1. Introduction
Bovine protein is an important source of nutrients for the increasing human population. However, to produce meat, greenhouse gas emissions are emitted, such as nitrous oxide (N
2O) and methane (CH
4) [
1]. To mitigate greenhouse gas emissions (GHG), the adoption of strategies such as fertilization, pasture management, and animal supplementation with feeds that do not compete with human feed can increase animal production and improve forage usage efficiency, reducing environmental impacts [
2].
During the digestion of the forage in the rumen, microorganisms hydrolyze the starch and polysaccharides of the plant cell wall, producing sugars and volatile fatty acids (VFAs) such as acetate, propionate, butyrate, CO
2, and hydrogen (H
2). Sugars and proteins are fermented and converted to VFAs, ammonia (NH
3), H
2, and carbon dioxide (CO
2). During the feed fermentation, the methanogenic
Archaea present in the rumen uses H
2 to reduce CO
2 to CH
4 [
3].
Methane has a global warming potential 28 times greater than CO
2 [
4] and, in livestock systems, is produced mainly by enteric fermentation through eructation and excreta. CH
4 is the second-largest GHG contributor to global warming [
5,
6]. After the excretion of dung, CH
4 continues to be produced by the anaerobic activities of the methanogenic
Archaea present in the dung. The magnitude of CH
4 production is driven by factors such as moisture, temperature, and soil compaction. They control the ability of
Archeas to use carbohydrates, H
2, and CO
2 for CH
4 production [
6].
According to the IPCC [
4], N
2O has a global warming power 265 times greater than CO
2 in the 100-year horizon and can be emitted from nitrogen fertilizer sources and animal excreta. Nitrous oxide production from animal excreta occurs due to microbial activities during the nitrification and denitrification processes in the soil, primarily driven by the amount of N content in the excreta [
7]. Cardoso et al. [
2] reported that, with increased grazing intensity, N
2O emissions also increase due to the rise of N being returned to the soil through the excreta (mainly from urine). In addition, variables such as soil temperature and moisture drive the N
2O production process [
8].
Nitrification is a microbial process that oxidizes ammonia (NH
3) to nitrate, while denitrification is an anaerobic process that reduces nitrate to gaseous dinitrogen (N
2) [
9,
10]. Butterbach-Bahl et al. [
11] stated that N
2O produced from the nitrification process can be used in denitrification or diffused in the next soil layer before being released into the atmosphere, which may decrease the N
2O emissions.
The “hole in the pipe” model describes all the phases and key drives variables that regulate the nitrification and denitrification processes to produce N
2O. The amount of N
2O produced is regulated by the N availability, while the soil moisture and temperature drive the microorganism activity [
12,
13]. The lack of synchronism between N and energy in the ruminant diet may lead to increases in N-ammoniacal production and its excretion via urine, which increases N availability in the soil pasture to be lost as N
2O [
7,
10,
14].
According to Carvalho et al. [
10], volatilization is the loss of soil N in the form of NH
3; it comes from fertilization with nitrogen fertilizers and N mineralization of animal excreta. NH
3 will be volatilized directly into the atmosphere from ammonium (NH
4+). Supplementation of the diet of animals can improve the utilization of forage and the efficiency of nutrient use and reduce the production of NH
3 and CH
4. For example, Ferrari et al. [
15] found CH
4 emissions around 33% lower than the IPCC emission factor for young Nellore bulls backgrounded in Marandu palisade grass pastures submitted to continuous stocking management and variable stocking rates to maintain a 25 to 30 cm pasture height and supplemented with energy supplements.
Increases in the green leaves proportion in the forage sward and high soluble protein levels in the forage have been found in Marandu pastures managed with variable stocking rates at a pasture height target of 25 cm, 50% grazing efficiency, and 150 kg N applied during the growing season [
16,
17]. Therefore, these higher levels of soluble protein may lead to increases in N losses through volatilization and N
2O emissions in pastures. One option to address this environmental impact is supplementation with supplements that increase the efficiency of N usage by the animals. Animal supplementations have been shown to affect feed digestibility and N metabolism in the animal, providing a better ruminal fermentation and absorption of nutrients such as nitrogen (N), thus reducing emissions from animal excreta [
10,
18,
19].
Supplements that are a source of energy (e.g., corn) increase the energy in the rumen for the growth of the microbial population, leading to an increase in the use of ammoniacal nitrogen and thus decreasing the concentration of nitrogen in the blood, consequently reducing N excretions in urine and the emissions of N
2O and CH
4 [
3,
14]. Furthermore, the efficiency of microbial protein synthesis increases. Hence, the soluble protein of pastures associated with energetic supplementation provides the substrate for protein synthesis, improving the efficiency of use by microorganisms and the performance of the animal [
15].
Ammonia volatilization occurs from nitrogen fertilization and the excretion of dung and urine in the soil. N volatilization is the primary source of N loss to the environment when not used by the ecosystem (soil, plant, and animal) [
14]. The type of nitrogen fertilizer; animal excreta; and variables such as the temperature, precipitation, soil temperature, moisture, nitrification, and pH influence the volatilization [
13].
Energy supplementation can increase the metabolizable energy in the rumen and substrate for the consumption of microorganisms that promote microbiota growth associated with condensed tannins and increase the flux of metabolizable protein from feed, improving the nutritional profile, energy balance, and animal performance [
14]. Tannins form complexes with proteins through hydrogen bonds in the rumen. The potential to form this complex is affected by the proline content, protein’s isoelectric point, pH, ionic strength, and solution composition [
20].
Studies have suggested that providing diets containing tannins to ruminants can reduce protein degradation in the rumen, improving the efficiency of N use and thereby reducing N excretions via urine, reducing the production of CH
4 and ammonia concentration in the rumen and potentiating animal weight gain [
21,
22,
23]. The tannin–protein complex becomes resistant to the action of microorganisms, which reduces the degradation of protein in ammoniacal nitrogen and changes the N excretion route, reducing N in urine and increasing excretion via dung, leading to a reduction in the N
2O production. Tannins can also bind to starches but do not have the same affinity as the complex formed with proteins [
3,
24].
Using tannin extracts or feed that has tannins may alter the excretion pathways of N and binding proteins and reduce the ruminal protein degradation [
19,
25]. In addition, tannins act on ruminal microorganisms, reducing bacterial hydrolytic activities, which may decrease N excretion via urine and can be an alternative to reduce N
2O emissions. Still, this hypothesis requires further studies [
3,
19,
26].
Gemeda and Hassen [
27] conducted studies on the effect of condensed tannins on CH
4 production. They identified that this compound reduces the production of H
2, inhibiting the activity of the population of methanogenic
Archaea. Furthermore, as Aboagye and Beauchemin [
28] reported, tannins affect the population of methanogenic microorganisms acting on fibrinolytic bacteria, reducing the degradation of the fibrous fraction and acting as a sink for H
2, thus reducing the production of CH
4.
The influence of tannins on the efficiency of N use and GHG emission by ruminants, direct or indirect, in tropical areas still needs to be addressed [
29,
30]. Thus, we hypothesized that supplementations with different energy sources associated with tannin will reduce excretions of N through excreta, mitigate N
2O and CH
4 emissions, and reduce the volatilization of NH
3 from urine when compared with supplementation without a tannin source.
Therefore, we aimed to quantify the emissions and possible forms of mitigation of N2O, CH4, and NH3 from young Nellore bulls supplemented with energy sources associated or not with sources of condensed tannin (soybean hull, sorghum grain, and peanut skin) and, moreover, to identify the key variables affecting the CH4 and N2O flux from soil (ammonium, nitrate, moisture, and temperature) and climatic variables (precipitation and air temperature).
2. Materials and Methods
2.1. The Experimental Site
This study was carried out in the Forage sector of the State University of São Paulo (Unesp), Jaboticabal Campus, located at 21°13′ S and 48°17′ W at 595 m altitude, with an average annual precipitation of 1424 mm and an average yearly temperature of 22.3 °C. The rainy season is distributed from September to April, and the dry season occurs from May to August.
The experiment was conducted in a pasture of
Urochloa brizantha cv. Marandu (syn.
Brachiaria brizantha cv. Marandu), which was established in 2001. Eighteen paddocks (1 ha each) were used. The paddocks were equipped with feeders to offer supplements to the animals and drinkers. The soil of the studies was classified as Red Latosol or Ferralsol (word reference base for soil classification), with a clayey texture in the surface layer (0–20 cm) [
31].
The length of the experiments was 127 days (background phase), from 22 December 2020 to 27 April 2021. In the beginning, in December 2020, soil samples were collected at a depth of 0–20 cm with soil probes for the subsequent chemical analysis (
Table 1). From the soil test results, maintenance fertilization was applied as 60 kg of potassium and 80 kg of phosphorus per hectare, according to the recommendations of Bulletin 100 (Sao Paulo State Official Fertilization Guide) [
32]. N fertilization was split into 3 applications at the beginning of December, the end of January, and the beginning of March, using 50 kg of N per application in the form of urea.
Nellore bulls, with an initial average weight of 240 kg, were supplemented with energy supplementation at 0.3% body weight, and the average stocking rate among the treatments was 5.8 AU/ha (1 AU = 450 kg BW) [
33]. Pasture management was carried out in a continuous stocking system with variable stocking rates using the “put-and-take” method [
34]. The management target was to keep the pasture height at 25 ± 5 cm [
35,
36,
37].
The meteorological data observed during the experiment were obtained from the database of the Agrometeorological Station of the Department of Exact Sciences of Unesp, Jaboticabal Campus, located 700 m from the sampling area. The analyses of greenhouse gases and total N were made at the LANA (Animal Nutrition Laboratory) and Forage Laboratory, located at the same university.
2.2. Treatments and Pasture Management
The first experiment was conducted in a completely randomized design to evaluate the greenhouse gas emissions. The treatments were (1) energy supplementation with soybean hull at 0.3% body weight (BW) without a tannin source; (2) energy supplementation with 0.3% BW sorghum grain, around 3% tannin; and (3) energy supplementation with peanut skin 0.3% of BW, around 6% tannin. Six replicates were used in each paddock, three paddocks were used per treatment, and the supplementation was provided daily in the feeder.
The chemical composition of Marandu palisade grass in a companion study is shown in
Table 2 [
33]. The chemical composition of the ingredients of the supplements used in this study is presented in
Table 2.
The tannins in the supplements were determined by the Folin–Denis reaction in spectrophotometry from the extract made with supplements and distilled water samples (
Table 3). The composition of the Folin–Denis reagent was sodium tungstate, phosphomolybdic acid, and phosphoric acid 85%. After the reaction with saturated sodium carbonate solution, a 760 nm spectrophotometry was performed [
38].
After 15 days of adaptation to the energy supplements, fresh dung and urine were collected from approximately 15 Nellore bulls (an average of 5 animals for each treatment). The animals were contained in a corral with a containment trunk under circular management to collect urine and dung immediately after excretion. The excreta were stored frozen in freezers, homogenized, and later applied [
30].
The dung samples were dried in an oven at 105 °C to determine the dry matter contents [
39], and the dung was then grounded in a Willey mill to continue the analyses. The total N of the dung and urine samples and dung carbon content was determined by dry combustion in a LECO CR-412 analyzer [
40]. Nitrate (NO
3−) and ammonium (NH
4+) analyses of the urine samples were performed separately using the same soil analysis methodology. The results of the chemical analyses of the excreta are presented in
Table 4.
2.3. Experiment 1—Quantification of N2O and CH4 Emissions
The experiment was divided into three phases for gas collection to ensure the occurrence of dung and urine excretion at least once in each experimental unit (static chamber, distributed six units per paddock). The CH4 and N2O emissions evaluations began on 22 December 2020; the first excreta application occurred on 30 January 2021 and the second on 15 March 2021. Six chambers were used in each experimental paddock to guarantee the maximum representativeness of the paddock and random distribution of the experimental units. Two were arranged close to the feeders, two centrally, and another two randomly. This strategy was used to better determine the results according to soil topography. Furthermore, this distribution was chosen to ensure the three main areas within a grazing area.
In the first phase, gas sampling occurred for 40 days, and there was no added dung and urine in the chambers, and the collections occurred once a week. The second phase started after the first application of excreta. It lasted 44 days, and the samplings occurred on days 1, 3, 5, and 7 after the excreta application and then once a week until the next application. In the chambers located in the central region and on the sides of the paddock, 1.5 kg of dung was added in one chamber and 1.5 L of urine in another chamber. In the chambers allocated near the feeder, the two excreta (1.5 kg of fresh dung + 1.5 L of urine) were added to better represent the behavior of the excretions of the animals raised in the pasture [
37]. The third phase also lasted 44 days, and new applications of bovine excreta were performed. As a result, the chambers that previously received dung had added urine and vice versa.
Greenhouse gas samplings were performed from 9 a.m. to 10 a.m. during the last two periods [
41,
42,
43]. The N
2O and CH
4 fluxes were quantified using standardized static closed chambers [
30] composed of a metal base and upper part with rectangular shapes inserted 7 cm deep in the soil 15 days before the first collection. The upper part was composed of a plastic container 30 cm high. The metal base was 28 cm in diameter, coated with a thermal insulator, and was positioned on top of the base only at the time of gas collection.
Gas samples were collected with sterile 50 mL polypropylene syringes and stored in vials (Shimadzu vials), with 20 mL sealed, and evacuated at −800 Pa. Samples were collected at 0, 20, and 40 min after chamber closure. The concentrations were quantified by gas chromatography (Shimadzu Green House Gas Analyzer GC-2014; Kyoto, Japan) and calibrated for N2O reading with the injector at 250 °C, column at 80 °C, using N2 as the carrier gas (30 mL min−1), and an electric capture detector at 325 °C. For CH4, H2 was used as a carrier gas (30 mL min−1) flame ionization detector at 280 °C. During gas collection, the temperatures inside and outside the chamber were measured with digital thermometers.
The fluxes of N
2O (N-N
2O in μg N-N
2O m
−2 h
−1) and CH
4 (C-CH
4 in μg m
−2 CH
4 h
−1) were corrected for the NTPs (Normal Temperature and Pressure Conditions) and calculated according to changes in the gas concentration inside the chamber during the incubation period, according to the following equation:
where δgas is the increase in gas concentration during the incubation period (L
−1 μL), δT is the incubation period (h), M is the molar mass of the gas in N or C, Vm is the molecular volume corrected by temperature and pressure at the sampling time (L mol
−1), V is the volume of the camera (m
3), and A is the area that the chamber covers (m
2).
The observed values were multiplied by 24, making it possible to determine the daily emissions since the day has 24 h, and integrated by interpolation, thus calculating the cumulative emission. The calculations included negative fluxes to avoid sampling bias in the results [
44].
In January 2021, samples of 10 cm in diameter and 5 cm in height were collected to determine the soil density. Soil samples were collected near the chambers 0 to 10 cm in depth for the moisture analysis, NO3− and NH4+. Approximately 10 g of soil was dried at 105 °C for 48 h to determine the moisture by the gravimetric method. From the gravimetric method, the volumetric humidity of the samples and the water-filled pore space (% WFPS) were determined using the soil density and particle density (2.65 g cm−3).
To quantify NO
3− and NH
4+, 10 g of fresh soil was mixed at 50 mL 2 M KCl, stirred for 40 min at 240 rpm (rotations per minute), and filtered. The filtered solution was frozen until the N-nitrate (plus N-nitrite) and N-ammonium were determined. To determine the ammonium (NH
4+) and nitrate (NO
3−), the Berthelot reaction [
41] and reduction by vanadium chloride-III [
42] were used, respectively.
Climatic variables such as precipitation and the minimum, average, and maximum air temperature were collected from an agrometeorological station installed near the experimental area (approximately 700 m). At the time of sampling, the soil temperature was evaluated using digital thermometers.
2.4. Experiment 2—Ammonia Volatilization Assessment
Simultaneously, a second experiment was conducted in an isolated area accessed by the animals. In this second experiment, the N losses in NH3 were evaluated by applying animal dung and urine according to the supplementation sources. The experimental design was a randomized block. The treatments were arranged into 2 factors: (1) type of excreta (dung and urine) and (2) type of supplementation (sorghum, soybean hulls, or corn) and treatment to evaluate background NH3 emissions without excretion application.
Semi-open chambers were introduced in a place with the same characteristics as the paddocks of the experiment; however, there was no grazing. After applying bovine excreta, NH3 samplings were performed at 5 p.m. on the first, third, fifth, ninth, fourteenth, and twenty-first days. The length of the experiment was chosen based on the capture of N volatilized. After the ninth day, NH3 volatilized dropped to the background level in this study.
The volatilized NH
3 was collected using static-free semi-open chambers (SALE) described by Araújo et al. [
45], elaborated with a transparent plastic bottle of ethylene polyethylene (PET) with a volume of 2 L and an area of 0.008 m
2. An ammonia absorber system was installed in the chamber composed of a foam slide 0.3 cm thick, 2.5 cm wide, and 25 cm long and soaked with an acid solution (H
2SO
4) plus glycerin (2%
v/
v) inside a container with a volume of 50 mL, hanging vertically. The N-NH
3 retained in foam was determined by distillation and titration, according to the Kjeldahl method [
46].
2.5. Statistical Analyses
Data were analyzed for the homoscedasticity of the variances and normality of the residues. Then, ANOVA was performed, and when significant at 5% probability, the means were compared using Tukey’s test. In Experiment 1, the total N
2O and CH
4 emissions were compared. The statistical model used for ANOVA was:
in which Y
j is the observed value of the gas production in treatment
i (i = 1, 2,..., I) and in repetition
j (j = 1, 2,..., J); m is the overall average (of all observations) of the experiment; t
i is the effect of the supplementation type
i; e
ij is the error associated with the Y
ij observation or effect of uncontrolled factors on the Y
ij observation.
In Experiment 2, the total percentage of volatilized N was compared following the statistical model:
where
μ = general mean; S
i = effect of the excreta type
i; E
j = effect of the supplementation strategy
j; SE
ij = interaction effect between excreta type
j and supplementation strategy
i; ε
ijk = random error associated with each observation.
Finally, to identify the drivers of GHG production, single and multivariate linear correlation analyses between gas and moisture fluxes, temperatures, ammonium, and soil nitrate were performed to identify the key variables that explain emissions. All statistical analyses were performed using Statistical Program SAS (version 9.2).