“Increasing CO2 threatens human nutrition” boldly proclaims a new paper from last week in Nature by Samuel Myers of Harvard and 19 co-authors (press release here). I don’t have time for this, but something needs to be said and if I don’t do it, I betcha nobody will.
Papers can be bad for various reasons, obviously. Logically enough, this is most often due to poor methodology. But it can also be because the science is more or less OK, as far as it goes, but the importance of the principal claim(s) of the paper does/do not really follow from the study’s actual findings. By “principal claim” I mean the one or two main points most emphasized, perhaps in the concluding paragraph, or the abstract, or even just the title itself, like in this case. This paper has both problems, but especially the latter.
Nature publishes a lot of bad science frankly, but I’m surprised more than normally at the audacity of this one. This thing shouldn’t have been published in Nature in the first place. I get the strong impression that Nature and other glamour journals are counting on people just reading the title and skimming the paper, without any real critical evaluation. Why? Because if you have any basic sense of the issues and read this paper, there’s no way you can accept that title as stated based on the study performed, even if the paper’s methodology were entirely sound, which it is not. Who exactly do they think is reading these papers, and indeed, who is reading them? Not skimming, I mean really reading closely. Well we have no idea, because critical, in-depth commentary on papers is rare indeed.
Think about the claim in the title for just a second; better yet, try to imagine that you’d never heard about carbon dioxide in relation to climate change. This is a very major claim, and a fairly novel one, so we should naturally ask questions like “How exactly, by what mechanism?”, “How much?”, “Under what kind of time frame?”, “What portion of humans are threatened?”, “What aspect of nutrition, exactly”, and any number of similar, critical questions.
I won’t even get into the plant physiology and/or genetics involved here (at least not now), but undoubtedly many issues could be raised thereon, because the effects of elevated CO2 on plant physiology and productivity is a very much unsettled topic with conflicting evidence from various types of studies, molecular to system-level. T
he lead author of the paper is not a plant biologist as far as I can tell. It is not clear to me whether or not the lead author is a plant physiologist. But the essence of the paper is plant physiology, not human nutrition, and moreover, it’s a meta-analysis combining some of their data and some from previous studies. And meta-analyses are frequently invitations to analytical problems because of the different methods, assumptions and purposes between the various studies comprising them. I’ll keep this simple, but it would still help if you looked it over–the tables and figures (in both the paper and the supplement) are freely available via the above link, as is the abstract. You can also of course request the full paper from the lead author, as I did.
Their basic meta-analysis finding is in fact very modest: that zinc, iron and/or total protein will decline by somewhere between 0 and 10 percent in the edible portion of multiple cultivars of three major cereal grain (wheat, rice, corn), two legume (soybean and garden pea), and one sugar/ethanol (sorghum) crops, in response to an increase in [CO2] to levels approximately double (i.e. 550-580 ppm) that of pre-industrial (~278-280 ppm). In most cases these declines are in the 4 to 8 percent range and in some cases there are in fact increases, not decreases.
The key communication problem is the way they expand this rather minor finding into a major, sweeping generalization, and the key methodological issue is the apparent failure to take into account one of the simplest of potentially complicating factors. Specifically, if you decrease the percentage composition of some component x (here, nutrient concentrations) in some material y (grain and seed), but you also increase the total amount of y, then even if the concentration of x has in fact decreased…well so what, you have a partially to completely compensating mechanism via the increase in total seed mass. That’s not a difficult concept and you’ve got twenty authors here–and not one of them (nor any of the three reviewers) can spot this issue? That should be one of the very first things you should think about, especially given that evidence for biomass increases with increased atmospheric [CO2] in C3 plants (e.g. wheat, rice, soy), is abundant.
I could get into at least 5 to 10 more points but will keep it to just three here: (1) How sure are they of the sources from which they get their claim that 2.3 billion people in the world suffer zinc and/or iron deficiences, putatively causing the loss of “63 million life-years annually”, (2) Even if they are right about these deficiencies, and their plant physiology work is robust and dependable, how can they assume this will in fact necessarily “threaten human nutrition”–do they think there are no other possible solutions to zinc and iron deficiencies in those claimed 2.3 billion people that are already deficient? (ever heard of supplements for example?), and (3) Their computations of how long it will take to reach ~ 560 to 580 ppm CO2 are off. We’re at about a 2 ppm/yr increase rate now (generously), there is no definite evidence for a strongly positive carbon cycle feedback, and so that rate equates to something like 80-90 years to 560-580 ppm, not their stated 40-70 years. They’re exaggerating, and then Harvard press release amplifies this by using the lower end of that range, claiming that we’ll be there by 2050. Two minutes for hooking and another two for embellishment, rules the referee.
And if you want to know the growing conditions used in the various FACE experiments involved, well they’re not summarized–you’re instead directed to five other papers scattered throughout the literature to get that information. How many people do you think are going to do that?
This paper should have been titled something like “Doubled CO2 may decrease zinc, iron and protein concentrations in several major crop species” and been submitted to Crop Science or Plant Physiology or similar journals. And, if an increase in total grain amount did in fact accompany the decline in nutrient concentrations, then the title should have reflected this reality.
My default position toward Nature (and PNAS for that matter), at least for environmental science papers (including climate science and ecology), has shifted from innocent until proven guilty, to roughly the opposite. I just don’t believe what they claim until I’ve read the paper involved closely, and since I don’t have time to do that, that means I basically don’t accept what they claim. I’ve just seen too much bad science and I don’t trust them to be fully objective and place scientific veracity over hype and headline. Sorry.
p.s. Please don’t just take my word on all this–look at the available tables and graphs at the Nature website and make your own decision, and better yet, read the paper.