Scientists did not mislead but lacked transparency in the way they used tree ring data to reconstruct historic temperatures. Credit: NOAA

The "rigour and honesty" of scientists embroiled in the climate change e-mail affair are "not in doubt" — according to an independent review of the matter released today. However, the scientists have been criticized for a lack of openness that risked "the credibility of UK climate science".

In November 2009, more than 1,000 e-mails and documents were hacked from the Climatic Research Unit (CRU) at the University of East Anglia in Norwich, UK, and posted on the Internet. They prompted allegations from climate-change sceptics that CRU scientists withheld, concealed and manipulated data in an attempt to boost the case for human-induced climate change. The review, led by Muir Russell, the former vice-chancellor of the University of Glasgow, UK, was charged with investigating the scientists' behaviour. In a 160-page report, the five-person review committee says that it found no evidence of malicious intent, but rather a "consistent pattern of failing to display the proper degree of openness" both among researchers and in the university's leadership in handling the affair.

"They were unhelpful in response to legitimate requests," says Russell, adding that scientists "need to have in the forefront of their minds the importance of the credibility of the knowledge base they are generating and of not losing public trust".

The review has not impressed vocal climate-change critics, such as Andrew Montford who maintains the blog Bishop Hill. "I find the review pretty appalling," he says. "I do not think they have gone deep enough."

Under scrutiny

In exploring allegations that the CRU withheld or tampered with data, the committee scrutinized e-mails concerning the selection of weather-station data in research published in Nature1, led by former CRU director, Phil Jones — who stepped down from his position while the investigation was under way.

To check the paper's conclusion that rising temperatures could not be caused by the local "urban heat island effect" — in which cities tend to be warmer than surrounding rural areas — but should rather be attributed to global climate change, the review panel downloaded the source data directly from publicly accessible sites. "It became very clear, very early on that anyone can get the took us literally minutes to download," says committee-member Peter Clarke, a professor of physics at the University of Edinburgh, UK.

Russell says that the review panel analysed this independently obtained data and produced "essentially the same shapes of graph" as that reported by the CRU scientists, regardless of choice of weather stations — indicating that the reported results can be trusted.

"Was there something funny going on with the data? No," says Russell. However, the review committee found that CRU scientists should have been more transparent about which stations they had used, to allow others to directly check their analyses.

Climate scientists must be open to scrutiny, says committee-member Geoffrey Boulton, a professor emeritus of geology at the University of Edinburgh who added that they cannot "make papal announcements: 'this is this and that is that and you have to accept it'."

Divergent data

Another allegation revolved around the handling of contentious tree-ring-width data — used as a temperature proxy — in a graph that contributed to the Fourth Assessment Report (AR4) of the Intergovernmental Panel on Climate Change (IPCC) and to a similar graph used in a 1999 report to the World Meteorological Organization (WMO). The width of tree-rings fail to closely track direct records of temperature beyond around the 1960s. CRU scientists did not include this divergent tree-ring data from later years, instead splicing in direct temperature measurements that show a sharp temperature rise.

In one of the most notorious leaked e-mails, Jones, referring to the WMO report graph, described how he had "just completed Mike's trick of adding in the real temps to each series for the last 20 years". Jones was referring to the fact that climatologist Michael Mann of Pennsylvania State University in University Park had used direct temperature measurements to reconstruct temperatures over the past 20 years or so in a graph in an earlier Nature paper2. However, while Mann and his colleagues had clearly labelled which temperature lines were derived from direct measurements and which referred to proxy data, the graph submitted by Jones for the WMO report did not.

Regarding the splicing, the review found "no evidence" that the data presented in the IPCC report is misleading. Had the tree-ring data been left in, it would not have implied that recent temperatures have been decreasing, but only that the proxy data no longer tracked direct temperature records, says Clarke. The committee criticized the researchers for not clearly labelling on the graph how proxies had been used in its construction. The committee also concluded that the scientists did not actively select certain tree-ring data in an attempt to bias the graph presented in the IPCC report.

"In an email, Mann told Nature he was satisfied with the results of the review, adding, "It is my hope that we can now put this bogus, manufactured scandal behind us."

Ross McKitrick, an environmental economist at the University of Guelph in Ontario, Canada, argues the review missed the point of criticisms regarding the deletion of tree-ring data after the 1960s. "It gave the false impression that all proxy graphs follow the twentieth-century records," he says. "But given that they do not, the question is how accurate were they in tracking temperatures in past centuries — did they miss warming in the past too?"

The committee also investigated allegations that Jones and his colleagues had tried to suppress two other scientific papers3,4 from inclusion in AR4. Both papers appeared in the final version of the report, and the committee found no evidence of attempts to subvert peer review. However, the committee noted that peer review "should not be overrated as a guarantee of the validity of individual pieces of research".

Russell made clear that it was not within the remit of the review to assess the validity of CRU's climate science — that is still open to legitimate debate. The affair has also highlighted the increasing role of the blogosphere in directing debate, says Russell — something that scientists in all disciplines will have to be prepared to deal with openly. "We should take this as a lesson in what not to let happen," he says.

Montford says the report does not address some of the allegations relating to the e-mails. "The scientific and political establishment would rather this all just went away now — but I think it is too important for that to happen," he says.