Scott K. Johnson
Update: In interviews with E&E News and ScienceInsider, Bates has denied being the whistleblower the House Science Committee has cited in the past, although his accusations are very similar to those cited by the Committee chairman.
Original article follows
On Sunday, the UK tabloid Mail on Sunday alleged a seemingly juicy (if unoriginal) climate science scandal. At its core, though, it’s not much more substantial than claiming the Apollo 11 astronauts failed to file some paperwork and pretending this casts doubt on the veracity of the Moon landing.
The story’s author, David Rose, has published a great many sensational articles over the years, falsely claiming to present evidence undermining the threat of climate change or the human cause behind it. But this latest article is noteworthy in that it appears to reveal the supposed “whistleblower” who has been cited by the US House Science Committee in its ongoing clash with climate scientists at the National Oceanic and Atmospheric Administration (NOAA).
The committee’s Twitter account, as well as the account of Committee Chair Lamar Smith (R-Texas), has gone hog-wild tweeting about the story. For example, the committee account tweeted, “@NOAA obstructed the committee's oversight at every turn. Now we know what they were hiding.”
The research at the center of the supposed scandal was a 2015 paper published in Science by a group of NOAA researchers. The paper presented an update to NOAA’s global surface temperature record using newly updated land and ocean databases—each of which had previously been published in peer-reviewed scientific journals. The update resulted in slight changes to NOAA’s numbers for some recent years, which slightly increased the short-term rise since 1998 (the adjustment brought NOAA's record closer to other major datasets).
The paper concluded that there was no evidence of a slowdown in global warming over the last decade or so, an idea that had been a focus of people who reject the seriousness of human-caused climate change.
Rather than engage with the science behind this paper, Rep. Lamar Smith has, without any evidence, accused the NOAA scientists of doctoring their results to exaggerate recent warming. Although NOAA provided Smith with the (publicly available) data and methods behind the paper and provided a personal explanation of the research, Smith subpoenaed the e-mails of the scientists and other NOAA staff. NOAA handed over staff e-mails but refused to make the researchers’ e-mails available for a fishing expedition, citing the importance of protecting scientists’ ability to communicate freely while trying to understand their data.
Rep. Smith claimed that a whistleblower at NOAA had provided his office with information proving that the study had been inappropriately rushed for political reasons. The Mail on Sunday claims the same thing and presents NOAA scientist John Bates as a whistleblower.
The whistleblower
Bates recently retired from NOAA after a career working primarily on satellite measurements used for weather forecasting. Recently, he was also in charge of data-archiving efforts for satellite and surface temperature records. Bates alleges that NOAA's Tom Karl and the rest of the team behind the paper failed to adequately follow NOAA’s internal processes for archiving their data and stress-testing the updated databases they used.
Bates also questions the way in which some sea surface temperature measurements were adjusted to sync them up with the rest of the measurements, falsely claiming that the technique alters the warming trend.
In a blog post, Maynooth University research Peter Thorne—who worked on both the land and sea databases underlying the Karl paper but not the Karl paper itself—disputed many of Bates’ claims. First off, Thorne notes that Bates was not personally involved in the research at any stage. And while Bates claims that Karl made a series of choices to exaggerate the apparent warming trend, Thorne points out that this would be difficult for Karl to do since he didn’t contribute to the underlying databases. Karl’s paper simply ran those updated databases through the same algorithm NOAA was already using.
Ars talked with Thomas Peterson, a co-author on the Karl paper who has since retired. Peterson provided some useful context for understanding Bates’ allegations. The satellites that Bates worked with were expensive hardware that couldn't be fixed if anything went wrong after they were launched. The engineering of the software running those satellites sensibly involved testing and re-testing and re-testing again to ensure no surprises would pop up once it was too late.
Bates expected the same approach from his surface temperature counterparts, but Peterson explained that their work with weather station data was not nearly so high-stakes—problems could easily be fixed on the fly. The engineering-style process NOAA was using for endlessly double-checking the software for all dataset updates could drag on for quite a long time—years, in fact—and Bates opposed any attempt to speed this up. Peterson and other scientists were naturally anxious to incorporate changes they knew were scientifically important.
Bates alleges that the Karl paper was “rushed” for political reasons, but Peterson said the reality was that NOAA was well behind the times, waiting to include known improvements like additional recording stations in the rapidly warming Arctic. “I had been arguing for years that we were putting out data that did not reflect our understanding of how the temperature was actually warming—[for] literally years we slowed down to try to account for some of these processing things that we had to do,” Peterson said. (At the time of the Karl paper, NOAA’s dataset showed less warming in recent years than other datasets, like NASA’s.)
Bates also claims that there were bugs in the land station database software that were ignored in the Karl paper. But according to Peterson, the slight day-to-day variability seen in the software’s output was simply the result of the fact that new data was added every day. Stations that straddled statistical cut-offs might fall on one side of the dividing line today, and on the other side tomorrow. There was nothing wrong with the software, they realized. It was just silly to re-run it every single day.
There may also be something beyond simple “engineers vs. scientists” tension behind Bates’ decision to go public with his allegations. Two former NOAA staffers confirmed to Ars that Tom Karl essentially demoted John Bates in 2012, when Karl was Director of NOAA’s National Centers for Environmental Information. Bates had held the title of Supervisory Meteorologist and Chief of the Remote Sensing Applications Division, but Karl removed him from that position partly due to a failure to maintain professionalism with colleagues, assigning him to a position in which he would no longer supervise other staff. It was apparently no secret that the demotion did not sit well with Bates.
Let’s ask the data
Office politics aside, the claims in the Mail on Sunday article that the Karl paper exaggerated the warming trend fall down when you examine any of the other surface temperature datasets. In a paper we recently covered, a team led by Berkeley researcher Zeke Hausfather compared the updated sea surface temperature dataset to shorter but simpler and independent sets of measurements made by satellites and automated floats. That analysis confirmed that the updated dataset is more accurate than its predecessor.
In a post for Carbon Brief, Hausfather noted that NOAA’s updated dataset doesn't cause it to show more warming than the datasets run by NASA, the Berkeley team, and the UK Met Office. Instead, the update caused NOAA to stop showing less warming than everyone else.
The House Science Committee’s Twitter account has yet to tweet a link to Hausfather's story.
Hausfather also points out a glaring error in the Mail on Sunday article that illustrates its author’s lack of knowledge. The article includes a graph of both the NOAA and UK Met Office records. The NOAA data appears to be roughly 0.1°C warmer than the UK Met Office data across the entire time span—supposedly evidence of “flawed NOAA data showing higher temperatures.” Apart from the fact that a constant offset would have no impact on temperature trends, the offset is simply a mistake. The numbers in the two datasets are calculated relative to different baselines—the 1901-2000 average for NOAA, and the 1961-1990 average for the Met Office. Once you put them on a common baseline, the differences largely disappear.
The Mail on Sunday updated the caption to note that the datasets “are offset in temperature by 0.12°C due to different analysis techniques,” but the graph remains unchanged. And it's not clear how many Mail readers will understand the importance of this offset.
For a final example of its author's incompetence, the Mail on Sunday article claims that NOAA has seen the error of its ways and that it's suddenly working on a new version (version 5) of the sea surface temperature database that reverts the previous changes and reduces the cooling trend again. In reality, the new version is just the normal process beginning the next update, which will incorporate data from the growing fleet of autonomous floats, among other things. It's happening because version 4 is now over three years old. Hausfather notes that while the preliminary version 5 does indeed reduce the recent warming trend by about 10 percent, it still shows 50 percent more warming than the version Bates and Lamar Smith seem to like.
The article on the Daily Mail website is headlined “Exposed: How world leaders were duped into investing billions over manipulated global warming data," but the list of those “duped” seems to be limited to the author of the story and any readers who make the mistake of trusting it. Sadly, those believers include the head of the House Science Committee.