Published online February 16, 2018

Deriving and Using an

Equation to Calculate 11
Economic Optimum Fertilizer
and Seeding Rates
Scott Fausti, Bruce J. Erickson, David E.
Clay, and C. Gregg Carlson*
Chapter Purpose Key Terms
Producing a profit requires that returns are greater than the costs. However, Economic optimum rate, site-
optimization for individual decisions requires an analysis where the return for specific seeding rates, 2nd
a specific investment is compared with the input costs for that investment. order polynomial equation,
Seeding and fertilizer rates are treated differently than many other costs, be- derivative.
cause these rates can be varied. Rates for many pesticides cannot be varied, Mathematical Skills
because they must follow the labeled rates. This chapter derives an equation
that can be used to calculate economic optimum seeding and fertilizer rates. Deriving the economic opti-
This solution is based on the use of a second order polynomial equation. If a mum rate equation for a 2nd
different model is used, this solution will not be appropriate. Even though the order polynomial equation, de-
derivation of this equation requires calculus, the use of the equation does not termining economic optimum
require calculus. Examples are provided. rates for an input with dimin-
ishing returns such as nitrogen
Calculating Site-Specific Economic fertilizer rate or seeding rate,
and developing a site-specific
Optimum Rates seeding algorithm.
The economic optimum rate (EOR) is the point where an incremental change in
input costs equals an incremental change in the value of product produced (Bullock
and Bullock, 1994; Cerrato and Blackmer, 1990; Kyveryga et al., 2007; Fig. 11.1). Tradi-
tionally, this condition defines profit maximization where marginal revenue equals
marginal cost of a production activity. Mathematically, EOR is the ratio where the
change in revenue (from production) equals the change in input cost. Or where
marginal revenue equals marginal cost, where D denotes incremental change:
∆ value of product [11.1]
1=
∆ input cost

In agriculture, the input or total costs (TC) are defined as the unit input cost
multiplied by total input units per unit area [TC = price per unit purchased ×
number of units per acre], and the value of the product, or total revenue (TR)

S. Fausti, California State Monterey Bay, Seaside, CA 93933; B.J. Erickson, Purdue University, Agronomy Department, West Lafayette, IN
49707; D. Clay and G. Carlson, South Dakota State University, Brookings, SD 57007. *Corresponding author ([email protected])
doi:10.2134/practicalmath2016.0027
Copyright © 2017. American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America, 5585 Guilford
Rd. Madison, WI 53711, USA. Practical Mathematics for Precision Farming. David E. Clay, Sharon A. Clay, and Stephanie A. Bruggeman (eds.).

181
 per unit area. Per unit area is defined in dollars per unit of product or the
price per unit of output multiplied by total units produced per unit area
[TR = price per unit sold × number of units per acre]. In discussion below, it
is assumed that the crop is corn, the inputs are nitrogen (N) and corn seed
(S), and the unit area is one acre. In this solution, the input costs for multiple
inputs are not considered simultaneously, and therefore N application rate
costs are considered separately from the seeding rate calculations on a per
acre basis. A solution that considers seeding and fertilizer rates simultane-
ously is available in Clay et al. (2011). To simplify the discussion the unit
price of corn is denoted PC, unit cost of N is PN, and the unit cost of corn
seed is PS.
For example, if the yield is 100 bushels per acre and the return on each
bushel is $4.00 (PC), then the value per unit area is $400 acre-1 (TR). Math-
Fig. 11.1. Relationship between invest- ematically, total return (TR) is defined as:
ment (cost) and return. The economic
TR = PC × bushels per acre.
optimum rate is the point where the mar-
ginal investment and equals the marginal The total cost (TC) of the input is calculated by multiplying the unit cost
return. On the chart, two points are identi- of product purchased times the number of products purchased per unit
fied, investment < return and investment area. Mathematically TC is defined as:
> return. This chapter discusses how
TC = Unit input price × total input usage per unit area
to identify the point where investment
= returns. The location of the economic Or for N fertilizer
optimum rate depends on the cost of the
TC = PN × pounds applied per acre.
inputs and value of the products.
For example, if fertilizer cost $341 per ton and a pound of fertilizer contains 0.46
 $ 341 1 lb 
×
pounds of N, then the cost of N per pound is $0.37 or  2000 lb 0 . 46 lb N  . Assume the application rate is 150 lb N per acre,

 $ 0 . 37 150 lb N $ 55 . 50 
 × = 
then the cost per unit area is $55.50 per acre  lb of N acre acre  . Based on these definitions and the mathematical
relationships discussed above, the cost and revenue components of the EOR per acre are defined by the equation,
∆( value of product) ∆(TR ) PC
1= = =
∆(cost of inputs) ∆(TC) PN [11.2]
Eq. [11.2] states that the ratio of value of the products (corn grain) to the cost of the inputs (N fertilizer) are the marginal rev-
enue and marginal cost, respectively. Problems 11.1a and 11.1b provide simple examples of how to calculate TR and TC per acre.
The TR and TC values are used to determine the economic optimum rate (EOR) and they are used to determine
the return per acre, which is equal to the TR minus TC. Mathematically, the return or profit per acre is defined as:
Return = PC ×Y(N)– PN×N [11.3]
Where, Y is corn yield per acre and is defined as a function of nitrogen inputs Y(N). Notice that yield (Y) is
defined as a function of N, the sole input cost for corn production. A reasonable assumption is that the more N
applied to an acre of corn the higher the yield, however as the amount of N applied increases the return decreases

Problem 11.1a.
If PN is $0.50 per pound and you apply 100 lb N per acre, what is the TC per acre?

Answer:
TC = 100 lb × $0.5 = $50 per acre.
Problem 11.1b.
If PC = $4 per bushel and you harvest 150 bushels per acre, what is the TR per acre?

Answer:
150 bu per acre × $4 per bushel = $600 per acre

182
(Fig. 11.1). This is known as the “Law of Diminishing Returns.” The basic economic principle of the Law of Dimin-
ishing Returns can be mathematically incorporated into an equation that predicts outputs based on inputs (Fig.
11.1). The solution for determining the EOR that is discussed below, requires the use of the second-order polyno-
mial equation, which has the form,
Yield = a + b×N1+ c×N2 [11.4]
Where, “a” is the y-intercept and represents the yield when no N is applied, and coefficients “b” and “c” represent the
nonlinear relationship between nitrogen input and corn yield. The slope of the second-order yield equation describes
the relationship between the nitrogen inputs and yield. The slope of the second order polynomial equation is the first
derivative of the Yield with respect to the inputs (N). In the first derivative, the symbol dY/dN, represents the incremen-
tal change in yield resulting from an incremental change in N rate. For example, if the yield when 0 lb of N is applied
is 100 bushels per acre, and the yield when 0.000001 lb N per acre are applied is 100.00001, the change in yield (0.00001
bushels per acre) per change in N (0.000001), represents the incremental change in yield per incremental change in N.
The first derivative of Eq. [11.4] is defined as:
dY
= b + 2cN [11.5]
dN
The derivate dY/dN is called the marginal product of the input, N. The economic interpretation of this equation is that
it provides the transformation rate of N per acre into the yield per acre. Next, by combining equations [11.3] and [11.4] by
substituting Eq. [11.4] for Y(N) in Eq. [11.3], and taking the derivative of return with respect to N will result in Eq. [11.6]:
dReturn
= PC ( b +2cN )− PN [11.6]
dN
Note, that PC is the marginal revenue (selling price of the corn) and PN is the cost of the N fertilizer ($ per lb).
Equation [11.6], alters the EOR condition depicted in Eq. [11.1] In Eq. [11.1], DTR is equal to PC. Now DTR = Pc(b + 2cN)
where PC denotes marginal revenue and b + 2cN is the marginal product of N. Incorporating the yield—N relation-
ship into the return function Eq. [11.3] allows the defining of the economic optimum rate for N in the following
equation,
PN
( b +2cN ) = [11.7]
PC
where, b and c are based on the use of the yield—N production model (Eq. [11.4]) discussed above. Notice that
the EOR condition now requires that the transformation rate of N per acre into the yield per acre be equal to the
ratio of corn price per bushel to the cost per pound of N. The solution value for N, which is the optimal N rate per
acre, can be found by solving for N in Eq. [11.7]. The resulting solution is,
 PN   1   b 
    −  = Optimal Application Rate for N [11.8]
 PC   2 c   2c 
In this solution, the b and c values are the coefficients in Eq. [11.4], PN is the cost of each unit of N fertilizer and
Pc is the selling price of each bushel of corn.

Problem 11.2.
Assume the following estimated quadratic equation values for a corn return function includes the following
values: b = 0.78 and c= (-0.0024), PC = $4 and PN = $0.55. First step is to solve Eq. [11.7] for N. Then substitute
values given to solve for the EOR application rate for N per acre.
Answer:  P  1   b 
N =  N    −  =133 . 8 lb per acre
 PC   2 c   2 c 
Problem 11.3.
What happens if the cost of N increases to $0.60 per pound?
Answer:
Note that if the corn price remains the same and the N price increses, the return has decreased (133.8 vs. 131.2).

183
Problem 11.4.
Using the b and c values provided in Problem [11.3], calculate the return if N is $0.60 per pound, and the price
of corn increases from 4 to $5 per bushel?
Answer:
137.5 pounds per acre
In this example, more N is economical because the price for corn increased.
Problem 11.5.
Using values in Problem 11.3, determine the application rate of N if corn is selling at $6.00 per bushel and N
costs $0.70 per pound.
Answer:
138.2 pounds per acre
Problem 11.6.
Determine the economic optimum rate using the quadratic model, yield = 106+0.96x-0.00343x2. First take the
derivative of this equation and then use the information to determine the EOR of N.
The derivative is 0.96-2(0.00343)x = 0.96 -0.00686x,
b= 0.96 and
2c = -0.00686
In this problem, N is the nitrogen rate in pounds per acre, 106 is a, 0.96 is b, and -0.00343 is c. The fertilizer
is selling for $0.50 per pound (PN) and corn is worth $3.50 per bushel (PC).
Answer:
P  1   b 
N = N    −  =119 . 1 lb per acre
 PC   2c   2c 
Problem 11.7.
What is the quadratic equation for the hypothetical relationship between N rate and corn yield?
N rate yield
pounds per acre bushels per acre
0 100
60 130
120 160
180 170
240 171
Answer:
This equation can be derived in Microsoft Excel 2013. However, the equation is determined differently in different
versions. In Excel 2013, highlight the data, select insert, select recommended charts, click on scatter and
select OK. At this point the chart should be created, left click on the first point, select add “trendline”,
select polynomial, select display equation on chart and display R 2 on chart. The resulting equation is,
Yield = 98.7 + 0.67 (N rate) – 0.00155 (N rate)2 The R 2 value is 0.99 which means that 99% of the yield variabil-
ity was defined by this equation. In this equation the b can c coefficients are 0.67 and -0.00155.
The equation can also be derived using regression under Data and Data Analysis. To use the Data Analy-
sis program it must be activated. In Microsoft Excel 2013 select Developer, Add-Ins, and click on Analysis
ToolPak, and click on OK. If you use the data set above, the program will only provide a linear equation. To
produce a 2nd order polynomial equation, you must add an additional column containing the N rate squared
values. When using the data analysis tool, both columns (N rate and N rate squared must be selected). The
data must be provided in columns. If provided in rows, the program may not work.
Problem 11.8.
What is the economic optimum nitrogen rate for the quadratic model where yield = 108+0.948x-0.00356x2, where N is the
nitrogen rate in pounds per acre? In this calculation N is selling for $0.50 per pound and corn is worth $3.50 per bushel.
Answer:
 P  1   b 
N =  N    −  =113 . 08 lb per acre
 PC   2 c   2 c 

184
 Problem 11.8b.
Next derive yield, N cost, and corn revenue per acre at the economically optimum N rate in this example.

Answer:
Corn Yield = 106+0.96×(113.08)-0.00343×(113.08)2 = 169.68 bushels per acre.
N Cost = $0.5 × 113.08 = $56.54
Total Revenue (profit) = $3.50 × 169.68 - $56.54 = $537.33
Problem 11.9.
In the summit and footslope of a field, what is the economic optimum N rates for the two landscape posi-
tions? Corn is selling for $4 per bushel and N is selling for $0.55 per pound. Calculate profit for footslope and
Summit for EOR N rate.
N rate Footslope yield Summit yield
pounds per acre bushels per acre bushels per acre
0 100 60
60 130 90
120 160 120
180 170 130
240 171 131

Answer:
First, use excel to estimate two regression equations for both landscape positions as shown in Problem 11.7. The
second order polynomial equations are:
Footslope: yield = 98.7 + 0.67 (N rate) – 0.00155 (N rate)2

Summit: yield = 58.7 + 0.67 (N rate) – 0.00155 (N rate)2
Note that both equations have identical derivatives,
dy
= 0 . 67 −(2×0 . 00155)x
dx , thus they have identical b and c terms. The N rate per acre is given by the EOR equation:
Foot Slope yield = 168 bu. Summit yield= 128
Foot Slope Profit = TR – TC Summit Slope Profit = TR - TC
$4×168 - $0.55×171.78 = $577 $4×128 - $0.55×171.78 = $417

Economic Optimum Seeding Rate

The next topic is estimating the EOR optimal seeding rate. The methodology for estimating the optimal seeding
rate (S) is identical to the N example above. Therefore Eq. [11.3] and [11.4] can be modified as follows:
Return= PC ×Y (S )− PS ×S [11.9]

Yield=a +b×S 1 + c ×S 2 [11.10]

The EOR equation (Eq. [11.8]) is modified as follows (Eq. [11.11]) so that it can be applied to deriving the optimal seeding rate
that will maximize returns. Note the only change in Eq. [11.11] is that PS (cost of seed per 1000 seeds) has been substituted for PN.
 P  1   b 
S =  S    −  = Optimal Seeding Rate [11.11]
 PC   2c   2c 
Developing Site-specific Seeding Rate Algorithms
Since the advent of agriculture, farmers have always been interested in reducing their input costs. To help identify
appropriate rates, agricultural companies have conducted numerous seeding rate experiments across the world. For
a specific cultivar, data from multiple experiments can be aggregated to create site-specific seeding rate algorithms.
The following discussion assumes that water stress is the primary factor limiting yield. This assumption is based on
experimentation that has shown that water plays a major role in impacting yields especially across landscapes with undu-
lating topography (Mishra et al., 2008; Carlson et al., 2011). With corn, optimum yields can be achieved by increasing the
seeding rates in high yielding areas and reducing the rate in low yielding areas (Clay et al., 2009). Site-specific algorithms

185
 Problem 11.10.
You conduct an experiment and measure the resulting corn yields. In this experiment, populations of 18000,
24000, 30000, and 36000 plants per acre yielded 128, 150,162, and 167 bushels per acre, respectively. The result-
ing relationship between the number of seeds (×1000) and yield were, yield =12.95 + 8.525(seed) -0.1181(seed)2
(derived from Excel polynomial curve fit equation). The cost of the seed is $160 per 80,000 kernel unit (1 bag of
corn seed contains 80,000 seeds), and the corn selling price (Pc) is $3.00 per bushel. What is the optimum seeding
rate per acre, using 1000 seeds as a unit per acre? When solving these problems the units used to create the 2nd
order polynomial equation must be the same as those used in the calculations below.
Answer:
First, find the cost of 1000 seeds (80/$160=$2 per 1000 seeds). So it costs $2 per 1000 seed unit or PS=$2.00. The
following calculations are based on 1000 seed per seed unit:
P  1   b  2 1 8 . 525
S = S    −  = × − = 33 . 27 rounded to 33 . 30 seeds units per acre
 PC   2c   2c  3 2 × ( − 0 . 1181 ) 2 ( − 0 .11 181 )
Remembering that a unit of seed was 1000 seeds, the optimum seeding rate is 33,300 seeds per acre. Convert
this to seeds ha-1. (1 ha = 2.47 acres)
33, 300 seeds 2 . 47 acre 82, 200 seeds
× =
acre ha ha

Problem 11.10b
At a seeding rate of 33,300 seeds, what is the expected yield? What is the profit?
Answer:
Corn yield: 12.95+8.525(33.3) -0.1181(33.3)2 = 165.9 bushels per acre.
Profit for optimal seeding rate: = TR – TC= $3×165.9 - $2×33.3 = $431.10.
Problem 11.10c.
The planter was not calibrated correctly and seeding occurred at 27,000 seeds. What is the expected yield and profit?
Answer:
Corn yield: 12.95 + 8.525(27)-0.1181(27)2 = 157 bushels per acre
Profit = $3×157 – ($2×27) = $417; this means that there would be $14 per acre lost
Problem 11.10d
Assume the seeding rate was 40,000 per acre. What is the expected yield and profit?
Answer:
Based on the calculations above the yield would be 165 and profit would be $415 or a $16 per acre loss. Although
yields at the optimum and higher seeding rates are similar the cost of seed is higher, resulting in less profit than if
seeded at a lower rate with lower yield.

Fig. 11.2. The calculated site-specific corn seeding algorithm,

based on Northern Great Plains experiments conducted between
1998 and 2002 (Carlson et al., 2011). A second order polynomial
equation is not used because, each individual point was identi-
fied using the techniques described above.

should account for this variability. However, because plants with different genetic capacity have different response func-
tions, the data must be sorted appropriately (Cox, 1997). A program that converts data from a number of studies into a site
specific algorithm is available in Carlson et al. (2011). The data and the algorithm (shown in Fig. 11.2) is based on the eco-
nomic optimum rate (EOR) discussion above. Based on this equation, the optimum population level for a yield of 140 and
200 bu acre-1 yield are 27,300 ( = 16,900 + 74.2 × 140) and 31,700 ( = 16,900 + 74.2 × 200) plants per acre, respectively.

186
 Problem 11.11.
Based on the calculations above, you determine that the desired live plant population is 35,200 plants
acre-1 and what is the expected yield? If the germination rate is 95% how many seeds should be planted?
Answer:
Desired population at harvestt (plants per acre)
Seeding rate (seeds per acre)=
% emergence of planted seeds
100

Expected yield: 35,200 = 16900 + 74.2 (yield) = 35200 – 16900 = 74.2(yield) =

18300/74.2 = yield yield = 246 bushels per acre
For this calculation % emergence of planted seeds with the germination
35, 200 plants per acre
Seeding rate (seeds per acre) = = 37,100 seed per acre
95 % ge r mination rate
100
Problem 11.12.
What is the relationship between the economically optimum plant populations for the following data? In this
calculation, corn is selling for $3.00 per bushel and a bag of corn (80,000 seeds) is selling for $250 per bag. Foot-
slope areas are located in lower landscape positions and summit areas are located in upland areas.
Plant Footslope Summit
(×1000) per acre bushels per acre bushels per acre
20 125 95
25 140 112
30 160 125
35 175 123
40 170 120

Estimated footslope and summit yield equations are:

Footslope: yield= -30.3+10.2(population)-0.129(population)2; r2=0.96, b= 10.2, and c= -0.129
Summit: yield = -55.2 + 10.65(population) – 0.157(population)2; r2=0.98, b=10.65, and c = -0.157
Seed costs are $3.125 per 1000 seeds and plant populations are in 1,000 seeds per acre.
Answer:
Footslope:
The optimum seeding rate in the footslope is 1000 time 35.50 or 35,500 seeds per acre.
Summit:
The optimum seeding rate in the summit is 1000 time 30.60 or 30,600 seeds per acre.
Next: what is the cost difference between the two seeding rates?
$250/80 = $3.125/1000 seeds 35.5 × $3.125 = $111 for footslope and 30.6 × $3.125 = $95.6 for the summit.
Then, what the Yield and Profit for the footslope and summit?
Foot Slope yield = 169.2 bu. Summit yield= 123.7 bu
Foot Slope Profit = TR – TC Summit Slope Profit = TR – TC
$3×169.22 - $3.125×35.49 = $396.75 $3×123.68 - $3.125×30.6 = $275.42

Summary
In summary, this chapter derives an equation for determining the economic optimum rate for a second order polynomial
equation and provides examples for determining the EOR for N and seeding rates. Based on the analysis of individual
experiments, an analysis across experiments was conducted to derive a precision farming seeding rate algorithm.

Acknowledgments
Support for this document was provided by South Dakota State University, South Dakota Soybean Research and
Promotion Council, and the Precision Farming Systems communities in the American Society of Agronomy and
the International Society of Precision Agriculture. Additional support was provided by the USDA-AFRI Higher
Education Grant (2014-04572).

187
 Additional Problems
11.13. Will the equation provided in this chapter do equally as well for determining economic optimum rates
using with a linear or exponential model?
∆ value of product
11.14. What does this equation, 1= ∆ input cost mean?
11.15. Why does the N or seeding rate decrease with increasing cost?
11.16. How would landscape position influence the N recommendation model?
11.17. If seed cost $300 a bag (80,000 seeds) and corn is selling for $3.50 per bushel, what is the economic optimum
seeding rate?
Plant Yield
(×1000) per acre bu per acre
20 130
25 145
30 162
35 178
40 175
Calculate the profit using the optimum seeding rate and then 20% higher and lower than that rate. Would over-
seeding or underseeding result in less loss?

11.18. Given the information on N rate application for footslope and summit yields, calculate the optimal
N rate, yield, and profit associated with optimal N rate for footslope and summit. Then calculate yield and
returns if the application rate was 15% lower or higher than the economic optimum. Would over or under
application result in less loss compared with the optimum?

N Nsq Footslope yield Summit yield

0 0 90 75
60 3600 120 95
120 14400 155 120
180 32400 170 130
240 57600 171 131

References
Bullock, D.G., and D.S. Bullock. 1994. Quadratic and quadratic-plus-plateau models for predicting optimum nitrogen rate
of corn: A comparison. Agron. J. 86:191–195. doi:10.2134/agronj1994.00021962008600010033x
Carlson, G., D.E. Clay, and J. Schefers. 2011. A case study for improving nutrient management efficiency by optimizing the
plant population. In: D.E. Clay and J. Shanahan, editors, GIS in agriculture: Nutrient management for improved energy
efficiency. CRC Press, New York. p. 157–172.
Cerrato, M.E., and A.M. Blackmer. 1990. Comparison of models for describing corn yield response to nitrogen fertilizer.
Agron. J. 82:138–143. doi:10.2134/agronj1990.00021962008200010030x
Clay, D.E., S.A. Clay, C.G. Carlson, and S. Murrell. 2011. Mathematics and calculations for agronomists and soil scientists.
International Plant Nutrition Institute, Peachtree Corner, Georgia.
Clay, S.A., D.E. Clay, D.P. Horvath, J. Pullis, C.G. Carlson, S. Hansen, and G. Reicks. 2009. Corn (Zea mays) responses to
competition: Growth alteration vs limiting factor. Agron. J. 101:1522–1529. doi:10.2134/agronj2008.0213x

188
Cox, W.J. 1997. Corn silage and grain yield response to plant density. J. Prod. Agric. 10:405–410. doi:10.2134/jpa1997.0405
Kyveryga, P.M., A.M. Blackmer, and T.F. Morris. 2007. Disaggregating model bias and variability when calculating eco-
nomic optimum nitrogen fertilization for corm. Agron. J. 99:1048–1056. doi:10.2134/agronj2006.0339
Mishra, U., D.E. Clay, T. Trooien, K. Dalsted, D.D. Malo, and C.G. Carlson. 2008. Assessing the value of us-
ing a remote sensing based evaportranspiration map in site-specific management. J. Plant Nutr. 31:1188–1202.
doi:10.1080/01904160802134491

189