I once got curious about the relationship between yield and stability
of yield. In particular, it seemed reasonable to me that yield
should stabilize (become less variable among years on a given farm,
and among farms in a region) as yield increased, particularly when
so much of the yield increase in the last 40 years is quite directly
attributable to (about a 50:50 split) between a) breeding
mgt-responsive cultivars and hybrids, and b) applying the mgt, in the
form of N, biocides, tile-drainage, drying costs, irrigation (in some
areas), etc. I reasoned that as one removed the influence of
deficient N or soggy soils, then one should expect to see yield
become more predictable and reliable - as well as higher.
Imagine my surprise then, when I came across a fascinating text
edited by Anderson, J.R. and P.B.R. Hazell (1989) entitled
Variability in Grain Yields (Johns Hopkins Univ Press). Each chapter
covers a region or a specific crop*region combination, and they ask
very much the same question as myself, as well as many others. For
example, how much of the variability in yield in caused by weather
vs. other factors.
The upshot of it all was that as yield increased, the CV for yield
increased - significantly so for most US crops - and this pattern
was surprisingly consistent around the world. CV is the coefficient
of variability for yield, and is a measure of the variation around
the mean, or in practical terms, the inverse of reliability or
predictability of the yield. So, as yield increased, variability
increased - yield became less stable, not, as I had originally
expected, more stable. Of all the analyzed crops, only in wheat was
the increased variability found to have originated, at least in part,
due to the weather (remember, this was published in 1989, so
reflected weather trends prior to that date).
Just to satisfy my own curiousity, I analyzed yield trends for corn
and for mixed grains for Ontario for the interval between 1948-52,
1968-72, and 1988-92 (I did this in 5 year increments; published in
Clark et al., 1996. Can J. Plant Sci 76:603-610). Sure enough, corn
yield doubled over the whole 40 years, and the CV for corn yield also
doubled over the same interval, with all of the increase in
variability occurring in the last 20 years - since the advent of
widespread conservation tillage. It was noteworthy, to me, that
variability for yield DID NOT increase when yield increased between
1948-52 and 1968-72 - if anything, it went slightly down.
The trend for mixed grain was quite different, with yield increasing
only up until 68-72 and then flat, while CV for yield was flat
throughout the 40 year interval - no response. Yield levels for
mixed grains are, of course, about half that for corn, and reflect a
much lesser input of management of all kinds.
How to interpret this? Well, at least one interpretation is that my
original surmise had been correct up to a point, after which
continuing to push yield higher and higher (in corn) started to
destablize yield. Further, conservation could conceivably had
affected variability, although I've not seen this well documented,
and it wouldn't explain the generality of this trend across all US
crops.
Another interpretation, given in Anderson and Hazell, is what is
called human-induced yield variation. Fascinating idea. With
contributions from none other than Don Duvick of Pioneer, it was
noted that corn producers are increasingly similar in both
the corn genetics they are using and the management they are
applying. Forty years ago, a much wider range of corn (open
pollinated; diverse hybrids) might have occupied a different place in
the rotation on each farm in the county, on a different manure
history, on land with or without tiles, planted when it suited the
particular operator, etc. Now, everyone does everything the same,
and at the same time, using an astonishly small number of very
widely adapted corn hybrids. And, they all respond to the same
market signals - price of N; future demand for corn; grain reserves;
drought in China, etc.
So, in a good year - everyone has a good crop and yields were
consistently high across a wide geographic region (I'm talking in
generalities, here, recognizing there are always pockets in any
region). In a bad year - whether due to weather or to high N or low
expected returns or whatever - everyone has a lower yield. And of
course, because the yield potential is now so high, the DIFFERENCE
in yield among good and bad years is also greater in absolute as
well as relative terms. Thus, the reason for increased variability
among years as yield potential has increased was directly caused by
a) the very high, input-based yields attainable in the good years, b)
the drop in yields without inputs/or due to bad weather, and c) the
homogeneity among producers in genetics and management. This is what
is called human-induced variation in yield.
So, in answer to the original question, I would wonder if risk might
actually be reduced - not as inferred, increased - by lowering inputs
to production? Although it is perhaps unnecessary to add this
proviso to this particular virtual group, I would note that "reduced
input" agriculture doesn't mean contemporary agriculture without the
inputs. It means a whole different way of growing crops and
livestock, with lesser reliance on exogenous energy sources and much
greater reliance on holistic farm management to capture synergies
among enterprises etc.. I discuss some of this in a recent talk at
UBC, now posted on my homepage. Ann
ACLARK@plant.uoguelph.ca
Dr. E. Ann Clark
Associate Professor
Crop Science
University of Guelph
Guelph, ON N1G 2W1
Phone: 519-824-4120 Ext. 2508
FAX: 519 763-8933
http://www.oac.uoguelph.ca/www/CRSC/faculty/eac.htm
To Unsubscribe: Email majordomo@ces.ncsu.edu with "unsubscribe sanet-mg".
To Subscribe to Digest: Email majordomo@ces.ncsu.edu with the command
"subscribe sanet-mg-digest".