How overfed are we? (part 2)

Posted by William M. Connolley on January 11, 2009
(6)
More »

How overfed are we? refers, in which I express some doubt about the problems of food production. But Battisti and Naylor (Science 9 January 2009: Vol. 323. no. 5911, pp. 240 – 244; DOI: 10.1126/science.1164363) Historical Warnings of Future Food Insecurity with Unprecedented Seasonal Heat re-raise these problems, and they do it in Science, so lets have another look. You can also read it direct.

They say:

Higher growing season temperatures can have dramatic impacts on agricultural productivity, farm incomes, and food security. We used observational data and output from 23 global climate models to show a high probability (>90%) that growing season temperatures in the tropics and subtropics by the end of the 21st century will exceed the most extreme seasonal temperatures recorded from 1900 to 2006. In temperate regions, the hottest seasons on record will represent the future norm in many locations. We used historical examples to illustrate the magnitude of damage to food systems caused by extreme seasonal heat and show that these short-run events could become long-term trends without sufficient investments in adaptation.

As far as I’m concerned, the predictions of temperature increases are non-controversial, so I won’t discuss them other than to note that the pix in their article are not temperature increases, but increases in exceedance, which is why the tropics looks more affected by the mid-latitudes. Its not clear to me whether crops care more about absolute temperature change, or about change relative to current variance.

B+N don’t make any predictions at all about future food supply. They note that in the past, exceptional heat has been linked to decreases in a years crops, and they note that in the future the temperature is going to go up, but thats it. So there is no way to know where they stand related to the 25% crop reduction as an extreme scenario. Perhaps, as physical scientists, they would rather stay out of that debate.

There are two obvious problems with / counters to their analysis. First, that crops are more sensitive to precipitation than temperatures; and second, that we will adapt crop varieties. There are other issues, such as the expected advances in tech, that JA mentions in the comments to my first post; but thats not too clocely connected to their analysis. And they rather quickly skate over the expected improvements in yields in areas that are currently too cold.

Taking the second first, their historical analysis is based around individual warm years leading to problems. Since no-one predicts these warm years in advance, the farmers then couldn’t adapt their crops, much less their techniques, to an expected warmer year. Thats obviosly not true not, when we expect a fairly smooth increase in temperature. Most parts of the world, in 100 years time, will remain cooler than the warmest cropped regions now; and its not clear to me that we won’t be able to adapt, by switching crops, over that timeframe.

As to the first… B+N note the european summer 2003 as an example of high temperatures (a) killing people and (b) reducing crops. But I recall (wrongly? could be. Anyone know?) that it was also a dry year; and indeed the dryness contributed substanitally to the heat (less water to evap so less cooling). Future warming will not necessarily be associated with dryness in this way. B+N make some effort to get around this. In their analysis of historical heat waves, they note that 1972 was hot and dry and lead to poor crops in the USSR; but that previous dry summers hadn’t. The Sahel is less convincing. Famine there was originally from drought; B+N note that in recent years the rains have returned but crops haven’t fully recovered; I wouldn’t be too surprised if politics was part of that.

I think this bit of their analysis needs more work. The obvious naive question is, why didn’t they run their GCMs scenarios for precip as well as temperature? One obvious answer might be that they don’t really trust the precip from the GCMs. Another, that they aren’t feeding the GCMs into any kind of mechanistic crop model (but why not?). So they do their best to work their analysis in terms of temperature, and to claim that precip is less important. I’m not convinced.

Summary: there isn’t much new in this paper. The dependancy of crops on ppn and/or temp needs to be clarified.

[Update: ICE isn't impressed -W]

(6)
More »

Comments

  1. #1 cce
    2009/01/11

    Off topic.

    Grinsted et. al. projects 0.9 to 1.3m of SLR by 2100 for A1B.

    http://www.glaciology.net/Home/PDFs/grinstedclimdyn09sealevel200to2100ad.pdf

  2. #2 Joseph j7uy5
    2009/01/11

    A lot of the uncertainty in climate change models stems from the fact that some of the change is driven by economic activity. Future economic activity is enormously difficult to predict, as we are in a time of exceptional volatility.

    Perhaps a bigger problem arises from the fact that it is hard to know to what extent we we be able to adapt to the changes. Large scale water conservation, soil conservation, crop breeding, etc. all require capital investments. How much capital will be available for investment in the next few decades? Yes, there will be technological advances, but the technology does you no good if you can’t afford it, or if there is no fuel for the machines.

    It makes sense to try to warn people that crop production could become increasingly challenging, but I doubt that we can come up with a meaningful guess as to the magnitude of decline.

  3. #3 Magnus W
    2009/01/12

    What Joseph just said…

    And I think I said earlier, a absolute change that gives more precipitation does not necessary mean better farming conditions. (i.e. soil moisture, evaporation)

    There also is a difference between a steady increasing temperature and a warmer temperature, the first is harder to adapt to. Add to this the stress we putt out through increased population and i.e. increase in ozone (low).

    However you could always say that better tech will solve this but it just might not also… Saudi Arabia have money to de ionise ocean water… you could also draw the water from the air but as said… that costs money and how the developing world gets hit is hard to tell. And could we really model this now? How far away is it? Not that close I would guess… ?

  4. #4 ice
    2009/01/12

    haven’t read the paper yet (i should), but obviously temperature is not sufficient an information to make any kind of projection on crop production – the irony here is that the terrestrial tropics are precisely the region where precipitation projections are the most uncertain (even the sign) (as appear on the graph for instance (bottom):
    http://iceblog.over-blog.com/article-14438363.html )

    i find it frustating that Science papers about global warming impacts on agriculture are often rather simplistic, like this one, or the previous lobell et al 2008 paper, with mere regressions or qualitative discussions, while there are a whole lot of people strugling to come up with refined (but less “Science-sexy”) downscaling/modelling/upscaling (for example)quantitative approaches …

  5. #5 JCH
    2009/01/12

    Well, any analysis using annual precipitation would not be helpful. In farming it’s all about when there is or isn’t precipitation. As for adaptation, farming is hard in a stable climate. It’s just astonishingly naive to believe it will be less hard in an unstable climate. I think it’s going to be harder.

    Last week Monsanto announced they are about to market a new drought-resistant corn. If successful, the study is out the window.

    The one aspect of Ag I never see mentioned in these studies is sunlight patterns. Perhaps it has no significance. More hours of sunlight in July sounds good for corn – depending on precipitation.

  6. #6 Alexander Ac
    2009/01/13

    First sign of water stress is stomata closure (nothing new), thus creating posivive feedback to initial drought (i.e. even less transpiration). Could that be reason that 2003 heat wave was that extreme?

    It is clear that precip AND temperature are much more indicative of true state than precip. or temperature separately. My guess is that if prec. is included, results could be worse (probably because of less rain in lowlands (?))

    It seems also reasonable, that if temp increase is NOT assocciated with precip. increase, it leads to more drought –> less food (without irrigation), and this is more pronounced if decrease in precit occurs.

    Now, do we have ANY indication od future precip. trends in tropics(preferentially in lowlands?)

    [Any one of the models will tell you. But any two will tell you something different :-( -W]