[2:10 p.m. | Updated At the bottom of the post, I’ve appended fresh input from one of the paper’s authors, along with excerpts from news releases and other related material.]
I posted a quick riff overnight on new research, published in Nature, that links human-driven global warming and rising instances of extreme precipitation in observed parts of the Northern Hemisphere over the last half of the 20th century. (There’s a second related paper, focused on heavy flooding in England in 2000, but I’m confining this discussion to the broader analysis.)
The research is important because, as Gavin Schmidt noted this morning on Realclimate, it digs in on one of the most daunting tasks in climate science — attributing a specific change in climate conditions to the long-term global heat-trapping influence of accumulating greenhouse gases. It adds to vast volumes of work devoted to this question.
But is it big news?
From the burst of coverage (including in the news pages of The Times), sure. And that’s no surprise given the core conclusion, with nary a caveat, in the opening summary of the paper:
Here we show that human-induced increases in greenhouse gases have contributed to the observed intensification of heavy precipitation events found over approximately two-thirds of data-covered parts of Northern Hemisphere land areas.
In scientific literature you rarely see statements so streamlined and definitive. For climate science, this is the equivalent of a smoking gun. News indeed. Add in the extreme floods last year (a period not included in the study) and you have more relevance, although Roger Pielke Jr. this morning notes the importance of distinguishing between analysis of certain kinds of extreme precipitation events and disastrous flooding.
The problem is that the Nature paper is not definitive at all, as you’ll see below.
None of this detracts from the importance of this work, or the overall picture of an increasingly human-influenced climate, with impacts on the frequency of gullywashers.
But this does raise big questions about the standards scientists and journals use in summarizing complex work and the justifiable need for journalists — and readers — to explore such work as if it has a “handle with care” sign attached.
This is not about “false balance.” This is about responsible reporting.
A previous instance occurred in 2006, when a paper in Science on frog die-offs in Costa Rica included this firm and sobering statement:
Here we show that a recent mass extinction associated with pathogen outbreaks is tied to global warming.
Things were far more complicated, of course, as you can read in my 2008 piece on Vanishing Frogs, Climate and the Front Page.
In the policy arena, the eagerness to trim away caveats is even more pronounced, as was the case when climate treaty negotiators in Cancún erroneously oversimplified the core finding of the 2007 report from the Intergovernmental Panel on Climate Change. In that instance, the error was fixed.
So what’s the issue with the new study of bad storms and warming?
It opens with an extraordinary summation, which is echoed in the news release disseminated by the journal:
Here we show that human-induced increases in greenhouse gases have contributed to the observed intensification of heavy precipitation events found over approximately two-thirds of data-covered parts of Northern Hemisphere land areas.
Just in case there’s any doubt in the scientific community, this is the news element in this otherwise creditable, but unremarkable[*], work.
At the tail end of the full paper, capping a paragraph about a weak spot in the analysis — that the observed trend in extreme precipitation events exceeds what is produced by various climate models — comes a sentence about uncertainties:
There are, however, uncertainties related to observational limitations, missing or uncertain external forcing and model performance.
There’s no indication that those caveats (which included links to 9 cited papers) apply to the grand conclusion. Did the authors stress the uncertainties in discussions with journalists? It sure doesn’t look that way. Should the journalists have pushed harder when confronted with definitive language? To my mind, yes.
As I wrote recently, there seems to be an inverse relationship between the definitiveness of an assertion and its credibility. This doesn’t mean that everything definitive is wrong (only Joe Romm could find a way to interpret it thus). It means that a reporter, or citizen, confronted with a flat statement on a tough issue would do well to dig a bit deeper.
In a note I sent to several climate researchers this morning, I put my feelings this way:
To me, as a reporter, the authors and journal are trying to have things both ways — including definitive statements in abstracts and summaries that draw the attention of the press and public, but then saying, no, this is not definitive… please note the uncertainties in the final line (even though that line isn’t actually necessarily linked, rhetorically, to the definitive statements).
This reminds me, sadly, of issues I’ve raised about the way the I.P.C.C. summaries for policy makers included definitive “headline” statements with the authors pointing, after the fact, to caveats deeper in the text.
Is there simply a different norm for the science literature that allows an abstract written like this one to be portrayed as accurately summarizing the results?
In journalism, you can’t have it both ways and keep your credibility. If a newspaper “lede” doesn’t reflect what’s below, it can be (appropriately) hammered for overstatement.
1:30 p.m. | Updated I sent a query to the paper authors this morning and just got a reply from Gabriele C. Hegerl of the University of Edinburgh, someone whose expertise I’ve drawn on many times over the years. Here’s my query, her answer, and then some excerpts from news releases, the paper and other related material (an ongoing exchange has developed with the other authors and with Gavin Schmidt of Realclimate and I’ll append this as that evolves):
“My main concern,” I wrote to the authors, “is that there is no clear rhetorical link between the line about uncertainties at the end (in a paragraph on a weak spot in the work – the models’ failure to attain the intensity seen in observed precip.) and the flat statement of a conclusive link in the abstract, opening summary and news release.
Here’s Hegerl’s helpful reply:
The paper is restricted in length, and that is particularly the case for the abstract (there are tough length guidelines). Therefore, it is not always possible to discuss findings in the detail that would be desirable IN THE ABSTRACT. As Myles [Myles Allen, an author of the other precipitation/climate paper] pointed out, ‘detection and attribution’ is code for a hypothesis test, meaning that we are checking this hypothesis at some confidence level. Every specialist in the area will understand this, and every scientist who looks at abstract AND figures will as well (common practice), and the remaining text is very explicit about methods and the accounted for uncertainties.
Figure 3 shows the results with 5-95% confidence ranges, showing that there is uncertainty in the findings.
The data uncertainties affect all results including the enhancement in observations relative to models.
So, of course there are uncertainties in the findings, as in any attribution and detection result, there is a remaining chance that the observed change is due to internal climate variability (5-ish %) particularly if the models would underestimate that variability. Natural forcings are expected to cause weak changes over that period.
The Nature paper is written for other scientists. The press release, which I am desperately trying to find, and Francis’ presentation make the results clear for the media. So there is of course no definitiveness in this result.
Ah here is the press release which is full of suggests and ‘may’s. Scientists try to communicate to other scientists in a different way from how we communicate to the media. It is important that the uncertainties are clearly communicated to the respective audience, and I think this is the case here.
[I agree that this release, unlike the one from Nature, does a good job.]
So: The study finds a fingerprint of anthropogenic influences on large scale increase in precipitation extremes, with remaining uncertainties – namely that there is still a possibility that the widespread increase in heavy precipitation could be due to an unusual event of natural variability.The intensification of extreme rainfall is expected with warming, and there is a clear physical mechanism for it, but it is never possible to completely separate a signal of external forcing from climate variability – the separation will always be statistical in nature. Further to that uncertainty, there are data and model uncertainties, which we believe to be important but not large enough to overthrow our conclusions. The main finding is that on large scales, precipitation changes seem to follow the intensification expected from greenhouse warming.
Gavin Schmidt followed up in a related e-mail this way:
“Here we show” statements are required by Nature and Science to clearly lay out the point of the paper. If you don’t include it, they will write it in. The caveats/uncertainties/issues all come later. I think the confusion is more cultural than anything. No one at Nature or Science or any of the authors in any subject think that uncertainties are zero, but they require a clear statement of the point of the paper within their house style.
As for headlines in real newspapers, how many times have we been told that they aren’t written by the journalist on the story and so no-one (in practice) can be held responsible for their inaccuracy? I strongly disagree with the notion that there is much accountability there!
I agree that there is inadequate accountability in the news business for oversimplifications that overplay the “front-page thought,” and have said so many times. One nice aspect of blogging is that I get to write the headlines. Keep in mind that my comments here are not about headlines, but about the cornerstone statements in a scientific paper, which are akin to the opening paragraphs of a news story. (Schmidt also took issue with my characterization above of the new work as “unremarkable,” and he’s right. It really is thorough, innovative and important.[*])
Here’s more background, starting with the opening summary section of the paper itself, which contains definitive language:
Extremes of weather and climate can have devastating effects on human society and the environment (1,2). Understanding past changes in the characteristics of such events, including recent increases in the intensity of heavy precipitation events over a large part of the Northern Hemisphere land area (3–5), is critical for reliable projections of future changes. Given that atmospheric water-holding capacity is expected to increase roughly exponentially with temperature—and that atmospheric water content is increasing in accord with this theoretical expectation (6–11) — it has been suggested that human influenced global warming may be partly responsible for increases in heavy precipitation (3,5,7). Because of the limited availability of daily observations, however, most previous studies have examined only the potential detectability of changes in extreme precipitation through model–model comparisons (12–15).
Here we show that human-induced increases in greenhouse gases have contributed to the observed intensification of heavy precipitation events found over approximately two-thirds of data-covered parts of Northern Hemisphere land areas. These results are based on a comparison of observed and multi-model simulated changes in extreme precipitation over the latter half of the twentieth century analyzed with an optimal fingerprinting technique. Changes in extreme precipitation projected by models, and thus the impacts of future changes in extreme precipitation, may be underestimated because models seem to underestimate the observed increase in heavy precipitation with warming (16).
The press release from Nature on this and the related paper on British flooding in 2000 provided no hint of the uncertainties:
Anthropogenic greenhouse gases have significantly increased the probability of heavy precipitation and local flood risk, report two papers in Nature this week. The findings are among the first formal identifications of human contribution to extreme hydrological events. It has previously been suggested that human-influenced global warming may be partly responsible for increases in heavy precipitation. However, because of the limited availability of daily observations, most studies to date have only examined the potential detectability of changes in precipitation through model-model comparisons. Francis Zwiers and colleagues studied rainfall from 1951 to 1999 in Northern Hemisphere land areas, including North America and Eurasia (including India). They show that human-induced increases in greenhouse gases have contributed to the observed intensification of heavy precipitation events found in approximately two-thirds of data-covered parts of Northern Hemisphere land areas.
The related “News and Views” commentary by Richard P. Allan of the University of Reading expressed the findings well, saying the authors “provide evidence that human-induced increases in greenhouse-gas concentrations led to the intensification of heavy precipitation events over large swathes of land in the Northern Hemisphere during the latter half of the twentieth century.”
The news release from the University of Victoria in British Columbia, home to one of the authors (a release I wasn’t aware of until just now) actually underplayed the findings a bit:
A new study co-authored by Francis Zwiers, the director of UVic’s Pacific Climate Impacts Consortium, suggests that human-induced global warming may be responsible for the increases in heavy precipitation that have been observed over much of the Northern Hemisphere including North America and Eurasia over the past several decades.
Here’s one more excerpt from the e-mail exchange I had with the authors, starting with me:
[It’s clear that] there is a deep cultural disconnect between the science community and the stakeholders and public, for whom norms require internal consistency – whether in a contract or a responsible news article (I’m a critic of that arena, too, for doing the same thing).
Something is definitive, or it isn’t.
Every reporter is looking for a way to justify putting this in print (or online or on a video screen).
Look at the resulting coverage and tell me how many media or other outlets accurately conveyed what’s been revealed by this important work?
And then Hegerl:
Communicating findings with uncertainty is really difficult, and we find ourselves often at a point where we try to explain something including the uncertainties only to find that in transition to the reporting in media, the caveats got lost. There are some wonderfully funny cartoons about that (one appended). It is very difficult to explain science in a generally understandable way and in a way that includes the uncertainties. It can be done but it isn’t easy, and it is very difficult to find the exact right line to tread how far something can be simplified without loosing important information, and still sounding comprehensible.
This demonstrates why Hegerl remains a vital source for me. Thanks, Gabi.