adapt-n increased grower profits and decreased

Adapt-N Increased Grower Profits and Decreased Environmental N Losses in

2012 Strip Trials

Bianca Moebius-Clune, Harold van Es, and Jeff Melkonian

Department of Crop and Soil Sciences

Due to the large influence of weather conditions on N losses during the early growing season,

interacting with soil and crop management practices (summarized by van Es et al. 2007), it is not

possible to accurately determine at the beginning of the growing season how much N fertilizer will be

needed for that year’s corn crop. Because of the uncertainty in the economic optimum N rate (EONR),

many growers apply “insurance fertilizer”, which in many years, particularly those with relatively dry late

spring weather, exceeds crop N needs. Over-application of N reduces farm profits and causes high

environmental losses. Seasonal corn N needs can be estimated much more accurately in the late spring

after critical weather events have occurred. Adapt-N is an online decision support tool (http://adapt-

n.cals.cornell.edu) designed to help farmers precisely manage nitrogen (N) inputs for grain, silage, and

sweet corn. It uses a well-calibrated computer model, and combines user information on soil and crop

management with high resolution weather information, to provide N sidedress recommendations and

other simulation results on nitrogen gains and losses.

In the first year of Adapt-N on-farm beta-testing (2011, published in What’s Cropping Up? by

Moebius-Clune et al, 2012), Adapt-N recommendations provided agronomic, economic and

environmental benefits compared to conventional N rates in almost all NY and IA trials. Adapt-N

significantly reduced 2011 N application rates in all of the NY trials (by 15 to 140 lbs), with an estimated

reduction in N losses to the environment (by 5 to 120 lbs/ac). Grower profits increased in 86% of NY

2011 trials, by an average of $26/ac. Yield-, and some profit-losses occurred in corn after soybean trials,

due to a model inaccuracy that was corrected in the 2012 Adapt-N version, explained below. In 2011

Iowa trials, Adapt-N reduced N rates in 7 of 9 trials (by an average of 45 lbs/ac) with generally

insignificant yield losses. Adapt-N recommended higher N rates that resulted in significant yield

increases in two trials where spring rainfall was greater. 2011 Iowa grower profits increased in 78% of

trials, by an average of $25/ac with use of Adapt-N.

Model improvements made for the 2012 season:

As a result of 2011 beta-testing, several improvements to Adapt-N recommendation calculations

were implemented prior to the 2012 growing season, including adjusted soil type inputs, previous crop

inputs and manure input options. Most importantly, model routines for soybean N contributions were

adjusted. The 2011 version of Adapt-N used a flat 30 lb soybean N credit, but also mechanistically

simulated the decreased immobilization of N in soybean residue, as compared with corn stover in corn-

after-corn rotations. This almost doubled the N credit for corn following soybean, resulting in artificially

low N recommendations and consequent yield losses in the 3 NY trials with soybean as the previous

crop. The 2012 version of Adapt-N modeled the soybean ‘N credit’ as an absence of the immobilization

penalty applied in corn-corn rotations, and provided a partial soybean N credits of 15 lb N/ac.

http://adapt-n.cals.cornell.edu/


http://blogs.cornell.edu/whatscroppingup/2012/03/28/adapt-n-increased-grower-profits-and-decreased-environmental-n-losses-in-2011-strip-trials/

A profit correction factor to optimize profits from N application (van Es et al, 2012) was also

incorporated in the 2012 version of the Adapt-N tool. This factor takes into account several key realities

that affect farmer profit: 1) the fertilizer to grain price ratio and its effect on the EONR, 2) the nonlinear

yield response to N inputs that causes a greater profit penalty for under-prediction than over-prediction

of the optimum N rate, and 3) the uncertainty in the optimum N rate, which is reduced with this

precisions tool in comparison with static approaches that do not incorporate location-specific weather

effects. Using a fertilizer to grain price ratio of 0.1 ($0.60/lb N: $6 bu grain) Adapt-N subtracted 8 lb N/ac

from predicted Agronomic Optimum N Rate (AONR) to determine the EONR, assuming an uncertainty

with a standard deviation of 20 lb N/ac for Adapt-N recommendations.

Methods We completed 42 replicated strip trials on commercial and research farms throughout New York and 18

replicated strip trials in Iowa (1 trial in Minnesota, included with “Iowa trials”) on commercial farms

during the 2012 growing season. The trials involved grain and silage corn, with and without manure

application, and different rotations (corn after corn, corn after soybean or other; Table 1). Sidedress

treatments involved at least two rates of nitrogen, a conventional “Grower-N” rate based on current

grower practice and an “Adapt-N” recommended rate. A simulation was run for each field just prior to

sidedressing to determine the Adapt-N rate. In 2012 NY trials, most Adapt-N rates were lower than

conventional N rates (by 20 to 138 lbs/ac; Table 1). In three trials Adapt-N rates were higher than the

conventional rate (by 10 to 23.5 lbs/ac). In 2012 Iowa trials, most Adapt-N rates were lower than the

conventional N rates (by 20 to 100 lbs/ac), while 2 Adapt-N rates were higher than the conventional N

rates (by 30 and 40 lbs/ac – the 40lb recommendation was erroneously applied as 70lb/ac). Growers in

IA and NY implemented field-scale strips with 2-4 replications per treatment. Additional rates of zero,

medium, or high N rates were included in some trials, but are not discussed here.

Yields were measured by weigh wagon, yield monitor, or in a few cases by representative sampling (two

20 ft x 2 row sections per strip). Partial profit differences between the Adapt-N recommended and

Grower-N management practices were estimated through a per-acre partial profit calculation:

Profit = [Adapt-N yield – Grower-N yield] * crop price – [Adapt-N rate – Grower-N rate] * price of N +

Sidedress savings/loss

Yields were used as measured, regardless of statistical significance, since the statistical power to detect

treatment effects is inherently low for two-treatment strip trials. For corn, a 2012 grain price of

$6.00/bu was assumed ($7.00/bu minus $1.00/bu for drying, storing and trucking from PA Custom

Rates; USDA, 2012). For silage, $50/T was used based on reported NY silage prices of $25-75/T. A N

fertilizer price of $0.60/lb N was used (reported prices ranged from $0.42 - $0.80/lb N in NY). If Adapt-N

recommended no need for sidedressing and the Grower-N rate was greater than 0 lbs/ac, a sidedress

savings of $8/ac was added to the profit. If Adapt-N recommended an application of N and the Grower-

N rate was 0 lbs/ac N, a loss of $8/ac was subtracted from the profit. [Total N losses to the environment

(atmosphere and water) and N leaching losses were estimated for each treatment by running model

http://scisoc.confex.com/scisoc/2012am/webprogram/Paper73195.html

simulations with all N inputs through December 15, 2012. TO BE COMPLETED] Agronomic, economic

and environmental outcomes of these trials were then used to assess Adapt-N performance.

Results

Agronomic, economic and environmental comparisons between Grower-N and Adapt-N treatments for

each trial are provided for NY and IA trials in Figure 1 and as averages in Table 2.

NY corn after corn trials

Adapt-N rates in all of the NY grain after grain trials resulted in N input reductions. In all but 1 trial, no

significant yield loss was measured with these reduced N rates. Adapt-N rates provided a profit

advantage over conventional N rates in almost all trials (by $1.62 to $88.20/acre). In 4 of the 5 instances

where Adapt-N rates resulted in profit losses, actual yield achieved with the higher N rate exceeded the

expected yield used in Adapt-N (by 8.53 bu/ac to 35.4 bu/ac) to estimate the sidedress rate. In each of

these four cases, a more appropriate expected yield would have been available from existing yield

records, and would have likely resulted in a sufficient recommendation.

NY corn after soy or other crops

Adapt-N rates in all NY corn grain after soy or other (wheat, oats, silage) trials resulted in profit increases

(of $7.64 to $105.30/acre). 2012 results demonstrate that modeling the soybean N contribution as a

partial credit plus the absence of corn stover N immobilization penalty provided more accurate N

recommendations than in 2011. Harvest data show that despite large reductions (average 56lb/ac) in N

rates applied for the Adapt-N treatments, reductions in yields were negligible in all trials and only

statistically significant in one case (Trial 34). The impact of field variability in these trials is illustrated by

the measured average of 4 bu/ac increase with lower N rates, and one trial where the Adapt-N rate was

46 lb/ac lower than the Grower-N rate, but resulted in a significant yield increase of 13 bu/ac (Trial 40).

NY silage

Adapt-N rates in 4 of 6 NY silage trials resulted in N input reductions (22.5 - 50 lbs[WU1]/ac). No

statistically significant yield loss was measured with these reduced N rates – on average measured yield

loss was 0.0 T/ac when N rates were decreased. Adapt-N rates provided a measured profit advantage

over conventional N rates in two trials (by $48 to $58/acre) when N rates were reduced by 50 lb/ac, but

profit losses were calculated in the other four cases, due to field variability, small N rate differences, and

drought as explained below.

Two of these trials (17 and 18) registered a profit loss due to small statistically insignificant yield loss,

and artificially low Grower-N rates as the grower is already reducing N rates to near Adapt-N rates (55-

75 lb/ac below standard recommendations that use current yield potentials). In comparison to standard

recommendations, these trials would constitute profit gains. In two trials, Adapt-N rates were higher by

10-11lb/ac, justified by the expected yield input of 22 and 24 T/ac respectively (Trials 14 and 15). Due to

drought, measured yields were below expected yields (by 5.3 and 11.8 tons), and due to field variability,

measured yields were lower in plots with the higher Adapt-N rate, resulting in a calculated profit loss.

Silage trials exhibited greater yield variability than non-manured grain trials, making it more difficult to

assess the performance of Adapt-N. Factors contributing to such variability are precision and accuracy in

manure testing and application, evenness of spreading, lack of available yield monitors for silage harvest

such that hand-harvesting had to be used, and the effects of drought in some trial locations.

A careful assessment of the currently available data in silage trials for 2011 and 2012 suggests that the

model is handling these well, and that profit gains are made particularly with large N use reductions.

However, due to variability and poor growing season conditions there is a need for further testing of the

efficacy of Adapt-N in predicting N needs in silage production.

IA corn grain trials

Adapt-N rates in all but two of the IA trials resulted in N input reductions (by 36 lb/ac on average). In all

but 1 trial, no significant yield loss was measured with these reduced N rates. Adapt-N rates provided a

profit advantage over conventional N rates in most of these trials (by $1.49 to $81.20/acre). Where the

Adapt-N rate was lower than the Grower-N rate, Adapt-N usually resulted in profit increases but two of

four trials where this was not the case were trials where actual yield exceeded expected yield (Trial 62).

In this trial, the Grower-N rate produced a yield 8.8 bu/ac higher than the expected yield (200 bu/ac),

while the Adapt-N rate did not exceed the expected yield, resulting in a profit loss despite N savings. In 2

trials (65 and 73) Adapt-N recommended a higher rate than the conventional N rate (by 30 lbs/ac and 40

lbs/ac, which was accidentally implemented as 70 lbs/ac). In both cases, the expected yield was not

attained (by a margin of 20 bu/ac and 113 bu/ac respectively), resulting in profit losses due to excessive

N application. 2012 Iowa trial results were impacted by abnormally droughty conditions for the majority

of trials. It is not possible for Adapt-N (or other N recommendation methods) to account for abnormal

weather events that occur after the date of sidedress.

It should be noted that in one of the Iowa trials, Adapt-N recommended 0 lbs/ac N acre as compared

with the conventional N rate of 75 lbs/ac N. Despite a yield reduction (9 bu/ac), the Adapt-N rate did not

decrease profit (+$1/ac), due to the large reduction in sidedress expense. This trial demonstrated that if

a grower applies a large standard rate of N, Adapt-N recommendations can provide profit gains even if

yields are somewhat reduced. Nevertheless, trials in which Adapt-N recommendations resulted in yield

reductions will be assessed to inform improvements to the model for the 2013 season.

Recommendation errors resulting in profit loss were largely made in instances where expected yield

either exceeded or underestimated actual yield. Drought conditions experienced in Iowa as well as NY

during the growing season resulted in abnormally low yields in several of the trials. Because sidedress

recommendations must be made before such post-sidedress weather has been determined, neither

Adapt-N nor other N recommendation tools can account for mid-to-late season weather impact on N

dynamics at this time.

Growers who tend to use high amounts of nitrogen will realize large savings. The profit losses resulting

from both overestimating and underestimating yield emphasized the importance of accurate yield

information in determining the Adapt-N recommendation. In a much wetter year, increased

profitability would come from appropriately applying more N at sidedress time in order to prevent

yield reductions from N losses. In the long term we expect that environmental losses will decrease in

both dry and wet years, because this tool provides strong incentives to shift N applications to

sidedress time.

Conclusions

From the testing on commercial farms in NY State and Iowa in 2011 and 2012 we have determined that

the value of the Adapt-N tool is substantial, resulting in N reductions and profit savings in 79% of all

trials. 2012 testing confirmed that the soybean N crediting method used by the model for 2011 soy-corn

rotation trials was appropriately adjusted for 2012. The tool was successful in adjusting for the effects of

early season conditions to recommend N fertilizer needs. The impact of expected yield input in

generating accurate N recommendations using Adapt-N was demonstrated in several trials.

Over the 2011-2012 growing seasons, 84 trials indicate that:

79% of Adapt-N recommendations provided increased grower profits over current rates.

N application rates were significantly reduced (15 to 140 lb/acre) in almost all cases, and thus

post-growing season losses of excess N to the environment were decreased substantially.

Yield losses were generally negligible (-1 bu/ac average across all trials)

Where Adapt-N recommended higher N rates, higher yields resulted when drought was not the

more limiting factor.

Grower profits increased on average by $27/acre in all trials, and by $32/ac in NY with the

improved 2012 model version.

User training to enable proper use of the Adapt-N tool is essential for appropriate N

recommendations, as model inputs must be carefully chosen.

For more information: The Adapt-N tool and training materials, including the recorded 3/21/2013

training webinar, are accessible to stakeholders through any device with internet access (desktop,

laptop, smartphone, tablet) at http://adapt-n.cals.cornell.edu/. Information on account setup is also

available. Adapt-N users can elect to receive email and/or cell phone alerts providing daily updates on N

recommendations and soil N and water status for each field location in Adapt-N.

Acknowledgements: This work was supported by funding from the NY Farm Viability Institute, the

USDA-NRCS Conservation Innovation Program, the International Plant Nutrition Institute, and MGT

Envirotec. We are grateful for the cooperation in field activities from Bob Schindelbeck, Keith Severson,

Kevin Ganoe, Sandra Menasha, Joe Lawrence, and Anita Deming of Cornell Cooperative Extension, from

Mike Davis at the Willsboro Research Farm, from Dave DeGolyer, Dave Shearing and Jason Post at the

Western NY Crop Management Association, from Eric Bever and Mike Contessa at Champlain Valley

Agronomics, from Mark Ochs at Mark Ochs Consulting, from Peg Cook at Cook’s Consulting in New York,

and from Shannon Gomes, Hal Tucker, Michael McNeill, and Frank Moore at MGT Envirotec in Iowa. We

also are thankful for the cooperation of the many farmers who implemented these trials on their farms.


In particular we would like to acknowledge Robert and Rodney Donald for implementing farm-wide

trials on most of their fields.

References

H. van Es, J. Melkonian, Moebius-Clune, B., C. Graham, and C.L. van Es. Building Uncertainty Into N Recommendations for Maize: Addressing Insurance Applications. 2012. ASA-CSSA-SSSA Meeting, Cincinnati, OH, Oct 22, 2012, http://scisoc.confex.com/scisoc/2012am/webprogram/Paper73195.html, Abstract. van Es, H.M., B.D. Kay, J.J. Melkonian, and J.M. Sogbedji. 2007. Nitrogen Management Under Maize in Humid Regions: Case for a Dynamic Approach. In: T. Bruulsema (ed.) Managing Crop Nutrition for Weather. Intern. Plant Nutrition Institute Publ. pp. 6-13. http://adapt-n.cals.cornell.edu/pubs/pdfs/vanEs_2007_Managing N for Weather_Ch2.pdf. [URL verified 2/22/12]. EPA, 2010. Inventory of U.S. Greenhouse Gas Emissions and Sinks (1990-2008). U.S. Environmental

Protection Agency, Washington, DC.

http://scisoc.confex.com/scisoc/2012am/webprogram/Paper73195.html

http://adapt-n.cals.cornell.edu/pubs/pdfs/vanEs_2007_Managing%20N%20for%20Weather_Ch2.pdf

http://adapt-n.cals.cornell.edu/pubs/pdfs/vanEs_2007_Managing%20N%20for%20Weather_Ch2.pdf

Table 1. 2012 Replicated field-scale strip trials implemented in NY, IA, and MN to compare current

Grower N application rates with N rates recommended by the Adapt-N tool. N rates represent total N in

lb/acre applied as inorganic fertilizer in 2012 for corn grain following grain (1a), corn grain following

soybean or other crops silage corn (1b) and corn grain in IA and MN (1c).

State Trial ID Harvest Manure Prior crop

Grower

(lb/ac)

Adapt-N

(lb/ac)

Difference (A-

G)

NY 1 grain no gra in 222 162 -60




NY 11 gra in no gra in 204 140 -64



















Trial Info Fertilizer N applieda)

*Adapt-N does not have previous crop inputs for small grains. Therefore, in these instances, a previous crop of

grain corn (NY 6), silage corn (NY 24) or soybean was selected (NY 28). If soybean was selected the 15 lb/ac N

credit was added to the Adapt-N recommendation.


Grower

(lb/ac)

Adapt-N

(lb/ac)

Difference (A-

G)

NY 6 grain no wheat* 177 144 -34

NY 9 grain no soy 206 144 -62

NY 10 gra in no soy 201 128 -73

NY 12 gra in no soy 189 89 -100

NY 22 gra in no soy 126 149 24

NY 23 gra in yes s i lage 178 114 -64

NY 24 gra in yes wheat* 177 145 -32

NY 28 gra in yes oats* 113 68 -45

NY 34 gra in no soy 170 105 -65

NY 38 gra in no soy 188 118 -70

NY 39 gra in no soy 224 138 -86

NY 40 gra in no soy 182 136 -46

NY 41 gra in no soy 230 161 -69

NY 14 s i lage yes s i lage 56 66 10

NY 15 s i lage yes gra in 92 103 11

NY 16 s i lage yes s i lage 71 21 -50

NY 17 s i lage no s i lage 125 96 -29

NY 18 s i lage no s i lage 125 103 -23

NY 35 s i lage yes sod 53 3 -50

Trial Info Fertilizer N appliedb)


Grower

(lb/ac)

Adapt-N

(lb/ac)

Difference (A-

G)

IA 59 gra in yes soy 140 100 -40

IA 60 gra in yes corn 80 40 -40


IA 62 gra in no soy 158 95 -63


IA 65 gra in no soy 125 155 30




IA 69 gra in no soy 142 122 -20

IA 71 gra in 2012 soy 120 80 -40

IA 73 gra in 2011 gra in 7 77 70

IA 76 gra in no gra in 252 157 -95

IA 77 gra in 2012 gra in 49 4 -45

IA 78 gra in 2010 soy 53 3 -50

IA 79 gra in 2009 gra in 183 148 -35

IA 82 gra in no soy 119 44 -75

IA 83 gra in no gra in 173 144 -30

MN 84 grain 2010 gra in 230 130 -100

Trial Info Fertilizer N appliedc)

Table 2. Agronomic and economic assessment of model performance. Values are average differences

resulting from Adapt-N use (Adapt-N minus Grower-N treatment), such that a negative number indicates

a decrease due to Adapt-N, a positive number indicates an increase due to Adapt-N. Profit calculations

assume $6.00/bu grain, $50/T silage, and $0.60/lb N.

NY Corn

grain after

corn grain

NY corn

grain after

soy or

other

NY silageAll NY

trials

All IA

trials

-82 -56 -22 -65 -36

-3 4 0 0 -1

$28.64 $57.30 -$7.68 $32.38 $16.63

Average Change due to Adapt-N Use in NY:

Profit ($/ac)Ave

rage

(A

dap

t-N

-

Gro

wer

-N)* N fertilizer input (lb/ac)

Yield (grain: bu/ac; silage: T/ac)

Figure 1. Yield and fertilizer use from Grower-N vs. Adapt-N treatments in NY and IA trials. Partial profit

gain (positive) or loss (negative) from using the Adapt-N recommendation, relative to grower’s current

practices. (*Yields were statistically different at p < 0.05; **Yields were statistically different at p < 0.01)

*

a)

b)

c)

d)

e)

f)

**

*

g) h)

i)

j)

**

Figure 2. Proportion of trials with profit gains (green) or losses (gray) as a result of using the Adapt-N

recommendation compared to current grower N management in 2011 and 2012 trials. Profit

calculations assume $0.60/lb N, $6.00/bu grain for 2012 and $5.50/bu grain for 2011. Profit calculations

assume $50/T silage, and $0.60/lb N for both years. Trials with profit losses due to underestimated yield

potential are included, so success rates reported here can be improved through proper tool use.

l)

k)

adapt-n increased grower profits and decreased

Documents