Please note that this blog no longer accepts comments (there was
too much spam coming in!). If you're reading this blog and want to
respond then please use the contact form on the site.
You can also follow me on twitter.
My latest Mind and Matter column in the Wall
Street Journal is on peak farmland, a more plausible prediction
than peak oil.
It's a brave scientist who dares to announce the turning point
of a trend, the top of a graph. A paper published this week does just that,
persuasively arguing that a centurieslong trend is about to
reverse: the use of land for farming. The authors write: "We are
confident that we stand on the peak of cropland use, gazing at a
wide expanse of land that will be spared for Nature."
Jesse Ausubel and Iddo Wernick of Rockefeller University, and
Paul Waggoner of the Connecticut Agricultural Experiment Station,
have reached this conclusion by documenting the gradual
"dematerialization" of agriculture. Globally, the production of a
given quantity of crop requires 65% less land than it did in 1961,
thanks to fertilizers, tractors, pesticides, better varieties and
other factors. Even corrected for different kinds of crops, the
acreage required is falling at 2% a year.
In the U.S., the total corn yield and the total corn acreage
tracked each other in lock step between 1870 and 1940-there was no
change in average yield per acre. But between 1940 and 2010, corn
production almost quintupled, while the acreage devoted to growing
corn fell slightly. Similar divergences appeared later in other
countries. Indian wheat production increased fivefold after 1970,
while wheat acreage crept up by less than 1.5 times. Chinese corn
production rose sevenfold over the same period while corn acreage
Yet the amount of farmland in the world was still rising until
recently. The reason is that increased farm productivity has been
matched by rising demand for food, driven by population growth and
swelling affluence. But the effects of these trends are waning.
Global population growth has slowed markedly in recent years-the
rate of change halving since 1970 to about 1% a year today. Growing
affluence leads people to eat more calories, and especially more
meat. Since it takes two to 10 calories of maize or wheat to
produce a calorie of meat, depending on the animal, carnivory
demands more cropland. But as a country gets richer, total calorie
intake soon levels off, even as wealth continues to rise, and the
change in meat consumption decelerates. Chinese meat consumption is
now rising less than half as fast as Chinese affluence; Indians
have grown richer without taking to meat much at all.
What the Rockefeller team did was plug some highly conservative
assumptions about the future into a model and see how much land
would be required for growing crops in 2060. Compared with current
trends, they assumed population growth will fall more slowly, that
affluence will increase faster and that the gluttony of people will
rise more rapidly. Conversely, they assumed that farm yields would
rise more slowly than they have been doing. This seems highly
implausible given that the gigantic continent of Africa seems to be
at last embarking on a yield-boosting green revolution as
far-reaching as Asia's was.
Even with these cautious assumptions, the researchers find that
over the next 50 years people are likely to release from farming a
land area "1½ times the size of Egypt, 2½ times the size of France,
or 10 Iowas, and possibly multiples of this amount."
Indeed, the authors find that this retreat from the land would
have already begun but for one factor so lunatic that they cannot
imagine it will not be reversed soon: biofuels. If the world had
not decided to subsidize the growing of energy crops on 3.4% of
arable land, then absolute declines in the acreage of arable land
"would have begun during the last decade." The prospect of "the
restoration of vast acreages of Nature" is enticing for nature
Predictions of peak oil have repeatedly proved wrong. But the
factors that made them wrong-productivity and technology-are
essentially the ones that make a prediction of peak farmland likely
to be right.