Effects of Concentrates Characterized by Different Carbohydrates on Methane Emission by Dairy Cows and Corresponding Effects on Methane Production in the Slurries

Animal production, particularly the keeping of ruminants, has been identified as the largest single source of anthropogenic methane emission. On average, 75% of the methane emitted is released from enteric fermentation while 25% is assumed to result from manure storage. There has been a strong interest for many years to reduce enteric methane, since it represents an energy loss for the ruminant, which can vary between 2 and 12% of the energy intake, and due to the growing awareness of enteric methane’s significant contribution to the global greenhouse gas emission. In animal nutrition, one of the major impacts on methanogenesis in ruminants is the type of carbohydrates fed, where easily-fermentable carbohydrates have been suggested to have a significant reducing effect compared to hard-fermentable carbohydrates. However, detailed knowledge about methane production of a wide range of carbohydrates is still missing. Recent findings on methane emission by dairy cows and their slurry, suggest that a reduction of methanogenesis in the rumen due to fatty acid feeding could result in an undesirable reduction in fiber digestion. The subsequently increased fiber excretion led to an increased methane production from the slurry, which partly compensated for the decreased methane emission from the animals themselves. With regard to effective methane mitigation strategies in animal nutrition, this would make it necessary to evaluate effects on methane release from slurry as well. Therefore, the aim of the present study was to investigate the effects of a wide range of different dietary carbohydrates on enteric methane production and methane production in the corresponding slurries. For the present investigation twelve dairy cows were fed six different diets in three experimental periods (n=6). The complete diets were calculated to meet maintenance and milk production requirements of the dairy cows, and to be iso-energetic and iso-nitrogenous. The roughage (maize silage, grass silage and hay) to concentrate ratio was 1:1. The six concentrates differed in their carbohydrate sources (lignified cellulose, non-lignified cellulose, pectin, fructan, sucrose, and starch), realized by the inclusion of different feedstuffs. After two weeks of adaptation to the respective experimental diets, feces, urine, and milk outputs were recorded and sub-samples were collected over eight days for laboratory analyses. Additional 10% of daily excreted feces and urine were stored at -20°C for a subsequent slurry experiment. On the fourth and fifth day of the collection periods, the gaseous exchange of the dairy cows was measured in open-circuit respiration chambers. For the slurry experiment, a total of 10 kg of feces and urine were mixed proportionately to the actual excretory proportions. Five kg of water were added to simulate a typical composition of slurry on farms. Methane release was measured by connecting the barrels containing the slurry to the open-circuit respiration apparatus. Depending on the diet used, the experimental data showed a high variation of enteric methane release between 461 and 600 L per cow and day. Surprisingly, the diet with high amounts of sucrose, an easily fermentable carbohydrate, did not decrease methane release compared to diets with high contents of structural carbohydrates. The lowest daily methane release was recorded with the diet having a high content of lignified cellulose. The total amount of methane released from the respective slurries within a period of 14 weeks varied from 89 to 159 L per cow and day, depending on the diet fed. The amount of methane released (L per kg slurry dry matter) was closely correlated (Y = 23 + 0.5 X; R=0.69) with the enteric methane emission of the animals (L per kg dry matter intake). This suggests that, contrary to previous findings where the effects of fatty acid were followed, there is obviously no compensation for methane release from slurry when enteric methanogenesis is reduced by changes in dietary carbohydrate sources. The close correlation between methane produced from animal and its slurry suggests that the modeling of total methane production from dairy farms could lead to a more accurate estimation, especially when applying different feeding regimes. The present experiment is one of the first studies done in this area and shows that more knowledge is needed on how for instance different roughage types and different amounts of concentrates influence the relationship between enteric methane production and methane produced in slurry. 1.0 INTRODUCTION The interest in reducing methane in domestic ruminants is twofold. Firstly, methane formation represents a loss of energy for the animal, which can vary between 2 and 12% of gross energy intake (Johnson and Ward 1996; GigerReverdin and Sauvant 2000). Secondly, in the wake of the Kyoto Protocol in 1997, there is an interest to reduce methane production from ruminants, since it is an important greenhouse gas. Ruminant livestock has been identified as the largest single source of antropogenic methane emission (Mathison et al., 1998). On average, 75% of the methane emitted is released from enteric fermentation, while 25% are assumed to result from manure storage. Enteric methane release has been estimated by using several equations reviewed by Wilkerson et al. (1995), but the factors taken into the equations widely differ and include either intake or digestibility of dry matter, energy, carbohydrates, nonfiber carbohydrates, acid detergent fiber, cellulose, hemicellulose, crude protein, or ether extract. The equations given by Holter and Young (1992) even considered of the amount of milk produced by the cow and its composition. The type of carbohydrate fed to the animal has been identified to have a major effect, most likely due to its effect on pH in the rumen and the microbial population (Johnson and Johnson, 1995). Mostly distinctions were only made between cellulose and hemicelluloses in the fiber fraction, and the easily-fermentable carbohydrates were treated as one unit (non-fiber carbohydrates; Moe and Tyrrell, 1979). The effect of a wider range of carbohydrates on methane release has not been investigated in vivo, but only been estimated through stoichiometric calculation or by compiling data from biliographies. In order to achieve a reduction in methane mitigation, more detailed knowledge on the methane production potential of a wide variety of differently composed diets is required. Most of the studies on enteric methane aim at reducing methane, but unfortunately this often also causes a reduction in fiber fraction. A reduced digestion in the animal would increase nutrient availability for the microorganisms present in the slurry, however, could increase the methane production from the slurry. This could even compensate for the reduced enteric methane production, as a recent study by Külling et al. (2002) has shown. Only very few investigations have been conducted, where both enteric methane and methane production from the slurry storage have been studied on the same animal, although this could be necessary to identify really effective nutritional mitigation strategies. The objectives of this study were to compare six concentrate ingredients characterized by different carbohydrates, and to determine their effect on enteric methane fermentation and the corresponding methane release in slurry stored for 14 weeks. 2.0 MATERIAL AND METHODS 2.1 DAIRY COW EXPERIMENT An experiment including twelve Brown Swiss dairy cows was carried out. Initially the cows were divided into six groups with the smallest possible between-group variation in milk yield. The cows were housed in individual tie stalls with free access to water. At the very start of the trial, the cows were allowed to adapt to the tie stall barn for eight days. Each of three experimental periods was preceded by 5 days in which the animals were gradually accustomed to the experimental diet, followed by a 14-day period adaptation and finally 8-day period of sampling. The six experimental diets (thereof three offered to each cow in the subsequent periods) consisted of roughage and concentrate in a ratio of 1:1. The roughage was composed of maize silage, grass silage and hay in proportions of 0.22, 0.45 and 0.33 of DM. The six concentrates contained either oat hulls (0.5 of total), soybean hulls (0.7), apple pulp (0.54), Jerusalem artichoke (0.68), molasses (0.18), or wheat (0.46), thus modelling lignified cellulose, non-lignified celllulose, pectin, fructan, sucrose and starch, respectively. The complete diets were calculated to be similar in metabolically available protein and net energy contents. This was achieved by including oat, crystalline fat, wheat straw meal, grass cubes, soybean meal or urea in different proportions in the concentrates (Table 1). The diets were fed in amounts which met the individual cow’s requirements for maintenance and milk production. The cows were fed with roughage at 6.30 am, 9.30 am, 2.00 pm and 5.30 pm, while concentrates were fed at 7.30 am, 9.30 am, 2.00 pm and 5.30 pm. During the sampling period, urine, feces and milk were collected quantitatively and sub-samples were taken for chemical analysis. On the fourth and fifth day the gaseous exchange was measured in an open-circuit respiration chamber. The roughage sub-samples from each week, the six concentrates and the bulked feces samples from each period were analysed for DM and total ash (automatically by TGA-500, Leco Corporation, St. Joseph, Michigan, USA) and feed was also analysed for ether extract. The nitrogen and carbon analysis (crude protein = 6.25 × N) were performed with fresh feces samples, acidified urine, milk and feed samples. Feed and feces samples were analysed for crude fiber (CF) (N 962.09; AOAC 1998), neutral detergent fiber (NDF), acid detergent fiber (ADF) and acid detergent lignin (ADL) (N 973.18 C and D; AOAC 1997) on the Fibertec (Fibertec System M, Tecator, 1020 Hot Extraction). The concentrates and roughage samples were furthermore analyzed for their contents of starch by the enzymatic colorimetric method (Bach Knudsen et al., 1987), while low molecular-weight sugars (sucrose, glucose, fructose) and fructan was measured by the enzymatic colorimetric method described by Larsson and Bengtsson (1983). Uronic acids (as constituents of pectin) were measured colorimetrically (Kontron Intryments, Uvikon 922, Milan, Italy) as described by Bach Knudsen et al. (1987). Jerusalem artichoke contains fructan. When applying the method described above for the analysis of low-molecular-weight sugar and fructan in Jerusalem artichoke the sucrose content can be easily overestimated while underestimating fructan, since the sucrose terminal is broken off at the main chain during hydrolysis. Therefore, sucrose in Jerusalem artichoke was further analysed by the HPLC method described by Bach Knudsen and Hessov (1995), and the difference between the sucrose measurement of the enzymatic low molecular-weight sugar and the HPLC-derived value was added to the fructan, while for sucrose the values from the HPLC analysis were used. Table 1. Ingredient composition (g kg DM) of diets. Dietary treatment Oat hulls Soybean hulls Apple pulp Jerusalem artichoke Molasses Wheat Forage Maize silage 112 112 111 112 112 113 Grass silage 224 223 222 224 223 225 Hay 164 164 163 165 164 165 Concentrate type Oat hulls 252 – – – – – Soybean hulls – 351 – – – – Apple pulp – – 271 – – – Jerusalem artichoke – – – 339 – – Beet molasses – – – – 92 – Wheat – – – – 188 228 Oat 99 8 164 – – – Crystalline fat 40 25 9 – 15 28 Wheat straw meal – – – 17 61 106 Grass cubes – 90 – 74 99 79 Soybean meal 109 24 54 60 43 54 Urea – 4 6 9 3 2 2.2 SLURRY STORAGE EXPERIMENT At the end of the feeding experiment a total of 10 kg feces and urine collected during the 8 day sampling period were mixed in the proportion of the daily output of each individual cow (n=6). Five kg of water were added and mixed to simulate the typical composition of slurry found on farm (Külling et al., 2002). Water, urine and feces were mixed in a 60-L barrels and a sub-sample of 200 g was taken for analysis. Methane, carbon dioxide and oxygen emission from the slurry were measured by connecting the barrels to the open-circuit respiration apparatus. An air-tight lid was tied to the barrel with an iron wire. The gas emission of the 36 barrels was measured once a week for nine hours, and this over a period of 14 weeks. At the end of the experiment the slurry was mixed thoroughly. A sub-sample was dried at 60 °C for 72 hours, ground and analysed for NDF and ADF. Slurry samples taken in week 0 and 14 were analysed right after sampling for contents of nitrogen, dry matter and organic matter. 3.0 RESULTS 3.1 ENTERIC AND SLURRY DEGRADATION OF NUTRIENTS Table 2 opposes intake and excretion of the cows fed the diets characterized by different feeds and carbohydrates. Cows fed Jerusalem artichoke caused a higher (P < 0.05) urine output compared to diets characterized by oat hulls, apple pulp and wheat, while differences in fecal output were insignificant. Table 2. OM intake and daily excretion of urine and feces of the dairy cows. Dietary treatment Oat hulls Soybean hulls Apple pulp Jerusalem artichoke Molasses Wheat SEM Pvalue OM intake (kg/d) 14.8 15.8 13.0 14.8 14.2 15.0 1.24 0.729 Total urine (kg/d) 14.0 b 18.2 ab 13.8 b 20.1 a 18.5 ab 14.5 b 1.19 0.002 Total feces (kg/d) 32.5 40.7 28.9 37.9 34.5 35.8 3.56 0.336 Feces OM (kg/d) 6.1 a 5.0 ab 3.9 b 3.8 b 4.0 b 4.4 ab 0.42 0.004 The apparent digestibility of fiber (NDF) by the cows was highest (P < 0.05) with soybean hulls compared to that in cows receiving apple pulp, Jerusalem artichoke, molasses or oat hulls. The oat hull diet resulted in an apparent NDF digestibility lower (P < 0.05) than in the five other diets (Figure 1).