Greenhouse gas and ammonia emissions from stored manure from beef cattle supplemented 3-nitrooxypropanol and monensin to reduce enteric methane emissions

The investigative material 3-nitrooxypropanol (3-NOP) can reduce enteric methane emissions from beef cattle. North American beef cattle are often supplemented the drug monensin to improve feed digestibility. Residual and confounding effects of these additives on manure greenhouse gas (GHG) emissions are unknown. This research tested whether manure carbon and nitrogen, and GHG and ammonia emissions, differed from cattle fed a typical finishing diet and 3-NOP [125–200 mg kg−1 dry matter (DM) feed], or both 3-NOP (125–200 mg kg−1 DM) and monensin (33 mg kg−1 DM) together, compared to a control (no supplements) when manure was stockpiled or composted for 202 days. Consistent with other studies, cumulative GHGs (except nitrous oxide) and ammonia emissions were higher from composted compared to stockpiled manure (all P < 0.01). Dry matter, total carbon and total nitrogen mass balance estimates, and cumulative GHG and ammonia emissions, from stored manure were not affected by 3-NOP or monensin. During the current experiment, supplementing beef cattle with 3-NOP did not significantly affect manure GHG or NH3 emissions during storage under the tested management conditions, suggesting supplementing cattle with 3-NOP does not have residual effects on manure decomposition as estimated using total carbon and nitrogen losses and GHG emissions.

Nitrous oxide. Between days 1 and 104, N 2 O fluxes averaged 0.5 g N m −2 d −1 from the composted manure and 0.4 g N m −2 d −1 from the stockpiled manure (Fig. 3a, b). From day 105 to 202, N 2 O fluxes averaged 0.03 g m −2 d −1 from the composted manure and 0.02 g m −2 d −1 from the stockpiled manure.
Daily N 2 O fluxes from composted manure were low throughout the trial. The exceptions were the measurements taken after compost turning where N 2 O fluxes from all additive treatments were between 6.2-to 24.8-fold higher when compared to measurements made the day prior to turning (Fig. 3a). From the stockpiled manure, daily N 2 O fluxes increased in all additives after day 28, but there were some differences in flux trends over time between additives: N 2 O fluxes from stockpiled 3-NOP manure were generally lower than the other additive treatments on day 40, and higher than other additive treatments on day 76 (Fig. 3b).
Despite some temporal discrepancies in daily N 2 O fluxes (Fig. 3a, b), there were no differences in cumulative N 2 O emitted from different handling methods, additive treatments, or any handling method by additive treatment interactions detected when emissions were presented as kg N m −2 , kg N pile −1 or the kg N tonne −1 initial N (Table 1).
Daily NH 3 fluxes from composted and stockpiled manure were high early in the trial, peaking within the first six days after pile construction in all additive treatments (Fig. 3c, d). Within the respective handling methods, www.nature.com/scientificreports/ daily NH 3 fluxes from all three additives were similar in magnitude throughout most of the trial and followed similar undulating trends. There were no significant differences in cumulative NH 3 emissions between the dietary additive treatments (Table 1). Cumulative NH 3 emissions averaged over all additives on a kg m −2 basis were 83% higher from the composted compared to the stockpiled manure (P < 0.001). Cumulative NH 3 emissions extrapolated over the initial surface area of the piles (kg pile −1 ) averaged over all additives were 158% higher from the composted www.nature.com/scientificreports/ compared to the stockpiled manure (P < 0.001). Compared to the stockpiled manure, the NH 3 kg N tonne −1 initial N was 142% higher (P < 0.001) from the composted manure.
Manure chemistry. When averaged over all storage methods and additive treatments, manure TC concentrations, TN concentrations, C:N ratio, inorganic N concentrations, pH and electrical conductivity (EC) differed between days 1 and 202 ( Table 2). All measured variables were higher on day 1 than on day 202 except pH and EC, which were higher on day 202. Manure storage affected TC concentrations; stockpiled manure TC concentrations at the end of the trial were lower (P < 0.10) than in initial samples, and composted manure TC Table 1. Cumulative gas emissions over the entire trial by additive treatment and pile handling method. The means are presented with ± standard error of the mean in parentheses for each additive treatment (control, no supplements; 3-NOP, 3-nitrooxypropanol; 3-NOP + Mon, 3-nitrooxypropanol and monensin, n = 3). Asterisks in the "All" columns *,**,*** indicate significant differences between handling methods (averaged over all additive treatments) at P < 0.05, P < 0.01, and P < 0.001, respectively. The P values are for the mixed effects ANOVA model results for pile handling method (P), additive treatment (T) and their interactions. www.nature.com/scientificreports/ concentrations at the end of the trial were lower (P < 0.10) than both the initial composted manure concentrations and the final stockpiled manure concentrations. Manure TC concentrations were significantly affected by the additive treatment (Table 2). When averaged over the entire trial, TC in the control (mean = 157.9 g C kg −1 ) and 3-NOP (mean = 158.1 g C kg −1 ) were lower (P ≤ 0.05) than in 3-NOP + Mon (mean = 170.5 g C kg −1 ).  (Table 2), but TN was not affected by additive treatment or manure handling ( Table 2).
Manure TC, TN and C:N ratios from the additive treatments were different between the beginning and end of the trial ( Table 2). The TN, TC and C:N ratio for all additive treatments were lower on day 202 than on day 1 (P ≤ 0.10). The TN and C:N ratio for 3-NOP + Mon were higher than 3-NOP (P ≤ 0.10) on day 1, but by day 202, there were no differences between the additive treatments. Table 2. Effects of manure handling method and additive treatments on manure chemical properties presented as mean ± standard error of the mean in parentheses. "All" represents an average of all additive treatments (n = 27). Treatments were: control, no supplements; 3-NOP, 3-nitrooxypropanol; 3-NOP + Mon, 3-nitrooxypropanol and monensin (n = 9). Day 1 samples were collected during pile construction, and day 202 samples were collected at the end of the trial from the composted and stockpiled manure. The P values are for the mixed effects ANOVA model results for pile handling method (P), additive treatment (T), day (D) and their interactions. Differences in capital letters represent differences between handling methods, days, and/or handling method by day interactions at P < 0.10. Differences in lower case letters represent differences between additive treatments, and additive treatment by day interactions at P < 0. 10 www.nature.com/scientificreports/ Stored manure carbon and nitrogen mass balance estimates. The mass balance estimates considered the changes in concentrations to compensate for losses in mass from the manure piles during the trial. Mass losses of DM as well as TC and TN tended to be higher in composted compared to stockpiled manure, however, these differences were not significant (Table 3). From the composted manure, the mass balance showed an equivalent of 90.9 kg C tonne −1 was lost from the control and 94.3 kg C tonne −1 was lost from 3-NOP + Mon, while only 55.9 kg C tonne −1 was lost from 3-NOP. Likewise, from the composted manure, an equivalent of 5.9 kg N tonne −1 was lost from the control and 5.3 kg N tonne −1 was lost from 3-NOP + Mon, while 3.2 kg N tonne −1 was lost from 3-NOP. While 3-NOP showed slightly lower mass losses of TC and TN compared to the control and 3-NOP + Mon in the composted manure, due to the high variability, these differences were not statistically significant ( Table 3). The combination of slightly greater TC and TN losses from 3-NOP resulted in greater (P < 0.10) DM losses in 3-NOP compared to 3-NOP+Mon (Table 3).
There were discrepancies between the losses calculated using the mass balance approach (Table 3) and the GHGs lost as a fraction of the initial TC and TN in the compost (Table 1). It appears that more TC was lost as CO 2 and CH 4 than were lost as a fraction of the initial TC when calculated using the mass balance approach. The mass losses of TN were higher than the cumulative losses of N as NH 3 and N 2 O (Table 1).
Across all additive treatments in the stockpiled manure, the mass balance calculations estimated that total losses of C averaged 58.3 kg C tonne −1 and total losses of N averaged 3.1 kg N tonne −1 ( Table 3). The total losses of C estimated using the mass balance approach from the stockpiled manure were similar to the cumulative losses of CO 2 and CH 4 expressed as kg C tonne −1 initial C from the stockpiled manure (equating to 65.2 kg C tonne −1 C). The TN losses calculated using the mass balance approach (Table 3) were higher than those determined by cumulative losses of N 2 O or NH 3 as kg N tonne −1 initial N from the stockpiled manure (Table 1).

Discussion
This was the first study to assess whether GHG and NH 3 emissions from composted and stockpiled beef cattle manure are affected by dietary supplementation with 3-NOP or a combination of 3-NOP and monensin. While 3-NOP and monensin supplementation can reduce enteric CH 4 production from ruminant animals 10,13,14 , we found that they did not significantly affect cumulative GHGs or NH 3 emissions from stored manure. Stored manure TC concentrations were affected by these dietary additives. However, the additives did not significantly alter mass balance estimates of TC and TN in composted or stockpiled manure.
In our study, a combination of 3-NOP and monensin resulted in significantly higher manure TC concentrations compared to the control and 3-NOP. This suggests there could be differences in the relative amounts of recalcitrant versus labile forms of C in manure from cattle supplemented with 3-NOP and monensin. A previous study noted 3-NOP-induced increases in excreted neutral (hemicellulose, cellulose and lignin) and acid detergent fibre (cellulose and lignin) in fresh manure from lactating dairy cows 14 , and monensin has been shown to alter digestibility of NDF and ADF in cattle 10 . Future studies should consider using more sophisticated methods to gather insights into these differences. When mass losses were considered, we did not find significant differences in TC and TN between the additive treatments suggesting that microbial activity during manure decomposition was not significantly affected by the additive treatments.
This was the first study to evaluate the potential impacts of 3-NOP on GHG and NH 3 emissions from stored manure. Therefore, there are no directly comparable data sets. Additive treatments did not significantly affect cumulative GHGs or NH 3 emissions. This is consistent with the lack of mass balance differences in TC and TN for stored manure between additive treatments. For GHG or NH 3 emissions to be affected by additives, differences in substrates or microbial communities and/or microbial activity within the manure pile would be required. Since changes in stored manure TC and TN are largely the result of microbial activity, there is no evidence that 3-NOP had a significant effect on these factors in our study.
It was expected that there would be differences between the composted and stockpiled manure. There was slightly greater thermophilic decomposition in the composted manure compared to the stockpiled manure as indicated by a longer duration of relatively higher internal pile temperatures 25,26 . However, we did not observe differences in mass losses of DM, TC or TN as a result of the different storage methods (Table 3). This is in contrast to the results of other studies which have noted greater losses of TC and TN from composted manure compared to stockpiled manure 17 . Our results may be a consequence of low manure moisture content initially and throughout which were less-than-optimal for manure composting 27 . They may also be attributed to the low compost turning frequency (three times) compared to the six to eight times in other studies as greater turning frequency leads to greater decomposition and therefore greater DM, TC and TN losses 28 .
Trends in daily fluxes observed during our trial were similar to those observed during other studies. Similar to our results, others have also reported CH 4 emissions from composted and stockpiled cattle manure spiked shortly after pile creation 18,29 and that CO 2 emissions from composted and stockpiled beef cattle manure were initially high and decreased over time 18,24 . Composting manure has been shown to result in high variability of N 2 O emissions 18,24 , and others have similarly reported N 2 O fluxes increased after NH 3 fluxes decreased 24,30 .
Composting manure resulted in higher cumulative CO 2 and NH 3 emissions compared to stockpiled manure which is consistent with other studies 22,28 . Compost turning increases aeration and aerobic decomposition of organic material and hydrolysis of organic N resulting in the production of CO 2 and NH 3 24,26 . Because turning introduces O 2 into the compost, it was a little surprising that we observed higher cumulative CH 4 emissions from the composted manure compared to the stockpiled manure.
Methane production requires the absence of O 2 as a precondition for production 31 and stockpiled manure can have high CH 4 emissions 20 . However, anaerobic conditions do not always form in stockpiled manure 18 and high CH 4 production within manure piles does not always equate to greater surface-to-atmosphere CH 4

Scientific Reports
| (2020) 10:19310 | https://doi.org/10.1038/s41598-020-75236-w www.nature.com/scientificreports/ Table 3. Mean ± standard error of the mean in parentheses showing the percent losses of dry matter, total carbon and total nitrogen from each pile handling method (P) and additive treatment (T) between the start and end of the trial (n = 18) determined by the mass balance analysis. The P values for the mixed effects ANOVA model results for P, T and P and T interactions for total % losses. Differences in lower case letters represent differences between treatments at P < 0.10.  23 . In the current study, low CH 4 emissions from our stockpiled manure can be partially attributed to the initial pile characteristics as the manure was mixed with straw bedding and the initial moisture content was ~ 0.34 kg kg −1 which is lower than the > 0.60 kg kg −1 initial manure moisture values reported in other studies that measured high CH 4 emissions from stockpiled manure 22,32,33 . These characteristics can contribute to low density and high porosity in the manure piles, which allows diffusion of O 2 into the pile from the atmosphere. This may have allowed aerobic conditions to predominate inside the stockpiled manure resulting in low CH 4 emissions 34 . Likewise in these conditions, CH 4 produced within the pile may be oxidized before being emitted to the atmosphere 23 . The lower cumulative CH 4 emissions from the stockpiled compared to the composted manure may also be attributed to the relatively smaller pile size of the stockpiles as CH 4 emissions have been found to be lower when manure piles are smaller 30 . We found no differences in cumulative N 2 O emissions between the composted manure and stockpiled manure. There was some evidence that reducing conditions were developing within the stockpiled manure after day 30 as shown by the increase in N 2 O fluxes, while in the composted manure most of the N 2 O was emitted after turning. During manure storage, nitrification on the pile surface leads to accumulation of NO 2 and NO 3 and these substrates are reintroduced into the compost pile during turning contributing to spikes in N 2 O fluxes from denitrification 23,26 . The data suggests the environmental conditions and processes occurring in the composted and stockpiled manure differed, however, they emitted similar amounts of N 2 O.
The mass balance estimates suggest more N was lost from both the composted and stockpiled manure than was accounted for as NH 3 or N 2 O. Some of the discrepancies could be related to gases that we have not accounted for (i.e. nitric oxide and nitrogen gas) 35 . The losses of TC from the stockpiled manure estimated by the mass balance (All = 58.3 kg tonne −1 ) were similar to the total losses of CO 2 + CH 4 kg C tonne −1 TC (All = 65.1 kg C tonne −1 ). However, mass balance estimates of TC losses (All = 80.4 kg tonne −1 ) and the losses of CO 2 + CH 4 kg C tonne −1 TC (All = 174.3 kg C tonne −1 ) were not comparable in the composted manure. Previous studies have found similarities between these values 36 . The mass balance estimates use the initial and final manure samples and mass balances are derived from an empirical model 37 . This method has been shown to sometimes cause erroneous values when manure samples are not representative of the entire pile, or when soil is introduced into the compost piles during turning 38 . These factors may have impacted our results. Despite these discrepancies, both the gas and mass balance data consistently showed no significant effects on C or N cycling in stored manure piles resulting from supplementing beef cattle with 3-NOP or 3-NOP and monensin. Future studies should consider placing manure piles on concrete pads to prevent mixing soil with the manure, and collecting multiple samples at various areas of the manure piles to ensure samples are representative.
Measurement of GHGs using static chambers can contribute to errors. There are inaccuracies associated with extrapolating 40 min static chamber measurements made one to three times a week over the entire 202 day trial. Temporal extrapolation of GHG fluxes from static chambers are associated with some error in cumulative GHG emissions when used on soils 39 . Likewise, the area covered by the chambers was small so we only measured a fraction of the total surface area of the manure piles.
Our study did not evaluate the potential effects of differences in 3-NOP and monensin dosage, and we also did not consider differences in cattle diets. The effects of 3-NOP on the digestibility of organic matter in manure have been found to vary slightly by dose 16 and enteric CH 4 production has been found to be related to diet 40 . Thus our results might not be replicated under different 3-NOP doses and different diets.
The presence of residual monensin in the manure was not considered during our study. Monensin in its active form can be excreted with manure, and supplementation with monensin could alter GHG emissions from manure 11,41 . However, both composting and stockpiling manure have been shown to reduce monensin concentrations in stored manure 42 . Future studies should consider incorporating extractions to test for monensin over time during manure storage to properly evaluate whether GHG emissions from stored manure are directly affected by monensin.
Management activities such as adding water to the manure piles, swapping straw for other materials 29,32 , changing cattle diet 43 and changing compost turning frequency 28 affect the degradation of residual dietary additives in manure, as well as manure components and GHG and NH 3 emissions. These factors have not been considered in the current study.
In summary, our study found some differences in manure chemistry as well as higher CO 2 , CH 4 and NH 3 emissions from composted compared to stockpiled manure, which can be attributed to differences in conditions resulting from differences in handling strategies and pile characteristics. We found that supplementing cattle with 3-NOP and monensin affected the initial manure TC concentrations, but they did not significantly influence estimates of TC or TN mass balances in stored manure. As a result, there were no significant differences between additive treatments for cumulative GHGs and NH 3 emissions from composted and stockpiled manure. The lack of differences between additive treatments in our trial suggests that manure from cattle supplemented 3-NOP does not require manure storage that differs from current strategies. Our trial results provide assurance that enteric CH 4 reductions achieved through dietary supplementation with 3-NOP are not negated by increases in GHG emissions during manure storage. This study can serve as a reference for policy makers that will want to know about any confounding effects that supplements have on GHG emissions and nutrient transformations in manure from beef cattle feedlots.

Methods
This experiment was conducted in semiarid Lethbridge, Alberta, Canada (49°42′03.3"N, 112°45′51.0"W). Precipitation data was acquired from an Agriculture and Agri-Food Canada weather station located < 1 km from the trial.

Scientific Reports
| (2020) 10:19310 | https://doi.org/10.1038/s41598-020-75236-w www.nature.com/scientificreports/ Manure collection. Manure mixed with cereal straw (manure:straw ratio of 4:1) was collected from a congruent feeding trial that was conducted from November 2015 to August 2016 44 . All cattle were fed the same highforage diet. Steers were fed a backgrounding diet for 105 days consisting of 65% barley (Hordem vulgare) silage, 5% supplement (vitamins, minerals and crude protein), and 30% dry rolled barley and then were transitioned to a finishing diet over 28 days. The finishing diet consisted of 8% barley silage, 5% supplements, and 87% dry rolled barley grain for 105 days. A full profile of feed ingredients and chemical composition of dietary ingredients are presented in Vyas et al. 44 .
In separate pens, three treatments were established based on the addition or exclusion of additives. The control group did not receive additional supplements, the "3-NOP" group received 3-NOP, and the "3-NOP + Mon" group were supplemented with 3-NOP and monensin. During backgrounding, 3-NOP was supplemented at 200 mg kg −1 dry matter (DM) and monensin was supplemented at 33 mg kg −1 DM. During finishing, 3-NOP was supplemented at 125 mg kg −1 DM and monensin was supplemented at 33 mg kg −1 DM. Experimental setup. Manure was removed from the feedlot on September 6, 2016, at the end of the associated feeding trial 44 . Manure from each treatment was separately transported from the feedlot to the experimental site. Each pile was separated by a buffer of 3.65 m in the east-west direction and 7.62 m in the north-south direction. The piles were divided into two blocks by handling method: composting and stockpiling. Three replicates of each handling method (composting and stockpiling) and treatment (control, 3-NOP and 3-NOP + Mon) were randomly distributed throughout their respective blocks for a total of 18 manure piles.
The initial masses of each compost pile and stockpile were determined during the pile creation from the manure spreader's on-board scale. Dimensions of each pile were determined by manually measuring the length, width, circumference and height of piles with a tape measure. This data was used to determine pile surface area.
For the stockpiles, manure was heaped into piles roughly conical in shape. For the composted manure, windrows were created. All manure piles were unconfined and situated on soil.
The piles were uncovered for the duration of the trial (202 days). The trial duration was dictated by the temperature, which decreases when manure decomposition slows. The manure remained stored until the springeven after decomposition had slowed-when it could be applied to the soil. The compost piles were turned on Manure properties and analyses. Air and internal manure pile temperatures at the top, middle and bottom of the piles were measured hourly during the trial using T-wire thermocouples recorded by a datalogger (CR800 with an AM16/32B channel relay multiplexer, Campbell Scientific, Logan, UT, USA). The thermocouples were briefly removed from the compost piles during turning. Pile temperatures were averaged into one temperature representative of each treatment and handling method.
At the start and end of the trial, each manure pile was cut perpendicular to its length to expose a vertical face. Manure was sampled to determine manure properties from the top, middle and bottom of each pile. The values were averaged for each treatment and handling method.
Gravimetric moisture content was determined by drying 100 g of manure at 60 °C for one week. The moisture content was expressed on a wet basis while manure C and N concentrations were expressed on a dry basis.
Manure pH and EC (ds cm −1 ) were obtained by mixing 15 g of fresh sample and 150 mL ultrapure water for 30 min, then centrifuging (10,000 rpm for 10 min). The pH and EC of the supernatant were measured on a Orion Star A215 pH & EC meter (Thermo Fisher Scientific, Waltham, MA, USA). The sample was then filtered (0.45 µm) and the supernatant solution was analysed for soluble inorganic N (ammonium, NH 4 + ; nitrite, NO 2 -; nitrate, NO 3 -) using ion chromatography (ICS-1000 Ion Chromatography System, Dionex, Sunnyvale, CA, USA). Manure TC and TN (kg tonne −1 DM manure) were determined from freeze-dried ground (≤ 0.15 mm) manure and measured using a CN analyser (Carlo Erba Instruments, Rodano, Italy) and expressed on a dry weight basis.

Mass balance estimates.
Mass losses (%) of DM, TC and TN were estimated from measurements of the initial and final samples from all additive treatments and both manure storage practices using equations outlined in Larney et al. 17 . Briefly, assuming an initial mass of 1000 kg of fresh manure, and using initial and final moisture measurements, DM losses were derived by assuming all mass losses were organic material and estimated according to an empirically derived relationship 37 . The TC and TN losses were derived by calculating the percent difference in their initial and the final concentrations relative to the initial and final masses of the manure piles.
Greenhouse gas and ammonia sampling. Ammonia and GHG fluxes were measured one to three times a week from September 7, 2016 (day 1) until November 29, 2016 (the first 83 days, the thermophilic phase when internal manure pile temperatures were > 45 °C) then three to six more times thereafter until March 28, 2017 (day 202). For each handling method and additive treatment, this resulted in a total of 14 NH 3 measurements. For stockpiled manure, we measured GHGs 15 times for each additive treatment, and from composted manure, we measured GHGs 18 times for each additive treatment. The three additional measurements made from the composted manure occurred immediately after the compost was turned.
Ammonia fluxes were measured using cylindrical PVC vented chambers (chamber area = 0.019 m 2 ) with bases inserted 3-5 cm into the manure.
The chambers were equipped with two 2.5 cm thick polyurethane foam disks cut fit-to-seal to the inner diameter of the PVC chambers and infused with 50 mL of phosphoric-glycerol solution to trap gaseous NH 3 . The "measurement disk" was situated above the pile surface to capture NH 3 emitted off the manure, and a Scientific Reports | (2020) 10:19310 | https://doi.org/10.1038/s41598-020-75236-w www.nature.com/scientificreports/ "scrubbing disk" was inserted between the measurement disk and the atmosphere to prevent contamination of the measurement disk. Each chamber was covered with a canopy to prevent precipitation from washing the acid solution from the disks. For measurements, the foam disks were brought to the laboratory in sealed bags, weighed, and the amount of solution absorbed into each disk was determined using before and after deployment weights. Disk NH 3 concentrations were extracted by saturating each disk with 100 mL of 0.5 M KCl solution for 30 min. Extractions were frozen at − 20 °C prior to analysis. The extracts were analysed using the indophenol blue method with a MultiSkan Go Microplate Spectrophotometer at 650 nm (Thermo Scientific, Waltham, MA, USA). Ammonia fluxes were calculated by dividing the extraction concentrations by the chamber area and deployment time.
Greenhouse gas samples were collected from vented static chambers (area = 0.07 m 2 , volume = 7 L). Chamber bases (inner diameter = 30 cm) were inserted 5 cm into the manure pile. The bases were briefly removed and reinstalled before and after compost turning.
For GHG measurements, chamber covers were clipped to the bases and samples taken at 0, 10, 20 and 40 min after covering. From each chamber, 11 mL of gas was extracted from the headspace using a polypropylene syringe and immediately injected into a pre-evacuated (− 1 atm) 5.8 mL Exetainer (Labco Ltd., Lampeter, United Kingdom). Gas concentrations were determined using a gas chromatograph equipped with electron capture, thermal conductivity and flame ionization detectors (Varian 3800, Varian Instruments, Palo Alto, CA, USA). The injector and column temperatures were kept at 55 °C. The carrier was P10 gas (10% methane, balance argon) for N 2 O and helium for CO 2 and CH 4 . The channel was maintained at a static pressure of 150 kPa.
Greenhouse gas fluxes from the surface of manure piles to the atmosphere were calculated using air temperature, the ideal gas law, chamber area and volume, and the change in gas concentration over time for each chamber. The change in gas concentration over time was assessed using both quadratic (QR) and linear regression (LR). Unless the second derivative of the model was ≤ 0 according to the LINEST function in Microsoft Excel (version 2019), fluxes were calculated with QR 45,46 . The minimum detectable fluxes were determined for CO 2 , CH 4 and N 2 O 47 . The minimum detectable fluxes for quadratic and linear fluxes were: ± 2215.7 and ± 641.5 µg C m −2 h −1 , respectively, for CO 2 ; ± 2.8 and ± 9.7 µg N m −2 h −1 , respectively, for N 2 O; and ± 4.9 and ± 16.8 µg C m −2 h −1 , respectively, for CH 4 . Fluxes below the minimum detectable flux were assigned a value of zero. Of the 387 fluxes measured, the QR method was used for 85, 81 and 55% of the CO 2 , N 2 O and CH 4 fluxes, respectively, the LR method was used for 12, 19 and 44% of the CO 2 , N 2 O and CH 4 fluxes, respectively, and 3, 1 and 1% of the CO 2 , N 2 O and CH 4 fluxes were below detection, respectively.
Daily NH 3 and GHG flux rates for the composted and stockpiled manures are presented by additive treatment as well as an average for all treatments together ("All") to compare flux rates between storage methods. The daily gas emissions are expressed as fluxes per unit area (g m −2 d −1 ).
The cumulative emissions from each pile were determined by integrating (trapezoidal method) between daily measurements and summing the values, which were then averaged based on additive treatment and storage method.
The cumulative emissions for each treatment and storage method are presented on a per unit area basis (kg m −2 ), averaged over the initial surface area of the manure piles (kg pile −1 ), and determined as a proportion of the initial manure TC and TN content (kg C or N tonne −1 ).
The GHG and NH 3 emissions were extrapolated from a per unit area to per surface area of the pile by multiplying the emissions (kg m −2 ) by the initial surface area of the piles (pile 1 as m 2 ).
The emissions as kg tonne −1 initial manure TC or TN were calculated by dividing the initial surface area of the piles (m 2 ) by the initial surface area, multiplying the emissions by the kg tonne −1 initial DM manure, then multiplying by the fraction of initial TC or TN in the manure.
Data handling and statistics. Manure chemical concentrations, % losses of DM, TC and TN from mass balance estimates, and cumulative GHG and NH 3 emissions were subject to statistical analyses. Data were analysed using R Statistics (version 3.4.3). Normality of each dataset was tested using a Kolmogorov-Smirnov test and homogeneity of variance assessed using Levene's test. None of the data sets were transformed. To test for differences in manure chemical concentrations by handling method (compost and stockpile), additive (control, 3-NOP and 3-NOP + Mon) and day (1 vs 202), a three-factor mixed model ANOVA was conducted (nlme package in R), and when appropriate, post-hoc analysis was completed using a Tukey's HSD Test (lsmeans package in R). To test for differences in cumulative GHG and NH 3 emissions and % mass losses of DM, TC and TN by handling method (compost and stockpile) and additive (control, 3-NOP and 3-NOP + Mon), a two-factor mixed model ANOVA was used. In the mixed models, replicate was treated as a random effect and all other factors included in the model were treated as fixed effects. Significance was evaluated at P ≤ 0.10.