On climate change, Gallup pollsters say Americans divide into three groups -- the "Concerned Believers," the "Mixed Middle," and the "Cool Skeptics." Believers have a slender plurality at 39 percent but skeptics make up a solid 25 percent. They think there’s little to worry about and that media reports on the topic are exaggerated.
Fox News host Steve Doocy gave the doubters some ammunition on June 24, 2014. In a segment on Fox and Friends called "News by the Numbers," Doocy drew viewers’ attention to the year 1934.
"That's the hottest year on record in the United States," Doocy said. "At least until NASA scientists fudged the numbers to make 1998 the hottest year to overstate the extent of global warming. The 1930s were by far the hottest decade in the United States."
A reader wondered if NASA really did cook the books (we love reader suggestions!), so we are checking Doocy’s claim about fudging the numbers.
We asked Fox News for their source and while they didn’t respond, a number of conservative news outlets have made much in recent days of a blog post from a man who writes under the pseudonym Steven Goddard. Goddard charged that until 2000, NASA reported that in the United States, 1934 was hotter than 1998 and that the country has been cooling since then.
"Right after the year 2000, NASA and NOAA dramatically altered U.S. climate history, making the past much colder and the present much warmer," Goddard wrote.
He provided this animated chart to prove his point (the chart marked "a" is the old version):
Climate science experts say not so fast
Doocy exaggerated the findings in this blog post when he applied it to global warming. The post itself only talks about U.S. land temperatures and what happens in the United States is separate from global shifts.
As far as what the blog actually claimed, while it accurately copied the changes in the government charts, experts in U.S. temperature measurement say it ignores why the charts shifted. There were major changes in how the country gathered temperature information over the decades.
Zeke Hausfather is a data scientist with Berkeley Earth, a research group that has expressed doubts about some of the reports on climate change coming from Washington and international bodies. Hausfather took Goddard to task when Goddard made a similar claim about numbers fudging earlier this month. The missing piece in Goddard’s analysis, Hausfather said, was he ignored that the network of weather stations that feed data to the government today is not the one that existed 80 years ago.
"He is simply averaging absolute temperatures," Hausfather wrote. "Absolute temperatures work fine if and only if the composition of the station network remains unchanged over time."
Weather stations that once were in a valley might now be on a hill top and vice versa. But the shift could be greater than simple elevation. Stations were moved from one part of a state to another. The number of stations within a given area shifted. All these differences, Hausfather and other experts said, will alter the typical temperatures gathered by government meteorologists.
Gavin Schmidt, director of the NASA Goddard Institute for Space Studies, said the raw data used in the blog post suffered from an equally troubling flaw. The temperatures were not measured at the same time of day.
"Over time, the U.S. network went from recording max/min temperatures at different points of the day, to doing it at midnight," Schmidt said.
In fact, volunteers staffed many of the stations. Before 1940, most followed Weather Service guidelines and recorded the temperature at sundown. Through the second half of the century, there was a gradual shift to recording morning temperatures. This change produced the appearance of a cooling trend when none existed.
Comparing apples to apples
Better instruments and more consistent methods have allowed scientists to collect more reliable data. But for climate studies, long-term trends are key and the challenge has been how to make the best use of the older readings.
In the mid 1980s, the government settled on a list of about 1,200 stations across the country to track temperature trends. Around 1990, climatologists began delivering computer programs to factor in the artificial changes that systematically pushed the readings one way or the other. Over time, they accounted for the impacts of equipment, location, the time of day of measurements and urbanization (more asphalt leads to higher surface temperatures).
There is no question that running the raw data through these programs changes the graphs of average temperatures. However, multiple researchers from a variety of institutions have fed into this process and come up with their own computer models. Results from different teams largely match up.
John Nielsen-Gammon is a researcher at Texas A&M University and is the Texas state climatologist. Nielsen-Gammon finds nothing nefarious in the government analysis of temperature trends.
"It is reasonable to expect the adjusted data record to change over time as the technology for identifying and removing artificial changes improves," Nielsen-Gammon said. "If there are any biases, they are caused by the quality of the underlying data, not by any biases intentionally introduced into the adjustment process."
All of the experts we reached or whose work we read rejected Goddard’s conclusions.
Mark C. Serreze, professor of geography at the University of Colorado-Boulder, said no fabrication has taken place.
"Goddard's results stem from an erroneous analysis of the data," Serreze said.
Anthony Watts, a popular skeptic of most climate change data, posted his objection to Goddard’s claim.
"I took Goddard to task over this as well in a private email, saying he was very wrong and needed to do better," Watts wrote.
Doocy with Fox News said NASA scientists fudged the numbers to overstate the extent of global warming. This exaggerated the thrust in the underlying blog post. It accused government scientists of altering the U.S. temperature record, not the record for the entire earth.
As for what the blog said, we found that experts across the spectrum found fundamental flaws in its analytic methods. By relying on raw data, it ignored that the number and location of weather stations and the methods of measuring temperatures across the United States have changed greatly over the past 80 years.
The experts we reached or whose work we read generally agree that the corrections for flawed data produce valid results. The bare bones approach used in the blog post provides no solution to the issues of weaknesses in the raw data.
We rate the claim Pants on Fire.
After the Fact
The man behind the science Fox quoted responds
Added on July 1, 2014, 2:35 p.m.
After our fact-check published we heard from Goddard, who later identified himself as Tony Heller. He noted that one of the experts we quoted in our initial piece has since revised his views of Heller’s research.
While there remains no evidence that "NASA scientists fudged the numbers to make 1998 the hottest year to overstate the extent of global warming," as Fox’s Doocy claimed, we did want to add additional perspective to this fact-check.
In short, as one of the experts in our fact-check noted, the adjusted data set from the government is imperfect and it changes as people work on it. However, the weight of evidence says the imperfections, or errors, have little impact on the broader trends.
Flaws in the data were the focus of blog posts from two of the experts we used in our fact-check.
"I was so used to Goddard being wrong, I expected it again, but this time Steve Goddard was right and my confirmation bias prevented me from seeing that there was in fact a real issue in the data," science blogger Anthony Watts wrote.
Watts zeroed in on temperature records from the United States Historical Climatology Network. The network takes readings from over 1,200 stations, chosen because they have been providing information for many decades on a good cross-section of places across the country. The network’s Web page says at the top, "USHCN temperature records have been ‘corrected’ to account for various historical changes in station location, instrumentation, and observing practice."
In the past, temperature readings might come from any time of the day. The modern standard is to take readings at midnight. By looking at many readings from many stations over time, analysts create a data set that adjusts for what they call a bias based on the time of observation. The original raw data remains available but researchers use the altered data set for long term comparisons.
Importantly for Watts, these corrections also include filling in estimated readings when none are available. Many of these adjustments and in-filling are done on older data but Watts found two examples posted by another blogger, one from Texas and one from Kansas, where current readings were changed in the data set.
"While the raw data file has the actual measurements, for some reason the final data they publish doesn’t get the memo that good data is actually present for these stations, so it ‘infills’ it with estimated data using data from surrounding stations," Watts wrote. "It’s a bug, a big one."
Judith Curry, a climatologist at the Georgia Institute of Technology, also originally downplayed Heller’s claims. But after seeing the data from Texas, Curry acknowledged that he might have had a point.
"I infer from this that there seems to be a real problem with the USHCN data set, or at least with some of the stations," Curry wrote. "Maybe it is a tempest in a teacup, but it looks like something that requires NOAA’s attention."
That said, Curry also tweeted that "what Goddard did to the data was bogus."
In the fact-check, Doocy accused the government of manipulating the data purposely to overstate the extent of global warming. Evidence of that does not exist.
In fact, researchers who have been skeptical of the government’s climate record have looked at the temperature data and found that it holds up (even if it contains errors). Zeke Hausfather, a data scientist, is a member of the group known as Berkeley Earth.
"Despite using different methods, and using about 8 times more raw station data, we ended up with nearly identical results," Hausfather said.
Hausfather provided PunditFact the following graphic. NCDC refers to the National Climatic Data Center, the agency home for the temperature readings. The blue line is Hausfather’s data, the red line is the NCDC’s.