In the organic field, five treatments were designed to compare: (1) the effect of cover crop species on N loss and (2) the effect of fertilizer scheme on N loss within systems using a hairy vetch cover crop (Table 1, Figure S1). For the first comparison, three cover crop species were planted: hairy vetch (Vicia villosa) seeded at 33.6 kg ha⁻¹, winter wheat (Triticum aestivum) seeded at 134.5 kg ha⁻¹, and a mix (“bi-culture”) of hairy vetch and winter wheat seeded at 22.4 and 67.3 kg ha⁻¹, respectively. To capture the effect of cover crop alone, each of the three cover crop treatments were compared under no fertilizer conditions.

To quantify the effect of fertilizer scheme, two organic N fertilizer treatments applied to a hairy vetch cover crop treatment were compared to a hairy vetch cover crop with no fertilizer added (the same hairy vetch plots used in the cover crop species comparison above). At the time of cover crop termination, hairy vetch biomass was sampled from all plots and was found to contain 70–80 kg N ha⁻¹. For the unfertilized treatment, the vetch cover crop N was the only source of N added (averaging 78 kg N ha⁻¹). For the organic N fertilizer treatment, 168 kg N ha⁻¹ of blended, pelletized, animal byproduct fertilizer (NatureSafe 13-0-0, Griffin Industries LLC, Cold Spring, KY) was added, giving a total of 242 kg N ha⁻¹ (74 kg N ha⁻¹ from the vetch+168 from the NatureSafe). Due to the atypically cold winter, hairy vetch biomass was low at the site (1,741 kg ha⁻¹ dry wt); therefore, for the fertilizer N-credit treatment, additional vetch biomass from an adjacent field was added to these plots to better mimic a more typical year (Smith et al., 1987; Cline and Silvernail, 2002). Vetch biomass sampling after this addition indicated there was 176 kg N ha⁻¹ present. To bring the level of added N in the fertilizer N-credit treatment in line with that of the organic N treatment (Sarrantonio, 2012), an additional 56 kg N ha⁻¹ of the same organic fertilizer was added. Therefore, the fertilizer N-credit treatment received 232 kg N ha⁻¹ in total (Table 1).

The cover crops were terminated via flail mowing on 20 May 2014 (when hairy vetch was at 50% flowering). The corn variety used was 71T77cnv from BlueRiver Organics (114 day, untreated conventional, non-GMO). Between row weed pressure was managed as needed with four mowing events using a BCS Model 853 walk-behind tractor (BCS America, LLC, Portland, OR) with a flail-mower attachment on 13 and 25 June, 10 July, and 22 August 2014. Within-row weed pressure was managed with a single hand cultivation event on 2 July 2014. No weed biomass was removed from the plots. Insect pests were managed with three applications of Bacillus thuringiensis kurstaki (Javelin®, Certis, USA, Columbia, MD) and spinosad (Entrust®, Dow Agrisciences, Indianapolis, IN) using a spreader sticker (Nu-film 17, Miller Chemical and Fertilizer Corporation, Hanover, PA) (Table S2).

In the conventional field, three treatments were designed to compare the effect of two different fertilizers on N loss in conventional systems with a winter wheat (T. aestivum) cover crop seeded at 134.5 kg ha⁻¹ (Table 1, Figure S1). The N treatments were 0 N, 168 kg ha⁻¹ N applied as urea 40-0-0 with a urease inhibitor (AgrotainUltra®), and 168 kg ha⁻¹ N applied as the same fertilizer used in the organic treatments. Animal derived N sources such as manure are often used as an alternative to ammonium nitrate or urea, but in order to eliminate adding additional phosphorous and/or potassium a 13-0-0 packaged source of organic fertilizer was used in this study (NatureSafe 13-0-0, Griffin Industries LLC, Cold Spring, KY). The cover crop was terminated via flail mowing followed by glyphosate application (Roundup Pro; Monsanto, St. Louis, MO, U.S.A.) on 20 May 2014. Corn was planted and fertilizer applied on 28 May 2014. The corn variety used in the conventional field was REV24BHR93 from Terral Seed, Inc. Weeds were managed with one additional application of glyphosate on 10 July 2014 and no insect pest management was deemed necessary upon weekly scouting (Table S2).

The static chamber method was employed to measure gaseous emissions (Parkin and Venterea, 2010). In each of the 24 plots (8 treatments, 3 replicates), a rectangular stainless steel anchor (16.35 × 52.70 × 15.24 cm) was inserted into the soil so that the top was nearly flush with the soil surface. Anchors were inserted at random locations in each plot 10 days after cover crop seed was broadcast in the fall. Anchors were removed prior to cover crop termination in order to avoid damage from heavy machinery. The day of corn planting, the anchors were re-inserted into the plots and were placed perpendicular to the corn rows so that they were between two corn plants and spanning soil surface area both within and between rows.

To measure nitrous oxide (N₂O) and ammonia (NH₃), a “cap” made from an identical stainless steel chamber, equipped with a vent tube and lined with Teflon© tape (Bytac©, Saint Gobain Performance Plastics, Paris, France) was attached to the anchor to create a sealed chamber. The chamber was connected to a photoacoustic spectroscopy gas analyzer (Innova Air Tech Instruments Model 1412, Ballerup, Denmark) via Teflon© tubing. Measurements were taken continuously for 10 min on the days of sampling and NH₃ and N₂O concentrations (mg L⁻¹) were recorded simultaneously. Gaseous flux was calculated using the equations described by Iqbal et al. (2013). Annual fluxes were estimated by interpolating linearly between sampling dates, calculating the area under the curve using the trapezoidal rule. During sampling periods, additional environmental parameters were also recorded, including soil moisture at 5 cm depth (DELTA-T HH2 moisture meter using a ML2x 6 cm theta probe, Delta-T Devices, Cambridge, England), soil temperature at 5 cm depth, and ambient air temperature (Taylor Digital Pocket Thermometer, Model 9878E, Taylor Precision Products, Oak Brook, IL).

Measurements commenced on 28 October 2013 and continued until 29 October 2014. Sampling intensity varied throughout the experiment, with increased intensity during periods when fluxes were expected to be affected by management practices. During the cover crop growing season (28 October 2013–20 May 2014) measurements were taken twice a month (cover crop growth period). Following corn planting and fertilization (28 May 2014), measurements were taken daily for a week and then every other day until 12 June 2014, on which date fluxes from fertilized treatments appeared similar to those from unfertilized treatments (post fertilization intensive period). During the rest of the corn-growing season, measurements were taken every seven to 10 days, returning to a twice a month sampling scheme in October (post fertilization period). Measurements were taken between 10:00 a.m. and 3:00 p.m.

Ion exchange resin lysimeters were used to measure NO₃-N and NH₄-N leachate, after Susfalk and Johnson (2002). Cation and anion exchange resins (25 g) (LANXESS NM-60, Klenzoid Equipment Company, Wayne, PA) were placed between Nitex® nylon cloth and sand layers that were enclosed in polyvinyl chloride tubes 5 cm in diameter. Lysimeters were deployed during cover crop growth from 10 October 2013 until 16 May 2014 and during corn growth from 16 May 2014 until 20 October 2014. Two lysimeters were installed at 40 cm depth in each plot under an undisturbed soil profile, ensuring good contact with the soil. The first round was placed randomly in each plot, but during the corn growing season, one was placed within and one between the corn rows.

When resin lysimeters were collected, inorganic ammonium and nitrate was extracted by shaking the resin in 100 ml of 2 M KCl for 1 h. Ammonium concentration was determined using a modification of the Berthelot reaction (Chaney and Marbach, 1962) and nitrate quantified via reduction to nitrite using a copperized cadmium reduction microplate device (ParaTechs Co., Lexington, KY) as described by Crutchfield and Grove (2011). Colorimetric analysis was conducted using a microplate reader (Molecular Devices, VERSAmax, Sunnyvale, CA).

To monitor soil N dynamics, cation and anion exchange resin bags were made with 10 g of resin (LANXESS NM-60, Klenzoid Equipment Company, Wayne, PA) tied in a porous material (Gibson et al., 1985). Each month from November 2013 to November 2014, three resin bags per plot were inserted at 15 cm depth and those from the prior month were removed for analysis of inorganic ammonium and nitrate. There were two instances (January to mid-February and mid-February to early April) when the resin bags were deployed for 6 weeks as opposed to 1 month, as the ground was frozen and it was not possible to remove the resin bags. After removal from the field, resin bags were rinsed with deionized water and extracted with 40 mL 2.0 M KCl. Extract was stored overnight at 4°C and analyzed colorimetrically as described above.

Beginning in May 2014, cover crop biomass residue was sampled monthly through September. The first sampling was conducted the day of cover crop termination. During the months of June through September, the decomposing cover crop residue on the soil surface was collected. Two samples were collected randomly from each plot (25 × 25 cm quadrat), were dried (55°C for 48 h), and then weighed. The same location was never sampled twice. Samples were processed on a grinding mill to pass through a 1 mm sieve (Cyclotec 1093, FOSS™, Eden Prairie, MN) and sub-samples were then ground on a ball grinder (Cianflone Scientific Instrument Corporation, Pittsburgh, PA) or a jar-mill (U.S. Stoneware, East Palestine, OH). Cover crop biomass was analyzed for C and N content via flame combustion (Flash EA 1112 elemental analyzer, CE Elantech Inc., Lakewood, CA).

Corn was harvested by hand on 6 October 2014. Yield was calculated from the grain produced by the inner three rows of each plot. Corn was dried at 55°C for 48 h and then allowed to air-dry for 5 weeks. Corn was shelled, weighed, and a grain sub-sample was collected and dried (55°C for 24 h) to correct for moisture content.

Biomass of the cover crops at termination was significantly greater in the vetch-wheat bi-culture and wheat plots than the hairy vetch alone (p < 0.0096, p < 0.0118, respectively), with dry weights of 4,968 (bi-culture), 3,590 (wheat), and 2,791 (hairy vetch) kg ha⁻¹. Nitrogen content of biomass followed similar trends and was 100 kg N ha⁻¹ in the bi-culture, 78 kg N ha⁻¹ in hairy vetch alone, and 27 kg N ha⁻¹ in the wheat (Figure 1A). During cover crop growth (October–May), soil moisture varied from a low of 13% on 28 October 2013 to a high of 35% on 18 February 2014, but was comparable across treatments, differing slightly (<6%) at only three time points (Figure S2). Similarly, overall soil temperature did not vary substantially across treatments (Figure S3).

Soil resin NO₃-N concentrations were highest in the hairy vetch treatment prior to cover crop termination in January/February, but NH₄-N concentrations were similar across treatments (Figures 2A,B). Nitrogen loss measured as NO₃-N leachate differed between cover crop species, with hairy vetch > bi-culture > wheat (Table 2). Though N₂O-N fluxes were slightly higher in the wheat treatment on 3 March and 17 and 29 April 2014 (Figure 3A), linear contrasts indicated there were no significant differences in either N₂O-N or NH₃-N loss across treatments when analyzed across the entire cover crop growing season (Table 2).

During cover crop growth, low soil temperatures and N availability likely contributed to low gaseous loss for all treatments (Parsons et al., 1991). Thus, the primary N loss pathway during this time was NO₃-N leaching. Our result of hairy vetch having the greatest NO₃-N loss via leachate is consistent with other work that has evaluated legume vs. grass cover crop effects on N leaching (McCracken et al., 1994; Ranells and Wagger, 1997). The low biomass produced by the hairy vetch treatment (only 56–78% of the biomass of the other treatments) and the release of fixed N to the soil through rhizodeposition (Fustec et al., 2011) likely contributed to greater leaching. The fact that the bi-culture had the greatest biomass accumulation, highest N content, and lower NO₃-N leachate than the hairy vetch treatment is consistent with other studies that have shown when hairy vetch is grown in bi-culture with wheat, biomass production and the winter hardiness of the legume is increased (Teasdale and Abdul-Baki, 1998; Snapp et al., 2005). For organic production, higher cover crop biomass and N content in the bi-culture may provide a slow-release source of N for the subsequent corn crop.

Post-termination of the cover crops (May–October), N content and %N of hairy vetch and bi-culture residue declined, indicating N mineralization during decomposition, whereas the N content of the wheat residue stayed fairly constant at a low level and the %N slightly increased, indicating N immobilization (Figures 1A,B). These trends reflected significant differences in residue C:N (wheat > bi-culture > vetch; Figure 1C), and likely contributed to higher NO₃-N levels measured in August soil resin bags (Figure 2A), higher NO₃-N leachate values (Table 2), higher N₂O-N emissions during the post-fertilization intensive measurement period (Figure 3B), and higher corn yield (Table 2) in hairy vetch and bi-culture treatments vs. wheat. No significant differences between treatments were observed for total NH₃-N or N₂O-N emissions (Table 2), for N₂O-N emissions during the post-fertilization period (Figure 3C), or for NH₃-N emissions during specific measurement periods (Figures 4A–C).

During the corn growing season, N leachate concentrations were lower than during the cover crop growing season, but N₂O-N emissions were greater. A combination of increased N availability in the system after cover crop termination, warmer soil temperatures, and adequate soil moisture (Figures S2, S3) were likely causal (Colbourn, 1993; DeKlein and VanLogtestijn, 1996). Similar to the research of Wisal et al. (2011), who found that soils amended with low C:N ratio residues undergo decreased N immobilization and increased N₂O-N emissions, cover crop C:N appeared to influence gaseous N emissions. N₂O-N fluxes increased after cover crop termination for the vetch (from 57 to 232 μg N m⁻² h⁻¹) and bi-culture treatments (62–155 μg N m⁻² h⁻¹), but decreased in the wheat system (from 70 to 46 μg N m⁻² h⁻¹), as the high C:N wheat cover crop biomass residue induced N immobilization rather than mineralization.

Although the wheat cover crop had the lowest N loss (primarily due to low N leaching), N immobilization into the residue after termination (Wagger, 1989; Wyland et al., 1995; McSwiney et al., 2010) resulted in significantly lower corn yields. Though capable of reducing N loss, systems cover cropped with wheat would require significant quantities of N fertilizer to obtain the yields necessary for a viable production system, perhaps negating its N loss benefit. The fact that the bi-culture treatment had less NO₃-N leachate loss during cover crop growth and similar post-termination leaching, gaseous N loss, and corn yield compared to the hairy vetch treatment suggests that legume-grass cover crop mixtures have potential for reducing N loss in organic conservation agricultural systems. Further cover crop management research is essential as all treatments in this system were unfertilized and most likely N limited, as corn yields were half of those produced in the fertilized organic treatments also evaluated in this experiment. With higher fertilizer N applications, the effect of cover crop species on N loss pathways and yield may change (Alvarez et al., 2017). Additionally, longer-term studies are needed as soil properties are affected by the addition of organic residues over time (Saha and Mishra, 2009; Saptoka et al., 2012; Mbuthia et al., 2015).

As all treatments were planted with hairy vetch and fertilizers had not yet been applied, measured soil resin N concentrations, N leachate, and gaseous N₂O-N loss parameters did not differ during the cover crop growing season (Figures 2C,D, 3D, Table 2). Though there were slight differences in NH₃-N emissions on three dates (Figure 4D), treatments were not significantly different across the entire measurement period (Table 2) as gaseous N losses were low at this time as, most likely limited by soil temperatures (Figure S3; Parsons et al., 1991; Colbourn, 1993; DeKlein and VanLogtestijn, 1996). N leaching was the dominant N loss pathway.

Due to the unusually low winter temperatures experienced at the site, hairy vetch biomass production was low overall (Figure 1D). To simulate a more representative year, supplemental vetch biomass (~2,000 kg ha⁻¹) was harvested from an adjacent organic field and added as residue to the fertilizer N-credit treatment. This resulted in hairy vetch biomass and N content being significantly greater in the fertilizer N-credit treatment than the unfertilized or organic treatments at cover crop termination (Figure 1D). Had this biomass addition not occurred, the fertilizer N-credit approach would have been nearly identical to the organic (fertilizer alone) treatment, and not representative of typical vetch production at the site, thus limiting our understanding of the N dynamics of a conservation agriculture system predominantly fertilized with vetch biomass.

Although initial cover crop N content was greatest in the fertilizer N-credit treatment, by June, there was no difference between treatments (Figure 1D), as the hairy vetch residue underwent rapid decomposition (Hadas et al., 2002; Acosta et al., 2011). N content and %N of residue declined similarly in all treatments during decomposition (Figures 1D,E,F). Post-fertilization, soil NO₃-N and NH₄-N dynamics differed between treatments. Overall, soil resin NO₃-N and NH₄-N tended to be greatest in the organic N fertilized treatment, followed by the fertilizer N-credit treatment, and lowest in the unfertilized treatment (Figures 2C,D). Despite observed treatment differences in post-fertilization soil resin extracted N, there were no differences in NO₃-N or NH₄-N leachate values between treatments (Table 2). However, there were differences in gaseous N emissions.

During the corn growing season, N gaseous loss pathways were influenced by the quantity and form (vetch biomass vs. organic fertilizer) of applied N. Overall, the organic N treatment had the highest N₂O-N emissions (Table 2), but the fertilizer N-credit treatment had higher initial emissions (Figure 3E). A greater quantity of cover crop residue in the fertilizer N-credit treatment led to soil moisture being consistently greater (by 2–6%) during the post-fertilization intensive measurement period (29 May−12 Jun) vs. the unfertilized or organic fertilized treatments (Figure S2), most likely via prevention of evaporative loss from the soil surface. These conditions may have stimulated microbial activity, N mineralization, and subsequent, denitrification and N₂O fluxes (Colbourn, 1993; Schomberg et al., 1994; DeKlein and VanLogtestijn, 1996; Dobbie et al., 1999; Hu et al., 2013). Periods of high soil moisture continued to induce denitrification throughout the growing season; N₂O-N gas flux peaks in all treatments were often linked with rain events, particularly in the organic N treatment. Following a rain event on 4 June 2014, emissions increased dramatically in the organic N treatment and then remained greater than emissions from the fertilizer N-credit and unfertilized treatments for the remainder of the corn growing season (19 June−31 October; Figure 3F).

N in slow release, organic fertilizers, such as that applied in this study, commonly undergo mineralization followed by denitrification after rain events (Sistani et al., 2011; Venterea et al., 2011; Halvorson and Del Grosso, 2012, 2013). While maintaining a source of available N over the course of the growing season is ideal for plant growth, temporal asynchrony between microbial activity and plant N uptake, such as can occur during wet-dry cycles, can result in significant N₂O-N loss (Augustine and McNaughton, 2004; Schwinning and Sala, 2004; Dijkstra et al., 2012; Parkin and Hatfield, 2014).

NH₃-N emissions were also greatest in the organic fertilizer treatment, especially during the post-fertilization intensive measurement period, with significant fluxes occurring on 6 June 2014 (Figures 4E,F), following a rain event. Similar to rainfall effects on N₂O-N fluxes, precipitation most likely induced fertilizer mineralization, increasing soil NH₃ concentrations and, as soil moisture declined, volatilization occurred (Rochette et al., 2009). By subtracting the unfertilized gaseous N emissions calculated over the corn growing season from the fertilized treatments and dividing by the total amount of N applied in the fertilized treatments, the percentage of the applied N that was diverted into N₂O-N or NH₃-N loss pathways was estimated (Parkin and Kaspar, 2006). This approach estimated that 1.4% of the applied N in the organic N treatment was lost as N₂O-N, whereas only 0.6% of applied N was lost as N₂O-N in the fertilizer N-credit treatment (Table 2). For gaseous NH₃-N emissions, ~18.5% and <3% of applied N lost as NH₃-N from the organic N and fertilizer N-credit treatments, respectively (Table 2).

These results suggest that more N was released from the organic fertilizer than the fertilizer N-credit treatment. Yet, corn yield was similar between these two treatments (and was significantly greater than the unfertilized treatment; Table 2). Though more inorganic N was released in the organic fertilizer treatment, it may not always have been accessible for corn plant uptake or, as other studies have shown, at a higher level of N supply the corn in the organic fertilizer treatment may have used N less efficiently (Moll et al., 1982). It appears that greater N availability in the organic fertilizer treatment later in the season stimulated gaseous emission losses rather than plant uptake and yield. This research suggests that increased incorporation of N supplied by a cover crop into fertilization schemes may decrease N loss and promote more efficient N use by improving the synchronicity of N release with corn plant demand compared to organic fertilizer applications.

At termination, wheat cover crop biomass averaged 6,410 kg ha⁻¹ and contained ~70 kg N ha⁻¹. There were no significant differences across treatments in biomass or N content as all were planted with the same cover crop species (Figure 1G). Similarly, soil NO₃-N and NH₄-N concentrations did not differ across treatments during this time period, as fertilization had not yet occurred (Figures 2E,F). However, during February/March, soil NH₄-N was slightly lower in the organic N fertilizer treatment (urea N, unfertilized > organic N; Figure 2F). There was no difference in N loss via leachate or N₂O-N emissions (Figure 3G), but differences in gaseous NH₃-N loss did occur (Table 2). On several dates, NH₃-N emissions were slightly elevated in the urea N and unfertilized treatments (Figure 4G), reflecting slightly higher soil NH₄-N in these treatments at those times (Figure 2F). However, when pre-fertilization total NH₃-N emissions were calculated, all three treatments differed from each other (urea N > unfertilized > organic N; Table 2). Soil moisture during this time was, on average, 2% greater in the urea N vs. the organic N treatment (Figure S2), though a means comparisons test failed to identify specific dates where the treatment effect was significant.

Higher soil resin NH₄-N and gaseous NH₃-N losses in the urea treatment during this time period are difficult to explain, as all treatments were planted in the same cover crop species (wheat) and had not received any fertilizer. The only difference apparent between treatments at this time was the presence of volunteer big flower vetch (Vicia grandiflora): it was most abundant in the urea treatment (average percent cover ~7%), followed by the organic fertilizer (~5%), and the unfertilized (~4%). Perhaps the presence of this weed or other unknown factors contributed to observed treatment differences.

Post-termination of the winter wheat cover crop and post-fertilization of the organic N and urea N treatments (May–October), N dynamics in the decomposing cover crop residue did not significantly differ between treatments, though residue N content varied across the time period; %N steadily increased (indicating immobilization of N) and C:N ratios declined (Figures 1G–I). However, soil N dynamics differed between treatments. In June, soil resin NO₃-N concentrations were significantly different between all three treatments (urea N > organic N > unfertilized), and in July both fertilized treatments had greater NO₃-N concentrations than the unfertilized treatment (Figure 2E). On average, NH₄-N tended to be greater in both fertilized treatments vs. the unfertilized treatment, but means comparisons failed to identify specific months during which soil NH₄-N significantly differed, finding only that urea N was marginally greater than the unfertilized treatment for the month of June (p = 0.0552).

Gaseous N loss and leaching were strongly affected by the quantity and type of fertilizer applied, with gaseous and leaching N losses in the unfertilized treatment being minimal compared to fertilized treatments. The urea N treatment had 5x more NO₃-N leachate loss than either the organic N or the unfertilized treatments during this time period (Table 2). Some urea hydrolysis likely occurred prior to corn plant germination; thus, N may have leached down into the soil profile preceding corn plant establishment (Quisenberry and Phillips, 1976), which is supported by the high resin NO₃-N and NH₄-N concentrations measured in the urea treatment in June (Figure 2E). To reduce NO₃ leaching loss associated with urea N applications, a split application or application after corn plant establishment might be necessary (Meisinger and Delgado, 2002).

During the post-fertilization intensive measurement period (29 May–12 June), N₂O-N emissions were greater in both the fertilized treatments vs. the unfertilized (Figure 3H). On 6 June 2014, N₂O-N emissions were more than 8x greater in the urea N treatment than the unfertilized treatment, and on 12 June 2014 they were more than 9x greater in the organic N treatment than the unfertilized treatment. On several dates during the remainder of the corn growing season (19 June–31 October), N₂O-N emissions from the organic N treatment remained significantly higher than the other treatments (Figure 3I). However, these date-specific trends were not strong enough to significantly influence the entire post-fertilization N₂O-N flux estimates (29 May–31 October), as no differences in N₂O-N emissions between treatments were found (Table 2). Approximately 2% of fertilizer was lost as N₂O-N in the organic N treatment, whereas only 1% was lost from the urea N (Table 2).

Higher N₂O-N loss in the organic N treatment was not unexpected as the fertilizer was not only 13% N, but 40% C (Kirk Carls, Nature Safe Natural and Organic Fertilizers, personal communication, 14 March 2015). Numerous studies have found that fertilizers that provide a C source in conjunction with N stimulate microbial activity and, subsequently, N loss (Limmer and Steele, 1982; Hayakawa et al., 2009; Chantigny et al., 2010; Sistani et al., 2011; Mitchell et al., 2013). N₂O-N emissions from the organic fertilizer treatment in our study were similar to those calculated by Sistani et al. (2011) in a Kentucky no-till corn system applied with poultry litter. The higher N₂O-N flux measured in the urea treatment on 6 June 2014 may be the result of urea hydrolyzing more readily than the organic fertilizer, enriching the soil with inorganic N prior to the rain event on 4 June (Figures 2E,F; Parkin and Hatfield, 2014). After this date, the slower N release nature of the organic fertilizer caused higher N₂O-N emissions from this treatment for majority of the remainder of the study.

NH₃-N emissions were also stimulated by fertilizer additions (Figure 4H). Temporally similar to N₂O-N trends, the urea N treatment had greater emissions during the first part of the post-fertilization intensive measurement period while the organic N treatment had greater emissions during the latter half of the measurement period. Despite substantial fertilizer effects early on, treatment differences did not persist across the longer-term post-fertilization period (Figure 4I). When summed across the entire post-fertilization period (29 May–31 October), NH₃-N emissions were significantly greater in the organic N vs. unfertilized treatments (Table 2). Approximately 29% of fertilizer was lost as NH₃-N in the organic N treatment, whereas only 18% was lost from the urea N (Table 2).

Christianson et al. (2012) reported that up to 30% of N can be volatilized after broadcast urea N applications, and, in systems where organic residue is present, broadcast urea may be particularly vulnerable to volatilization as the residue acts as a barrier between the soil surface and the fertilizer (Mohr et al., 1998; Rochette et al., 2009; San Francisco et al., 2011). The urease inhibitor included in our urea fertilizer treatment was likely responsible for mitigating volatilization to some degree prior to precipitation events that incorporated the fertilizer into the soil profile (a rain event occurred the night following application, ~0.25 cm, and another on 4 June 2014, ~4.72 cm). After contact with the soil, volatilization was likely minimal, as the chemistry at this field site is not considered particularly susceptible to volatilization (Table S1). NH₃-N volatilization of urea N subsided after the rain event on 4 June 2014, whereas this same event caused a spike in NH₃-N emissions in the organic N treatment. NH₃-N emissions in both fertilized treatments were relatively low after 12 June 2014 (Figure 4I).

Despite differences in N loss, corn yield was statistically equivalent between fertilized treatments and was at least double the unfertilized yield (Table 2). Our results illustrate that significant amounts of N can be lost to the environment in systems using both urea and organic fertilizers, but predominant N loss pathways differ; the organic fertilizer treatment had greater gaseous loss while the urea treatment had greater N leaching (Table 2). A nitrification inhibitor may be advantageous in both systems as it could reduce the quantity of NO₃-N vulnerable to leaching in the urea treatment and to denitrification in the organic fertilizer treatment.