Last week we discussed a new reconstruction of atmospheric CO2 and temperature over the last 485 million years that claimed to have found a strong correlation even though it admitted there wasn’t one in the entire Mesozoic Era that occupies almost the entire middle part of the period. And this week we want to look at how the Washington Post and others covered the story because, while we’re always glad to see people taking the historical view on an issue, the Post did a number of suspicious things typical of commentators, starting with the usual bit about the settled science being wrong and therefore right that commands a certain admiration for its mental gymnastics if not its intellectual soundness.

The Post declares:

“An ambitious effort to understand the Earth’s climate over the past 485 million years has revealed a history of wild shifts and far hotter temperatures than scientists previously realized – offering a reminder of how much change the planet has already endured and a warning about the unprecedented rate of warming caused by humans.”

Which is basically all wrong. Previous long reconstructions including the one shown on our stripes mug had already revealed wild shifts and hotter temperatures. The nature of the past serves as a warning about natural variability and proof that what’s happening now is not “unprecedented” either in pace or scale. The exact opposite of what the reporter said.

Or rather what the scientists, to their shame, said. The Post said of Emily Judd, “a researcher at University of Arizona and the Smithsonian specializing in ancient climates and the lead author of the study” that:

“At no point in the nearly half-billion years that Judd and her colleagues analyzed did the Earth change as fast as it is changing now”.

Bosh. Insolent bosh. How can you possibly know, in telling us for instance that the average temperature of the planet rose from 20C around 420 million years ago to 34 degrees about 410 million years ago, that it never went up 1.5 degrees in 150 years? None of your proxies begin to capture shifts, or fluctuations, on such a short time-scale so long ago.

There’s also a glaring falsehood in the story:

“in keeping with decades of past research on climate, the chart hews closely to estimates of carbon dioxide in the atmosphere, with temperatures rising in proportion to concentrations of the heat-trapping gas.”

Nothing of the sort is true. It’s certainly not true of other reconstructions of the two that show a long slow decline of CO2 from the Jurassic-Cretaceous boundary while temperature held steady for tens of millions of years in the glory days of the dinosaurs. But it’s also not true of their own reconstruction, driven by a possibly excessive conviction that CO2 and temperature must track one another, because while their own tends to minimize the dramatic late-Jurassic rise in CO2 shown in other reconstructions that coincides with temperature plunging into what looks like a close brush with an ice age, the study does concede that there’s no correlation between CO2 and temperature throughout the entire Mesozoic Era, consisting of the Triassic, Jurassic and Cretaceous Periods so it begins with a massive extinction (including of Dimetrodon, the cool one with the sail on its back) followed by a flowering first of reptiles then of dinosaurs, and ends with the asteroid that wiped out the non-avian dinosaurs and ushered in the age of mammals. Aka a very long time indeed. But it does show that dip in temperature as dramatic.

Now the Post story, in keeping with the canons of modern journalism, did not show a graph superimposing the two even though it was supposedly the main point of the study. And nor did it contain a link to the study. But we found it and would like to quote a few key passages, in which GMST stands for “Global Mean Surface Temperature”. For instance:

“CO2 concentrations scale with the climate states defined in Fig. 3 (Fig. 4C), albeit with large ranges related to the uncertainties associated with the CO2 reconstruction, and real variability in CO2 and GMST between stages.”

So first they say they match and then they admit they don’t. How bad is it? This bad:

“there is no discernible relationship between CO2 and GMST during the Mesozoic”.

Also:

“median hothouse CO2 values (871 ppmv) are somewhat low compared to estimates from well-studied greenhouse intervals, such as the early Eocene (~1500 ppmv). This discrepancy is caused by the perplexingly low (~775 ppmv) Cretaceous hothouse CO2 estimates, despite ample evidence of extreme warmth. Omitting Mesozoic stages increases the median hothouse CO2 value to 1199 ppmv, which is in better agreement with the Cenozoic evidence (Fig. 4C).”

Perplexingly, no less. And yes, there are lots of ways to get data to do what you wanted it to by omitting the stuff that doesn’t. But since the Mesozoic stretches for over 180 million years, from 251.9 to 66.0 million years ago or so, taking it out of a study that covers 485 million years because CO2 and temperature don’t correlate and you want them to could be mistaken for cheating. And calling it an “enigma” or “discrepancy” doesn’t fix the problem.

So where’s the fabled journalistic skepticism? Or is there no discernible relationship between common sense and writing for a newspaper now?

Another enthusiastic publication, Ars Technica, which claims it “has been separating the signal from the noise for over 25 years” and boasts of being “founded in 1998 when Founder & Editor-in-Chief Ken Fisher announced his plans for starting a publication devoted to technology that would cater to what he called ‘alpha geeks’: technologists and IT professionals” hails the study as:

“A record of the Earth’s temperature covering half a billion years
With one exception, a strong link between carbon dioxide and global temperatures.”

Yeah. One exception comprising close to half of the whole period right in its middle. And it also lets a sabre-tooth, or Sphenacodon, out of the bag with:

“Climate models provide a way of converting these proxies, which typically come from a single geographic location, to a global temperature. By using details like the continental configuration and carbon dioxide levels, the models can estimate which reasonable global temperatures are consistent with the proxy data, meaning a specific temperature at a specific location on the globe. The researchers used an ensemble of climate models so that the results weren’t dependent on any particular implementation of atmospheric physics.”

Because the laws of physics can apparently differ from model to model. And extrapolating temperature from one spot on the globe to the whole planet introduces, shall we say, the hint of a shade of the spectre of some error bars.

Lay those atop similar uncertainties about CO2 reconstruction and the researchers should tread carefully, especially if you have a strong sense of how you want them to come out and are good at computer programming. As one might perhaps expect “alpha geeks” to do since one thing that does often correlate is alpha-geekery and math skills. Also logic? Because indeed, another more subtle aspect of this problem is the statement that the actual paper says the correlation is too strong:

“The consistent relationship between CO2 and GMST across the record is somewhat surprising given that on Phanerozoic timescales, we expect that non-CO2 forcings, including changes in solar luminosity and other greenhouse gases, play a role in driving climate change.”

Almost as if you told the computer to find a strong relationship and it took you too literally. Especially when you say that maybe the problem is “an incomplete knowledge of how different proxies encode past CO2 information” when they don’t correlate, but it’s all fine when they do.

What to make of it all while reporting it? Well, Ars Technica quietly concedes this one too, saying the strong correlation “is somewhat surprising, given the other changes the Earth has experienced over this time.” And indeed if your careful, ruthless modeling says the sun, continental drift, massive volcanic activities and everything else just bounce off that CO2 blanket, you should be suspicious that your model is parroting back what you told it to say not actually reconstructing the climate. Is it so hard to notice even when you’re on their team not that of readers seeking to be properly informed?

It’s worth adding that we do not know as much as we wish we did. And we do not know nearly as much as some people say we do. Including, in this Judd et al. study of the Phanerozoic, whether this supposedly robust correlation of CO2 and temperature over the long run has anything to do with what temperature might have done or not done on a much shorter time-scale. Was there no century in the Jurassic where temperature rose by 2C and CO2 just sat there being plant food? We don’t know and neither do they.

So apparently journalists, from the popular press to the geekosphere, have suspended disbelief and good reporting for propaganda. Think they’ll put that on their “about” page?