Facebook Twitter Tumblr Close Skip to main content
A Project of The Annenberg Public Policy Center

No Data Manipulation at NOAA

Top Republicans on the House science committee claim a former National Oceanic and Atmospheric Administration scientist “confirmed” that his NOAA colleagues “manipulated” climate data for a 2015 study. But that scientist denies that he accused NOAA of manipulating data.

Rep. Lamar Smith, the chairman of the House Committee on Science, Space, and Technology, and two subcommittee chairmen issued a Feb. 5 press release — “Former NOAA Scientist Confirms Colleagues Manipulated Climate Records” — as part of an ongoing dispute over the validity of a paper published in the journal Science in June 2015 by NOAA scientists.

The NOAA study was one of several peer-reviewed studies published in recent years that found the so-called global warming “slowdown” did not occur. The studies countered the U.N. Intergovernmental Panel on Climate Change’s 2013 assessment report, which suggested a slowdown in global warming had occurred.

The Republican press release was issued a day after John Bates, a former NOAA scientist not involved with the study, published a blog post that accused the paper’s lead author, Thomas R. Karl, former director of NOAA’s National Centers for Environmental Information, of having “his ‘thumb on the scale’— in the documentation, scientific choices, and release of datasets—in an effort to discredit the notion of a global warming hiatus.”

Prompted by Bates’ blog post, the press release issued by Republicans accused Karl and his co-authors of data manipulation. “I applaud Dr. Bates’s efforts in uncovering the truth of this data manipulation,” Illinois Rep. Darin LaHood, chairman of the oversight subcommittee, said in the press release.

The committee also repeatedly tweeted about the incident, citing a misleading article published in the British tabloid the Daily Mail that made the same accusations, and retweeted similarly misleading articles by Fox News and the National Review (see update below).

But in interviews with the Associated Press and E&E, an online energy and environmental news outlet, Bates said he had not accused his colleagues of data manipulation.

Bates told the AP on Feb. 6 that there was “no data tampering, no data changing, nothing malicious” involved with his colleagues’ study. “It’s not trumped up data in any way shape or form,” he said.

Rather, Bates claimed Karl and his group hadn’t followed NOAA protocol in “the way data was handled, documented and stored, raising issues of transparency and availability,” the AP reported, adding that Bates thought the study was rushed “to influence the December 2015 climate treaty negotiations in Paris.”

During a Feb. 7 hearing on the use of scientific information at the Environmental Protection Agency, Smith repeated the accusations while questioning Rush Holt, the CEO of the American Association for the Advancement of Science, which publishes the journal Science.

After Holt cited Bates’ statement to E&E — “The issue here is not an issue of tampering with data,” Bates said — Smith responded, “I encourage you to talk to Dr. Bates because everything I have read that he has said about the Karl report suggests to me that NOAA cheated and got caught.” Smith added, “They did falsify data to exaggerate global warming.” In his opening statement, Smith also said, “Science should retract the Karl study,” based on Bates’ accusations.

We won’t get into whether Karl and his co-authors adhered to all of NOAA’s protocols or rushed their paper to publication. (We’ve written about the latter previously.)

But we will explain why there’s no evidence to support the committee members’ accusations of data manipulation. We’ll also explain why, regardless of whether or not there was a slowdown in warming in the early 21st century, the long-term warming trend remains unabated.

Update, Sept. 25: An organization that regulates the newspaper and magazine industry in the U.K. ruled in July that the Daily Mail article that was repeatedly shared on Twitter by the House Science Committee was inaccurate and misleading. The article portrayed Bates’ criticisms of NOAA scientists as providing “irrefutable evidence” that the paper had been grounded in “misleading, ‘unverified’ data,” which led to world leaders being “duped” about global warming. “These claims by the newspaper went much further than the concerns which Dr Bates had detailed in his blog or in the interview,” the Independent Press Standards Organisation, wrote on July 7. In a Sept. 16 statement on the ruling, the Daily Mail acknowledged, “In characterising Dr Bates’ claims in this way the newspaper had failed to take care over the accuracy of the article.”

The ‘Most Accurate’ Ocean Data

Karl and his co-authors’ 2015 analysis of surface temperature exhibited “more than twice as much warming as did the old analysis at the global scale” between 1998 and 2012, consequently discounting the IPCC report’s global warming “slowdown.”

To come to this conclusion, the group published two reanalyzed sets of data in the 2015 paper — surface temperature data from the land and from the ocean — dating back to 1880. We’ll discuss the ocean data first.

The researchers state in their paper that the doubling in warming was “clearly attributable” to the reanalysis of the ocean data. This is due, in part, to the fact that the oceans cover 71 percent of Earth’s surface, so changes in ocean temperature data analysis significantly impact the overall analysis of the global warming rate.

But why did the data need to be adjusted?

Since ocean temperature data used in the Science study came from two different sources — buoys and ships — the full data set needed to be adjusted to accommodate for differences in how each source measures temperature.

Several studies have noted that “ship data are systematically warmer than the buoy data” and that buoy data are “more accurate,” the researchers write. Since buoys have increased in use over time and ships have decreased in use, there is a “time-dependent bias” in the global ocean temperature record, they explained.

The previous analyses, including from the IPCC report, didn’t fully correct for this bias, which prompted the idea of a global warming “slowdown.” But with the corrections used in the Science study, the slowdown largely disappeared.

Bates did not accuse his former colleagues of manipulating ocean temperature data. But he did argue that the group didn’t adhere to NOAA’s protocols involving the release and archiving of the ocean data.

However, any purported protocol issues haven’t prevented other scientists from replicating the Science study’s analysis of ocean data.

Zeke Hausfather, a data scientist at the University of California, Berkeley, and others replicated the NOAA scientists’ analysis of ocean data in a study published in the journal Science Advances in January.

The researchers found that NOAA’s analysis “generally agree[s]” with other independent ocean temperature data sets, such as those compiled from satellite radiometersArgo floats and other sources.

This is important because the ability to verify NOAA’s analysis using different lines of evidence — in this case, ocean temperature data from various independent sources — suggests that it’s accurate, Hausfather told us over the phone.

In fact, Hausfather and his group also compared ocean temperature records from other institutions, such as the U.K.’s Met Office Hadley Centre for Climate Science and Services and the Japanese Meteorological Agency, to these independent data sets and found that NOAA’s analysis is the “most accurate” measure of global ocean surface temperate change among the different international agencies.

In short, Hausfather and his colleagues’ study bolsters the Science study’s finding that previous analyses, used by the IPCC, for example, underestimated the rates of surface warming in the early 21st century.

Land Data and the Long-term Trend

As for the land temperature data, Bates did argue in his blog post that his colleagues used “experimental,” and thus, not fully verified, land temperature data.

He mainly argued that his colleagues should have labeled their land data analysis “experimental” in their paper so that other researchers know it’s not a “routine, operational update” that would have fixed software “flaws.” In this way, his colleagues defied “NOAA scientific integrity guidelines,” Bates said.

But Bates still didn’t accuse them of data manipulation, as the House science committee claimed.

Regardless, Karl and his co-authors explain in their paper that, compared with the ocean data, their reanalysis of land temperature data showed relatively minor changes, and thus, contributed relatively little to changing the early 21st century warming trend.

Karl and his group also note that when it comes to the overall long-term warming trend — that is, from 1880 to the present — their reanalysis had “essentially the same rate of warming as that of the previous analysis.”

This means that in the large scheme of things, the rate of global warming remains unchanged, whether or not there was a “slowdown” in the rate of global warming in the beginning of the 21st century.

To top it off, Karl and his co-authors point out that “it is also clear that the long-term trend would be significantly higher … without corrections” to the raw data. In other words, compared with the raw data, the adjusted data show less warming over the long-term.

Why? Because ocean surface temperature data greatly impacts the overall global warming trend and a lot of 20th century ocean data comes from ships, which are systematically too warm. As a result, scientists have to actually adjust the long-term trend downward to accommodate for this bias.

As Hausfather said when we spoke to him over the phone, “If scientists are cooking the books, they’re cooking them in the wrong direction.”

Editor’s Note: SciCheck is made possible by a grant from the Stanton Foundation.