Incubation experiments using nitrogen isotope discrimination to estimate ammonia emission from amended sheep manure treatments

Abstract

Two 10-day in vitro experiments were conducted to investigate the relationship between nitrogen (N) isotope discrimination (δ15N) and ammonia (NH₃) emissions from sheep manure. In Exp. 1, three different manure mixtures were set up: control (C); C mixed with lignite (C + L); and grape marc (GM), with 5, 4 and 5 replications, respectively. For C, urine and faeces were collected from sheep fed a diet of 550 g lucerne hay/kg, 400 g barley grain/kg and 50 g faba bean/kg; for C + L, urine and faeces were collected from sheep fed the C diet and 100 g ground lignite added to each incubation system at the start of the experiment; for GM, urine and faeces were collected from sheep fed a diet consisting of C diet with 200 g/kg of the diet replaced with GM. In Exp. 2, three different urine-faeces mixtures were set up: 2U:1F, 1.4U:1F and 1U:1F with urine to faeces ratios of 2:1, 1.4:1 and 1:1, respectively, each with 5 replications. Lignite in C + L led to significantly lower cumulative manure-N loss by 81 and 68% in comparison with C and GM groups, respectively (P = 0.001). Cumulative emitted manure NH₃-N was lower in C + L than C and GM groups by 35 and 36%, respectively (P = 0.020). Emitted manure NH₃-N was higher in 2U:1F compared to 1.4U:1F and 1U:1F by 18 and 26%, respectively (P < 0.001). This confirms the relationship between manure δ15N and cumulative NH₃-N loss reported by earlier studies, which may be useful for estimating NH₃ losses.

Introduction

Ammonia (NH₃) emission from livestock manure has major negative effects on the environment including causing acid rain, eutrophication of surface waters, fine particulate matter formation in the air (Hristov et al., 2011), and respiratory diseases in humans (Hristov et al., 2011). Emitted NH₃ can be converted to nitrous oxide (N₂O), a greenhouse gas (GHG), which contributes to global warming (Chen et al., 2015). Therefore, there is an urgent need to assess and mitigate NH₃ emissions from livestock manure globally.

Previous studies showed that feed manipulations, such as the inclusion of grape marc (GM; i.e., grape pomace) can reduce livestock urinary nitrogen (N) (UN):faecal N (FN) ratio (UN:FN) in dairy cattle (Greenwood et al., 2012; Wu et al., 2022) and this may reduce NH₃ emission from manure (Lynch et al., 2007). It is known that GM has a moderate to high tannin and fat content (Spanghero et al., 2009) and this can increase the ruminal undegradable protein supply, resulting in a shift in the site of N excretion from urine to faeces. Faecal N is more stable and less readily converted to NH₃ than UN. Chen et al. (2015) and Sun et al. (2016) demonstrated that application of 3 to 6 kg/m² lignite to a feedlot cattle pen surface can reduce manure NH₃ loss by 30 to 66%. Lignite has three major chemical characteristics that may be related to the reduction of manure NH₃ emission: (1) low pH (3.69), (2) high cation exchange capacity (CEC; 96.8 centimole (+)/kg, and 3) high labile carbon content of up to 200 g/kg (Husted et al., 1991; McCrory and Hobbs, 2001; Chen et al., 2015). These chemical and physical characteristics limit the conversion of manure-N into NH₃ and leaves NH₃ in the NH₄⁺ form that is not emitted.

One of the major challenges to manage NH₃ emissions from livestock manure is the availability of accurate and practical methods to estimate emissions. Previous reviews clearly demonstrated the direct methods to accurately quantify NH₃ emission (e.g., micrometeorological methods), but they are heavily influenced by many environmental factors, such as temperature and wind speed, and they are often costly and labour intensive in large scale operation (Hristov et al., 2011). Therefore, the research focus has shifted to explore indirect methods to quantify NH₃ emission from manure. This includes biomarkers such as manure-N to potassium (K) ratio and N isotopic discrimination (δ ¹⁵N (‰) = [(¹⁵N/¹⁴N) sample – (¹⁵N/¹⁴N) air]/[¹⁵N/¹⁴N] air × 1000) (Hristov et al., 2009; Lee et al., 2011). Due to physical isotopic discrimination, NH₃ emitted from manure is highly depleted in ¹⁵N, which resulted in manure becoming progressively enriched in ¹⁵N over time. This change in manure ¹⁵N was useful to describe cumulative NH₃ emissions from dairy cow manure over a 15-day in vitro incubation (Hristov et al., 2009). However, a follow up study from the same group showed that the positive relationship was only sustained for the first 6 days of in vitro incubation (Lee et al., 2011). The reason for the discrepancy is not clear, but it may be related to different manure properties (Tamminga, 1996) or the incubation environment (Ndegwa et al., 2008; Hristov et al., 2009). Further research is needed to confirm when and how δ ¹⁵N can be used to predict NH₃ emissions from livestock manure.

To the best of authors' knowledge, no study has so far explored the relationship between δ ¹⁵N and emitted NH₃ with sheep manure. Therefore, this study aimed to use lignite application and GM feeding to induce changes in sheep manure NH₃ emissions and investigate the relationship between δ ¹⁵N and emitted manure NH₃ in sheep manure over a 10-day in vitro incubation. Also, the study aimed to investigate how different ratios of sheep urine to faeces can affect NH₃ emissions and the relationship between δ ¹⁵N and emitted manure NH₃ in sheep manure over a 10-day in vitro incubation. The authors hypothesized that the use of lignite application in sheep manure and the inclusion of GM into sheep diets can significantly reduce manure NH₃ emissions.

Materials and methods

Manure preparation in experiment 1

The background experiment was described by Wu et al. (2022) and the summary dietary information for the current in vitro experiment is presented in Table 1. Four sheep were used as donors for urine and faeces. Total excreted urine and faeces were measured and collected from two sheep fed with control (C) and two sheep fed GM diet, and no preservative was used for urine and faeces collections (the sheep were adapted for a period of 14 days on each diet and then, the urine and faeces were collected in a period of 6 days). To reduce NH₃ emissions from excreted urine, the temperature of the sheep urine was kept below 10°C, by cooling excreted urine with ice blocks during the collection process. The collected urine and faeces were kept at −20°C prior to analysis. Fourteen incubation systems were set up, with three different manure mixtures (Table 1): C with 5 replications; C mixed with lignite (C + L) with 4 replications; and GM with 5 replications. For C, urine and faeces were collected from sheep fed a diet of 550 g lucerne hay/kg, 400 g/kg barley grain, and 50 g bean/kg; for C + L, urine and faeces were collected from sheep fed the C diet and 100 g ground lignite was added to each incubation system at the start of the experiment; and for GM, urine and faeces were collected from sheep fed the C diet with 200 g/kg of C diet replaced with GM (fresh matter basis). Faeces samples were removed from the freezer and processed twice using a juicer (Breville BJE410CRO The Juice Fountain Max juicer, Breville, China) to break up faecal particles; faeces were then re-frozen prior to reconstruction with urine to form a manure mixture. Urine and faeces were thawed and immediately reconstructed using a blender (300-W of 600-ml electric portable mini blender with a glass jar, ANKO Ltd, China) based on the urine to faeces volume ratio excreted by the animals for each dietary treatment. A porcelain pestle and mortar (115 mm/4½” diameter, weight 820 g, and glazed finish) was used to grind the lignite. Then, lignite was passed through a 500 μm mesh size sieve.

Table 1. The design and input material for in vitro Experiment 1 and Experiment 2

C, urine and faeces from sheep fed control diet (550 g lucerne hay/kg, 400 g barley grain/kg, and 50 g bean/kg); C + L, urine and faeces from sheep fed control diet (550 g lucerne hay/kg, 400 g barley grain/kg, and 50 g bean/kg) and mixed with 100 g lignite; GM, urine and faeces from sheep fed grape marc diet (control animal feed ration, 200 g/kg replaced with grape marc);

2U:1F, ratio of urine to faeces = 2:1; 1.4U:1F, ratio of urine to faeces = 1.4:1; 1U:1F, ratio of urine to faeces = 1:1; UN, urinary nitrogen; FN, faecal nitrogen.

All presented units are based on each incubation system.

Manure preparation in experiment 2

The design and input material for Exp. 2 are presented in Table 1. Six sheep were used as donors of urine and faeces. Urine and faeces samples were collected without preservatives. The collected urine and faeces were kept in a freezer at −20°C prior to analysis. Faeces samples were processed twice using a domestic blender to break up faecal particles; faeces were then re-frozen prior to reconstruction with urine to form a manure mixture. Urine and faeces were thawed and immediately reconstructed using a blender (300-W of 600-ml electric portable mini blender with a glass jar, ANKO Ltd, China) based on the urine to faeces volume ratio excreted by the animals per dietary treatments. Fifteen incubation systems were set up, with replicates of three different volumes of urine-faeces mixture (Table 1): (1) 2U:1F (urine to faeces = 2:1) with 5 replications; (2) 1.4U:1F (urine to faeces = 1.4:1) with 5 replications; and (3) 1U:1F (urine to faeces = 1:1) with 5 replications. Urine to faeces ratios (g N/g N) of 1.52 and 1.07, and 0.76 were used for 2U:1F, 1.4U:1F, and 1U:1F, respectively. The final manure volume of each incubation system, for all treatments, was 600 g.

Experimental settings

Ten-day laboratory experiments were conducted to incubate sheep manure to quantify NH₃ emissions, using an acid trap set up. In brief, the incubation system (adopted from Misselbrook et al. (2005)) consisted of an air pump (Aqua One 110, Stellar Ltd, China), an airflow meter (Darhor LZB-3WB 0.15–1.5 l/min, Hangzhou Darhor Technology Ltd, China), a water container (a 900-ml clip container; 12.5 cm high, 13.5 cm wide, 10.5 cm diameter, ANKO Ltd, China), a manure container (a 2.3-l clip container; 16.6 cm high; 16.4 cm diameter, ANKO Ltd, China), and an acid jar (Quickfit Flask Erlenmeyer 500-ml 29/30, Fisher Scientific Ltd, UK). The airflow meter adjusted flows to 1 l/min to maintain pressure, transfer moisture from the water container to the manure container and pass the manure gases into the acid trap. Daily prepared 0.5 M sulphuric acid (H₂SO₄; 500-ml) was used to capture the released NH₃.

The ambient temperature was measured by a laboratory thermometer 3 times/day at 10 am, 4 pm and 10 pm. All incubation systems were also checked for leakage at 10 am, 4 pm and 10 pm by placing an airflow meter before the acid jar for approximately one minute. Fifteen grams manure per incubation system was collected randomly from five different locations, using a straw, and the samples were stored in a freezer at −20°C. At the same time, 2 g manure samples were taken and mixed with 4 ml purified water (pH = 7) and shaken for 30 min prior to measuring pH (pH/ISE and EC/TDS Benchtop Meter-IC-HI5521-02, HANNA Ltd, Australia). Acid traps were replaced daily at 10 am and 15 ml of the acid solutions were sampled and stored in a freezer at −20°C.

Sample analyses

All manure samples were taken from the freezer and freeze-dried (Christ Freeze Dryer GAMMA 1-16 LSCplus, Christ Ltd, Germany) for 5 days. Then, they were ground through a 2-mm screen by tissuelyzer (QIAGEN Ltd, Germany) with a pulse frequency of 30 for 40 s. Manure samples (3 ± 0.5 mg) were weighed directly into tin capsules (pressed, standard weight 8 × 5 mm, Sercon Ltd., Gateway, UK) and analysed for N (g/kg) and ¹⁵N (‰, ¹⁵N comparative to total ¹⁴N plus ¹⁵N) on a 20–20 Europa isotope ratio mass spectrometer (Europa Scientific Ltd., Crewe, Cheshire, UK). On completion of the 10-day incubation periods, the analysis of acid trap samples was performed. Briefly, 15 ml of each acid sample was neutralized with 6 M sodium hydroxide (NaOH) to obtain a pH between 4.5 and 6. Then, NH₄⁺-N concentrations of the acid samples were determined using a Segmented Flow Analyzer (SFA; San + +, Skalar, V 3.2). The limit of quantification was 0.2 mg/l and any values below this were not recorded. Values more than 20 mg/l were obtained using appropriate dilution and recalculation. All used chemicals in the current experiment were of analytical reagent grade, and all fresh acid solutions were prepared with distilled water. The composition of N isotope of manure mixture was expressed as δ ¹⁵N (‰) and calculated as:

$$\eqalign{\delta ^{ 15}{\rm N} &= ( {{\rm R\ sample\ - R\ standard}} ) {\rm /R\ standard, \;\ where}\cr {\rm R}&= \ ^{ 15}{\rm N/}( ^{ 14}{\rm N}{\rm} + ^{ 15}{\rm N)} }$$

The corrected δ ¹⁵N based on sample data at day zero (Δ¹⁵N; ‰) was also introduced as a possible biomarker in these experiments and expressed as Δ¹⁵N using the formula:

$${\rm \Delta }^{ 15}{\rm N} = {\rm \delta }^{ 15}{\rm N}_{{\rm each\ day}}- \delta ^{ 15}{\rm N}_{{\rm day\ zero}}$$

Statistical analyses

A one-way analysis of variance (ANOVA) was conducted to test for statistically significant differences among treatments. In experiment 1, the ANOVA with unequal sample sizes were used, as replication units were uneven. Fishers protected least significant difference (LSD) test were used to compare the mean values of treatments. The significant differences were set at P < 0.05 and trends were declared at 0.05 < P < 0.10. As the incubation systems of each treatment were the replication units in these experiments, data per day from each treatment were analysed using repeated measurements, with treatment as treatment structure and replication as block. The statistical package of GenStat (version 16; VSN International Ltd., Hemel Hempstead, UK) was used for all statistical analyses.

Results

Experiment 1

The results from the first experiment are shown in Table 2. Manure-N (g/kg of DM) differed significantly among treatments (P < 0.001; Table 2). Manure in C and GM had approximately 21 and 24% higher N% than C + L, respectively (P < 0.001; Table 2). In addition, manure pH in C + L was lower than in C and GM (P < 0.001; Table 2). Manure temperature was approximately 3% higher in C + L and GM compared to C (P = 0.006; Table 2). Daily manure-N losses (g) varied among treatments (0.108, 0.028 and 0.090 g for C, C + L and GM, respectively; P < 0.001; Table 2). Added lignite in the C + L led to significant lower manure-N losses in comparison with C and GM groups (P < 0.001; Table 2). Moreover, cumulative manure-N loss (as g/100 g of manure-N) was lower in C + L than C and GM by 81 and 68%, respectively (P = 0.001; Table 2). Daily emitted manure NH₃-N (g) was significantly different among treatments (P < 0.001; Table 2), with far less emitted manure NH₃-N from C + L (0.013 g) compared to C (0.093 g) and GM (0.078 g) treatments. Cumulative emitted manure NH₃-N (as g/100 g of manure-N loss) was lower in C + L than C and GM by 35 and 36%, respectively. (P = 0.020; Table 2). Manure δ ¹⁵N _{(day zero)} was highly significantly different among treatments (P < 0.001; Table 2) with far more enrichment in C + L (10.15‰) compared to C (1.95‰) and GM (2.95‰) treatments. Manure δ ¹⁵N _{(last day)} in C + L (19.12‰) was also significantly enriched than C (8.06‰) and GM (6.78‰) (P < 0.001; Table 2). Manure Δ¹⁵N _{(last day – day zero)} was highly significantly different among treatments (P < 0.001; Table 2) with the highest enrichment in C + L (8.98‰) and the lowest in GM (3.84‰).

Table 2. Manure composition, pH, temperature, nitrogen losses, ammonia emissions, and nitrogen isotopic discrimination in three treatments of Experiment 1; C, C + L and GM

C, urine and faeces from sheep fed control diet (550 g lucerne hay/kg, 400 g barley grain/kg, and 50 g bean/kg); C + L, urine and faeces from sheep fed control diet (550 g lucerne hay/kg, 400 g barley grain/kg, and 50 g bean/kg) and mixed with 100 g lignite; GM, urine and faeces from sheep fed grape marc diet (control animal feed ration, 200 g/kg replaced with grape marc); s.e.m., Standard error of means; DM, dry matter; N, nitrogen; Cumulative manure-N loss: as g/100 g of manure-N; NH₃-N, nitrogen content in the form of ammonia; Cumulative emitted manure NH₃-N: as g/100 g of manure-N loss.

The presented numbers for manure DM, manure-N, manure pH, and manure temperature are based on the average over 10-day incubation period.

Cumulative emitted manure NH₃-N increased (P < 0.001) non-linearly (R ² = 0.99) over the 10 days from 0.156 g, 0.010 and 0.151 g to 0.932, 0.132 and 0.691 g for C, C + L and GM treatments, respectively. Manure δ ¹⁵N also increased non-linearly during the incubation (P < 0.001; R ² = 0.96) in 10 days from 3.9 and 3.4‰ to 8.1 and 6.8‰ for C and GM, respectively. However, a reduction in manure δ ¹⁵N occurred for C + L during the incubation from 20.5 to 19.1‰ (R ² = 0.48; P < 0.001). The relationship between cumulative emitted manure NH₃-N and manure δ ¹⁵N (P < 0.001) was positive and strong for both C and GM (R ² = 0.96 and R ² = 0.93, respectively; Fig. 1a). However, the correlation between cumulative emitted manure NH₃-N and manure δ ¹⁵N was non-significant for C + L (R ² = 0.44; P = 0.128; Fig. 1b). A combined equation for treatments C and GM (Y = 0.0077 X² + 0.1102 X + 0.0645) also showed that the relationship between cumulative manure emitted NH₃-N and manure δ ¹⁵N was positive and highly significant (R ² = 0.88, s.e. = 0.120, P < 0.001). Moreover, a strong positive relationship was found between cumulative emitted manure NH₃-N and manure Δ¹⁵N (P < 0.001) for C and GM (R ² = 0.96 and R ² = 0.93, respectively; Fig. 2a). A moderate negative, but significant, correlation (R ² = 0.44; P < 0.05) between cumulative emitted manure NH₃-N and manure Δ¹⁵N was found for C + L (Fig. 2b).

Figure 1. Relationship between cumulative emitted ammonia-nitrogen (NH₃-N) from manure and manure N isotopic discrimination (δ ¹⁵N) during Experiment 1. (a) C, urine and faeces from sheep fed control diet (550 g lucerne hay/kg, 400 g barley grain/kg, and 50 g bean/kg); GM, urine and faeces from sheep fed grape marc diet (control animal feed ration, 200 g/kg replaced with grape marc); (b) C + L, urine and faeces from sheep feed control diet (550 g lucerne hay/kg, 400 g barley grain/kg, and 50 g bean/kg) and mixed with 100 g lignite. The error bars show standard error (s.e.).

Equations for Fig. 1 (a): C, Equation: Y = −0.023 X² + 0.4817 X – 1.4103, R ² = 0.96, s.e. = 0.059, P < 0.001.

GM, Equation: Y = 0.0123 X² + 0.0205 X – 0.0537, R ² = 0.93, s.e. = 0.055, P < 0.001.

Combined equation of C and GM: Y = 0.0077 X² + 0.1102 X + 0.0645, R ² = 0.88, s.e. = 0.120, P < 0.001.

Equation for Fig. 1 (b): C + L, Equation: Y = 0.0043 X² – 0.2112 X + 2.5772, R ² = 0.44, s.e. = 0.750, P = 0.128.

Figure 2. Relationship between cumulative emitted ammonia-nitrogen (NH₃-N) from manure and manure N isotopic discrimination corrected at day zero (Δ¹⁵N) during Experiment 1. (a) C, urine and faeces from sheep fed control diet (550 g lucerne hay/kg, 400 g barley grain/kg, and 50 g bean/kg); GM, urine and faeces from sheep fed grape marc diet (control animal feed ration, 200 g/kg replaced with grape marc); (b) C + L, urine and faeces from sheep fed control diet (550 g lucerne hay/kg, 400 g barley grain/kg, and 50 g bean/kg) and mixed with 100 g lignite. The error bars show standard error (s.e.).

Equations for Fig. 2 (a): C, Equation: Y = −0.023 X² + 0.3923 X – 0.5624, R ² = 0.96, s.e. = 0.070, P < 0.001.

GM, Equation: Y = 0.0123 X² + 0.0927 X + 0.1127, R ² = 0.93, s.e. = 0.053, P < 0.001.

Combined equation of C and GM: Y = 0.0077 X² + 0.1102 X + 0.0645, R ² = 0.88, s.e. = 0.085, P < 0.001.

Equations for Fig. 2 (b): C + L, Equation: Y = 0.0043 X² – 0.1232 X + 0.8805, R ² = 0.44, s.e. = 0.033, P < 0.05.

Experiment 2

The results from the second experiment are shown in Table 3. Total concentrations of manure-N (g/kg of DM) differed significantly between 2U:1F and 1.4U:1F (P = 0.046; Table 3). Manure pH in 2U:1F varied from 1U:1F (P = 0.010; Table 3). Manure pH in 2U:1F was significantly lower than 1U:1F (P = 0.010; Table 3). Manure temperature was approximately 6% higher in 1.4U:1F and 1U:1F compared to 2U:1F (P = 0.008; Table 3). Daily manure-N losses (g) were 0.098, 0.108 and 0.114 g for 2U:1F, 1.4U:1F and 1U:1F treatments, respectively (P = 0.021; Table 3). Moreover, cumulative manure-N losses (as g/100 g of manure-N) were significant among groups; 41.6, 39.7, and 36.9 for 2U:1F, 1.4U:1F and 1U:1F treatments, respectively (P = 0.002; Table 3). Daily emitted manure NH₃-N (g) was significantly higher in 2U:1F (0.097 g) compared to 1.4U:1F (0.090 g) and 1U:1F (0.089 g) (P = 0.003; Table 3). Cumulative emitted manure NH₃-N (as g/100 g of manure-N loss) was significantly higher in 2U:1F compared to 1.4U:1F and 1U:1F by 18 and 26%, respectively (P < 0.001; Table 3). Manure δ ¹⁵N _{(day zero)} was not significantly different among treatment (P = 0.271; Table 3). Manure δ ¹⁵N _{(last day)} was significantly more depleted in 1.4U:1F (9.33‰) compared to 2U:1F (9.88‰) and 1U:1F (9.71‰) (P = 0.028; Table 3). Manure Δ¹⁵N _{(last day – day zero)} was not significantly different among treatments (P = 0.133; Table 3).

Table 3. Manure composition, pH, temperature, nitrogen losses, and ammonia emissions in three different treatments of Experiment 2; 2U:1F, 1.4U:1F and 1U:1F

2U:1F, ratio of urine to faeces = 2:1; 1.4U:1F, ratio of urine to faeces = 1.4:1; 1U:1F, ratio of urine to faeces = 1:1; s.e.m., Standard error of means; DM, dry matter; N, nitrogen; Cumulative manure-N loss: as g/100 g of manure-N; NH₃-N, nitrogen content in the form of ammonia; Cumulative emitted manure NH₃-N: as g/100 g of manure-N loss.

The presented numbers for manure DM, manure-N, manure pH, and manure temperature are based on the average over 10-day incubation period.

Cumulative emitted NH₃-N from manure increased non-linearly (P < 0.001; R ² = 0.99) over the 10 days (0.828 g, 0.763 and 0.749 g for 2U:1F, 1.4U:1F, and 1U:1F treatments, respectively). Manure δ ¹⁵N also increased non-linearly (P < 0.001; R ² = 0.98) over the 10 days (7.8, 6.8 and 7.0‰ for 2U:1F, 1.4U:1F and 1U:1F treatments, respectively). The relationship between cumulative manure emitted NH₃₋N and manure δ ¹⁵N was positive and highly significant (P < 0.001) for all treatments (Fig. 3). A combined equation using results from all three treatments (Y = 0.094e^{0.2193 X}) also showed that the relationship between cumulative manure emitted NH₃-N and manure δ ¹⁵N was positive and highly significant (R ² = 0.95, s.e. = 0.066, P < 0.001). Additionally, positive relationships between cumulative manure emitted NH₃-N and manure Δ¹⁵N were highly significant (P < 0.001) for all treatments (R ² = 0.99, R ² = 0.95; R ² = 0.95, respectively for 2U:1F, 1.4U:1F and 1U:1F; Fig. 4).

Figure 3. Relationship between cumulative emitted ammonia-nitrogen (NH₃-N) from manure and manure N isotopic discrimination (δ ¹⁵N) during Experiment 2: 2U:1F, ratio of urine to faeces = 2:1; 1.4U:1F, ratio of urine to faeces = 1.4:1; 1U:1F, ratio of urine to faeces = 1:1. The error bars show standard error (s.e.).

2U:1F, Equation: Y = 0.0823e^{0.2399 X}, R ² = 0.99, s.e. = 0.043, P < 0.001.

1.4U:1F, Equation: Y = 0.0999e^{0.2255 X}, R ² = 0.95, s.e. = 0.068, P < 0.001.

1U:1F, Equation: Y = 0.094e^{0.2193 X}, R ² = 0.95, s.e. = 0.074, P < 0.001.

Combined equation of all three treatments: Y = 0.094e^{0.2193 X}, R ² = 0.95, s.e. = 0.066, P < 0.001.

Figure 4. Relationship between cumulative emitted ammonia-nitrogen (NH₃-N) from manure and manure N isotopic discrimination corrected at day zero (Δ¹⁵N) during Experiment 2: 2U:1F, ratio of urine to faeces = 2:1; 1.4U:1F, ratio of urine to faeces = 1.4:1; 1U:1F, ratio of urine to faeces = 1:1. The error bars show standard error (s.e.).

2U:1F, Equation: Y = 0.0896e^{0.2355 X}, R ² = 0.99, s.e. = 0.103, P < 0.001.

1.4U:1F, Equation: Y = 0.1266e^{0.2255 X}, R ² = 0.95, s.e. = 0.073, P < 0.001.

1U:1F, Equation: Y = 0.1248e^{0.2193 X}, R ² = 0.95, s.e. = 0.077, P < 0.001.

Combined equation of all three treatments: Y = 0.1216e^{0.2147 X}, R ² = 0.95, s.e. = 0.090, P < 0.001.

Discussion

Manure composition effects on manure ammonia emissions and nitrogen losses

In Exp. 1, the application of lignite to manure reduced daily emitted NH₃-N (g) approximately 86% compared to C. This agrees with previous results. Chen et al. (2015) and Sun et al. (2016) showed that lignite application could reduce manure NH₃ emissions by 66 and 29.5% from beef cattle pens, respectively. Impraim et al. (2020) also observed that lignite-amended cattle manure retained 350 to 540 g/kg of N by avoiding NH₃ loss compared to manure that received no lignite application. Lignite's ability to mitigate NH₃ emissions from manure may be related to three major chemical characteristics: low pH (3.69), high CEC (96.8 cmol (+)/kg), and high labile carbon content (up to 200 g/kg) (Whitehead and Raistrick, 1993; McCrory and Hobbs, 2001; Chen et al., 2015). The low pH and high buffering capacity of lignite alters the ratio of NH₄⁺/NH₃ towards NH₄⁺, which is non-volatile, leading to a reduction of NH₃ emissions. It has been proposed that the inclusion of GM into ruminant diets can also be useful for reducing manure NH₃ emissions as GM changes manure property (Lynch et al., 2007). Tannin is present at high concentrations in GM (Spanghero et al., 2009; Nudda et al., 2015), and it has been shown that a ruminant diet supplemented with tannin from grape seed reduces NH₃ emissions (Waghorn et al., 2002; Grainger et al., 2009). However, Scuderi et al. (2019) observed that the inclusion of GM in a dairy cattle ration did not change N parameters (e.g., UN and FN). Our results showed that even though GM treatment had a higher manure-N content, a significant reduction in manure-N loss occurred without significant increase in manure NH₃-N emission.

In Exp. 2, the higher manure pH in 1U:1F in comparison with 2U:1F may have resulted in higher daily manure-N loss from 1U:1F (0.114 g vs. 0.098 g). The average ambient temperature during the experiment was 11°C. As Hristov et al. (2011) highlighted, estimating the influence of diets on potential gas-emission from manure depends on ambient temperature; however, as previously mentioned, in vitro methods have a limitation in that they do not account for the effects of environmental factors such as wind speed and turbulence over the manure surface. A significant lower cumulative manure-N loss (g/100 g) in 1U:1F compared to other two treatments can be partly explained by the lower proportion of urine to faeces volume. It is possible that manure-N could be emitted more in forms other than NH₃, such as through denitrification. This may relate to the lesser contribution of faeces than urine in the manure in 1.4U:1F and 1U:1F. It is generally accepted that manure NH₃ mainly derives from urea exposed to faecal urease (Wilkerson et al., 1997). Thomsen (2000) investigated UN v. FN influences on ¹⁵N in solid sheep manure during both anaerobic and aerobic (composted) storage. In both situations, UN contributed most to total N losses. Lee et al. (2009) investigated UN v. FN effects on gaseous N emission from stored dairy cattle manure. The results demonstrated that in the first 10 days of manure storage, the main source of emitted NH₃ was from UN (i.e., 90 g/100 g). The same results were achieved in the study by Burchill et al. (2019) in cattle. The cumulative NH₃ emissions increased linearly with increasing urine N rate and emission factors.

A high cumulative manure-N loss (g/100 g) and cumulative NH₃-N emissions (g/100 g) occurred in Exp. 1 (except C + L) and Exp. 2. This result was due to the rapid increase in manure NH₃-N concentration during the 10-day incubation. Lee et al. (2011) mentioned that the rapid increase in NH₄⁺ concentration in the manure was due to urinary-urea hydrolysis. In both experiments (except for manure mixture in C + L in the Exp. 1), the recapture of total N loss as NH₃-N from manure in the acid trap was high (~87 g/100 g and ~88 g/100 g for C and GM treatments, and ~98 g/100 g vs. ~83 g/100 g and ~78 g/100 g for 2U:1F, 1.4U:1F and 1U:1F, respectively). This suggests that the acid trap captured NH₃-N emitted from manure effectively. This result is in contrast with the result by Ndegwa et al. (2008), who reported that the efficiency of NH₃-N trapping was reduced with an increase in emitted NH₃-N. The scenario is different for the manure mixture in C + L. The effectiveness of the recapture of NH₃-N from the manure mixture in the acid trap was only moderate (~57 g/100 g), which could simply be the sensitivity of the NH₃-N analysis, as there were much lower losses with this treatment. It seems likely that nitrous oxide emissions may also have occurred for manure C + L, due to nitrification and denitrification. Earlier reports (Bussink and Oenema, 1998; Harper et al., 2000) showed that reduction of nitrate to N₂O and dinitrogen gas (N₂) might be significant sources of N loss from lagoons/retention pond. Jones et al. (2000) showed that several chemical and biological mechanisms might exist for N₂ formation during the storage of manure.

Manure ammonia-nitrogen and nitrogen isotopic discrimination changes over experiment period

In Exp. 1, despite an increase of N in the manure mixture in C + L compared to manure-N in C (2.98 g vs. 2.35 g, respectively), the cumulative emitted manure NH₃-N in C + L was approximately 6 times less than C, which might be due to a lower proportion of water in the C + L (i.e., higher manure DM [g/kg]), and lignite lowering the pH. This result showed the effectiveness of lignite to reduce manure NH₃-N despite more N being present in the C + L mixture. Despite the increase of manure-N in GM compared to C (2.80 g vs. 2.35 g respectively; Table 3), and the manure content being 72 g higher in GM compared to C, the cumulative manure NH₃-N was approximately 44% less than C during 10-day incubation. As UN/FN and the manure-N concentration in GM were less than C (~40%), this might be a reason for the lower cumulative emitted NH₃-N from GM compared to C manure. Another possible reason for the reduced cumulative emitted NH₃-N from GM manure might be because of the reduced ratio of urine to faeces in GM than C. In Exp. 2, the highest cumulative emitted NH₃-N for 2U:1F manure and the lowest cumulative emitted NH₃-N from 1U:1F manure were more likely due to the highest and lowest proportion of urea to NH₃ content of the manure in 2U:1F and 1U:1F manure, respectively, compared to the other treatments. As Tamminga (1996) and Ndegwa et al. (2008) demonstrated, urine to faeces ratio is one of the major factors that influences manure NH₃ emissions.

The results from Exp. 2 suggested that the emitted manure NH₃-N might depend partly on manure-N content and partly on bacterial enzymes in the manure. In general, the levels of NH₃-N emitted in both experiments (except for C + L group) were in the broad range compared to Hristov et al. (2009) and within the range identified by Lee et al. (2011) in large ruminant experiments. All relationships have a similar slope (i.e., Fig. 5), suggesting that for a given amount of NH₃-N release, there was a similar amount of discrimination of N isotopes. The differences in starting values of δ ¹⁵N among experiments highlighted the point that part of the discrimination of N isotopes may derive from other N losses than NH₃. For instance, the loss of N₂O or N₂ from denitrification, which would also lead to discrimination (Bussink and Oenema, 1998; Harper et al., 2000). This is likely reflected in C + L of Exp. 1, which had much less NH₃-N loss. In both experiments, an increase in manure δ ¹⁵N, was similar to the results described by Hristov et al. (2009) and Lee et al. (2011). However, in the Exp. 1, manure δ ¹⁵N for C + L increased from 20.5 to 22.2‰ (day 1 to day 2) and then decreased from 22.2 to 19.1‰. We do not have an explanation for this observation. However, it may be related to (1) some reactions between lignite and manure, causing the discrimination of N isotopes; (2) N loss in gases other than NH₃ in C + L; and/or (3) other N losses involving the discrimination of N isotopes.

Figure 5. Relationship between cumulative emitted ammonia-nitrogen (NH₃-N) from manure and manure N isotopic discrimination (δ ¹⁵N) during experiments : Hristov et al., 2009 (in cattle); Lee et al., 2011 (in cattle); Current experiment (Experiment 1: in sheep): C, urine and faeces from sheep fed control diet (550 g lucerne hay/kg, 400 g barley grain/kg, and 50 g bean/kg); C + L, urine and faeces from sheep fed control diet control (550 g lucerne hay/kg, 400 g barley grain/kg, and 50 g bean/kg) and mixed with 100 g lignite; GM, urine and faeces from sheep fed grape marc diet (control animal feed ration, 200 g/kg replaced with grape marc); Current experiment (Experiment 2: in sheep): 2U:1F, ratio of urine to faeces = 2:1; 1.4U:1F, ratio of urine to faeces = 1.4:1; 1U:1F, ratio of urine to faeces = 1:1.

Manure ammonia-nitrogen emissions in relationship with manure nitrogen isotopic discrimination

Our experiments mainly aimed to test the relationship between δ ¹⁵N of manure and cumulative emitted manure NH₃-N. The values of cumulative emitted NH₃-N and manure δ ¹⁵N relationship for treatments C and GM in Exp. 1 and all treatments in Exp. 2 were in the upper range shown by Hristov et al. (2009) and in the range described by Lee et al. (2011). A strongly positive relationship between cumulative emitted NH₃-N and manure δ ¹⁵N in both experiments (except for C + L) is consistent with the results of other reports in large ruminants (Hristov et al., 2006; Hristov et al., 2009; Lee et al., 2011). Excluding results from the C + L treatment, an increase in manure δ ¹⁵N coincided with an increase in cumulative NH₃-N emissions. Lee et al. (2011) indicated that the rapid increase in δ ¹⁵N might be due to loss of depleted ¹⁵N of NH₃-N, due to urea hydrolysis to NH₄⁺. Therefore, the NH₃-N emissions and the discrimination of N isotopes could directly link. However, the ¹⁵N measurement of NH₃-N was not investigated in these experiments. In Lee et al. (2011) this rapid increase in δ ¹⁵N happened at the beginning of the manure incubation. Lignite δ ¹⁵N measurement (i.e., for C + L) was highly variable in the current experiment, but less variation was observed in N (g/kg) for lignite; therefore, caution is needed when measuring δ ¹⁵N from lignite.

Conclusion

These two laboratory experiments confirmed that manure-N content and manure properties are major factors determining manure NH₃ emissions. The use of lignite application in sheep manure and the inclusion of GM into sheep diets can significantly reduce manure NH₃ emissions. A non-linear positive relationship between δ ¹⁵N of manure and NH₃ emissions was observed during 10-day incubations of manure in both experiments, except for manure treated with lignite. These experiments confirmed previous reports that manure δ ¹⁵N and Δ¹⁵N may be valuable biomarkers for estimating NH₃ emissions from sheep manure.