A Look into Yesteryear on the Origin of the Phrase
"Site-Specific Farming"
Editor's Comment
Years before the
first commercial yield monitor was available, a group of Purdue scientists were
discussing and developing an alternative form of farm management.
"Site-specific" farming was presented by Howard Doster of Purdue
Agricultural Economics at the "Purdue On-Farm Computer Conference -
November 21, 1983". This often debated phrase, was being used at Purdue
three years before the famed AgChem VRT Ortlip patent, which was issued in
1986.
The following paper
provides an interesting look at site-specific farming plans seen through the
eyes of Purdue researchers in the early 1980's. It is useful to compare current
and past theories surrounding this once novel idea and its reality today.
Greg Blumhoff, SSMC Information Systems
Manager
Computer Assisted Big Ten Crop
Farming
Howard Doster,
Robert Nielson, Sam Parsons, and Tom Bauman
Purdue On-Farm
Computer Conference, 1983
Introduction
How would you like to increase your net
income, say $10.00 per acre? Suppose
this is possible if only you do a better job of selecting and applying your
fertilizer, seed, chemicals and tillage according to your soil type. To do this better job, suppose you must:
First,
determine your soil types.
Second,
determine your in-field location "on the go" as you travel across
your fields.
Third,
be able to locate each soil type as you travel across your fields.
Fourth,
be able to switch rates and perhaps even varieties "on the go" as you
apply fertilizer, seed, or chemicals and switch tillage methods as you move
onto a different soil type.
Fifth,
be able to measure crop yield and moisture content "on the go" so as
to identify your actual yield by soil type.
Sixth,
collect weather information at several locations on your farm.
Seventh,
identify and record disease, week, tillage or other stress agent by site found.
Eighth,
submit your input-output data to a central data bank where your results will be
compared to the results other farmers obtained on the same soil types with
different weather.
Ninth,
receive statistically significant input-output relationships from a central
data bank describing what happens on each of your soil types when (a) farmers
use the same practices as you but experience a range of weather situations and
(b) farmers use various other production practices and experience a range of
weather situations.
Tenth,
select and carry out the best combination of practices for your farm.
We are gathered here today at a computer
conference. I assume we all feel a
computer might somehow improve our information and skills. This is the third year your committee has
developed a program for crops and machinery.
We have done considerable dreaming.
We have also searched the country to see what others are doing.
I think it is fair to say that many
persons went away from our first session two years ago dissatisfied. We did not present them with any computer
programs that would immediately make them healthy, wealthy or wise. In developing that first session, your
committee did conclude that a computer could be used to store, retrieve and
interpret crop information. Your
committee also concluded this information should be site specific. Because crops responded differently in
different soil types, records should be kept, plans should be made, and
performance should be carried out for each soil type. A method for identifying and storing information by 1.3 acre
cells or sites was presented. The cell
size and location was intentionally set to coincide with the current USDA
"about to be computerized" soil map files for Indiana counties.
In 1982, your committee set out to
determine how farmers and others were using computers for crop and machinery
information. Apparently, not many people
were using their computers for this purpose.
We did find persons who were using their computers to file crop input
and output information by field and by landlord. We found persons who were recording machinery repairs by
machine. However, no one was very
excited about what he or she was doing.
The records were no different than "by-hand" systems. While they did provide a document that the
events occurred, the information was not being used for any exciting purposes.
In spite of the limited immediate
usefulness of the material presented, the session last year was perhaps the
most exciting session I have ever moderated!
Why? First, the speakers and the
audience came to realize the limited value for decision making of one person's
performance records. So what if a
person keeps account of his inputs and outputs on a field? The field contains different soil types, the
weather and other stress will not be repeated, and new inputs will appear on
the market. So what if a person keeps
account of tractor breakdowns? Of what
use can he make of the information? He
has only one or a few tractors. Since
he didn't know it was likely to need an engine overhaul at around 2500 hours,
it didn't occur to him to have the job done in the off season just ahead of this
expected occurrence.
Why was last year such an exciting
session for me? I helped introduce many
persons to probability theory and statistical inference as procedures for
drawing conclusions about what happened and what will happen in the
future! Given a large number of
observations, the probability of specific responses can be estimated. Farmers can use information from similar
farming situations to make better decisions for their own farm.
We even had a spontaneous joke. Someone asked our seed corn company
representative speaker whether the yield of a hybrid with a test plot yield of
194 bushel/acre was statistically different from another company's hybrid which
yielded 192 bushels. Our speaker's
response -- If my hybrid yielded 194 bushels, it's different. If my hybrid yielded 192 bushels, there's no
statistically significant difference between them!
Given the apparent randomness of
weather events and the continued introduction of new input alternatives, your
committee feels that farmers would improve their management skills tremendously
if they could more easily measure their own input-output performance, compare
it with the performances of other farmers with the same soil types, and, by
statistical inference, select the best corn or other crop growing
processes. Computerized measurement and
machinery control, information storage, retrieval, and interpretation may make
this improvement in management skills possible. Let me explain how this might be done.
TYPES
OF INFORMATION AND SKILLS NEEDED
What Makes Corn Grow?
A farmer must
know what is "supposed to happen" for each recipe before he can
select the recipe he wants to happen.
Further, after he is carrying out a recipe, he must know what was
supposed to happen before he can decide if it is really happening or if the
crop is "sick".
Which of the
many possible recipes for growing corn is most profitable for each site on his
farm? This question will never be
completely answered. Yet, each of us
will continue to try to get a better answer.
How will we do that? We will
collect information that is more precise as we attempt to determine what is
happening on each site each year.
To really do
this, we must:
a. inventory the
soil characteristics of each unit (say, acre) of our farm. These characteristics include: fertility,
water holding capacity, pH, trafficability, etc.
b. measure and
record weather; i.e., identify the temperature, rainfall, evaporation,
etc. Rainfall in particular should
probably be measured separately at many sites on each farm.
c. measure and
record the date and amount of each input and each crop yield and crop moisture
content for each acre.
d. identify and
estimate yield effect of each disease, weather stress, or stress caused by some
poorly carried out production practice.
No single
farmer, or no single research for that matter, can expect to experience all the
likely weather or disease stresses on his own sites and no one person will grow
more than a few of the possible recipes for growing corn. Farmers, university researchers and farm
input suppliers now share their experiences.
This experience sharing business would be much more helpful if the
information handling procedures could become more standardized. Your committee members are currently
considering how to do that. We have a
felt need for standardized information collection storage, retrieval, exchange
and interpretation procedures.
With better
information handling procedures, we expect to get more useful information. We then expect to more precisely predict
what is supposed to happen when a particular recipe is grown. We also expect to use this information to
signal when a plant is "sick", i.e., not performing as expected.
How Do You Grow Corn?
Further, when
performance is different from expectations, he needs to be able to identify and
separate weather effects, disease effects, and the effects of his own unusual
performance.
Being able to
get the job done is certainly a necessary corn growing skill. Once a corn growing recipe is selected for
each site, a farmer must carry out the various tasks of tilling, fertilizing,
planting, observing for and controlling pests, and harvesting the crop.
First, a
farmer must identify each site on his farm and find each site as he performs
tasks on adjacent "sites" in the same corn row. When he can do this, he can consider a new
level of performance for his corn growing tasks.
We know that
the best recipe for one soil type is often quite different from the best recipe
for other types of soil located in the same field. Tillage, fertility, planting rate, perhaps even variety,
herbicide type and/or rate, could be different. Often, it is costly to make "fields" out of each soil
type. Turn rows yield less and turning
takes time. Therefore, methods for
controlling application rates, adjusting rates "on the go", and
recording rates applied offer the potential for greatly enhancing a farmer's
corn growing skills.
Equipment for
measuring and recording crop yields and moisture content "on the go"
by site is a part of this needed technology.
Other equipment for improving a farmer's corn growing skills might
include process changing devices such as on/off switches, and devices for
changing planting depth, press wheel tension, header height, ground speed,
cylinder or fan speed, concave setting, etc.
How Do You Grow Corn Profitably?
Presumably a
farmer wants to select a recipe for corn or other crops for each site that will
make him the greatest net return to his total set of resources. Since he won't be able to plant or harvest
all his sites on the same date, he will want to consider recipes for some sites
which are based on "less timely" operations.
As stated
earlier, no one farmer can expect to get all the corn or other crop growing
information he needs if he relies only on his own experience. However, if several farmers and researchers
were to use a well developed and standardized information collection, storage,
retrieval, and interpretation system, a farmer could more quickly discover likely
recipes by using statistical inference procedures to compare the results from
the recipes used by the other farmers with similar soil but different weather
and other stress.
With a large
number of persons supplying similar information a farmer could (a) learn the
likely consequences of growing corn by his recipe for several different weather
situations and (b) learn the best recipe used this year for several different
weather situations for each of his soil types.
Summary
Earlier, we asked if you would like to
increase your net income by $10 per acre.
We think that's reasonable estimate of the benefits you could realize if
you had the crop management and control system just described. In various ways, computers will be used to
perform many of the jobs in this system.
Please help us improve the ideas and implement the practices.
Actually, you and your neighbors are
already using many parts of this system.
We have asked some of them to tell us what they are doing. We will spend the remainder of this session
visiting with them.
Editor's Comment
Why is this material
significant? The plan outlined in the presentation is as sound today as it was
18 years ago (pretty unusual, wouldn't
you say?).
We
will end you with an insightful quote:
"
What has been is what will be, and what has been done is what will be done;
There is nothing new under the
sun." -- Ecclesiastes 1:9.
-- Greg Blumhoff,
SSMC Information Systems Manager