Use common sense when looking at research results
By Rosalie I. Tennison
If adding up the predicted returns from crop inputs looks too good to be true,
they probably are.
Agricultural science, like all
other branches of science, is based on a system that builds on itself. When
one discovery is made, researchers focus on perfecting and fine-tuning the information.
Fertilizer boosts yields by 10 percent, hybrids by eight percent, grass weed
control by 15 percent and so on. Soil scientist, Don Flaten of the Faculty of
Agricultural and Food Sciences at the University of Manitoba dubs traditional
scientific discovery as a series of 'single input' experiments. He cautions
that assuming the single input results are cumulative is dangerous.
"There are almost no comprehensive studies that take into account all
inputs," says Flaten. "If growers were to spend the money on the seed,
nutrition and crop protection as recommended by science, their economic return
would not come close to their expectations based on experiments where only one
input is investigated."
One plus one does not equal two
To that end, Flaten delved into the issue and came up with a plan for 'integrated
management' of cropping systems. While not intending to target any one input
unfairly or specifically, for the sake of discussion, Flaten uses nitrogen as
"Typically, we use marginal returns analyses to determine the optimum
rate of applying a single input, such as nitrogen," begins Flaten. "The
first step is to gather information on expected yield response to that input.
Most growers will calculate the economic return of nitrogen at increasing incremental
rates to achieve a maximum economic yield, which is where the spread between
revenues and costs is the greatest."
To prove his point that single input experiments can be misleading, he developed
a hypothetical canola scenario, where he added up all the yield advantages of
additional inputs and their associated costs.
In Flaten's hypothetical scenario (see Table 1), the response for adding each
input individually is based on single input research. As the costs continue
to build and the expected overall yield becomes unrealistically high, the cost
of inputs will inevitably exceed the expected return by a large margin.
|Table 1. Hypothetical yield increases and
economic returns for production of canola at $5 per bushel.
|Yield source or response||
|Check or base yield*||15||$75||$55||$20|
|Treated hybrid seed||10||$50||$25||$25|
|*Non-certified seed, machine operating
and repair, no fertilizer, no pesticide.
In real field trials conducted by Agriculture and Agri-Food Canada's Dr. Byron
Irvine at the Brandon Research Station, Flaten's hypothetical scenario proved
to be true. Irvine's actual data shows how the principle of using all recommendations
as given may not provide as much profit as expected (see Table 2).
"The misuse of yield responses to single inputs occurs when the individual
economic responses are assumed to be meaningful for a farmer's field,"
explains Flaten. "This overlooks the need for other inputs, such as crop
protection or pest control, which can add significant costs to an integrated
crop production system."
By examining Tables 1 and 2, growers can see the unrealistic results when response
data for single inputs are used for oversimplified economic analyses. What most
farmers do instinctively, suggests Flaten, is to employ caution and view inputs
interdependently, which is how it should work. The challenge to this type of
approach to achieve the necessary comprehensive data is the cost to mount such
a complex research project. Further, setting the parameters that will ultimately
pay off with useful information adds to the complexity of the research. Therefore,
science rarely provides the crop production industry with that kind of information
and growers are left to guess. Flaten suggests that, as a consequence, "The
results of single input studies cannot be applied directly to whole cropping
|Table 2. Average benefit of canola crop
inputs when yield responses to those inputs are added individually or as
a part of a complete cropping system at Brandon in 2001, 2002 and 2003 (canola
priced at $7/bu, data from Dr. Byron Irvine, AAFC Brandon Research Centre).
|Yield source or response||
|Base yield (med. genetics, no fertilizer
|Variety response (with all other inputs at high levels)**||4.9||$34||Zero||$35|
|Pesticide response (with all other inputs
at high levels)
|Fertilizer response (with all other inputs
at high levels)
|Theoretical yield if all inputs were
|Real measured yield with all inputs at
|*Costs for base yield include pre-seeding
glyphosate, seed, machinery, fuel, repairs and other basic costs.
|**The higher expense for the high yielding
canola variety was offset by planting at a lower seeding rate than for the
medium yielding variety and using less seed treatment.
|***The benefit of the lower seed treatment
costs for the high yielding canola variety is included twice in the theoretical
addition of input costs, under-estimating those total input costs, compared
Blend science with common sense
In the absence of data on fully integrated crop production systems, Flaten recommends
that researchers and marketers follow the lead of producers and adopt a more
cautious and realistic approach to determining and describing the expected value
of crop inputs in a more integrated way.
Flaten suggests that farmers, agronomists and scientists look at crop production
inputs as a series of interdependent groups. These would include: Essentials
(basic seed, fuel, taxes or costs that cannot be avoided in order to grow a
crop); Enhancements (fertilizer and superior genetics that increase yield potential);
Repairs (crop protection products that help fix a problem and potentially save
yield); and Maintenance (soil conserving equipment, fertilizers added to balance
crop removal). Using this approach to planning for maximum economic return allows
growers to build on a base of basic needs rather than 'adding on' as needed
because no allowance was made for unexpected extras.
Lastly, Flaten recommends that farmers set budgets for a complete cropping
system based on low, medium and high input production systems. These budgets
should be developed for local growing conditions and should account for all
the independent inputs required to reach the goal.
"This examination of how to determine the level of inputs needed to achieve
maximum economic yield is not a call to reduce fertilizer use or stop buying
certified seed," comments Flaten. "Merely, it is a suggestion to take
an approach to planning crop production that integrates what we have learned
from single input studies and apply it in real life cropping situations. The
benefits and costs for the whole management system should be considered as an
While most growers know this instinctively and many practice a form of integrating
all the information science provides to them, the reality is that it is a somewhat
ad hoc approach to meshing all the research into a useful production system.
Now that science has fine-tuned many discoveries and has proven that we cannot
grow crops successfully without sound crop nutrition, herbicides, high quality
seed and good use of mechanical tools, research is needed on how to put everything
that has been learned into a useful package.
The integrated crop production package that is selected will vary from farm
to farm and year-to-year; there will not be a simple recipe for all situations.
However, the process of planning, budgeting and adjusting the entire cropping
system will help to reduce the guess work in crop production and improve the
odds of maximizing overall net returns. And taking the best knowledge that science
has to offer, integrating it into a total crop production package is what farmers
know as the 'art' of farming.
Print this page