Environmental Protection Agency Administrator Scott Pruitt speaks with coal miners in Sycamore, Pa., in April. (Justin Merriman/Getty Images)

In a sign of growing tensions between scientists and the Trump administration, researchers published a scientific paper Wednesday that was conceived and written as an explicit refutation to an assertion by Environmental Protection Agency Administrator Scott Pruitt about climate change.

The study, in the journal Nature Scientific Reports, sets up a direct test of a claim by Pruitt, made in written Senate comments following his confirmation hearing, that “over the past two decades satellite data indicates there has been a leveling off of warming.”

After reviewing temperature trends contained in three satellite data sets going back to 1979, the paper concludes that the data sets show a global warming trend — and that Pruitt was incorrect.

“Satellite temperature measurements do not support the claim of a ‘leveling off of warming’ over the past two decades,” write the authors, led by Benjamin Santer of Lawrence Livermore National Laboratory. Santer co-authored the study with three Livermore colleagues and scientists from MIT, the University of Washington in Seattle and Remote Sensing Systems, which keeps one of the three satellite temperature data sets.

“In my opinion, when incorrect science is elevated to the level of formal congressional testimony and makes its way into the official congressional record, climate scientists have some responsibility to test specific claims that were made, determine whether those claims are correct or not, and publish their results,” said Santer in an interview, when asked about the framing of the research.

The study wades into an ongoing and highly fraught debate over how to interpret the temperature records of the planet’s lower atmosphere, or troposphere, provided by polar orbiting satellites.

Such data have often been cited by climate change doubters so as to suggest that there is no global warming trend, or that global warming has recently slowed down, and therefore to contradict thermometer-based measurements taken at the planet’s surface (which show a clear warming trend).

The head of the Environmental Protection Agency, Scott Pruitt, says he is not convinced carbon dioxide from human activity is the main driver of climate change and wants Congress to weigh in on whether CO2 should be regulated. (Reuters)

But the new study finds that all of the three satellite data sets — kept by Remote Sensing Systems, the Center for Satellite Applications and Research at the National Oceanic and Atmospheric Administration, and the University of Alabama at Huntsville — show a long-term warming trend in the middle to upper part of the troposphere. After correcting for a cooling-down of the stratosphere (the layer above the troposphere), the paper finds that the trend is roughly 0.36 degrees Fahrenheit per decade for the first two data sets, and 0.26 degrees Fahrenheit per decade for the third.

The study further examined whether any shorter temperature trend in these data sets could be described as a “leveling off,” as Pruitt had put it. It did so by examining 20-year periods in the data sets and comparing those with the predictions of climate simulations that reflected the natural variations of the climate but excluded human-caused greenhouse gas emissions. These models were thus meant to represent what the climate would do on its own if humans were not altering it.

The study finds warming trends for all the 20-year periods, including the “last two decades” referred to by Pruitt, although it acknowledges that the trend is somewhat lower over these later periods. But it attributes this to natural climate variations, including a very strong El Nino event in 1997 and 1998 that caused dramatic warmth around the beginning of the 20-year window that ends in the present.

Even in these periods that saw somewhat less warming, the study finds that it was still far more warming than would be without human perturbations of the climate. “The probability that internal variability could produce warming exceeding that observed over the last 20 years is only 1.6 %, 3.1 %, and 6.3% (respectively)” in the three data sets, the authors find.

“Pruitt is not correct in saying that warming has leveled off,” Santer said. “It hasn’t in any of the satellite data sets, and indeed, in older and newer versions of the three satellite data sets, we judge the most recent warming to be statistically significant — to be larger than the warming that our current model-based estimates tells us that we should see due to internal variability alone.”

The EPA did not immediately respond to a request for comment.

“Another solid piece of work by Santer et al. that demonstrate multi-decadal satellite-derived global tropospheric temperatures are increasing far more than we would expect from natural causes,” said Thomas Karl, a longtime climate researcher who formerly headed NOAA’s National Centers for Environmental Information.  “Other satellite instruments, which measure temperatures closer to where we live, work and grow our food show at least as much, or more warming, in recent decades.”

Gavin Schmidt, who heads the Goddard Institute for Space Studies at NASA, said by email that when it comes to measurements of the Earth’s troposphere by satellite, “the trends over the whole period are clear.”

“This doesn’t however imply that a) there aren’t still issues with the satellite retrievals (there may well be), and b) that models did a perfect job over this time period,” Schmidt cautioned.

John Christy, a researcher at the University of Alabama at Huntsville who keeps that data set and whose work has been often cited by climate change “skeptics,” agreed there is a warming trend in the satellite data overall but said that climate models predict that it should be larger. “The datasets are still significantly cooler than the model average,” he said by email.

Christy also argued that the other two data sets, which are warmer than his, are “outliers regarding the magnitude.”

“I wouldn’t get too excited about this study,” Christy said.

But it is not as though a scientific study refuting one of his statements to the Senate holds much risk for Pruitt, said Sarah Binder, a senior fellow at the Brookings Institution and a political scientist at George Washington University.

“It’s significant in the sense that it shows the limits of the confirmation process, especially when the president’s party controls the Senate and senators can no longer filibuster nominees. In other words, it’s possible to float factually inaccurate statements and yet not ding your chances of confirmation,” Binder said. “Of course, the climate change issue is highly partisan: Republicans tend to disagree with a general scientific consensus that the earth is warming. So the idea that a Republican EPA nominee might give [a] factually contested statement on climate change and not pay a price is not terribly surprising.”

In the end, Santer argued, scientists should fact-check politicians even if they’re at a disadvantage when it comes to how long it takes to do so.

“These claims were made in the U.S. Senate, in a confirmation,” said Santer. “It takes time however to set the record straight, to do due diligence, to do the research necessary to address the claims. And one would hope that the scientific response receives at least some token amount of attention, and that the original incorrect claim does not dominate the public discourse on these critically important issues.”