Effect of Seasons on Enteric Methane Emissions from Cattle Grazing Urochloa brizantha

The objective of this study was to evaluate the effect of seasons under a tropical climate on forage quality, as well the effect of an Urochloa brizantha cv. Marandu grazing system on enteric methane (CH4) emissions from Nellore cattle in the Southeast region of Brazil. Sixteen Nellore steers (18 months old and initial weight 318.0 ± 116.59 kg of LW; final weight 469 ± 98.50 kg of LW) were used for a trial period of 10 months, with four collection periods in winter (August), spring (December), summer (February) and autumn (May). Each collection period consisted of 28 days, corresponding to the representative month of each season where the last six days were designed for methane data collection. Animals were randomly distributed within 16 experimental plots, distributed in four random blocks over four trial periods. CH4 emissions were determined using the sulphur hexafluoride (SF6) tracer gas technique measured by gas chromatography and fluxes of CH4 calculated. The forage quality was characterized by higher CP and IVDMD and lower lignin contents in spring, differing specially from winter forage. Average CH4 emissions were between 102.49 and 220.91 g d (37.4 to 80.6 kg ani yr); 16.89 and 30.20 g kg DMI; 1.35 and 2.90 Mcal ani d; 0.18 and 0.57 g kg ADG and 5.05 and 8.76% of GE. Emissions in terms of CO2 equivalents were between 4.68 and 14.22 g CO2-eq g ADG. Variations in CH4 emissions were related to seasonal effect on the forage quality and variations in dry matter intake.


Introduction
In Brazil, cattle represent 83.9% of all livestock production (of which 89% is beef cattle and 11% dairy cattle).Extensive production systems predominate and main national herd is composed by Zebu cattle (B.indicus), of which Nellore is the most numerous breed (80%), raised in predominantly extensive systems (Lima et al., 2010).The main food source is tropical forages, especially, the genus Urochloa, which occupy about 50% of the cultivated pastures area due to the low production cost compared to systems using confined animals and grains in the diet (Berchielli et al., 2012).As Brazil has the second largest cattle herd in the world (FAO, 2013) and the feed system is based on tropical forages, the country has been indicated as a methane-producing potential especially when the diet consists in low nutritional value forage (Canesin et al., 2014) which favours the low performance of animals and the increased production of greenhouse gases (GHGs), mainly enteric methane (CH 4 ).However, these emissions could be reduced with cattle supplementation on pasture and proper management of grassland ecosystem which acts in favour of carbon sequestration (IPCC, 2007).The Urochloa brizantha cv.Marandu is a perennial forage grass of cespitose growth habit, forming clumps of up to 1.0 m in diameter and tillers with height of up to 1.5 m.It is well adapted and has good forage production in natural fertile soils; excellent performance in sandy soils; deep root system which allows to obtain water during dry periods; is more palatable than the other species of the genus, and therefore is widely used (Costa et al., 2004).
Cattle farming is a significant source of methane (CH 4 ) gas emissions, an important contributing factor towards global warming (IPCC, 2007).Methane is produced as a result of the natural digestive process of ruminant herbivores.This process occurs in the rumen as result of a symbiotic relationship between ruminant and ruminal microbiota consisting of bacteria, protozoa and fungi.Fermentation that occurs during metabolism, especially of ingested carbohydrates as vegetative matter, is an anaerobic process carried out by ruminal microorganisms which convert cellulosic carbohydrates into short-chain fatty acids, mainly acetic, propionic and butyric acid (Lima et al., 2006).During the fermentative process heat is dissipated over the body surface and carbon dioxide (CO 2 ) and CH 4 gas are expelled into the atmosphere via eructation and respiration to avoid the prejudicial effect of excessive hydrogen (H 2 ) production to animal health (Dukes et al., 1977).Methane emission represent a loss of between 5% and 8% of gross energy intake according to Blaxter and Clapperton (1965), and between 2% and 12% according to K. A. Johnson and D. E. Johnson (1995), while IPCC provides estimates of 3 to 6.5% (IPCC, 2006).Considering that CH 4 production varies in accordance to the physiological state and type of animal (Lima et al., 2006) as well as, with the quantity and quality of ingested food (USEPA, 1990a(USEPA, , 1990b)), various cattle production systems will result in different factors of methane emission.
Global CH 4 emission from enteric fermentation by ruminants are estimated to be in the range of 76 to 92 million tonnes per year (Dlugokenky et al., 2011), approximately 16% of the total from anthropogenic sources, while those originating from animal manures are estimated at 25 million tons of CH 4 per year (Mosier et al., 2004), approximately 5% of the total from anthropogenic sources.Methane emissions from livestock production in Brazil were estimated at around 11.5 million tonnes per year (Lima et al., 2010), taking into account ruminal and manure emission in 2005.Beef and dairy cattle alone account for 97% of enteric CH 4 emissions from agricultural sources in the country, with the other 3.0% attributable to other categories of animals (buffalo, mules, goats, asses, horses, swine) (Lima et al., 2010).This scenario demonstrates the importance of studies that provide data on the real contribution of ruminants under tropical conditions to greenhouse gas emissions as well as providing indications for reducing CH 4 emissions from livestock production via strategic nutritional and feed management (Tamminga, 1992;Holter & Young, 1992).
Bearing in mind this scenario, the objective of this study was to evaluate the effect of seasons on forage quality and consequently enteric CH 4 emission under the climatic conditions of the Southeast region of Brazil over an one-year period, using the SF 6 technique with Nellore cattle grazing Urochloa brizantha cv.Marandu pasture.

Location and Experimental Area
The experiment was conducted at the Institute of Animal Science and Pastures (IZ) that belongs to the São Paulo State Department of Agriculture and Food Supply (SAA) and interacts through the São Paulo Agency of Agribusiness and Technology (APTA).IZ is located in the municipality of Nova Odessa at an average altitude of 560 m with geographical coordinates of 22°42′S latitude and 47°18′W longitude, with soil characterized as dark red yellow latosol (Oxisol).Tropical climate predominates with a mean annual rainfall of 1367 mm and average annual temperature of 21.7 °C.The dry season is characterized by a cold period comprising autumn season (March to May, with a minimum average temperature of 15.5 °C and a maximum average of 28.0 °C and average rainfall of 102.86 mm) and winter season (June to August, with a minimum average temperature of 11.4 °C and maximum average of 25.8 °C and average rainfall of 27.86 mm).While the wet season is characterized by a hot period covering spring season (September to November, with a minimum average temperature of 15.6 °C and a maximum average of 28.8 °C and average rainfall of 109.9 mm) and summer (December to February, with a minimum average temperature of 18.8 °C and average maximum of 30.0 °C and average rainfall of 215.3 mm) (CEPAGRI, 2015).
Evaluation was carried out in an area of 48 hectares, divided into 48 paddocks of 1 ha each.The experimental plot was represented by an area of 3 ha, composed of three paddocks totalling therefore 16 experimental plots.All experimental plots were equally managed so that seasonal effects should remain homogenous across the 16 plots.Grazing was rotated within the experimental area with rest and occupation periods of 56 and 28 days respectively during the dry season and of 42 and 21 days during the wet season.Pasture used for the trial consisted of the species Urochloa brizantha (syn.Brachiaria brizantha) cultivar Marandu (seeded 12 years ego).The only feed supplement daily provided and ad libitum was a complete mineral mix specifically for beef cattle in the stocker or backgrounding phase, with no added protein and energy content.

Animals and Experimental Design
Sixteen Nellore steers of 18 months old (initial weight 318.0 ± 116.59 kg of live-weight (LW); final weight 469 ± 98.50 kg of LW) were used for a trial period of 10 months from July to May with four collection periods, represented by each season, winter (August), spring (December), summer (February) and autumn (May).Each collection period consisted of 28 days, corresponding to the representative month of each season where the last six days were designed for methane data collection.Animals were weighed at the beginning and at the end of each trial period and remained at grazing rotational grazing during the whole experimental period.The animals were randomly assigned to each experimental plot, along with other regulatory animals which were required to maintain adequate management of the pasture.
An experimental design of randomized blocks was used, represented by 16 animals, each one in a separate paddock, totalling 16 trial plots distributed in four blocks over four trial periods, applied in an Urochloa brizantha cultivar Marandu pasture grazing system.In this manner the model takes into account block effects (4) and treatment effects (seasons of the year).

Evaluation of Availability and Quality Analysis of Forage
Forage availability was estimated using the square method (Gardner, 1967) on the first day of data collection.Forage samples were taken at randomly allocated points within the trial areas before animals were released into the plot and shortly after their removal by cutting manually with garden shears at a height of approx.5 cm above the ground.Material was collected by paddock and by season before being dried in a forced air heater at 65 °C for 72 hours.Samples were ground through a 1.5 mm screen for later determination of dry matter content (DM, Method 934.01;AOAC, 1990); mineral content (MM, Method 923.03;AOAC, 1990); crude protein content (CP, Method 920.87;AOAC, 1990); ether extract content (EE, determined gravimetrically after extraction using ethyl ether in a Soxhlet extractor -Method 920,85; AOAC, 1990); and neutral-detergent fibre (NDF), acid-detergent fibre (ADF) and lignin content following Van Soest (1994).Nutritional analyses were conducted at the Animal Nutrition Laboratory of the Institute of Animal Science and Pastures (IZ), Nova Odessa, SP (Brazil).

Dry Matter Intake
Dry matter intake (DMI) of forage and of total digestible nutrients (TDN) was estimated for each animal using the Cornell Net Carbohydrate and Protein System (5.0) program.In vitro dry matter digestibility (IVDMD) was considered to be equal to TDN, and consequently it was considered that 1kg of dry matter consumed was equivalent to 4.44 Mcal of digestible energy (DE) following NRC (1996).

CH 4 Measurement
The internal tracer sulphur hexafluoride (SF 6 ) technique was employed for the measurement of CH 4 , as described by K. A. Johnson and D. E. Johnson (1995) and adapted in Brazil by Primavesi et al. (2004b).
Five days prior to the start of the first gas collection with collection canisters (container in the form of a yoke fabricated from 60 mm class 20 PVC tubing with an internal pressure of close to zero atmospheres), permeation tubes loaded with SF 6 (566.7 mg), calibrated and identified, with known constant release rates (2.06 ± 0.71 mg of SF 6 d -1 ), were introduced into the rumen of the animals to remain until completion of the final trial period.The CH 4 sampling was conducted during winter (August), spring (December), summer (February) and autumn (May) in animals equipped with air-sampling apparatus consisting of a halter (with stainless steel capillary tube) and collection canister (in the form of a yoke) coupled to a metal shut-off valve and quick-connect.A high vacuum pump was employed for this technique with two stages and a digital manometer with a range of 0 to 203 kPa (2 atm.or 29.4 psi or 1,520 mm Hg; on the scale of 0 to 2 atm), which permitted the measurement of the initial and final pressure of each canister during the data collection period.
Before use, the halters were calibrated so as to reach half an atmosphere of pressure after 24 hours of gas collection by using 0.127 mm internal diameter stainless steel capillary tube attached to the halter.The calibration was determined by the length of the capillary tubing.
After the animals had adapted to the presence of the collection canisters (during days 18-22 of each trial period), measurements of CH 4 production using the tracer gas SF 6 were carried out at 24-hour intervals for six days from the 23rd day of grazing.Concentrations of CH 4 and SF 6 were determined by gas chromatography at EMBRAPA Meio Ambiente, in Jaguariúna, São Paulo State (Brazil), using a HP6890 gas chromatograph (Agilent, Delaware, USA), equipped with Flame ionization detector (FID) to 280 °C, megabore column (0.53 mm × 30 m × 15 µm) Plot HP-Al/M (for CH 4 ), electron capture detector (ECD) to 300 °C and megabore column (0.53 mm × 30 m × 25 µm) HP-MolSiv (for SF 6 ), with two 0.5 cm 3 loops coupled to two 6-port valves.The gas chromatography oven was maintained at 50 °C during the analysis.Directly after the sample collection period and before the determination of CH 4 and SF 6 concentrations, the collection canisters were pressurized with nitrogen 5.0 to pressures of 1.3 to 1.5 psi (g), with initial and final dilution readings taken using a portable digital manometer (±0.01), certified for a reading scale of -1 to +2 bar (g) (Druck, model DPI705), to obtain the dilution factor.The calibration curves were established based on certified gas standards (White Martins), with ppm concentrations for CH 4 (4.85 and 20 ppm) and ppt for SF 6 (34 ± 9, 91 ± 9 and 978 ± 98 ppt), as related by Johnson et al. (2007).
Emission rate of CH 4 was calculated based on the known release rate of the tracer in the rumen and concentrations of methane and the tracer in samples from the environment and in the gas samples collected from the animals.Later, the primary data obtained was used to calculate the emission potential in grams of CH  (Holter & Young, 1992); percentage of digestible energy lost in the form of methane -CH 4 (%DE), with digestible energy estimated from the digestive percentage of crude energy; grams of CH 4 per kilogram of average daily gain -(CH 4, g kg -1 ADG) and emission intensity in CO 2 equivalents per kilogram of average daily gain (g CO 2 -eq -1 g -1 ADG).The variables for CH 4 obtained were expressed based on the average CH 4 emissions for the season in which the data was collected.

Statistical Analysis
Results for forage availability and quality as in vitro dry matter digestibility, intake of dry matter (DMI), crude protein (CPI) and neutral detergent fibre (NDFI), as well rates of enteric emissions of CH 4 were submitted to analysis of variance and its effects were evaluated by the Tukey test (P < 0.05) using the Statistical Analysis System program (Version 9.2, 2010).The SAS´s MIXED procedure was used for these variables.The model considered the effect of the season and the block effect as fixed and random variables respectively.

Results
The season of the year was found to have a significant effect on forage availability (P < 0.05) measured in kilograms of dry matter per hectare (kg DM ha -1 ) and on each of the variables representing its chemical composition (Table 1).
Forage availability was confirmed to be greater in winter than spring and summer but not different from autumn (P < 0.05).This greater availability was probably due to the management strategy of deferment (leaving pasture unoccupied to grow before the dry winter season) which the pasture had been subjected to before the trial periods began.However, the forage itself contained higher levels (P < 0.05) of DM, ADF and Lignin than the other seasons.Likewise, NDF content was higher in both winter and autumn than in summer and spring (the last two also differing from each other).CP, MM and EE content and in vitro dry matter digestibility were lower (P < 0.05) in winter compared with the other seasonal trial periods.As for dry matter intake, expressed in kilograms per day (kg d -1 ) or as a percentage of live weight (% LW); crude protein intake -(CPI kg d -1 ); neutral-detergent fibre intake -(NDFI kg d -1 ) and kilograms of digestible dry matter (DDM kg -1 ), as presented in Table 2, there were significant differences (P < 0.05) in terms of seasonal effects for each of these analysed variables.This occurred as a result of the chemical variation in the forage material which was the primary cause of differences in dry matter intake and consequently protein and neutral-detergent intake.Note. 1 DMI = dry matter intake, expressed in kilograms per day or as a percentage of live weight, CPI = crude protein intake, NDFI = neutral-detergent fibre intake, DDM = digestible dry matter; 2 a,b Different letters in the same line differ significantly (P < 0.05) using the Tukey test (PDIFF).
Significant differences in CH 4 emissions were observed among seasons (Table 3).Emissions were highest during summer compared to the other seasons for all evaluated variables, except for CH 4 emissions expressed per kilogram of average daily gain.Summer differed from the other seasons (P < 0.05) in terms of CH 4 emission in grams per day -(CH 4 , g d -1 ), with spring not differing from winter and autumn and with the lowest emissions recorded in winter.

Availability and Quality of Forage
During the wet season (spring and summer) the forage was found to present lower DM, NDF, ADF and lignin content and higher CP content as well as increased IVDMD when compared to winter forage.Euclides et al. (2009) recorded similar results when evaluating the nutritive value of forage in Urochloa brizantha pasture throughout the year.These authors suggested that when FDN and acid-detergent lignin (ADL) content is lower the CP content of the forage and in-vitro digestibility of organic matter (IVOMD) is higher and vice-versa, irrespective of the cultivar analysed.Likewise, the authors cite that, independent of the experimental year of the forage quality analysis, CP content is higher and NDF content lower during the wet season than during the dry season.As such, variation in the nutritive value of U. brizantha cultivars during the year was a consequence of the climatic variations which occurred (Euclides et al., 2008(Euclides et al., , 2009)), as well as the different flowering times of these cultivars (Valle et al., 2004) as these factors are determinants of the potential nutritive value of the forage source.

Nutrient Intake
Dry matter intake in kilograms per day was greater during spring than winter (P < 0.05) with these seasons not differing from summer and autumn for this variable.When dry matter intake was expressed in relation to live weight, once again there was an increase in spring, differing in turn from the other seasons with the lowest intake observed in the autumn.
The DM intake value as a percentage of live weight (% LW) of the grazing animals was close to the 2.5% value recommended by the NRC (1996) for beef cattle during the spring period, but values for winter (2.0%), summer (1.81%) and autumn (1.73%) were considered to be low.This could be attributed to a crude protein deficiency (Table 1) in the forage during these seasons, with CP contents lower than 7%, a percentage which guarantees adequate microorganism activity in the rumen (Van Soest, 1994).
A seasonal effect was observed (P < 0.05) for CPI and NDFI, with both being at their lowest level during winter, while CPI was highest during spring.Likewise, an increase in NDFI was observed during autumn in relation to winter, with these seasons not differing from summer and spring.This effect can be related to the forage chemical composition (Table 1) during each season and consequently to the obtained DMI value.
During spring, summer and autumn greater mass of digestible dry matter was observed (P < 0.05) than in winter.This was to be expected as the winter forage presented higher lignin content (Table 1), and was characterized as lower quality when compared to forage from other seasons.This resulted in a reduction in DMI and consequently in CPI and NDFI, probably due to a lower passage rate and greater retention time in rumen reflected as lower in-vitro dry matter digestibility.

Methane Emissions
The low CH 4 emission rate obtained during winter is related to a DM consumption decrease during this season due to an inferior forage when compared to the other seasons (Kurihara et al., 1999).The forage is characterized as inferior mainly because of its higher fibre content and lower digestibility.This shows a direct relationship among forage quality, DM intake and consequently CH 4 emissions.Various studies mention that the quantity of consumed feed is an important determinant of daily CH 4 emissions in cattle and as such it has been included in all indicators of daily CH 4 production (Blaxter & Clapperton, 1965;Benchaar et al., 1998).Kurihara et al. (1999), found that Brahman heifers, consuming low quality tropical grasses (Angleton grass, Dichanthium aristatum), showed lower DM consumption (3.58 kg of DM d -1 ) and lower CH 4 emissions (113 g d -1 ).However, when animals had access to a higher quality forage source (Rhodes grass, Chloris gayana), consumption increased (7.07 kg of DM d -1 ) and consequently so did CH 4 emission (235 g d -1 ).A study by Nascimento et al (2016), evaluating CH 4 emission from Nellore cattle feeding on Urochloa brizantha hay harvested at different stages of maturity observed CH 4 emission values of between 132.7 and 138.3 g d -1 for the different treatments, values close to those observed in this study.
For CH 4 emissions expressed in g kg -1 DMI, values ranged from 16.9 to 30.3 g kg -1 DMI, with highest emissions during summer (P < 0.05) and lowest emission during spring in relation to the other seasons.Similar values were observed by Possenti et al. (2008), when studying the effects of two different percentages of Leucaena hay content in the diet (20 and 50% DM), obtaining emissions of 20.5 and 16.9 g kg -1 DMI, respectively.Likewise, Nascimento et al. (2016) cited CH 4 values of between 17.4 and 23.4 g kg -1 DMI using Urochloa brizantha harvested at different stages of maturity.
referring to the intensity of CH 4 emissions in CO 2 equivalents per gram of average daily gain (g CO 2 -eq -1 g -1 ADG) for beef cattle, highlighting the importance of continuing with studies determining emission intensities throughout the productive cycle, in different seasons, to supply accurate and precise information on the interaction between animal performance and greenhouse gas emissions.

Conclusions
The results indicate differences in CH 4 emissions resulting from variations in forage quality, dry matter consumption and seasons.Thereby, the high quality forage plants have lower fibre content and higher digestibility, resulting in increased daily dry matter intake, increased weight gain and consequently lower methane emission (g CO 2 -eq -1 g -1 ADG).
Future experimentation should try to account for the large number of variables involved, considering enteric methane emissions per kilo of product (meat) throughout the animal´s productive cycle, whilst also taking into account the importance of measuring emissions of methane and nitrous oxide from the soil as well as carbon sequestration, which could positively compensate for methane emissions.

Note. 1
MW = metabolic weight, ADG = average daily gain, CH 4 , g d -1 = emission potential in grams of CH 4 per day, CH 4 , g kg -1 DMI = grams of CH 4 per kilogram of dry matter intake, CH 4 , g kg -1 DDM = grams of CH 4 per kilogram of digestible dry matter, DEI, Mcal ani -1 d -1 = megacalories of digestible energy per animal per day, CH 4 , Mcal ani -1 d -1 = megacalories of CH 4 per animal per day, CH 4 (%GE) = percentage of gross energy lost in the form of methane, CH 4 (%DE) = percentage of digestible energy lost in the form of methane, CH 4 , g kg -1 ADG = grams of CH 4 per kilogram of average daily gain, g CO 2 -eq -1 g -1 ADG = emission intensity in CO 2 equivalents per kilogram of average daily gain. 2 a,b Different letters in the same line differ significantly (P < 0.05) using the Tukey test (PDIFF).

Table 1 .
Effect of the seasons of the year on availability and quality of Urochloa brizantha cv.Marandu pasture, expressed as percentage of dry matter (% DM)

Table 2 .
Effect of season on dry matter, crude protein and neutral-detergent fibre intake in Nellore cattle grazing Urochloa brizantha cv.Marandu

Table 3 .
Effect of the seasons on enteric methane emissions from Nellore cattle grazing Urochloa brizantha cv.Marandu