In 2015 congressman Lamar Smith instigated an investigation against scientists at NOAA that is still ongoing, claiming that they had manipulated global temperature data. The scientists, led by Tom Karl, had published a paper which claimed that global warming since 2000 had been underestimated, and that claims of a 'hiatus' in warming were therefore wrong. Smith demanded access to the email discussions and notes of the scientists, hoping to find evidence of fraud. They provided the data and methods to verify their claims, but refused to provide emails. Scientists often argue against their own positions—this is a key part of scientific skepticism—so email discussions offer lots of material which can be taken out of context to give a misleading impression. We saw the same thing happen with the “climategate” emails.

In 1989 two scientists, Martin Fleischmann and Stanley Pons, published extraordinary claims about the possibility of room temperature nuclear fusion. The immediate response of the scientific community was to test those claims, with many groups trying to and failing to repeat their experiment. The consensus of the community was that the claims were wrong. The scientific community did not ask for Fleischmann and Pons' emails. There's a good reason for this; reading a scientist's emails will tell you, at most, whether they believe their own claims. It won't tell you whether those claims are right. Scientists frequently believe they are right when in fact they are wrong, as science is hard and many initial results do not hold up as robustly when independently replicated. And even when scientists are wrong, we often learn something from finding out how they were wrong.

The real test of a scientific claim is to compare it to observations—in other words to replicate that claim by experiment. Replication is key to science, though it is often under-appreciated and poorly incentivized in a publication environment focused on novelty and media interest. Scientists often have to explore many blind alleys to get to the right answer, and it often takes someone else to spot the errors in a study. As a result, scientists generally do not fully accept a result until it has been independently verified by another group. All the focus on the replication crisis in psychology and medicine over the past few years only reinforces the need for academic journals to recognize that the replication and independent validation of influential research is in many ways just as important as the original research itself.

In a paper just published in Science Advances, we set out to replicate the work of Karl and colleagues. We evaluated their results by producing three different ocean temperature records of recent years, using data from buoys, satellites, and Argo floats. These "instrumentally homogenous" records have the benefit of requiring little adjustment or calibration, as they all come from the same type of instrument. The new NOAA record agrees quite well with these instrumentally homogenous records, while we found a strong and significant cool bias in the old NOAA record and a modest but still significant cooling bias in records from the U.K.'s Hadley Centre and the Japanese Meteorological Association. Our results suggests that the new NOAA record is likely the most accurate of the various sea surface temperature records during the past two decades, and should help resolve some of the criticism that accompanied the original NOAA study.

These checks took a few weeks, and were largely done in our spare time (although writing them up for publication took much longer). Compare this with the months of effort by Smith and his lawyers, whose hourly rates are many times those of paid scientists, and funded by the taxpayer. And all of this was to answer the wrong question, since at best Smith could only have determined what the scientists believed. Only the data can determine whether their claims match reality. More worrying is the chilling effect of Smith's investigation on science itself. The message Smith's actions send to scientists is that if politicians don't like our results, they will do everything in their power to make our lives miserable. Smith's actions send a clear message to scientists that we should produce results which are convenient to political narratives, rather than which accurately reflect reality. 

There are some scientific results we can be very confident of, and there are some where we are less certain. The balance of evidence supports the new NOAA temperature record, although it is hard to determine temperature trends very precisely over periods of less than 30 years. Does this mean that there never was a 'hiatus'? That is a different question, and one that is still being hotly debated in the scientific community, in part because there is little agreement on the meaning of the term. It is as much a social question as a scientific one. However the scientific question of how fast the Earth has been warming over the past two decades can be answered by replication from the scientific community, not by a political investigation. And the best evidence we have says that NOAA got it right.