Pigeon Pea Intercropped with Tropical Pasture as a Mitigation Strategy for Enteric Methane Emissions of Nellore Steers

Simple Summary Intercropping tropical grasses with legumes has the potential to mitigate greenhouse gas emissions in livestock production systems. Here, we evaluate pigeon pea (Cajanus cajan (L.) Millsp.) intercropped with Urochloa spp. for feeding Nellore steers and compared it with a degraded and recovered pasture system during the rainy and dry seasons of the year. We found that including pigeon pea in grazing systems met the nutritional requirements of the animals to obtain higher gains, improving their performance while also reducing the intensity of enteric CH4 emissions, thus contributing to the sustainability of ruminant production based on pastures. Abstract In this study, we evaluate the effects of intercropping pigeon pea (Cajanus cajan (L.) Millsp.) with tropical pastures for feeding Nellore cattle and compared animal performance and enteric CH4 emissions with other pasture-based systems during the dry and rainy seasons of 2021. Thirty-six Nellore steers (with a body weight of 221 ± 7 kg and an age of 15–16 months) were randomly distributed in three treatments with three replicates (in paddocks of 1.5 hectares each): (1) a degraded pasture of Urochloa spp. (DEG); (2) a recovered and fertilized pasture of Urochloa spp. (REC); and (3) pigeon pea intercropped with Urochloa spp. (MIX). Enteric CH4 emissions were estimated using the sulfur hexafluoride (SF6) tracer gas technique, and dry matter intake (DMI) was determined using internal (iNDF) and external (TiO2) markers. Forages were collected by hand plucking after observations of ingestive behavior, and feces was collected after voluntary defecation. The proportion of grass and legume intake was estimated by C stable isotopes, and the forage nutritional quality was determined, while animal performance was monitored monthly, and the stocking rate was adjusted by the “put and take” technique. The results indicated that intercropping pigeon pea with tropical grasses is an interesting strategy for sustainable livestock production based on pastures. The MIX treatment was able to meet the nutritional requirements of the animals, which presented higher performance. In addition, there was a reduction in CH4 emissions up to 70% when expressed per average daily weight gain in comparison to the DEG treatment.


Introduction
Agricultural activity is facing many challenges such as the need to increase food production to meet the growing world population [1] while adapting to environmental this study evaluates the effects of introducing pigeon pea in tropical grass pastures as an intercropped system to feed Nellore cattle, and thus compares animal production para meters and enteric CH 4 emissions with other commonly used systems based on pasture, during the dry (May-October) and rainy (November-April) seasons of the year.

ha in total).
A total of 36 Nellore steers from the herd of Embrapa Pecuária Sudeste (of a body weight of 221 ± 7 kg and an age of 15-16 months) were used as experimental animals randomly distributed in the grazing units (12 animals per treatment-nine noncannulated (tested animals) steers and three rumen-cannulated steers). In each treatment, nine animals were monitored for performance evaluation, six were used for CH 4 measurements using the SF 6 tracer gas technique, and three were monitored for dry matter intake (DMI) measurements. The protocols used for the experimental animals followed the guidelines of the Committee for the Use and Care of Institutional Animals (CEUA) of Embrapa (no 05/2016) and College of Veterinary Medicine and Animal Science of University of São Paulo (no 6228200521).
All pastures were managed under continuous grazing, and a variable number of nonexperimental (regulator) animals were used to adjust the stocking rate by the "put and take" technique as described by Mott and Lucas [19], aiming to maintain a specific intermediate pasture height (of a maximum of 30 cm and a minimum of 15 cm ) as an indirect assessment of forage mass availability [20] in each grazing unit of REC and MIX treatments. When it was not possible to include regulatory animals to adjust the stocking rate, only experimental animals were kept in these pastures. Monthly, the number of animals in each pasture and their weight were monitored to estimate the stocking rate (expressed as the number of animals or animal units (AU) (of a body weight of 450 kg) per ha), animal performance and productivity parameters. The DEG treatment was not managed to maintain a minimum height; however, when the maximum height of the pasture was reached (30 cm), regulator animals were placed in the grazing units. These DEG pastures were classified as degraded using the criterion proposed by Oliveira [21], in which pastures can be considered degraded when there are areas larger than 2 m 2 in the canopy that are covered by invasive plants, in this case, Paspalum notatum Flüggé.
The pastures were established in 1996 with U. brizantha cv. Marandu but were later infested by U. decumbens cv. Basilisk. In 2011, the area was divided into nine grazing units, and six paddocks randomly converted into REC or MIX treatments. The REC pastures were established by liming, fertilizing and pasture management; moreover, in the MIX pastures, pigeon pea was overseeded and replanted every three years due to the decline in plant population over the years. Currently, the pigeon pea stand is around 180,000 plants per hectare; however, this number may decrease over the years due to animal trampling and grazing, adverse weather conditions, plant senescence, and other reasons [16]. Soil sampling was carried out annually in 0-0.2 and 0.2-0.4 m depths, with 12 subsamples in each paddock (the grazing unit). The REC and MIX pastures received the same application of dolomitic limestone EVN (effective neutralizing value = 70) and fertilization with ordinary superphosphate (OS; 18% of P 2 O 5 ) and potassium chloride (KCl; 60% of K 2 O) to reach 70% base saturation, and concentrations of 12 mg dm −3 of phosphorus (P) and 3% of potassium (K) in the soil cation exchange capacity [22]. Nitrogen fertilization (200 kg of N-urea per ha, divided into five applications during the rainy season) was applied only for the REC system. The DEG pastures were not corrected nor fertilized.
During the experimental period that lasted from July 2020 to July 2021, samples for assessing forage quality, dry matter intake (DMI) and CH 4 emissions were collected in two seasons: rainy (January) and dry (July). For performance evaluations, the animals were monitored monthly, and the average daily gain (ADG) data was grouped into dry (July-September 2020 and April-June 2021) and rainy (from October 2020 to March 2021) seasons. In the dry season, the average temperature was 20.4 • C and the average cumulative rainfall was 156 mm, while during the rainy season the average temperature was 22.9 • C and the average cumulative rainfall was 868 mm according to the climatic data obtained from an automatic weather station located near the experimental site (http://www.cppse.embrapa.br/meteorologia/index.php?pg=automatica accessed on 23 August 2021) ( Figure 1). A mineral supplement was provided ad libitum throughout the year (Table 1). During the dry season, the MIX animals received the mineral formulation, while the REC and DEG animals received an adaptation formulation for 14 days and then a mineral-energetic-protein supplementation. During the rainy season, all animals received only the mineral formulation.

Forage Sampling and Chemical Analysis
For collecting samples of forages during the two harvest dates of the rainy (January) and dry (July) seasons, the methodology of a grazing simulation with observations of ingestive behavior described by Sollenberger et al. [23] was used. The forage samples (pastures and pigeon pea) were separately collected by hand plucking (±150 g of fresh matter) in three consecutive days, observing the animals for approximately 24 min, and using scissors to cut the portion of forages the animals were consuming. Samples were stored in paper bags (18 cm × 44 cm), weighed, and then dried in a forced ventilation oven at 65 • C for 72 h (pigeon pea samples were dried at 40 • C until the sample weight became constant to not compromise the tannin analysis), milled to 1 mm in a Willey-type mill and subjected to chemical analysis at the Animal Nutrition Laboratory of Embrapa Pecuária Sudeste.
The chemical analysis of the forages was based on the content of DM (DM at 105 • C; Method 934.01, AOAC [24]). Concentrations of mineral matter (MM; Method 923.03), crude protein (CP; Method 920.87) and ether extract (EE; Method 920.85) were determined according to the AOAC [24]. Neutral detergent fiber (NDF), acid detergent fiber (ADF) and lignin (Lig) concentrations were analyzed according to Goering and Van Soest [25]. Gross energy (GE) was determined using a bomb calorimeter (IKA WERKE ® , model C 500). Condensed tannins (CT) concentrations were evaluated using the methodology proposed by Makkar [26]. The isotope ratio of C ( 13 C/ 12 C) of the forage samples were determined using a continuous-flow isotope ratio mass spectrometer (Delta Plux, ThermoFisher Scientific ® , Bremen, Germany) coupled to an elemental analyzer (CHN-1110, Carlo Erba, Rodano, Italy) at the Laboratory of Isotope Ecology of the Center for Nuclear Energy in Agriculture (LEI-CENA/USP), and calculated as: where R is the ratio of 13 C/ 12 C and Pee Dee Belemnite is the internationally recognized standard.

Dry Matter Intake and Dry Matter Digestibility
The total DMI (kg DM/day) was estimated by the sum of forages and mineral supplements consumed by the animals: where DMI = total dry matter intake (kg DM/day); DMI f = forage dry matter intake (kg DM/day); DMI s = mineral supplement intake (kg).
The mineral supplement intake was estimated by the difference between the amount provided and the amount of supplement leftovers in the trough after five days. For this measurement, a digital scale (1-10,000 g) was used, and the calculation followed the equation:

Total Weight
where DMI s = mineral supplement intake (kg/kg of BW per day); DMI sSupplied = total supplement provided (kg); DMI sLeftovers = mineral supplement leftovers after 5 days (days); Total Weight = total weight of animals with access to that (kg).
To determine the forage DMI (DMIf), indirect methods with external (titanium dioxide, TiO 2 ) and internal (indigestible neutral detergent fiber, iNDF) markers were used. TiO 2 in small paper capsules was instilled with the aid of an oral applicator. The external marker was administered for 10 days in the amount of 15 g per animal per day. In the last 5 days of TiO 2 administration, feces samples were collected after spontaneous defecation in the paddocks. The feces samples were frozen in properly identified plastic bags, then thawed, homogenized, and dried at 65 • C for three days.
After drying, the samples were ground in a Willey-type knife mill with 2 mm sieves. Subsequently, an analysis of iNDF and TiO 2 was performed using the technique described by Mertens [27] and Titgemeyer et al. [28], and DMI f was calculated according to the equation: where DMI f = forage dry matter intake (kg DM/day); fecal output = TiO 2 supplied/TiO 2 recovered in feces (kg/day); iNDF (feces) = feces content of indigestible neutral detergent fiber (%); iNDF (forages) = forage content of indigestible neutral detergent fiber (%). The dry matter digestibility (DMD) was calculated through an indirect method, using the following equation adapted from Givens et al. [29]: where DMD = dry matter digestibility (%); DMI total = total dry matter intake (kg); and fecal output = TiO 2 supplied/TiO 2 recovered in feces (kg/day). Feces samples were also analyzed for their C isotopic composition as previously described, and the principle of isotopic differences between C 3 and C 4 plants was used to estimate the intake proportion of each forage (tropical grasses-C 4 ; and pigeon pea-C 3 ) following the equation described by Norman et al. [30] and Ovani et al. [31]: where A = δ 13 C value in feces; B = δ 13 C value of the C 4 plant; C = δ 13 C value of the C 3 plant.

Animal Performance
To determine the performance variables, the animals were weighed on a hydraulic trunk with a built-in scale (Parede Móvel Hidráulico/idBeck 3.0-BechHouser ® , 2009) after 16 h of fasting, and this was repeated every 28 days. The individual performance was evaluated by the animals' average daily gain (ADG) obtained by dividing the body weight (BW) difference between two successive weighing measurements performed in an interval of days between the measurements, according to the equation: where ADG = average daily gain (kg); fBW = final BW, most recent animal weight (kg); iBW = initial BW, animal weight from previous weighing (kg); IW = interval between weighing (days). The stocking rate (SR) was expressed in animal units (AUs) and the number of animals per hectare, assuming that one AU is equivalent to 450 kg of animals of the Zebu breed according to the equation: where SR = stocking rate (AU ha −1 ); BWtotal = total body weight of tracer and regulator animals present in the experimental area (kg); AU = animal unit (450 kg); Area = experimental unit area (ha). The feed conversion ratio (FCR), and feed efficiency (FE) were calculated using the following formulations: where FCR = feed conversion ratio (kg DMI/kg ADG); DMI = dry matter intake (kg DM/day); ADG = average daily weight gain (kg); FE = feed efficiency (kg ADG/kg DMI).

Enteric CH 4 Emission
The SF 6 tracer gas technique [32][33][34][35][36] was used for measuring enteric CH 4 emissions from rumination, eructation and breathing. Fourteen days before gas sampling, the animals were fitted with gas collection halters to allow acclimatization in an adaptation period. Seventy-two hours prior to the sampling period, a small brass permeation tube was placed in the rumen allowing the tracer gas to equilibrate in the ruminal environment. Each animal was sampled daily (24 h) for five consecutive days. The gas samples were obtained continuously through a capillary tube connected to a collecting container placed on the neck of the animal. A halter with a 0.127 mm stainless steel capillary tube and a 15 µm in-line filter was placed on the animal's head and connected to an evacuated sampling vessel. Before the experiment, collection canisters made of polyvinyl chloride (PVC) were attached to a vacuum pump in the laboratory to create a negative pressure (of around −13.15 psi). As the vacuum in the sampling vessel slowly dissipated, the negative pressure continuously drew the air sample around the animal's mouth and nose.
Additional PVC canisters were placed near the experimental pastures to monitor the ambient daily concentration of CH 4 and SF 6 during each sampling period. Sampling was performed daily at 07:00 h when the animals were removed from the paddocks and transferred to the working facilities of Embrapa Pecuária Sudeste. After gas sampling, pure nitrogen was added to the canisters and then CH 4 and SF 6 were measured using gas chromatographs (Agilent HP-6890, Wilmington, DE, USA; and Shimadzu ® GC-2014, Kyoto, Japan).
The CH 4 flux was calculated following the equation: where QCH 4 = CH 4 emission rate per animal; QSF 6 = known SF 6 emission rate from the capsule in the rumen; (CH 4 ) Y = CH 4 concentrations in the collection device; (CH 4 ) b = basal concentration of CH 4 ; (SF 6 ) Y = SF 6 concentration in the collection device; and (SF 6 ) b = basal SF 6 concentration in the ambient. The gross energy intake (GEI) was calculated by multiplying the DMI (kg) and diet GE (MJ/kg) content, and the CH 4 conversion rate (Ym, the percentage of GEI converted to CH 4 ) was calculated using the following equation, considering 55.6 MJ/kg the heat value of CH 4 :

Statistical Analysis
For the statistical analysis, the grazing units (paddocks) were considered the experimental units and data were analyzed using the statistical software SAS 9.4 (SAS Institute Inc., Cary, NC, USA). Before the analysis, outliers were identified, and the normality of residuals was tested (Shapiro-Wilk). When the normality assumption was not accepted, the logarithmic transformation was applied, and then the data were analyzed using the mixed procedure (PROC MIXED) testing different covariance structures and choosing the best fitting model based on the lowest value of the corrected Akaike information criterion (AICC) [37]. The statistical model included the three pasture-based grazing systems and seasons (dry and rainy) as fixed effects, and the interaction between treatment and season was tested. Fixed effects were considered significant at 5%, and in the face of the treatments × seasons interaction the effects of one factor on the other were evaluated using the SLICE command of PROC MIXED. Finally, all means were estimated according to the least squares test (LSMEANS) and the multiple comparisons were performed using the GLIMMIX procedure applying the Fisher's test through the PDIFF LINES option.

Pigeon Pea and Tropical Grasses Chemical Composition
The chemical composition and content of condensed tannins of pigeon pea and Urochloa spp. sampled during the dry and rainy seasons through hand plucking following the methodology of grazing simulation are presented in Table 2. Crude protein (CP), neutral detergent fiber (NDF), acid detergent fiber (ADF), lignin (Lig), ether extract (EE), mineral matter (Ash), neutral detergent insoluble nitrogen (NDIN), acid-detergent insoluble nitrogen (ADIN), gross energy (GE) expressed as calories per gram (cal/g) and condensed tannins (CT), expressed as eq g leucocyanidin/kg DM (dry matter), and carbon isotope ratio (δ 13 C). SEM: standard error of the mean.
The isotopic results (δ 13 C value in feces = −13.8 ± 0.23‰ in the rainy season, and −18.7 ± 1.33‰ in the dry season) indicated that there was little intake of the legume during the rainy season (4.1 ± 0.01); however, during the dry season the intake of pigeon pea reached around 41% (40.7 ± 0.11) of the diet in the MIX treatment. The CP, NDF and ADF content of the forages in the different pasture-based systems, considering the proportion of Urochloa spp. (C 4 ) and pigeon pea (C 3 ) intake for the MIX treatment estimated by stable isotopes are presented in Table 3.
Treatment affected CP, NDF and ADF, while NDF was also affected by the different seasons (p < 0.05). When compared to DEG and REG, the MIX treatment presented 45 and 25% greater values of CP (p = 0.0016). Additionally, MIX presented approximately 12% lower values of NDF (p = 0.0013) and 14% lower values of ADF (p = 0.0012) compared to the other treatments. During the dry season, a 10% lower value of NDF was found compared to the rainy season (p = 0.0004). The decomposition of the treatment × season interactions is present in Figure 2. Similar results of EE, CT and DMD were found among the treatments during the rainy season (Figure 2b,d,e) (p > 0.05). However, during the dry season, greater values of these variables were found for MIX compared to the REC and DEG treatments (Figure 2b

Forage and Mineral Supplement Intake
The average values of the forage and mineral supplement DMI during the experimental period are presented in Table 4. When the DMI was expressed as %ABW, no effect of treatment (p > 0.05) was found for the forage, total, and supplement DMI. Nevertheless, when expressed as kg DM/day, a 43% lower supplement DMI was found in the MIX treatment (p = 0.0068). In addition, when expressed as kg DM/day, 70% higher values of forage (p = 0.0118) and total DMI (p = 0.0132) were found in the rainy season, while when expressed as %ABW, a supplement DMI of 2.4 times greater was found in the dry season (p = 0.0420).

Animal Performance and Stocking Rate
The average values of iBW, fBW, ADG, FCR and FE in the three different pasture-based treatments in the rainy or dry seasons during the experimental period are presented in Table 5. Approximately 16 and 9% greater values of fBW (p = 0.0165), and 57 and 21% greater values of ADG (p = 0.0008) were found in the MIX treatment compared to DEG and REC. The rainy season presented a 72% lower value of FCR (p = 0.0175). A FE value of three time greater (p = 0.0063) and an ADG value of six times greater (p <.0001) were found during the rainy season. In fact, all treatments presented greater ADG values in the rainy season, and the MIX treatment showed greater ADG value (p = 0.0546) compared to DEG and REC in both seasons. Interactions between season and treatments were found for the stocking rate parameters (p < 0.0001) (Figure 3). For the MIX treatment, a higher stocking rate value was found in the dry season of the year (p < 0.0001), which could be related to the greater forage biomass usually obtained when including Cajanus cajan (L.) Millsp. in pasture systems. However, during the rainy season, the highest stocking rate was found in the REC treatment (p < 0.0001).

Enteric CH 4 Emissions
The average values of enteric CH 4 emissions per hectare, AU and GEI using the SF 6 tracer gas technique in the different treatments and seasons, as well as the statistical pro-abilities, are presented in Table 6. No effect of treatment or season was found when expressed per hectare and AU (p > 0.05). When expressed as CH 4 /GEI, 59% greater emission was found in the rainy season of the experimental period (p = 0.0212). Interactions between season and treatments were found for the enteric CH 4 emission parameters and Ym (p < 0.05) (Figure 4a-e). No differences among treatments were found in the rainy season of the experimental period (p > 0.05). During the dry season, lower values of enteric CH 4 emissions and Ym were found for MIX compared to both the DEG and REC treatments (p < 0.05) (Figure 4a-e). Additionally, during the dry season, lower emissions of CH 4 per ADG were found for REC compared to the DEG treatment (p > 0.05) (Figure 4b). In comparison to the DEG treatment, there was a reduction in CH 4 emissions in the MIX treatment of up to 70% when expressed per ADG during the dry season of the experimental period (3699.62 vs. 873.04 g CH 4 /kg ADG; Figure 4b).

ER REVIEW 13 of 20
and REC treatments (p < 0.05) (Figure 4a-e). Additionally, during the dry season, lower emissions of CH4 per ADG were found for REC compared to the DEG treatment (p > 0.05) (Figure 4b). In comparison to the DEG treatment, there was a reduction in CH4 emissions in the MIX treatment of up to 70% when expressed per ADG during the dry season of the experimental period (3699.62 vs. 873.04 g CH4/kg ADG; Figure 4b).

Discussion
The hand plucking technique was used for separately sampling the forage in the different treatments, and the isotopic analysis allowed the estimation of the proportion of Urochloa spp. and pigeon pea intake in the MIX treatment. A recent review article by Castro-Montoya and Dickhoefer [17] pointed out that there are 18 in vivo trials with pigeon pea being fed to ruminants, and to the best of our knowledge, this is the first study reporting the nutritional quality of a diet composed by Urochloa spp. and pigeon pea in an intercropped pasture-legume system for feeding Nellore cattle in Southeast Brazil.
Efficient digestion by ruminal microorganisms requires at least 7% of CP [38], and during both seasons the CP values of all treatments were above the minimum and were consistent with the values reported for fertilized and unfertilized Urochloa spp. pastures [39]. Additionally, the CP content of Cajanus cajan (L.) Millsp. sampled during both dry and rainy seasons ( Table 2) were in line with those reported by Miano et al. [40], Hampel et al. [41] and Valadares Filho et al. [42] (17 to 24% CP), resulting in an average CP content

Discussion
The hand plucking technique was used for separately sampling the forage in the different treatments, and the isotopic analysis allowed the estimation of the proportion of Urochloa spp. and pigeon pea intake in the MIX treatment. A recent review article by Castro-Montoya and Dickhoefer [17] pointed out that there are 18 in vivo trials with pigeon pea being fed to ruminants, and to the best of our knowledge, this is the first study reporting the nutritional quality of a diet composed by Urochloa spp. and pigeon pea in an intercropped pasture-legume system for feeding Nellore cattle in Southeast Brazil.
Efficient digestion by ruminal microorganisms requires at least 7% of CP [38], and during both seasons the CP values of all treatments were above the minimum and were consistent with the values reported for fertilized and unfertilized Urochloa spp. pastures [39]. Additionally, the CP content of Cajanus cajan (L.) Millsp. sampled during both dry and rainy seasons ( Table 2) were in line with those reported by Miano et al. [40], Hampel et al. [41] and Valadares Filho et al. [42] (17 to 24% CP), resulting in an average CP content of the MIX treatment being 25-45% greater than that of the REC and DEG during the experimental period, which may have contributed to the better animal performance of this treatment. In forage diets, the NDF content is one of the determinants of forage intake [43], and the DEG and REC treatments showed NDF and ADF values in line with those reported for Urochloa spp. under tropical conditions [39,44,45]. For the MIX treatment, NDF content was lower than those found for pastures intercropped with pigeon pea [41]. Additionally, the lowest NDF and ADF values of the MIX treatment were in line with those reported by Alves et al. [46] and Pereira et al. [47] when evaluating the nutritional quality of Cajanus cajan (L.) Millsp. However, the value of ADF in the MIX treatment was higher than that reported for a consortium of Panicum maximum Jacq. and pigeon pea [41]. On the other hand, the Lig values of the MIX treatment were lower than those found for this consortium [41]. The mean EE values of the DEG and REC treatments were similar to those reported by Sá et al. [48] for C 4 pastures composed mainly of Urochloa spp, and the EE values of MIX treatment are similar to those found by Vitti et al. [49] and Castro-Montoya and Dickhoefer [17] when evaluating the nutritional quality of pigeon pea. In both seasons, the GE content of DEG and REC treatments were similar to those found by de la Mora et al. [50], while the GE of the MIX treatment was similar to that for a pigeon pea green forage in the Brazilian Tables of Feed Composition for Cattle [42].
In both seasons, the CT content of the DEG and REC treatments were higher than the content of other tropical grasses reported by Bueno et al. [51], while the MIX treatment presented values higher than those found by Pereira et al. [47] for pigeon pea, and values lower than those found in a tropical pasture intercropped with pigeon pea [41]. Some studies have shown that feed consumption by ruminants can be reduced when the concentration of TC exceeds 50 g CT/kg DM, due to the reduction in acceptability and conditioned aversion [52,53]. As the level of CT found here for all treatments was below this value, no negative effect was seen on the consumption of the diet, as other authors have shown when using diets with similar CT contents, irrespective of the plant used [54][55][56][57]. In addition, according to Perna Junior et al. [58], values of around 20 to 45 g CT/kg DM are sufficient to interfere in the digestive process of ruminants.
The DMD of the MIX treatment during the dry season was higher than that reported for a pigeon pea green forage [42], while the DMD value of REC was similar to those found by Dias et al. [59] and Euclides et al. [60]. During the dry season the DMD of MIX was greater than that of both the DEG and REC treatments, a similar result to that found by Epifanio et al. [61] when evaluating Urochloa spp. intercropped with the legume Stylosanthes spp., which revealed an increase in digestibility compared to pastures composed only with grasses. A possible explanation for the higher DMD value of MIX is some of the associative effects between forages on feed digestion [62]. Increased digestion when a low-quality forage is supplemented by a legume with high nitrogen content can be attributed to the stimulation of the microbial activity and modification of digestive processes in the rumen, including proteolysis and CH 4 production when secondary metabolites such as tannins, saponins or polyphenol oxidase are present in low quantities [62].
During the dry season, the forage and total DMI were lower than those found in the rainy season. These results can be justified by the structure of the vegetation, lower acceptability, presence of antinutritional compounds, lower passage rate of food through the gastrointestinal tract and lower forage availability in the dry season of the year, in addition to factors inherent to the animals such as breed, sex and age [63,64]. The DMI of REC was lower than that found by Meo-Filho et al. [65]. For all treatments, the DMI during the rainy season was approximately 1 kg lower than that described by Barioni et al. [66] in DMI tables for Nellore steers under grazing conditions. In the dry season, the DMI values for all treatments were similar to those reported by Barioni et al. [66]. In a meta-analytical approach evaluating zebu animals grazing Urochloa spp. with mineral and energy/protein supplementation [67], the DMI results were lower than those found in this study, with the average performance of animals consuming only mineral supplementation being similar to those found in the DEG treatment. In addition, energy/protein supplement consumption was around 1 kg per animal [67], a value above that found in the DEG and REC treatments. The weight gain of the animals receiving the energy/protein supplement, in amounts of around 580 g per day, was greater than that found in the treatments of this study.
Daily DMI is a very important factor in ensuring the release of nutrients for maintenance and production. Tulu et al. [68] found considerable variations in DMI among pigeon pea genotypes. Usually during the dry season, tropical grasses present low nutritional quality and forage availability, and these could explain the lower forage and total DMI found in this study during this season. Additionally, in the dry season, animals preferentially consume more supplements to enhance the use of diet substrates and optimize animal performance and feed efficiency by ameliorating the pasture's nutritional composition [69,70], and a higher intake of the supplement was found during the dry season when expressed as %ABW. However, despite the different composition of mineral supplements throughout the year among the treatments, a lower supplement DMI was found for the pasture with pigeon pea (MIX). This could be attributed to some of the pigeon pea characteristics since it is a legume that reaches its reproductive phase and improved acceptability of its pods and oldest leaves during the dry season of the year, being consumed as an important source of protein [16,71], thus reducing the need for energetic-protein mineral supplements [72]. In times of scarcity and high prices for protein mineral supplements, the introduction of this legume in pasture systems is even more relevant.
The similar iBW evidenced the animals' weight uniformity among the treatments, while higher fBW and ADG values in the MIX treatment compared to DEG and REC indicate greater performance in the pasture intercropped with pigeon pea. A higher performance of cattle on pastures intercropped with legumes was also found by Machado and Sales [73] when comparing these to pastures with Urochloa spp exclusively. Both forage DMI and ADG values were in line with those described by Oliveira et al. [16] for a consortium system using pigeon pea.
It is important to consider that pigeon pea can fix N content and add organic matter to the soil, factors that can contribute to greater forage nutritional quality and availability to the animals. This legume also contributes to the recovery of degraded pastures [16], which represent approximately 70% of pasture areas in Brazil [74,75]. In the DEG treatment, which represents a pasture with some level of degradation, the stocking rate expressed both as the number of animals per hectare and AUs per hectare was lower than that in the other treatments, a fact that could be related to the low persistence and biomass production of the tropical grass in a soil without proper nutritional management [39]. The REC treatment that received nitrogen fertilization showed a higher stocking rate during the rainy season, with values similar to those found by Meo-Filho et al. [65] when evaluating a fertilized intensively managed pasture under rotational grazing with a liming application. However, during the dry season, the REC treatment had a lower stocking rate than MIX. During dry seasons, the seasonality of production and nutritional quality of tropical grasses are observed [63,64], reducing a pasture's support capacity, while it is in this period that pigeon pea begins to be consumed more as an important source of forage for animals, enabling a higher stocking rate [16]. Considering the seasons, a higher feed conversion ratio was found during the dry season, and this is justified by the poorer nutritional quality of the forages. In the same way, the greater feed efficiency found in the rainy season is justified by the better nutritional quality of the forage to which the animals had access during this season [76].
Decreasing the emissions of enteric CH 4 from ruminant production is a strategy to limit the global temperature increase to 1.5 • C by 2050 [77]. During the dry season, when higher pigeon pea intake was observed, CH 4 emissions expressed per animal, per ADG, per ABW and per DMI were lower in the MIX treatment, which can be attributed to some of its nutritional quality and CT content. The effect of tannins on the reduction in enteric CH 4 production is usually related to its direct action by inhibiting the activity of methanogenic microorganisms and/or reducing the digestibility of rumen fiber fractions [78]. Additionally, it is important that the benefits of the reduction in the emission of CH 4 do not hide the possible harmful effects of tannins on nutrient digestibility and production parameters [79]. Further in vitro studies using tannin-binding agents (e.g., polyethylene glycol) evaluating the effects of pigeon pea on diet degradability, ruminal fermentation parameters, ruminal microorganisms and the potential of CH 4 mitigation may contribute to elucidating the results found in this study. In this sense, Berhanu et al. [80] evaluated the in vitro potential of mitigating CH 4 emissions from several legumes, including pigeon pea, and found a lower production of total gases as well as of CH 4 .
When expressed per ADG, the highest CH 4 emission was found for DEG during the dry season of the year, which can be explained by the reduced performance results of this treatment. During the dry season, the MIX treatment showed higher performance results, which contributed to the lower emission intensity found in the system with the inclusion of pigeon pea. When expressed as a percentage of the gross energy intake (Ym), similar values among treatments were found during the rainy season. However, in the dry season, the lowest Ym was found in the treatment with pigeon pea, once again indicating the potential that this intercropped system has in contributing to the sustainability of livestock production based on pastures.
Finally, the results of this study highlight the fact that the inclusion of pigeon pea in pasture-based systems can represent an advantage not only for cattle farmers in raising animals with greater performance, but also for Brazil as a country, which made a commitment to reduce CH 4 emissions by 30% by 2030 during the 26th UN Climate Change Conference of the Parties (COP26), in Glasgow, Scotland.

Conclusions
The introduction of pigeon pea in tropical pastures as an intercropped system used to feed Nellore cattle was able to meet the nutritional requirements of the animals in the MIX treatment. In this treatment, the animals presented a lower intake of the mineral supplement, greater average daily gain, and reduced intensity emissions of enteric CH 4 compared to other pasture-based systems commonly used in Brazil. These results confirmed our initial hypothesis that pigeon pea intercropped with tropical grasses may improve the sustainability of livestock production based on pastures, especially in the dry season of the year, contributing to the reduction of GHG emissions and increasing animal productivity. Future research may consider the inclusion of this legume during backgrounding and finishing beef cattle on tropical pastures, as well as its effects on the ruminal environment, potential carcass production and meat quality.

Data Availability Statement:
The data presented in this study are available upon request from the corresponding author.