Market prices in 1993 and the results of this study indicate that there are opportunities for increased profits through variable rate fertilization.

by Dan S. Long, Gregg R. Carlson', Gerald A. Nielsen2, and Gerard Lachapelle3

Northern Agricultural Research Center, Havre. MT 59501, 'MSU Department of Plant. Soil, and Environmental Sciences, Bozeman, MT 59717. and University of Calgary, Department of Geomatics Engineering. Calgary. AIl'erta.

INTRODUCTION

Most farm fields are managed as if they were homogenous units, receiving a single uniform rate of fertilizer, herbicide and other treatments. However, fields commonly have smaller parts that differ in crop requirements and productivity. Consequently, conventional uniform management may rob growers of profitability by overtreating the low yielding parts of a field and undertreating the high yielding parts (1, 3).

Site-specific management is the practice of dividing fields into smaller parts based on the crop productivity differences. Variable application is used to maximize return in each part of a field by adjusting treatments to productivity differences (2, 5). To apply the correct treatment rates, it is necessary to know the actual location of a farm implement as it moves through a field. The global positioning system (GPS) meets the positioning needs of site-specific management, because it can provide position data that are accurate to within three feet.

This article describes a study in which we investigated the profitability of variable versus uniform nitrogen fertilization and determined optimum nitrogen rates for variable fertilization.

MATERIALS AND METHODS

We selected a 102 acre production field in northern Montana. Soil nitrogen was sampled at depths of 0 to 2 feet and 2 to 4 feet. Samples were taken at 62 nodes of a 300 foot grid. Zero, 20, 40, 60, 80, and 100 lb. of nitrogen per acre were broadcast as urea fertilizer in strips 60 feet wide and about 4,000 feet long (Figure 1). The nitrogen rate design was replicated three times bringing to 18 the total number of strips. The field was then seeded to spring wheat. Phosphorous fertilizer was applied during seeding in a band six inches deep and 12 inches between rows at a uniform rate of 45 lb. P205 per acre. We harvested using a Case-IH 1660

combine equipped with a 25 foot header, an AgLeader 2000 yield sensor, ultra-high precision GPS receiver, moisture sensor and antenna. The GPS equipment allowed researchers to recognize which yield reading belonged to a specific field location. The combine cut each strip into two parallel, 25-foot-wide swaths. The crop observations of each swath in a strip were randomly allocated to one of two independent groups of data, termed either uniform fertilization or variable fertilization.Figure 1. Nitrogen rates of urea fertilizer that were applied in strips across the 102 acre field

We computed nitrogen recommendations for the field with the equation:

Recommended N (2.5 x Yield Goal) - Soil Test N

The yield goal for the whole field was 35 bushels per acre. The recommended nitrogen equaled the total crop nitrogen requirement, or 2.5 pounds of nitrogen per bushel of yield. times this yield goal, minus the soil test nitrogen. Recommendations were computed for both uniform (Figure 2A) and variable fertilization (Figure 2B) based on the soil test nitrogen in the top two feet and in the top four feet. An average value of soil test nitrogen was computed from the test results of all samples in the field to arrive at a nitrogen recommendation for uniform fertilization.

For variable fertilization, the field was first divided into smaller areas based on differences in crop tone appearing in a vertical color-infrared air photo taken three weeks before harvest (see cover illustration). Lighter areas tended to be higher drier sites. darker areas lower and wetter sites. An average value of soil test nitrogen was then computed from the test results of the available samples in each smaller area of the field to arrive at nitrogen recommendations for variable fertilization.

The very small. dark toned area near the middle of the field. not sampled for soil test nitrogen, was grouped with the small, dark toned area in the bottom of the field. This brought to five the total number of smaller field areas.Figure 2. Nitrogen recommendations for uniform fertilization (A) and variable fertilization (B) based on soil test nitrogen and nitrogen rates that matched the recommended rates (c)

We analyzed the yield observations of nitrogen rates that closely matched (Figure 2C) the recommended nitrogen rates (2A and 2B). The nitrogen rate representing uniform fertilization involved the yield observations from three continuous swaths within the entire field. The nitrogen rates representing variable fertilization involved the yield observations from two to three swaths within each smaller area in the field. Swaths between neighboring areas differing in nitrogen recommendation were not continuous in one strip, because nitrogen rates were applied at constant rates across the field.

Net returns were computed by subtracting the 1994 cost of fertilizer, soil testing and fertilizer application from the gross return. Variable fertilizer application costs included one composite soil sample for each of the five smaller field areas and labor for sampling soils and mapping soil test results. The gross return was computed by multiplying the wheat yield (bushels per acre) times the average 1983 to 1993 Portland market quotes (dollars per bushel) for dark northern spring wheat at protein levels between 13 and 15 percent (Table 1). We averaged the available samples of grain protein in each strip of the field and in each strip of the smaller areas of the field. These average protein values were used to establish the market prices of wheat yielded from uniform or variable fertilization swaths. Economic returns were analyzed using the General Linear Model procedures of SAS (4).

Table 1. Average 1983 to 1993 market quotes for dark northern sprlng wheat
Protein	Quote
%	$ per bu
13.0	4.085
13.5	4.202
14.0	4.319
14.5	4.398
15.0	4.473

RESULTS AND DISCUSSION

Yield map. The yield map of the field (see cover illustration) was derived from about 25,000 observations of grain yield using the combine equipped with a yield sensor and GPS receiver. Grain yield varied from 18 to 60 bushels per acre resulting from spatial variation in nitrogen rates, soils and topography. In general, the yield map looks like the crop patterns on the air photo. Low yield corresponds with light photo tone in high and dry areas, whereas high yield corresponds with dark photo tone in low and wet areas.

Nitrogen recommendations for uniform fertilization. Soil test nitrogen in the top two feet of the soil profile averaged 62 pounds per acre and varied from 12 to 210 pounds per acre. Based on this average soil test nitrogen and a 35 bushel per acre yield goal, the nitrogen recommendation was 26 pounds per acre, or 20 pounds per acre when rounded down to the nearest applied nitrogen rate (Figure 2A). Mean-while, soil test nitrogen in the top four feet of the soil profile averaged 117 lb. per acre and ranged from 30 to 560 pounds per acre. Based on this average soil test nitrogen and a 35 bushel per acre yield goal, the nitrogen recommendation was zero pounds per acre (Figure 2B).

Nitrogen recommendations for variable fertilization. Maps showing the nitrogen recommendations for variable fertilization are illustrated in Figure 2B. These maps have been derived from the color infrared air photo and soil test nitrogen values. Recommended nitrogen rates varied between 0 and 40 lb. per acre when based on soil test nitrogen in the top two feet of the soil, and between 0 and 20 pounds per acre when based on soil test nitrogen in the top four feet of the soil.

Profitability of uniform vs. variable fertilization. About 2,500 of the 25,000 yield observations were analyzed for each of uniform fertilization and variable fertilization. Gross return, treatment cost, and net return for uniform and variable fertilization are reported in Table 2.

Table 2. Costs and returns for uniform and variable fertilization ($/acre)
	Gross return	Cost*	Net return**
	Top 2 feet
Uniform (20 lb. N/acre)	149.57	6.69	142.88a
Variable (0 to 40 lb. N/acre)	161.90	13.04	1 48.87b
	Top 4 feet
Uniform (0 lb. N/acre	141.82	0.34	141.48a
Variable (0 to 20 lb. N/acre)	145.83	9.88	138.18b
* Nitrogen $0. 185 per pound sampling 102 acre field @ $35 per composite sample and fertilizer application charge of $2.65 per acre (uniform) or$5 per acre (variable). **Values not followed by the same letters differ signaficantly (P<0.05).

Gross return from variable fertilization was about $12 per acre more compared with uniform fertilization when nitrogen rate was based on tests from the top two feet of soil. Despite about $6 per acre greater treatment cost, variable fertilization netted about $6 per acre more than uniform fertilization. This improved profitability was because nitrogen was put where it did the most good. More nitrogen was applied to areas of inadequate soil nitrogen, which boosted grain yield and protein. and no nitrogen was applied to areas of excessive soil nitrogen.

Variable fertilization grossed about $4 per acre more than uniform fertilization when nitrogen rates were based on tests of the top four feet of soil. However, treatment costs of variable fertilization were about $9 per acre more than that of uniform fertilization resulting in variable fertilization netting about $3 per acre less than uniform fertilization.

In this field, variable fertilization was profitable when recommended nitrogen rates were based on soil test nitrogen in the top two feet of the soil profile. This suggests that maximizing returns with variable fertilization depends on nitrogen recommendations that meet the actual crop nitrogen requirements.

Profitability from optimal nitrogen rates. Figure 3A plots the net return, computed from all 25,000 observations of grain yield, versus the six nitrogen rates. An optimum net return of about $153 per acre coincides with a nitrogen rate of 40 lb. per acre. The recommended nitrogen rates of 0 and 20 lb. nitrogen per acre. based on soil test nitrogen from the top two feet or top four feet. underestimated this optimizing nitrogen rate by 20 to 40 pounds per acre.

Figure 3B plots the net returns versus the six nitrogen rates within each smaller area of the field. Optimal net returns and corresponding nitrogen rates are identified by the peaks of the curves for each smaller area. Nitrogen rates between 20 and 60 lb. per acre would have optimized profitability of variable fertilization of the field. The recommended nitrogen rates have underestimated these optimal rates by up to 40 lb. nitrogen per acre.

Table 3 reports the gross return, treatment cost and net return of uniform and variable fertilization resulting from optimal nitrogen rates. Improvement in gross return from variable fertilization was about $5 per acre compared with uniform fertilization. Evidently, this yield-derived increase in productivity resulted from increasing nitrogen fertilization in areas low in soil nitrogen. Treatment costs associated with. variable fertilization, however, offset any yield-derived increase in gross return such that there was little difference in net return between both types of fertilization. Greater profitability could be realized from variable fertilization if associated treatment costs could be reduced.

Gross Net

Nitrogen/A Return Cost* Return**

$ per acre

Table 3. costs and returns from fertilIzation msulting from economically optimum nitrogen rates
Nitrogen/A	Gross return	Cost*	Net return**
	$/acre
Uniform 40 lbs	163.60	10.39	153.21a
Variable 20, 40, or 60 lbs	168.61	16.25	152.36a
* Includes cost of Nitrogen @ $0.185 per lb, sampling 102 acre field @ $35 per composite sample, and fertilizer application charge of $2.65 per acre (uniform) or $5.00 per acre (variable). **Values not followed by the same letters differ significantly (P<0. 05).

CONCLUSIONS

Variable fertilization offers to improve nitrogen input efficiency for Montana grain producers. Gross returns based on average 1983 to 1993 market quotes indicate opportunities for increased profit exist. These results need verification in other locations having contrasting environmental conditions and crop management practices. These opportunities appear more favorable when recommended nitrogen rates meet actual needs of crop and treatment costs (soil sampling, mapping and fertilizer application) are minimized.

It may be possible to improve profits by reducing soil sampling intensity. However, such streamlining may not accurately predict crop nutrient requirements. In this study variable fertilization was based on dividing the field into smaller areas using an air photo of the same year's crop as the test yield data. Therefore, the question remains whether this method would work when a prior year 5 photograph is used to divide the field into treatment areas. Mapping costs would be reduced if a prior photo could be used.

We believe that conventional equipment can be used for applying variable rates. Since application rates on truck-mounted fertilizer applicators can be manually changed, and GPS receivers can be programmed to guide vehicles in the field, it should be possible to use conventional equipment for applying variable rates. This would substantially reduce the cost of variable nitrogen application.

Obstacles to implementing this variable nitrogen application include the need for producers to acquire and learn how to use the new technology. However, a few Montana producers already have purchased GPS equipment and are investigating the potential usefulness of variable rate fertilization.

ACKNOWLEDGMENTS

This research was funded by the Montana Fertilizer Tax Committee and the Montana Agricultural Experiment Station. The production field was provided by Willard and Karla Vaughn. Mark Peterson helped set up the field, provided the combine and harvested the crop. Technical assistance and GPS data processing were provided by Doug Roberts of University of Calgary.

REFERENCES

1. Carr, P.M., G.R. Carison, J.S: Jacobsen, G.A. Nielsen, and E.O. Scogley. 1991.Far:ning soils. not fields: A strategy for increasing fertilizer profitability. Journal of Production Agriculture. 4:57-61.

2. Fiez, T.E., B.C. Miller, and W.L. Pan. 1994. Assessment of spatially variable nitrogen fertilizer management in winter wheat. Journal of Production Agriculture. 7:86-93.

3. Larson, WE. and P.C. Robert. 1991. Farming by soil. p. 103-112. In R. Lal and F.J. Pierce (ed.) Soil Management for Sustainability. Soil and Water Conservation Society of America, Ankeny, IA.

4. SAS Institute Inc. 1988. SAS procedures guide, Release 6.08 edition. Cary, NC.

5. Wibawa, W.D., D.L. Dludlu. L.J. Swenson, D.G. Hopkins, and W.C. Dahnke. 1993. Variable fertilizer application based on yield goal and soil map unit. Journal of Production Agriculture. 6:255-261.