended in 1960.

Figure 1. Motherjones starting point. Red- as plotted by MJ. Black- Briffa et al 2001 version plus 13 deg C.

As she noted, Jaeah could not locate the decline in this data. This is not surprising. The decline had been deleted from the data. (Looking forward a little here, the Climategate letters were the first time that data after 1960 became available for the Briffa et al 2001 reconstruction – as reported shortly after Climategate at CA here.)

Frustrated, Jaeah turned to Mann for clarificationm who, as noted above, told her that they were looking at the “wrong tree ring sets”, directing them to Briffa and Osborn (Science 2002). Jaeah extracted this data set and showed it in her video as shown below, observing:

When we plotted this data set, we saw a clear decline around 1960 but still had no idea what happened after 1960.

Figure . Screenshot of Mother Jones version of Briffa series in Briffa and Osborn 2002.

here Jaeah is running into the Gang Who Can’t Shoot Straight. The two versions are actually identical. The Briffa and Osborn (Science 2002) version is a gaussian smooth of the Briffa et al 2001 data. If the first version was “wrong”, then so was the second “version”. And the reason why the Briffa and Osborn 2002 version didn’t show the decline was that the decline had been deleted from it as well.

Jaeah, like many of us, was stumped. Then, in language that precisely tracks the very first mention of Climategate, Jaeah says: “a miracle happened.” The miracle was that they “received a letter from Briffa”. An explanation from the Team – a miracle indeed. Cynics might observe that this sort of miracle only seems to happen to the faithful.

Jaeah recounts her brush with the holy as follows:

He explained that the particular sample did show a decline after 1960 possibly due to increased exposure to carbon dioxide. But that sample data was computed in a way that wasn’t scientifically useful. Meaning basically that all the fuss over the decline came from one obscure data set showing tree ring densities in high latitude regions.

At this point, Jaeah, apparently overcome by the miracle, seems to have swooned and ceased checking facts. To my knowledge, no one has ever suggested that additional carbon dioxide caused a decline after 1960. Nor is there any scientific reason why it should. On the contrary, there is considerable evidence that carbon dioxide increases plant growth; indeed, the anomalous growth of Graybill’s bristlecones – which are integral ingredients in the Mann et al 19998-99 and many subsequent reconstructions – was attributed by Graybill and Idso to increased carbon dioxide.

Nor was the Briffa reconstruction based on “one obscure” data set. The decline was observed in an extremely large data set (389 sites in the Schweingruber network) covering the Northern Hemisphere that had been selected ex ante as being temperature dependent. In fact, the data sets that didn’t show the decline were the ones that were more vulnerable to small and problematic data sets i.e. Yamal and bristlecones, described in the past at CA as respectively cocaine and heroin for paleoclimatologists, who are addicted to these proxies.

Jaeah continues:

When that data was computed in one specific way, that formula gave scientists the wrong idea about the earth’s climate. When they realized it, they stopped using that formula on tree ring for temperatures after 1960. And relied on things like say actual recorded temperatures.

Jaeah’s final graphic is entitled “Briffa post 1960 data finally” and, as shown below, includes data after 1960. Presumably this was included in the “miracle” email from Briffa, though Jaeah didn’t say. As CA readers know, post-1960 Briffa data was first obtained in a Climategate email shortly after the IPCC Lead Authors meeting in Aruba (this version is plotted below.) The Mother Jones version looks related, but the scale looks a little different.

Jaeah’s conclusion is not warranted from the facts. She states:

so really the mystery of the decline was no mystery at all. Just a messy debate resulting from poor communication and misinterpretation.

Jaeah’s original puzzlement was why the IPCC report didn’t show the decline – an issue that was not investigated by any of the inquiries. (Thus, Jaeah could not locate an explanation in the inquiries.) While the deletion of adverse data can be traced back to Briffa and Osborn 1999 and Jones et al 1999 (both published in first half 1999), the Climategate emails show serious concern at the IPCC Lead Authors about the Briffa version in the Zero Order Draft (which had both a decline and lower centennial variability). Folland had observed:

But the current diagram with the tree ring only data [i.e. the Briffa reconstruction] somewhat contradicts the multiproxy curve and dilutes the message rather significantly.

Folland added caveats that they wanted the “truth”. Mann replied:

everyone in the room at IPCC was in agreement that this [the Briffa reconstruction] was a problem and a potential distraction/detraction from the reasonably concensus viewpoint we’d like to show w/ the Jones et al and Mann et al series.

Mann continued that he didn’t want to give “fodder” to skeptics:

So, if we show Keith’s series in this plot, we have to comment that “something else” is responsible for the discrepancies in this case. [Perhaps Keith can help us out a bit by explaining the processing that went into the series and the potential factors that might lead to it being "warmer" than the Jones et al and Mann et al series?? We would need to put in a few words in this regard] Otherwise, the skeptics have an field day casting doubt on our ability to understand the factors that influence these estimates and, thus, can undermine faith in the paleoestimates. I don’t think that doubt is scientifically justified, and I’d hate to be the one to have to give it fodder!

The “solution” to the problem eventually adopted by IPCC (whether it was Mann or Briffa or both is not material to the fact that it was some combination) built on the tactic already tested in Briffa and Osborn (Science 1999) – delete the inconvenient bits. This has nothing to do with “poor communication and misinterpretation”. Critics are not misinterpreting what they did. They understand precisely what they did and do not accept it as valid scientific methodology, with Richard Muller speaking sharply on this point.

Jon Stewart’s analysis remains the most acute;

It means nothing…He’s just using a trick to… hide the decline. It’s just scientist-speak for using a standard statistical technique for calibrating data in order to …. trick you…into not knowing about ….the decline.

No comments:

Post a Comment