Illinois
Fertilizer Conference Proceedings
January 27-29, 1992
| Home | 1992 Index | Search |
| Home | 1992 Index | Search |
Daryl D. Buchholz1
Agronomy Department, University of Missouri-Columbia
Over the past three years nearly 10,000 acres of cropland in the Mississippi delta region of southeast Missouri have been soil sampled on a 330 lineal foot grid for use in applying variable rates of fertilizer across fields. This work has been conducted with the leadership of - Mr. Bill Holmes, - a farmer near Oran, MO, and with financial and/or cooperative support through Missouri Department of Natural Resources, Soil Conservation Service, University of Missouri Extension, Delta Growers Cooperative, Southeast Cooperative, Soil Teq, Inc., and others. In this work, only phosphorus (P) and potassium (K) have been varied. Nitrogen (N) has not yet been varied within the fields. Limestone applications have not been made to date, but numerous fields require varied rates of lime. As a largely commercial venture, some estimate of economic return to fertilizing at variable rates across the field is warranted. Without actual yield data, only theoretical evaluations can be made.
Most will agree that applying varied rates of fertilizer within a field based on soil sampling is a sound practice. Just as we moved from the average rate a dealer spread for everyone in the trade area to using a soil sample to dictate a rate for a single field, now we use soil samples to dictate needed changes within a field. Those points are hard to argue. However, is it worthwhile? There is a cost for sampling, analyses, interpretation, mapping, spreading, and data management. There may also be environmental returns to variable spreading that must be considered in some field situations.
The purpose of this paper is to outline the successful procedures used in
developing the field maps, present examples of field variation, and to propose
a theoretical method to evaluate return from varied application of P and K.
Soil samples were taken using a grid sampling technique every 330 feet (2.5
acre). The boundary distances of the field were determined and transect's were
laid out beginning at a point 165 feet from two field borders. A sample was
taken at that point as a composite of 12 subsamples within a 10 foot radius
of that point. Using radar technology, an all terrain vehicle was then driven
330 feet down the field to obtain the next sample. This procedure continued
until the entire field was sampled. Samples were taken from the 0-6 inch depth
and then analyzed by routine methods of the MU Extension Soil Testing Laboratory.
Results were mapped and contours of nutrient variation established using computer
software assistance. The variations were grouped into similar soil test interpretation
areas and recommendations developed for each area using MU Extension soil test
interpretations and recommendations. Some fields were variably spread by flagging
areas and making up varying blends spread with a conventional fertilizer applicator.
In fall 1990, a truck outfitted with the Soil Teq, Inc. equipment technology
was leased by the two cooperatives to spread additional sampled acreage.
Applying fertilizer according to the average soil nutrient status for a field results in misapplication that is either in excess of crop needs or insufficient to establish maximum crop potential. But, does that misapplication result in economic loss or environmental concern? An attempt has been made to theoretically evaluate yield loss and/or excess fertilizer application on three fields in S.E. Missouri when comparing the average spread to a better system of variable spreading. The three fields are not the same fields shown in Table 1. Limitations of no more than five different spreads were evaluated since the current computer software with Soil Teq, Inc. equipment is limited to five spreads per field.
In comparing average nutrient recommendations to the better system of variable recommendations within a field, it is assumed the farmer would be using this average recommendation. However, in most situations the farmer was not using this average recommendation. Instead, the previous applications may have been vastly different. This was because the soil test data from a field composite sample taken previously was vastly different, due to field variability and sampling technique, or that the farmer did not follow soil tests from that field as a guide. Thus, one can see an implied advantage to the detailed soil sampling. Simply, your data set from which to draw fertilizer recommendations is far superior to anything used previously.
The theoretical technique to establish yield response to the variable spreading employs the use of Missouri's fertility index equation (Figure 1) as developed by Dr. T. R. Fisher (Fisher, 1974). The equation takes the form:
FI = (2Xo/Xd) - (Xo2/Xd2)
Where: FI = fertility index or relative yield
Xd = Desired level where no yield loss is expected
Xo = actual or observed level
The Xd and Xo, are used either as soil test levels or fertilizer application rates of phosphorus or potassium. The Xd level implies maximum yield with no potential for yield loss. The interpreted value used for these calculations was the P or K soil test level where Missouri recommends no additional nutrient applied for the given crop. In the examples to be given those values are 68 lbs P/a from the Bray & Kurtz No. 1 extract and 1.5(220 + 5(CEC)) lbs K/a from the 1N NH4OAC extract. The fertilizer F.I. value uses the recommended P or K from the variable recommendations as Xd and the average recommended P or K for the Xo.
Those two F. L's must then be combined to evaluate actual yield loss from applying the incorrect amount of each nutrient. This can be accomplished with the equation:
Combined F.I. = ((1 - Soil test F. I.) (Fertilizer F.I.)) + (Soil test F.I.)
A weighted averaged combined F.I. would predict yield loss from the field due to using an average fertilizer application rate.
Example Field 1 (Table 2)
On the field used in the original extension demonstration project (Table 2), the 3 expected yield from average application was 0.897 of the variable application. On this field assuming 150 bushel corn yield potential with variable spreading, one would expect 134.5 bushels per acre with average application, or a loss of 15.5 bushels per acre. At a corn price of $2.50 per bushel the gain from variable spread application would amount to $38.75 per acre.
Additionally, phosphate fertilizer was applied at 20 pounds per acre on 34.3 acres that did not require the nutrient. That cost at $0.25 per pound of P2O5 would average $2.14 per acre. The total gross return to variable spreading the phosphate on this 80 acre field would be $3721.20 or $40.89 per acre.
Example Field 2 (Table 3)
In this example, both P and K were varied within the field. Assuming 150 bushel average yield, gains from variable spread of phosphate would be 4.5 bushels per acre and from potash only 0.6 bushels per acre for a total of 5.1 bushels per acre or $12.75 per acre.
Additional unneeded fertilizer savings through variable spreading amounted to $2.76 per acre for phosphate and $0.87 per acre for potash. Total gain to variable spreading would be estimated at $1346 or an average of $16.38 per acre on the 82.2 acre field.
Example Field 3 (Table 4)
This field had relatively good soil P and K levels. Average gain to phosphate, assuming 150 bushel yield average, was 1.8 bushels per acre. Potash gains were 0.6 bushels per acre, for a total of 2.4 bushels or $6.00 per acre.
Over fertilization cost an average of $2.96 per acre for phosphate
and $0.18 for potash. Total gain would be estimated at $1587 or $9.14 per acre.
One can return to the original question, "is it worthwhile?". Considering a variable spreading of P and K on 1 year gross returns, the answer is "yes, maybe, and probably not". Estimated costs for the data collection, management, and spreading run from $10 to $15 per acre. That cost varies widely with dealers. One must remember that in the examples only P and K were varied and therefore evaluated. Potential to vary N application will also improve returns. Most fields in the acid soil regions will have varying lime requirements, also. As information begins to be gathered on a geographic basis within fields, the spinoff uses will begin to happen. Yield mapping will improve our ability to accurately predict N requirements, optimum plant population, appropriate hybrids or varieties, and maintenance P and K needs. Spot spreading of micronutrient also becomes a reality.
We may not all agree on the potential benefits of this technology on possible
surface and ground contamination. However, public perception and acceptance
of fertilizing specific field areas with tailored application rates will be
positive. This management just makes good sense.
Figure 1. Relationship of soil test level to fertility index
Table 1: Examples of field soil nutrient variations from sampling on a 330 foot grid system
Fisher, T.R. 1974. Some considerations for interpretation of soil tests for phosphorus potassium. Missouri Agricultural Experiment Station Bulletin 1007.
1Associate Professor and State Extension Agronomist - Soil Fertility, Agronomy Extension, 214 Waters Hall, Columbia, MO 65211
