The Trouble with Temperature

Press still trying to set the record straight on cool weather

Articles about climate change legislation quickly piled up on Tuesday morning as the Senate began debating a proposal to cap greenhouse-gas emissions. The bill may be imperfect, and it may fail, but the deliberations are nonetheless the culmination of years of effort to convince to the government to address the threat of global warming.

Over the last couple of months, however, a number of reporters have had to return to the most basic scientific question underlying the Senate’s work: Is the Earth actually heating up? Over the long run, the answer is an unequivocal yes. But the short-term picture is less certain. For almost ten years, the average global temperature has been relatively stable, leading to one of the most enduring arguments against man-made climate change: that there hasn’t been any warming since 1998.

Skeptics have seized upon this rationale again and again in an effort to deny the human signature on climate change. Each time they do, reporters and bloggers try to explain that average global temperature can plateau or decline for many years, but the long-term trend is toward a warmer world. The global-warming-stopped-in-1998 meme has proved exceedingly tough to dispel, however.

The latest attempt at that feat came Monday in the form of an enterprising article by Associated Press science reporter Seth Borenstein, headlined “Statisticians reject global cooling.” The impetus for his work, as explained in the piece, was a gush of Internet chatter about cooling, most of which had stemmed from the book SuperFreakonomics, released last week, and a BBC article published a few weeks ago.

Both items created storms of controversy on the Web by suggesting that global warming stopped ten years ago. Professional and amateur pundits assailed their respective authors for misrepresenting temperature data to support their conclusions, but what seemed like an equal number touted their writing as proof of holes in the scientific consensus about climate change. In an effort to settle the argument, Borenstein decided to try something different.

“In a blind test, the AP sent temperature data to four independent statisticians and asked them to look for trends, without telling them what the numbers represented,” his story explained. “The experts found no true temperature declines over time.”

That is, of course, the same conclusion that many others have already reached. I asked Borenstein why he thought it was necessary to consult the four statisticians (a full explanation of the AP’s methodology can be found here) after so many cogent rebuttals of the cooling argument had already been written.

“Simple. Better them than me,” he replied. “They’re experts in looking for trends. Plus, I think the concept of it being blind for the statisticians, not knowing they were looking at temperature data, takes out any claims of bias. My issue was how do you look objectively for a trend? How about if you don’t know what you are looking at.”

That is a reasonable response in a world where punditry on both sides of the climate debate is often visceral and vitriolic. The question is: Will Borenstein’s article do anything to improve public understanding of recent and short-term temperature trends? Unfortunately, the likely answer is no.

SuperFreakonomics and the BBC article were by no means the first pieces of writing to mislead readers about global cooling. We went through this debate last year with Politico and The Washington Post’s editorial page. Nor is Borenstein the only reporter who has tried to deliver a more accurate characterization of how the last ten years fit into the bigger picture of climate change.

“The plateau in temperatures has been seized upon by skeptics as evidence that the threat of global warming is overblown,” New York Times reporter Andrew Revkin wrote in late-September article dedicated to the subject.

The skeptics had been bandying about the work of Mojib Latif, a climate scientist who works at the Leibniz Institute of Marine Sciences in Germany. In 2008, Latif and his colleagues published a research paper predicting that global temperature may not increase in the next decade due to cyclical ocean circulation patterns. The paper also noted that North America, Europe, and the North Atlantic might cool slightly. The authors stressed that these would be temporary trends superimposed on the long-term pattern of warming. When Latif made similar remarks at United Nations climate conference in early September, however, the story quickly turned into “Scientists pull an about face on global warming.” Revkin’s article was an attempt to provide the missing context, but it also epitomized just how difficult that can be.

Ironically, skeptics such as George Will touted Revkin’s story while pundits like Climate Progress’s Joe Romm accused him of “pushing the global cooling myth (again!).” Indeed, Revkin got the exact same response during the winter of 2008, when he penned a similar piece headlined “Skeptics on Human Climate Impact Seize on Cold Spell.”

True, Revkin’s more recent article contained at least one minor error. A sentence that originally read “The recent spate of years with cool temperatures…” was rightly changed to read “The recent spate of years with stable temperatures…,” because the last decade comprises eight of the hottest years on record. But it goes to show that covering short-term temperature trends can be a damned-if-you-do-damned-if-you-don’t scenario. At the very least, with criticism of articles like Revkin’s coming from both the right and the left, it is no wonder that the cooling meme ensues.

One issue that neither Borenstein nor Revkin’s piece addressed is the unpredictability of short-term climate modeling. The subject often proves too cumbersome for the daily news, but can still be instructive when contemplating recent temperature trends and forecasts., a blog run by a group of climate modelers, recently featured a useful post on this topic. In it, NASA’s Gavin Schmidt explained that the “dominant source” of uncertainty in short-term climate forecasts is natural variability rather than the influence of human industry.

In other words, climate models aren’t very good at predicting what the temperature will be five, ten, or twenty years from now because that will depend on the weather, which can amplify or drown out the subtler effects of manmade global warming. Those models can reliably predict long-term temperature trends, however, because over longer periods of time, those obfuscating weather patterns (whether hot or cold) “get averaged out” and let the signal from man-made warming shine through.

This relationship between weather and climate is why actual temperature data (over the last ten years, for instance) comes out looking like a saw-toothed pattern that tilts upward over time. It is also why scientists like Latif can predict short-term cooling and long-term warming without being inconsistent. On a yearly basis, weather will produce much more drastic temperature swings than anything greenhouse gases can do. But those emissions will have their day in the end, slowly but surely elevating the baseline temperature of those weather-related swings.

The trick for modelers, then, involves gaining a better understanding of the relationship between cyclical weather patterns—primarily those related to the oceans and the sun—and temperature. The RealClimate post about decadal modeling began by criticizing a September article in New Scientist by Fred Pearce about Latif’s temperature-plateau hypothesis. But in 2008, Pearce and Michael Le Page co-authored a feature for the magazine that remains one of the few cogent explanations in popular press of efforts to improve decadal modeling.

“[M]aking forecasts is all about figuring what dominates the state of the atmosphere on various timescales,” Pearce and Le Page wrote. “Some things, like accumulating greenhouse gases, matter over many decades while other things, like warm and cold fronts, dominate over days and months. Over periods of a few years, there’s growing evidence that the oceans are the key – and this is encouraging researchers to attempt short-term forecasts.”

That’s exactly the kind of succinct explanation that would greatly improve other articles about temperature trends. Even then, however, it’s doubtful that skeptics will stop hammering the argument that global warming has stopped. James Inhofe brought up the recent temperature plateau on the Senate floor on Tuesday, citing the BBC article mentioned above. A column by The Washington Post’s Dana Milbank noted that fellow Republicans on the Senate’s environment committee don’t agree with Inhofe’s assessment of the science, so maybe there is hope for the legislature, but reporters should not be complacent. Bad temperature analyses are like zombies in their capacity for reanimation.

Curtis Brainard writes on science and environment reporting. Follow him on Twitter @cbrainard.