Agronomy Journa l • Volume 109, I s sue 5 • 2017 1811

Variability in crop N requirements across states, regions, seasons, crops, and even fields complicates the selection of an appropriate N rate, and has led to

concerns of the environmental impacts that may result from over-application. Nitrogen from agricultural sources can be lost through volatilization or denitrification, but nitrate (NO3) leaching constitutes the largest percent of loss and can conse-quently contaminate ground and surface water (David et al., 1997; Jaynes et al., 2001; Randall and Goss, 2001). Aside from the well documented negative health effects of NO3 in drink-ing water (Spalding and Exner, 1993; Sogbedji et al., 2001), excess NO3 can also cause eutrophication and hypoxia of lakes and coastal waters (Spalding and Exner, 1993; Randall and Goss, 2001), a consequence that can be devastating to ecosys-tem biodiversity (Kronvang et al., 2001).

Minimizing N rates can lower leaching losses and protect water quality while maintaining crop yields, but identifying appropriate N rates can be a challenge. In North Carolina, the realistic yield expectations (RYE) database provides recom-mended N rates for 32 different crops by soil series (North Carolina Nutrient Management Workgroup, 2003). The database was developed and validated through N rate trials across the state. More recently maize N recommendations were updated to reflect changes in N use efficiency related to new hybrid releases and farmer management practices (Rajkovich et al., 2015). Nationally, the “4Rs” (right rate, right placement, right timing, and right source) have been promoted by private and public organizations (Natural Resources Conservation Service, 2005; The Fertilizer Institute, 2011; International Plant Nutrition Institute, 2012) as a platform for farmers to think about their N applications in a logical, interconnected way.

Beyond typical soil-based N rate recommendations, farmers also have access to fertilizer N-loss prevention amendments intended to reduce the amount of N lost; essentially to bolster the “right source” aspect of an N management plan. These N-loss prevention amendments, which can be added to N fertilizers, have various modes of action and reported levels of effectiveness in the literature. The N-loss prevention amend-ment NBPT+DCD (AgrotainPlus; Koch Agronomic Services LLC, Wichita, KS), contains the urease inhibitor N-(n-butyl)-thiophosphoric triamide (NBPT at 30–70 g kg–1 a.i.) and the

Evaluation of Nitrogen-Loss Prevention Amendments in Maize and Wheat in North Carolina

Shelby Rajkovich, Deanna Osmond,* Randy Weisz, Carl Crozier, Daniel Israel, and Robert Austin

Published in Agron. J. 109:1811–1824 (2017) doi:10.2134/agronj2016.03.0153 Available freely online through the author-supported open access option

Copyright (c) 2017 American Society of Agronomy 5585 Guilford Road, Madison, WI 53711 USAThis is an open access article distributed under the CC BY license(https://creativecommons.org/licenses/by/4.0/)

ABSTRACTTo reduce environmental losses of N and increase crop use, it is critical to optimize N fertilization rates and determine if N-loss prevention amendments increase yields. Research objectives were to: (i) determine N-release patterns of three N-loss amendments (urea ammonium nitrate [UAN] treated with NBPT+DCD, nitrapyrin, or an organo-Ca) and UAN through a laboratory incubation; (ii) determine effectiveness of these four products for maize (Zea mays L.) and winter wheat (Triticum aestivum L.) produced in two to three regions of North Carolina; and (iii) determine agronomic optimum N rate for wheat and corn compared to state-recommended rates. Nitrogen release was measured in three soils (coastal plain, piedmont, and moun-tains) during the incubation experiment. Field experiments were randomized complete block designs (four replications of six maize N rates and five wheat N rates), with each rate applied as one of four product treatments (UAN and UAN+ one of three N-loss prevention amendments). In the incubation experiment, soils treated with UAN+nitrapyrin or UAN+NBPT+DCD delayed nitrification longer than soils treated with UAN or UAN+organo-Ca. There was no significant effect of product on maize grain yield (coastal plain and mountains) and wheat yield (coastal plain and piedmont). A year × product interac-tion occurred for maize grain yield in the piedmont. Agronomic optimum N rates mostly aligned with current North Carolina N fertilizer recommendations. Despite positive laboratory results, N-loss amendments did not have a significant effect on yield in 9 of 10 site-years, indicating that proper N rates are a more effec-tive nutrient management strategy.

Dep. of Crop and Soil Sciences, North Carolina State Univ., Raleigh, NC 27695. Received 14 Mar. 2016. Accepted 8 June 2017. *Corresponding author (deanna_osmond@ncsu.edu).

Abbreviations: AONR, agronomic optimum nitrogen rate; CEC, cation exchange capacity; DCD, dicyandiamide; NBPT, N-(n-butyl)-thiophosphoric triamide; NUE, nitrogen use efficiency; RYE, realistic yield expectations; UAN, urea ammonium nitrate.

Core Ideas• Fertilizer additives to decrease N losses did not provide consis-

tent yield advantages.• Plots treated with N-loss products did not increase N use ef-

ficiency or N uptake.• Agronomic optimum N rates observed in the field aligned with

North Carolina recommendations.

AGRonoMY, SoILS & envIRonMenTAL QUALITY

Published July 27, 2017

1812 Agronomy Journa l • Volume 109, Issue 5 • 2017

nitrification inhibitor dicyandiamide (DCD) (>600 g kg–1 a.i.). While Sistani et al. (2014) found no effect of NBPT+DCD on maize grain or stover yield or whole plant nutrient uptake in Kentucky, Espindula et al. (2013) did find a grain yield advan-tage in winter wheat grown in Brazil. Previously, NBPT alone had been broadly tested in the South and southeastern United States. A review of those trials found NBPT alone did not improve yields in Alabama, New Mexico, Arizona, or Texas for maize, cotton, or wheat (Mitchell and Osmond, 2012).

An N-loss prevention amendment described as an “N man-agement aid” is marketed by AgExplore International (Parma, MO) as “NZone”. It contains a mixture of alkylarylpolyoxy-ethylene glycols (22%), Ca aminoethylpiperazine (6%) and Ca heteropolysaccharides (5%). Since other agricultural chemical products which contain alkylarylpolyoxyethylene glycols are marketed as nonionic surfactants, low foam agents, wetters, and spreaders, some potential effect on penetration of plant cuticles or soils is suspected; this N-loss prevention amendment is referred to as an organo-Ca in this paper. Examples include Brewer 80-20 (Breer International, Vero Beach, FL), Chemwett Plus (Coastal Agrobusiness, Inc., Greenville, NC) and Spredde 90/10 (AgExplore International, Parma, MO). Additional active ingredients are Ca polymers, which according to the manufacturer “cause a short term pH increase that causes H+ to be released from the exchange complex, allowing newly formed ammonium (NH4–N) ions to attach”, thus protecting the N from nitrification (AgExplore International, 2015). In a laboratory incubations study, Goos (2011) found that this organo-Ca N amendment did not delay N release.

Finally, nitrapyrin ([2-chloro-6-(trichloromethyl) pyri-dine]) in a microencapsulated N-loss prevention amendment (InstinctII; Dow AgroSciences LLC, Indianapolis, IN) is the active ingredient and slows nitrification from NH4 to NO3. This N-loss amendment has been registered for use in the United States by the EPA. Burzaco et al. (2014) conducted a meta-analysis of maize trials conducted from 1988 to 2011 that included nitrapyrin added to anhydrous ammonia (n = 112 treatment means). The reported overall mean grain yield response to nitrapyrin was 116 kg ha–1 (p = 0.09). However, in an accompanying field trial in Indiana, nitrapyrin did not sig-nificantly affect yield across 3 site-years (Burzaco et al., 2014).

Given the reported variability in the effectiveness of N-loss prevention amendments to improve yields across cropping systems and regions, our objectives were to: (i) determine and compare the N-release patterns in the laboratory of UAN and three N-loss amendments including UAN+NBPT+DCD, UAN+nitrapyrin, and UAN+organo-Ca; (ii) determine the effectiveness of adding N-loss products to UAN for maize and winter wheat in three different physiographic regions across North Carolina; and (iii) determine the agronomic optimum nitrogen rate (AONR) for maize and winter wheat to verify, validate, and refine the North Carolina RYE database.

MATeRIALS AnD MeTHoDSLaboratory Incubations

The procedure for this aerobic incubation experiment to measure conversion of NH4–N to NO3–N was adapted from Cahill et al. (2010a). A composite soil sample from each 2015 maize trial site was collected at depth of 0 to 20 cm from

unfertilized plots and mixed in a large plastic bucket. Soils consisted of a Lynchburg fine loamy sand (fine-loamy, siliceous, semiactive, thermic Aeric Paleaquult) at the coastal plain field station, a Lloyd clay loam (fine, kaolinitic, thermic Rhodic Kanhapludult) at the piedmont field station, and a Comus silt loam (coarse-loamy, mixed, active, mesic Fluventic Dystrudept) at the mountain field station. Each soil was dried and passed through a 2-mm sieve. The upper and lower bounds of plant-available water for each soil were estimated by pressure plate (Dane and Hopmans, 2002) utilizing disturbed samples.

Soil from each site (200 g dry weight) was mixed in trip-licate with UAN, UAN+ NBPT+DCD, UAN+nitrapyrin, or UAN+organo-Ca in 0.025-mm thick polyethylene bags (16.5 by 15 cm) (Gordon, 1988) until fertilizer N was evenly distributed throughout the soil. A control treatment received no N fertilizer. All four fertilizer products were added at an N rate equivalent to the highest N application rate for maize (224 kg ha–1) and the additive amounts were added based on label information, which were then scaled for each soil type. Distilled water was added to each bag to bring the moisture level to 80% of field capacity, after which they were then sealed and placed in an incubator at a relatively constant temperature of 23 to 26°C. Bags were aerated once weekly, which previously was demonstrated to provide sufficient carbon dioxide and oxygen exchange (Gordon 1988), and weighed to determine moisture loss (Cahill et al., 2010a; Gordon 1988). Additional distilled water was added to bags with a weight decrease of 5% or more to maintain ~80% field capacity.

Soil samples (10 g on a dry weight basis per bag) were taken from each bag at 0, 2, 7, 14, 28, 42, 56, and 84 d. At each sam-pling date, 25 mL of 1 M potassium chloride (KCl) was added to a 10 g (dry weight) sample and placed on a shaker for 30 min. The extractant was filtered through Whatman no. 2 filter papers into vials and frozen until analyzed on a Lachat flow injection autoanalyzer (model Quick Chem 8000, Lachat Inc., Loveland, CO; QuikChem methods 10-107-04-1-A and 10-107-06-2-A) at the NC State University Environmental and Agricultural Testing Service (EATS) lab (Raleigh, NC). Extractant was analyzed for NO3–N and NH4–N concentrations. In the few circumstances when the Lachat returned values that indi-cated the N concentration was below the detection limit of 0.10 mg L–1, the concentration was assumed to be 0.10 mg L–1. The concentration of NH4–N and NO3–N from each sample was corrected for extraction volume and expressed on a soil dry weight basis. Nitrate-N and NH4–N concentrations in the control soil was subtracted from concentration in the N fertil-izer treated soil at each sampling date to account for mineraliza-tion of soil organic matter when calculating net nitrification of applied fertilizer N. If this correction resulted in a negative value, the concentration was set to zero. Applied N recovery was calculated from the sum of extracted NO3–N plus NH4–N concentrations and expressed as a percentage of N applied.

Field Study

Maize and winter wheat N rate and N amendment tri-als were conducted from 2013 to 2015. Maize trials were conducted yearly at three research stations in three differ-ent physiographic regions across North Carolina. Each year included one trial in each region (coastal plain, piedmont, and

Agronomy Journa l • Volume 109, Issue 5 • 2017 1813

mountains) for a total of 6 site-years. Yearly winter wheat tri-als were conducted at research stations in the coastal plain and piedmont region for 2 yr, and thus consisted of 4 site-years.

Soil series utilized for maize sites included: Lynchburg fine loamy sand in 2014 or Portsmouth sandy loam (fine-loamy over sandy or sandy-skeletal, mixed, semiactive, thermic Typic Umbraquult) in 2015 in the coastal plain; Lloyd clay loam in the piedmont in two separate fields; and Comus silt loam in 2014 or Codorus silt loam (fine-loamy, mixed, active, mesic Fluvaquentic Dystrudept) in 2015 in the mountains. Wheat was planted into Johns sandy loam (fine-loamy over sandy or sandy-skeletal, sili-ceous, semiactive, thermic Aquic Hapludult) for the 2013/2014 growing season or Portsmouth sandy loam in the coastal plain for the 2014/2015 growing season, and Lloyd clay loam in the piedmont, which were different fields for each growing season.

The experimental design was a randomized complete block (RBC) with four replications. Included in the RBC was four N products (UAN and UAN with one of three N-loss pre-vention amendments). For the purpose of this paper we will use the term product to refer to either UAN solution alone or UAN solution plus an N-loss prevention amendment. In North Carolina, UAN solution applied alone is the most commonly used N fertilizer, and serves as a reference point as well as the manufacturers’ recommended method of applying the N-loss prevention amendments. The N-loss prevention amendments (NBPT+DCD, nitrapyrin, or an organo-Ca) were mixed into the UAN and applied at their label rates (NBPT+DCD at 7.6 kg amendment t–1 UAN; nitrapyrin at 2.4 kg amendment ha–1; or an organo-Ca at 0.81 kg amend-ment t–1 UAN). Each N product was applied at six N rates for maize, including a zero-N control. (Control plots did not receive any N or N-loss amendment.) Total N rates for maize were zero-N (control), 45, 90, 135, 180, and 224 kg N ha–1. All plots except zero-N received 45 kg N ha–1 applied at plant-ing with the remainder applied as a side dress application at growth stage V5–V6; thus, each rate, except zero-N, received one of four different products (UAN or UAN treated with one of three N-loss prevention amendments at both planting and V5–V6). The products were dribbled onto the soil surface, as is common for most producers in North Carolina who utilize UAN, using a custom built backpack sprayer and boom (R&D Sprayers, Opelousas, LA) that was calibrated before each applica-tion and triple-rinsed with soapy water between each product.

Each product was applied at five N rates for wheat. The at-plant N application rate for wheat was 34 kg N ha–1 followed by 0, 45, 90, 135, 180 kg N ha–1 as spring applied N prior to forma-tion of the first joint (Feekes stage 4–5 or Zadoks stage 30; Weisz et al., 2001); thus there was no true zero-N application. At-plant N was applied as UAN and spring N was applied as UAN or UAN treated with NBPT+DCD, Nitrapyrin, or an organo-Ca. The N-loss prevention amendments were applied at their label rates (see above paragraph). All products were dribbled using the same backpack sprayer as was used in the maize trial.

MaizeIn 2014 and 2015, maize (Pioneer 1690YHR hybrid) was

planted 11 and 14 April in the coastal plain, 13 and 24 April at piedmont, and 5 and 12 May in the mountains in plots approx-imately 3 by 10 m, which provided four rows per treatment.

Mean plant population was calculated by extrapolating the population from a 3 m span of the inner two rows at harvest to a hectare basis: 74,237 ha–1 on 91 cm rows in 2014 and 58,885 ha–1 on 76 cm rows in 2015 at the coastal plain fields; 81,547 ha–1 in 2014 and 61,032 ha–1 in 2015 on 76 cm rows at the piedmont fields; and 79,305 ha–1 in 2014 and 65,324 ha–1 in 2015 on 76 cm rows in the mountain fields.

Coastal plain and piedmont fields were conservation-till following soybean, while the mountain fields were disked to a depth of 30 cm approximately 1 wk prior to planting and fol-lowed a winter wheat cover crop. In 2014, the coastal plain field received no lime, whereas in 2015, the field received 1344 kg ha–1 lime and 112 kg ha–1 0–0–60 based on soil test results. In both years, paraquat (1,1’-Dimethyl-4,4’-bipyridinium dichloride) and atrazine (1-Chloro-3-ethylamino-5-isopropylamino-2,4,6-triazine) herbicides were applied just prior to planting, and Halex GT herbicides {S-metolachlor [RS)-2-Chloro-N-(2-ethyl-6-methyl-phenyl)-N-(1-methoxypropan-2-yl)acetamide], glyphosate [N-(phosphonomethyl)] glycine], and mesotrione [2-[4-(Methylsulfonyl)-2-nitrobenzoyl]cyclohexane-1,3-dione]} was applied 40 d (d) post-plant. At the piedmont, fields received 504 kg ha–1 0–25–25 per soil test recommendations, and herbi-cides glyphosate, and atrazine both years. Lime was applied at the rate of 2240 kg ha–1 in 2015. Due to extremely dry conditions in 2015 at the piedmont site, 25.4 mm of irrigation was applied 19 June (68 DAP) and 30 June (79 DAP) for a total of 50.8 mm. In the mountains, the fields received 347 kg ha–1 0–26–26 per soil test recommendations and were treated with Trizmet II (atrazine and metolachlor) and paraquat prior to planting both years.

Ears from the same 3 m length of the inner two rows used to determine population were harvested by hand, weighed, and measured for moisture. Grain yield was normalized to a stan-dard moisture content of 155 g kg–1. A subsample of six ears was dried, shelled, and ground for nutrient analysis. Stover yield was determined from six randomly selected plants harvested from each plot, weighed, chopped, and subsampled. The subsample was dried to indicate moisture content and ground to measure nutrient content. Grain and stover N concentrations were used to calculate grain N, stover N, and total N content. Apparent nitrogen use efficiency (NUE) was then determined from total N content at each N rate. Apparent NUE for each N application rate was defined as total plant (stover+grain) N content in fertil-ized treatment minus total N content (stover+grain) in control treatment (0 kg N ha–1) divided by the application N rate.

wheatPrior to planting wheat in the coastal plain (2014),

336 kg ha–1 10–0–20 with 12% S and 5% Mn was broadcast based on soil test recommendations. Broadleaf weed and grass control consisted of Osprey Herbicide {2-{[(4-methoxy-6-methyl-1,3,5-triazin-2-yl)amino]-oxomethyl]sulfamoyl}benzoic acid methyl ester)} and Harmony Extra SG {methyl 3-[[[[(4-methoxy-6-methyl-1,3,5-triazin-2-yl)amino]carbonyl]amino] sulfonyl]-2-thiophenecarboxylate } (thifensulfuron methyl and tribenuron methyl). In both years, the piedmont fields were treated with 336 kg ha–1 10–20–20 broadcast per soil test recommenda-tions, followed by Finesse {chlorsulfuron and 2-{[(4-methoxy-6-methyl-1,3,5-triazin-2-yl)amino]-oxomethyl]sulfamoyl}benzoic acid methyl ester} and glyphosate herbicides post-plant.

1814 Agronomy Journa l • Volume 109, Issue 5 • 2017

Wheat trials in the coastal plain were conventionally tilled and seeded at the rate of 3.7 million seeds ha–1 with plots measuring 1.2 by 9 m, which resulted in eight rows per plot. Conservation tillage was used in the wheat trials in the pied-mont with the same seeding rate as in the coastal plain, but plot size was 1.5 by 9 m. In 2013, North Carolina Yadkin variety was planted in the coastal plain and Southern States 8350 was planted in the piedmont. In 2014, DynaGro Shirley variety was planted at both coastal plain and piedmont. Wheat was planted after corn at all field trials.

A Wintersteiger Delta plot combine (Wintersteiger Inc., Salt Lake City, UT) was used to harvest entire wheat plots. Moisture content and yield data, as well as a subsample of wheat grain, were captured from the combine. Grain yield was normalized to a standard moisture content of 130 g kg–1. The straw discarded

from the combine was collected, weighed, and subsampled to determine straw yield. Subsamples of grain and straw were weighed, dried, re-weighed and ground for nutrient analysis, including N concentration. Similarly to the NUE calculation for maize, wheat grain, and straw N concentrations were used to calculate grain N, straw N, total N content, and spring NUE.

Sample AnalysisSubsamples of maize and wheat grain were ground with a

Retsch Mill Model ZM-100 (Verder Scientific Inc., Newtown, PA). Stover and straw were ground with a Thomas-Wiley Mill Model ED-5 (Arthur H. Thomas Co., Philadelphia, PA). Samples were analyzed for N by the North Carolina Department of Agriculture & Consumer Services lab (Raleigh, NC) with an ele-mental analyzer (NA1500; CE Elantech Instruments; Lakewood,

Fig. 1. Average daily precipitation and average soil temperature for (A) coastal plain 2014, (B) coastal plain 2015, (C) piedmont 2014, (D) piedmont 2015, (E) mountains 2014, and (F) mountains 2015 regions.

Agronomy Journa l • Volume 109, Issue 5 • 2017 1815

NJ) and an inductively coupled-plasma (ICP) spectrophotometer (Optima 3300 DV ICP emission spectrophotometer; PerkinElmer Corporation, Waltham, MA). Wheat grain and straw samples were analyzed for N by the North Carolina State University EATS lab (Raleigh, NC) with an elemental analyzer (2400 CHN/S Elemental Analyzer; PerkinElmer Corporation, Waltham, MA).

ClimatePrecipitation in the coastal plain was well distributed over the

maize growing season in both 2014 and 2015 (Fig. 1A and 1B). Total precipitation varied, however, with nearly 280 mm more precipitation than the 5-yr average in 2014 and 127 mm less in 2015. Average monthly soil and air temperatures were consistent with historical averages. Both years the piedmont sites received less precipitation than the 5-yr average (Fig. 1C and 1D), but 2015 was extremely dry, with rainfall totals 460 mm below aver-age. In the weeks between planting and side dress N application, the field received only 66 mm of precipitation. The only month in 2015 to record above average rainfall was September by 2 mm.

Lack of rainfall paired with slightly higher than average air temperatures in June, July, and August was problematic for crop growth. In the mountains, precipitation was slightly less than average (Fig. 1E and 1F), but well distributed. Cooler average soil and air temperatures relative to other trial locations were condu-cive to higher observed average soil moisture.

Both growing seasons for wheat in the coastal plain had greater precipitation than the 5-yr average, though the patterns differed. The 2 mo following planting in 2013–2014 were wet, followed by two dry months and a wet spring. In contrast, the 2014–2015 season was consistently wetter than the 5-yr average with less total variation. In both years, average air temperature was fairly consis-tent with historical trends, though average air temperature in the winter did drop lower. In the piedmont, the 2013–2014 season had 92 mm more precipitation than the 5-yr average, but the 2014–2015 season was extremely dry with 301 mm less than the average. The lack of precipitation was distributed throughout the cropping year, with November as the only month to record more rainfall than average with a total of 102 mm.

Fig. 2. Effect of product on (A) NH4–N concentration, (B) NO3–N concentration, and (C) percent N recovered from four N product (UAN, UAN+NBPT+DCD, UAN+nitrapyrin, or UAN+organo-Ca) applications in aerobic incubation of coastal plain soil over an 82-d sampling period. UAN = urea ammonium nitrate, NBPT = N-(n-butyl)-thiophosphoric triamide, DCD = dicyandiamide, nitrapyrin +2-chloro-6-(trichloromethyl) pyridine, and organo-Ca = alkylarylpolyoxyethylene glycols Ca aminoethylpiperazine and Ca heteropolysaccharides. Bars denote standard deviation.

1816 Agronomy Journa l • Volume 109, Issue 5 • 2017

Statistical AnalysisThe laboratory incubations were a completely randomized

design. Previous experimentation (Cahill et al., 2010a) dem-onstrated that time was a factor and since our objective was to compare NH4–N and NO3–N at each time interval, standard deviations were used to compare products. Error bars for data points from the laboratory incubation are based on the stan-dard deviation of replicates of each product treatment at each sampling date. An error with the Latchat equipment resulted in indecipherable NH4–N and % N recovery data from Day 7 sampling of nitrapyrin and NBPT+DCD treated soil in the mountain series and thus was excluded.

All parameters measured or calculated (grain yield, stover or straw, % N concentration in the grain and stover or straw, total aboveground N content, and NUE) in the field trials were analyzed with PROC GLIMMIX in SAS version 9.3 (SAS Institute Inc., Cary, NC). Regions were separated for analysis given known variation in soil properties and climate and to test hypothesis that N-loss prevention amendments may perform differently at each region. An initial analysis of pooled data

confirmed that yields between sites for wheat (p < 0.0001) and maize (p < 0.0001) were significantly different. Region, year, N rate, and product were fixed effects, while replications within region-by-year combinations were random effects. Linear-plateau (LP) yield response models were generated with PROC GLM in SAS 9.3 to determine an AONR where an additional application of N fertilizer does not return a significant increase in yield. The LP model was not applied if LSMean yields did not change over N rates, or if yields continually increased with no evidence of a plateau (i.e., yields at highest N rate were statistically higher than yields at the previous rate). Additionally, when the LP model was used, the 95% confidence intervals for the estimated parameters were reviewed to check the goodness of model fit. The signifi-cance level applied to all data analysis was p < 0.05.

Due to a calculation error, the NBPT+DCD additive was applied at an incorrect rate in the 2013–2014 wheat trials and the data from those plots were deemed to be unusable. The NBPT+DCD additive was excluded from that year’s analysis, but NBPT+DCD amendment data was included in a separate analysis of the 2014–2015 data.

Fig. 3. Effect of product on (A) NH4–N concentration, (B) NO3–N concentration, and (C) percent N recovered from four N product (UAN, UAN+NBPT+DCD, UAN+nitrapyrin, or UAN+organo-Ca) applications in aerobic incubation of piedmont soil over an 82-d sampling period. UAN = urea ammonium nitrate, NBPT = N-(n-butyl)-thiophosphoric triamide, DCD = dicyandiamide, nitrapyrin +2-chloro-6-(trichloromethyl) pyridine, and organo-Ca = alkylarylpolyoxyethylene glycols Ca aminoethylpiperazine and Ca heteropolysaccharides. Bars denote standard deviation.

Agronomy Journa l • Volume 109, Issue 5 • 2017 1817

ReSULTS AnD DISCUSSIonLaboratory Incubation

At 0 d, N recovery from different product treatments ranged from 51 to 68% (Fig. 2C, 3C, and 4C). Ammonium N concen-trations generally increased from 0 to 3 d as the urea in UAN was converted to NH4, before steadily declining by 84 d (Fig. 2A, 3A, and 4A). Nitrate-N concentrations generally increased from 3 d as nitrification was occurring (Fig. 2B, 3B, and 4B).

In the coastal plain and mountain soils, nitrapyrin-amended UAN significantly retarded the conversion of NH4 better than other products, especially the organo-Ca or UAN on 14, 28, 42, 56, and 84 d (Fig. 2A and 4A). At 84 d, nitrapyrin coastal plain and mountain-treated soils had 58 and 47 mg kg–1 NH4–N remaining, respectively, while organo-Ca and UAN had essentially reached 0 NH4–N by 28 to 56 d. In the coastal plain soil, NBPT+DCD significantly retarded NH4 conversion to NO3 better than organo-Ca or UAN (Fig. 2A). A similar trend was observed in the mountain soil, where NBPT+DCD protected NH4–N significantly better than the organo-Ca and UAN throughout the experiment; UAN and organo-Ca had similar release curves except at 14 d (Fig. 4A).

Conversely, coastal plain and mountain soils treated with nitrapyrin had the lowest NO3–N concentrations over the sampling period (Fig. 2B and 4B). In the coastal plain soil, UAN was not significantly different from NBPT+DCD or organo-Ca treated soils, while in the mountain soil, NBPT+DCD appeared to delay the conversion of NH4–N to NO3–N on 14 and 28 d better than UAN alone or the organo-Ca, which essentially had the same release curves.

Overall, N recovery in the coastal plain and mountain soils increased from 0 to 3 d and remained fairly constant (Fig. 2C and 4C). In both soils, recovery percentages exceeding 100 were recorded, but are most likely due to the additive effects of sampling error. Cahill et al. (2010b) suggest a priming effect as a possible source of additional N in the study bags, but since soil organic matter content was very low and N was not limit-ing, this is unlikely to have played a significant role (Jenkinson et al., 1985; Chen et al., 2014).

The piedmont soils deviated from results obtained for the coastal plain and mountain soils (Fig. 3). At 3 d, NH4–N con-centration spiked for organo-Ca and UAN products, and then generally decreased similarly at each sampling date until reach-ing 0 mg kg–1 at 42 d (Fig. 3A). Soil treated with NBPT+DCD and nitrapyrin maintained constant NH4–N concentrations from 3 to 14d, with NBPT+DCD at a significantly higher concentration than nitrapyrin. Soil NO3–N concentrations initially decreased from the 33 to 42 mg kg–1 range on 0 d, representing NO3 immediately available with the addition of UAN, to <1 mg kg–1 by 3 d, then slowly increased (Fig. 3B). This dramatic drop may have been due to denitrification, which tends to be more dominant in fine-textured soils at higher moisture content (Bollmann and Conrad, 1998; Zhu et al., 2013). Nitrate concentrations for UAN and organo-Ca treated soils leveled off at 28 d to around 20 mg kg–1 and remained at that level until the end of the experiment, while nitrapyrin and NBPT+DCD N-loss prevention amendments remained lower and reached comparable levels at 56 d before peaking at the observed soil maximum NO3–N concentrations at 84 d of ~ 60 mg kg–1 (Fig. 3B). The decrease beginning at 28 d continued

Fig. 4. Effect of product on (A) NH4–N concentration, (B) NO3–N concentration, and (C) percent N recovered from four N product (UAN, UAN+NBPT+DCD, UAN+nitrapyrin, or UAN+organo-Ca) applications in aerobic incubation of mountain soil over an 82-d sampling period. UAN = urea ammonium nitrate, NBPT = N-(n-butyl)-thiophosphoric triamide, DCD = dicyandiamide, nitrapyrin +2-chloro-6-(trichloromethyl) pyridine, and organo-Ca = alkylarylpolyoxyethylene glycols Ca aminoethylpiperazine and Ca heteropolysaccharides. Dotted lines indicate missing data. Bars denote standard deviation.

1818 Agronomy Journa l • Volume 109, Issue 5 • 2017

to below 20 mg kg–1 by 84 d. In contrast to the other two soils, applied N recovery in the piedmont soil generally reached a maximum between 3 and 7 d and then decreased over the sam-pling period. By 84 d, only 22 to 46% of N was recovered. This likely reflects the initial loss of N through denitrification. If an average of 40 mg kg–1 of N was lost in the first 3 d, the remain-der of the total applied would fall into the observed recovery

range (Fig. 3C). Ensuring thorough mixture of fertilizer and N-loss prevention amendments, while maintaining constant moisture in the Lloyd clay loam soil was difficult and may have affected the results. Since the soil was sieved, structure was almost nonexistent and once water was added, pore space was likely limited.

Table 1. Results of ANOVA statistical analysis for maize N rate and product trials by parameter (grain and stover yield, nitrogen use ef-ficiency [NUE], grain and stover N content, and total aboveground N content) and by region (coastal plain, piedmont, and mountains) in North Carolina.

Parameter Year Product N Rate Year × Product Year × N Rate Product × N Rate Year × Product × N RateCoastal Plain Grain yield † ns‡ † ns * ns ns Stover yield † ns † ns ns ns ns NUE * ns ns ns ns ns ns Grain N † ns † ns ns ns ns Stover N † ns * * ns ns ns Total N * ns † ns ns ns ns

Piedmont Grain yield † ns † * * ns ns Stover yield † ns * ns ns ns ns NUE † ns ns ns ns ns ns Grain N † ns † ns † ns ns Stover N † ns † ns * ns ns Total N † ns † ns † ns ns

Mountains Grain yield † ns † ns † ns ns Stover yield † ns † ns * ns ns NUE † ns * ns † ns ns Grain N † ns † ns † ns ns Stover N † ns † ns ns ns ns Total N † ns † ns * ns ns

* p value of p < 0.05.† p value of < 0.0001.‡ ns indicates not significant (p > 0.05).

Fig. 5. Maize grain yield of year x product interaction at the piedmont sites. Products consisted of UAN, UAN and NBPT+DCD, UAN and Nitrapyrin, and UAN and Organo-Ca. UAN = urea ammonium nitrate, NBPT = N-(n-butyl)-thiophosphoric triamide, DCD = dicyandiamide, nitrapyrin +2-chloro-6-(trichloromethyl) pyridine, and organo-Ca = alkylarylpolyoxyethylene glycols Ca aminoethylpiperazine and Ca heteropolysaccharides. Significance of least-squares means at the 0.05 level indicated by different letters.

Agronomy Journa l • Volume 109, Issue 5 • 2017 1819

The effect of product transformations of N varied across the three soils, as both nitrapyrin and NBPT+DCD clearly showed efficacy in this mineralization study. Nitrapyrin appeared to effectively inhibit the nitrification activity in all three of the soils up to 42 d. The effectiveness of nitrapyrin in inhibiting nitrification was similar to that reported by Chen et al. (2010). The urease inhibitor component of NBPT+DCD did not sig-nificantly prevent the hydrolysis of urea, as indicated by NH4–N release patterns that are nearly identical to the other products. The nitrification inhibitor component of NBPT+DCD showed an effect in two of the three soils. The organo-Ca consistently followed the same N release patterns as UAN, and therefore was not different. The organo-Ca, which is claimed to open exchange sites for the binding of NH4, should theoretically perform well in clay soils with higher cation exchange capacity (CEC) where more exchange sites are available. This was not observed in the piedmont soil, with a CEC of 7.9, where favorable conditions were most likely to exist, or in the coastal plain fine loamy sand or the mountain silt loam textured soils with CECs ranging from 6.3 to 7.4. Additionally, the organo-Ca is a member of the adjuvant chemical family, and is more likely to influence N fertil-izer uptake when in contact with plant surfaces than through modification of soil N transformations. These results were simi-lar to Goos (2011).

Field ResearchMaize

The main effect of year was significant for all parameters at all regions, a result consistent with observed weather variations between years (Table 1). As expected over a wide range of N rates, the main effect of N rate was also significant for all param-eters in all regions, with the exception of apparent NUE at the coastal plain and piedmont sites. Also consistent with weather variations between years, the interaction effect of N rate × year was significant for grain yield at the coastal plain, and all param-eters at the other two regions with the exception of stover yield and NUE at the piedmont, and stover N in the mountains. It is important to note that the main effect of product, as well as N rate × product and year × product × N rate interactions were not significant for any parameter in any region (Table 1).

The only indication of a statistically significant product effect was in the year × product interaction observed for grain yield (p = 0.031) in the piedmont and stover N content (p = 0.006) in the coastal plain. The interaction effect in the pied-mont was largely influenced by drought-induced low yields in 2015, where there was no significant difference among products (Fig. 5); in 2014, nitrapyrin and the organo-Ca had signifi-cantly higher yields than UAN alone, while NBPT+DCD was not significantly different from the other products. In 2014 at the piedmont site, the separation of yield between UAN and the highest yielding product (organo-Ca) was 1216 kg ha–1 (Fig. 5).

Rainfall from the three regions for both years (Fig. 1) dem-onstrated greater rainfall amounts shortly after products were sidedressed in 2015 than in 2014, which in theory would have protected N and increased availability for growth and yield. Our incubation experiment using piedmont soil (Fig. 3) suggested that both nitrapyrin and NBPT+DCD protected NH4 for over 6 wk. If N leaching losses had occurred with either UAN or UAN+organo-Ca, yield losses might have occurred due to

reduced N availability. Since, however, the effect of product was only evident at Piedmont in 2014, it was the lack of rainfall in 2015 that had the greatest effect on yield rather than product and excess rainfall. The fact that product was not effective for coastal plain maize was not surprising in that rainfall amounts after sidedress were adequate (Fig. 1A and 1B). However, much greater rainfall immediately following sidedress N application at the mountain site in 2015 (Fig. 1E and 1F) could have increased N leaching losses where the N-loss amendments might have pro-tected N fertilizer and increased yield; yield differences, however, were only a function of N rate and not product.

At the coastal plain in 2014, grain yield increased with N rate up to 180 kg N ha–1 with a maximum yield of 13,457 kg ha–1 (Fig. 6A). In 2015, grain yield increased with N rate up to the highest application rate of 224 kg N ha–1, and maximum yields from both years were not statistically different from each other. There was a marked decrease in 2015 yields at the piedmont site due to prolonged periods without precipitation (Fig. 6B). Maximum yield in 2015 (6104 kg ha–1) was not significantly

Fig. 6. Maize grain yield of by year × N rate interaction at the (A) coastal plain, (B) piedmont, and (C) mountain sites. Significance of least-squares means at the 0.05 level indicated by different letters.

1820 Agronomy Journa l • Volume 109, Issue 5 • 2017

Table 2. Maize N content for grain, stover, and total aboveground biomass. There was significant year × N rate interactions by region (piedmont and mountains) in North Carolina.

Region YearN Rate, kg ha–1

0 45 90 135 180 224Grain N content, kg ha–1

Piedmont 2014 53c,d† 73f 104g 129h 137h,i 144i2015 26a 38b 46b,c 59d 61d,e 71e,f

Mountains 2014 71b 115c 148d 156d,e 170e 165e2015 36a 39a 83b 115c 124c 149d

Stover N content, kg ha–1

Piedmont 2014 26d,e 30e 35f 39f 10f 53g2015 11a 12a 16a,b 19b,c 20b,c 22c,d

Total aboveground N content, kg ha–1

Piedmont 2014 79c 103d 140e 168f 178f 197g2015 37a 50b 60b 78c 81c 93d

Mountains 2014 102b 156c 195d 204d,e 219e 217e2015 54a 60a 114b 157c 162c 186d

† Letters denote significance at p < 0.05 level of year × N rate interaction.

Fig. 7. Maize grain yield linear-plateau model for (A) coastal plain, (B) piedmont, and (C) mountain. Arrow indicates realistic yield expectation (RYE) recommended N rate for the soil series that the crop was produced on.

Agronomy Journa l • Volume 109, Issue 5 • 2017 1821

different from the previous year’s yield in plots receiving 0 kg N ha–1. In 2014, grain yield was nearly twice as high, plateauing at the 135 kg N ha–1 rate with a maximum yield of 11,004 kg ha–1. Grain yields at the mountain site plateaued at 15,929 kg ha–1 at the 90 kg N ha–1 rate in 2014 (Fig. 6C). In 2015, the highest yields at 180 and 224 kg N ha–1 were not statis-tically different. In both years, yields were exceptional, likely in response to well-timed rainfall and a cooler mountain climate.

Grain N content increased with N rate at each region, but was lower overall in 2015 at the piedmont and mountain sites (Table 2). Stover N content was significant for the year × N rate interaction at the piedmont, and by the main effects of year and N rate at the coastal plain and mountain sites. Total N content, of which the grain makes up a larger proportion, predictably followed the same significance trends as grain N at each site.

Apparent NUE was significant by year only at the coastal plain (p = 0.04) and piedmont (p < 0.0001, Table 1). In both regions, average apparent NUE was higher in 2014 than 2015. Average NUE was 67 and 61% at the coastal plain, and 59 and 28% at the piedmont in 2014 and 2015, respectively (data not shown). In the mountains, a year × N rate interaction was observed (p < 0.0001). Apparent NUE values (~75%) were not significant between years at N rates of 135 kg ha–1 and higher, but varied at lower rates (data not shown).

Grain yield data from each region and year was analyzed with a LP model to identify the AONR after it was determined that the LP model performed as well or better than the linear quadratic and quadratic (data not shown). The LP models were then compared to current N rate guidelines (RYE) for each physiographic region of study. The RYE recommended N rate was slightly higher than the observed AONR at the coastal plain site in 2014 (Fig. 7A), though the observed yields were 3100 kg ha–1 greater than predicted RYE yields (North Carolina Nutrient Management Workgroup, 2003). The yield equation for x < 150 kg N ha–1 was y = 5082 + 55.5x (r2 = 0.92 and p < 0.0001) and for x > 150 kg N ha–1 was y = 13,416 kg ha–1. In 2015, yields at each N rate were different

from the previous N rate, so the LP model was not fit as the response was linear. At the piedmont site, the RYE N rate was in excess compared to LP optimum N rate by 32 kg N ha–1 in 2014 (Fig. 7B). The yield equation in 2014 for x < 117 kg N ha–1 was y = 6029 + 43.0x (r2 = 0.64 and p < 0.0001) and for x > 117 kg N ha–1 was y = 11,071 kg ha–1, while in 2015 the yield equation for x < 148 kg N ha–1 was y = 2803 + 20.1x (r2 = 0.44 and p < 0.0001) and for x > 148 kg N ha–1 was y = 5790 kg ha–1. The RYE N rate was sufficient in 2015, though yields were extremely low due to drought, with the optimum yield reaching only 5704 kg ha–1. At the mountain site, the RYE recommended rate of 185 kg N ha–1 was well above the LP AONR of 66 kg N ha–1 in 2014 (Fig. 7C). This 121 kg N ha–1 discrepancy reflects a decision made by the North Carolina Nutrient Management Group in setting the RYE rate for Comus (historically high-yielding floodplain soil) higher than the LP average optimum N rate of around 80 kg N ha–1 (Rajkovich et al., 2015). The yield equation for x < 67 kg N ha–1 was y = 9547 + 101.9x (r2 = 0.64 and p < 0.0001) and for x > 67 kg N ha–1 was y = 16,337 kg ha–1. In 2015, the mountain trials were conducted on an upland soil and the optimum N rate, as predicted by the LP, was only 13 kg N ha–1 greater than the RYE rate. The yield equation for x < 191 kg N ha–1 was y = 3503 + 71.4x (r2 = 0.82 and p < 0.0001) and for x > 191 kg N ha–1 was y = 17,129 kg ha–1. With the exception of the coastal plain site in 2015 and moun-tain site in 2014, the difference between the AONR and RYE rate averaged 3 kg N ha–1, indicating that current RYE recom-mendations align with field observations.

wheatThe main effect of year was significant for all parameters at

both regions with the exception of apparent NUE at the coastal plain and grain yield at the piedmont (Table 3). The main effect of N rate was also highly significant, with the only nonsignificant parameter being apparent NUE at both regions. Year × N rate interactions were significant for all parameters except N straw

Table 3. Results of ANOVA statistical analysis for wheat N rate and product trials by parameter (grain and straw yield, nitrogen use efficiency (NUE), grain and straw N content, and total aboveground N content) and region (coastal plain and piedmont) in North Carolina.

Parameter Year Product N Rate Year × Product Year × N Rate Product × N Rate Year × Product × N RateCoastal Plain Grain yield † ns‡ † ns * ns ns Straw yield † ns † ns * ns ns NUE ns ns ns ns * ns ns Grain N † ns † ns * ns ns Straw N † ns † ns ns ns ns Total N † * † ns * ns ns

Piedmont Grain yield ns ns † ns * ns ns Straw yield † ns * ns * ns ns NUE † ns ns ns * ns ns Grain N * ns † ns * ns ns Straw N † ns † ns * ns ns Total N † ns † ns * ns ns

* p value of p < 0.05.† p value of < 0.0001.‡ ns indicates not significant (p > 0.05).

1822 Agronomy Journa l • Volume 109, Issue 5 • 2017

content at the coastal plain region while the year × N rate interac-tions were significant for all parameters at the piedmont (Table 3). Product was not significant as a main effect or in a two- or three-way interaction. Cumulative rainfall for the 28 d preceding spring N fertilization (early March) was 133, 98, 81, and 40 mm for wheat in the coastal plain (2013–2014 and 2014–2015) and piedmont (2013–2014 and 2014–2015), respectively. Despite an almost threefold difference in rainfall over years and sites, product did not affect yield. Although rainfall was particularly high at the coastal plain in 2013–2014 (133 mm), there was no discrimina-tion in yield between products.

In the coastal plain, grain yields in both years continued to increase with N rate (i.e., did not reach a plateau) with highest yields (6882 kg ha–1 in 2013–2014 and 8070 kg ha–1 in 2014–2015) achieved at the 180 kg N ha–1 spring application rate (Fig. 8A). The 2014–2015 crop yielded higher overall, reaching the previous year’s maximum yield at the 90 kg N ha–1 rate. The significant difference at each N rate between years was likely influenced by the variation in soil types between years. Trials were conducted on a Johns sandy loam in 2013–2014, with very low organic matter (12 g kg–1) compared to the Portsmouth sandy loam (120 g kg–1) utilized the following year. Grain yield did not reach a plateau within the range of included N rates in either year, and although there was evidence of continued yield response at the 180 kg N ha–1 application rate, the average total N application rate for wheat falls near the North Carolina recommended rate of 134 kg N ha–1 (Scharf and Alley, 1993;

Fig. 8. Wheat grain yield of year × N rate interactions by site (A) coastal plain and (B) piedmont. Significance of least-squares means at the 0.05 level indicated by different letters.

Fig. 9. Wheat grain yield linear-plateau model by year for the piedmont region. Arrow indicates realistic yield expectation (RYE) recommended N rate for the soil series that the crop was produced on.

Table 4. Wheat grain, straw, and total aboveground N content. There was significant year × N rate interactions by region (coastal plain and piedmont) in North Carolina.

Region YearSpring N rate, kg ha–1

0 45 90 135 180Grain N content, kg ha–1

Coastal Plain 2013–2014 40a† 59b 90d 112e,f 134g2014–2015 77c 104e 120f 133g 156h

Piedmont 2013–2014 73b 81b,c 109d 119e 122e2014–2015 63a 88c 107d 121e 138f

Straw N content, kg ha–1

Piedmont 2013–2014 7a 9a 14b 17b 14b2014–2015 16b 24c 26c,d 30d 38e

Total aboveground N content, kg ha–1

Coastal Plain 2013–2014 47a 70b 107d 132e 160g2014–2015 93c 127e 147f 163g 195h

Piedmont 2013–2014 81a 89a 122b 136c 136c2014–2015 79a 112b 133c 151d 177e

† Letters denote significance at p < 0.05 level of year × N rate interaction.

Agronomy Journa l • Volume 109, Issue 5 • 2017 1823

Scharf et al., 1993; Weisz et al., 2001), with the risks of lodg-ing increasing greatly beyond this rate (Scharf and Alley, 1993; Alley et al., 1996; Weisz, 2015).

At the piedmont sites, grain yields between years were only significantly different at the 90 kg N ha–1 rate (Fig. 8B). However, the data did meet the criteria for analysis with a LP model, which indicated optimum yield was reached at 113 kg N ha–1 in 2013–2014 and at 130 kg N ha–1 in 2014–2015 (Fig. 9). The yield equation in 2013–2014 for x < 113 kg N ha–1 was y = 4553 + 19.5x (r2 = 0.70 and p < 0.0001) and for x > 113 kg N ha–1 was y = 6746 kg ha–1, while in 2014–2015 the yield equation for x < 130 kg N ha–1 was y = 4775 + 14.2x (r2 = 0.65 and p < 0.0001) and for x > 148 kg N ha–1 was y = 6622 kg ha–1. The observed optimum N rates were 2 kg N ha–1 and 20 kg N ha–1 greater than RYE N rates in 2014 and 2015, respectively (Fig. 9). The optimum yields at these rates were nearly twice what the RYE database predicted (3283 kg ha–1), indicating wheat yield values, as well as N factors (kg grain yield per kg N fertilizer), in the RYE database may need to be updated.

Each site exhibited an increase in grain N content with increasing N rate (Table 4). Total wheat N content generally correlated with grain and straw yields. After an initial peak, a general decrease in apparent NUE was observed as spring N rate increased at both regions in both years (data not shown). However, at the piedmont site in 2013–2014 yr, NUE was much lower and increased from 24% at 45 kg N ha–1 to 47% at 90 kg N ha–1, before tapering off to 31% at the highest spring N rate. Both regions displayed NUE values that are typical to wheat (40–80%) (Thomason et al., 2002; Johnson and Raun, 2003).

A separate analysis of the effectiveness of NBPT+DCD com-pared to UAN, the organo-Ca, and nitrapyrin included only the 2 site-years in which it was applied: the coastal plain and piedmont sites in 2014–2015. Product, as a main effect or as an interaction, did not have a significant effect on grain yield, straw yield, apparent NUE, straw N content, or total N content at the piedmont. Only total N content was significant by prod-uct (p = 0.0402) in the coastal plain; all other parameters were nonsignificant for product. The N content of wheat grown in the NBPT+DCD treated plots was greater (120 kg N ha–1) rela-tive to the UAN treated plots (113 kg N ha–1) at the coastal plain.

ConCLUSIonSThe potential evidence of delayed nitrification observed with

nitrapyrin and NBPT+DCD N-loss prevention amendments in the incubation experiment did not correlate well with the results of the field trials of these same products in three physio-graphic regions (coastal plain, piedmont, and mountains) and with two crops (maize and wheat). Nitrapyrin, which has the most extensive history of inclusion in research trials, signifi-cantly delayed nitrification up to 84 d in the coastal plain and mountain soils, while both NBPT+DCD and nitrapyrin sig-nificantly delayed nitrification up to 84 d in the piedmont soil during the incubation study. Under field conditions applica-tion of N-loss prevention amendments (UAN+NBPT+DCD, UAN+nitrapyrin, or UAN+organo-Ca) did not significantly affect maize or wheat grain or stover/straw yield relative to UAN in 9 of 10 site-years; only N rate was significant. The slight advantage observed with some N-loss prevention amend-ments in 1 out of 10 site-years may not justify their additional

expense, as typically there was no increase in yield. Finally, the linear-plateau models for maize indicated that optimum N rates were very similar to the RYE N rates with two exceptions. For wheat grown in the coastal plain, yields never plateaued over the N fertilizer range tested, whereas piedmont N rates, as determined by linear plateau model, were very close to the RYE database N rates. The N rate analysis from this trial adds to and confirms the North Carolina RYE database for maize produced on some soils; the wheat database yields, N rates, and N factor may need to be scrutinized more carefully.

ACKnowLeDGMenTS

This research was generously supported by the North Carolina Agriculture Foundation, North Carolina State University Department of Soil Science, Small Grains Association of North Carolina, AgExplore International, and Koch Industries. Product donations from Dow AgroSciences were greatly appreciated. Special thanks to Wes Childres for technical field assistance. We want to sincerely thank the reviewers who provided many useful suggestions and made this a better paper.

ReFeRenCeS

AgExplore International. 2015. NZone Booklet. AgExplore Int., Parma, MO. http://www.mncag.com/LabelsMSDS/NZone_Booklet.pdf (accessed 10 Jan. 2016).

Alley, M.M., P. Scharf, D.E. Brann, W.E. Baethgen, and J.L. Ham-mons. 1996. Nitrogen fertilization for winter wheat: Principles and recommendations. Publ. no. 424-206. Virginia Polytechnic and State Univ., Blacksburg.

Bollmann, A., and R. Conrad. 1998. Influence of O2 avail-ability on NO and N2O release by nitrification and deni-trification in soils. Glob. Change Biol. 4(4):387–396. doi:10.1046/j.1365-2486.1998.00161.x

Burzaco, J.P., I.a. Ciampitti, and T.J. Vyn. 2014. Nitrapyrin impacts on maize yield and nitrogen use efficiency with spring-applied nitrogen: Field studies vs. meta-analysis comparison. Agron. J. 106(2):753–760. doi:10.2134/agronj2013.0043

Cahill, S., D. Osmond, D. Israel, and D. Osmond. 2010a. Nitrogen release from coated urea fertilizers in different soils. Commun. Soil Sci. Plant Anal. 41(10):1245–1256. doi:10.1080/00103621003721437

Cahill, S., D. Osmond, R. Weisz, R. Heiniger, R. Weisz, and D. Osmond. 2010b. Evaluation of alternative nitrogen fertilizers for corn and winter wheat production. Agron. J. 102(4):1226–1236. doi:10.2134/agronj2010.0095

Chen, R., M. Senbayram, S. Blagodatsky, O. Myachina, K. Dittert, X. Lin et al. 2014. Soil C and N availability determine the priming effect: Microbial N mining and stoichiometric decomposition theories. Glob. Change Biol. 20(7):2356–2367. doi:10.1111/gcb.12475

Chen, D., H.C. Suter, A. Islam, and R. Edis. 2010. Influence of nitrification inhibitors on nitrification and nitrous oxide (N2O) emission from a clay loam soil fertilized with urea. Soil Biol. Biochem. 42(4):660–664. doi:10.1016/j.soilbio.2009.12.014

Dane, J.H., and J.W. Hopmans. 2002. Water retention and storage- Laboratory methods. In: J.H. Dane and G.C. Topp, editors, Methods of soil analysis-Part IV, Physical methods. SSSA, Madi-son, WI. p. 675–720.

David, M.B., L.E. Gentry, D.A. Kovacic, and K.M. Smith. 1997. Nitrogen balance in and export from an agricultural water-shed. J. Environ. Qual. 26(4):1038–1048. doi:10.2134/jeq1997.00472425002600040015x

1824 Agronomy Journa l • Volume 109, Issue 5 • 2017

Espindula, M.C., V.S. Rocha, M.A. de Souza, M. Campanharo, and G. de Sousa Paula. 2013. Rates of urea with or without urease inhibitor for topdressing wheat. Chil. J. Agric. Res. 73(2):160–167.

Goos, R.J. 2011. Nitrogen fertilizer additives, which ones work? North Central Extension-Industry Soil Fertility Conference, Vol. 27, Des Moines, IA. 16–17 Nov. 2011.

Gordon, A.M. 1988. Use of polyethylene bags and films in soil incubation studies. Soil Sci. Soc. Am. J. 52(5):1519–1520. doi:10.2136/sssaj1988.03615995005200050063x

International Plant Nutrition Institute. 2012. 4R Plant nutrition manual. 1st ed. Int. Plant Nutrition Inst., Peachtree Corners, GA.

Jaynes, D.B., T.S. Colvin, D.L. Karlen, C.A. Cambardella, and D.W. Meek. 2001. Nitrate loss in subsurface drainage as affected by nitrogen fertilizer rate. J. Environ. Qual. 30(4):1305–1314. doi:10.2134/jeq2001.3041305x

Jenkinson, D.S., R.H. Fox, and J.H. Rayner. 1985. Interac-tions between fertilizer nitrogen and soil nitrogen- The soil-called “priming” effect. J. Soil Sci. 36:425–444. doi:10.1111/j.1365-2389.1985.tb00348.x

Johnson, G.V., and W.R. Raun. 2003. Nitrogen response index as a guide to fertilizer management. J. Plant Nutr. 26(2):249–262. doi:10.1081/PLN-120017134

Kronvang, B., J.P. Jensen, C.C. Hoffmann, and P. Boers. 2001. Nitro-gen in the environment: Sources, problems and management. Elsevier, Amsterdam, the Netherlands.

Mitchell, C.C., and D.L. Osmond. 2012. New technology and alterna-tive nitrogen sources for crops in the southern US. South. Coop. Ser. Bull. no. 416-0:1–26. Alabama Agric. Exp. Stn., Auburn.

Natural Resources Conservation Service. 2005. 590 Nutrient man-agement standard. NRCS, Washington, DC. https://www.nrcs.usda.gov/Internet/FSE_DOCUMENTS/nrcs143_022228.pdf (accessed 29 June 2017).

North Carolina Nutrient Management Workgroup. 2003. Realistic yields and nitrogen application factors for North Carolina crops. North Carolina State Univ., Raleigh. http://go.ncsu.edu/yields (accessed 23 Nov. 2016).

Rajkovich, S.R., C.R. Crozier, T.J. Smyth, D. Crouse, and D.L. Osmond. 2015. Updating North Carolina corn yields and nitrogen recommendations to match current production prac-tices and new hybrids. Crop. Forage, Turfgrass Manage. 1(1). doi:10.2134/cftm2014.0085

Randall, G.W., and M.J. Goss. 2001. Nitrogen in the environment: Sources, problems and management. Elsevier, Amsterdam, the Netherlands.

Scharf, P.C., and M.M. Alley. 1993. Spring nitrogen on winter wheat: II.A flexible multicomponent rate recommendation system. Agron. J. 85(6):1186–1192. doi:10.2134/agronj1993.00021962008500060018x

Scharf, P.C., M.M. Alley, and Y.Z. Lei. 1993. Spring nitrogen on winter wheat: I. Farmer-field validation of tissue test-based rate recommendations. Agron. J. 85(6):1181–1186. doi:10.2134/agronj1993.00021962008500060017x

Sistani, K.R., M. Jn-Baptiste, and J.R. Simmons. 2014. Corn response to enhanced-efficiency nitrogen fertilizers and poultry litter. Agron. J. 106(2):761–770. doi:10.2134/agronj2013.0087

Sogbedji, J.M., D.R. Bouldin, W.J. Cox, H.M. van Es, and S.D. Klausner. 2001. Spatial and temporal processes affecting nitro-gen availability at the landscape scale. Soil Tillage Res. 58(3-4):233–244. doi:10.1016/S0167-1987(00)00171-9

Spalding, R.F., and M.E. Exner. 1993. Occurrence of nitrate in groundwater—A review. J. Environ. Qual. 22(3):392–402.. doi:10.2134/jeq1993.00472425002200030002x

The Fertilizer Institute. 2011. 4R Nutrient stewardship system. The Fertilizer Inst. http://www.nutrientstewardship.com/ (accessed 23 Nov. 2016).

Thomason, W.E., W.R. Raun, G.V. Johnson, K.W. Freeman, K.J. Wynn, and R.W. Mullen. 2002. Production system techniques to increase nitrogen use efficiency in winter wheat. J. Plant Nutr. 25(10):2261–2283. doi:10.1081/PLN-120014074

Weisz, R., editor. 2015. North Carolina small grains production. North Carolina Coop. Ext. Serv., Raleigh. http://www.small-grains.ncsu.edu/ (accessed 23 Nov. 2016).

Weisz, R., C.R. Crozier, and R.W. Heiniger. 2001. Optimizing nitro-gen application timing in no-till soft red winter wheat. Agron. J. 93(2):435–442. doi:10.2134/agronj2001.932435x

Zhu, X., M. Burger, T.A. Doane, and W.R. Horwath. 2013. Ammo-nia oxidation pathways and nitrifier denitrification are signifi-cant sources of N2O and NO under low oxygen availability. Proc. Natl. Acad. Sci. USA 110(16):6328–6333. doi:10.1073/pnas.1219993110