"Seas will rise no more than 69 centimetres by 2100"--Wait, What??

When Science Media Coverage Fails Us

by Kai Chan
A recent NewScientist article claimed in its headline, "Seas will rise no more than 69 centimetres by 2100". Along with this rather spurious claim, the article is mute on the potentially dire implications of global mean sea-level rise of much less than 69 cm. To read the article is to immerse oneself in a bath of relief; the article starts with "It's not as bad as we thought." This kind of treatment of uncertainty (that is, suggesting there isn't any) and implications (omitting them)--especially in a highly reputable Science news outfit like NewScientist--calls for serious scrutiny of the media coverage of science.

This article covers a report from Ice2Sea that does discuss uncertainties and implications. The Ice2Sea report clearly states that their analysis is largely based on one global emissions scenario (A1B, "business as usual"), the uncertainty of which is discussed. The report includes an alternative expert-based approach to estimating uncertainty, based on which the report says, "there is only a low (1-in-20) likelihood that the ice sheets will contribute more than 84cm to sea-level rise by 2100." Critically, it it also says, "The collective opinion is that whilst high rates [of sea level rise] are unlikely, at present no absolute upper-bound can be put on the rate of sea-level rise by 2100." (italics added) The comparison with the NewScientist claim that sea level will rise by no more than 69 cm is stark, and problematic.

The 69 cm upper limit seems to be sum of the highest modelled outputs for each of the components (the Greenland Ice Sheet, West and East Antarctic Ice Sheets, etc.). What's missing from the Ice2Sea report is a clear expression of the uncertainty and how to interpret it, which has surely contributed to science journalists mistaking such a range of model outputs as the range of possible realities (a mistake that should be deeply troubling to most modellers).

Every model makes many assumptions that complicate the use of model outputs as truth, assumptions that are not represented in the range of uncertainty. In the Ice2Sea models, the level of greenhouse gas emissions is but one of many such assumptions. The Ice2Sea failure to represent these assumptions clearly and comprehensively in the interpretation of model output is not unusual. But it should be: as scientists, we need to be much better at communicating the bounds of our knowledge.

The focus of the NewScientist article on science overturning science (without discussion of the implications of the new projects for human suffering and upheaval) is also problematic. The article has no mention how much of the human-inhabited land would be uninhabitable with even 30 cm sea-level rise. it doesn't provide a basis for readers to compare the projected sea-level rises with recent historic ones (which have been much smaller). And it says nothing about the fact--well documented in the Ice2Sea report--that what  matters is not only mean sea-level rise, but also the storm-surge levels that will be exacerbated in other ways by climate change. Instead, the science-overturning-science theme risks over-inflating in readers' minds the lack of knowledge and the upheaval of understanding, which provides ample fodder for climate change deniers and skeptics. In covering science, researchers and journalists alike should remember that scientists are not the only audience. We should always ask ourselves, "How would a lay person take this? How could climate deniers, etc., use this?" With climate deniers being common in some circles including North American legislators, one can imagine this article's a "not so bad" theme serving as an excuse for further inaction on climate change.

Although I'm worked up about this NewScientist article, there's really nothing unusual about it. There are countless but critical ways that normal media coverage of science misrepresents uncertainty and distorts the societal implications of science. Research from students in my recent course reveals that when journalists report on events (like floods) that are, according to the best available science, expected to increase due to climate change by the best available science, these climate-change connections are rarely stated.

Science reporting must change. In its quest for novelty and brevity, science reporting must retain the context for readers to connect the news to its relevance and significance. There are three key ways this can be achieved. First, science that has major societal implications must never be reported without those implications. Second, events (like floods) or processes (like antibiotic resistance), should, if possible, always be connected to the likely underlying drivers (like climate change, and prophylactic use of antibiotics). Third, never say 'Never': natural science is not exact enough to support such certainty.

[This isn't the last you'll hear about these issues of modelling and uncertainty or media coverage of science. Students working with me (e.g., Edward Gregr, Natascia Tamburello, Devon Deckant) are working hard to explore these topics and identify solutions. So please watch this space, e.g., by subscribing to this blog and following @KaiChanUBC on Twitter.]

Kai Chan is an Associate Professor and Canada Research Chair at UBC. He wrote the Eco-Minded column in the Vancouver Metro for 3 years.