Methane emission intensity in a classical pasture versus an intensive tmr feeding program

Classical comparisons between cows grazing pasture and those fed total mixed rations (TMR) show that, on average, cows on pasture have higher methane (CH4) emission intensity (methane/milk). However, to assess greenhouse gas (GHG) emissions fairly from these two feeding strategies requires quantification of other emission ‘costs’ such as nitrous oxide (N2O), which can be much higher on pasture. In addition, the composition of the TMR must be considered. A TMR based on grain and silage will have a different CH4 emission intensity than a TMR which includes inedible by-products, as it is common in many parts of the world. Recent reviews have suggested that within either feeding strategy, targeting forage quality may have the biggest returns on reducing CH4 emission intensity. With the goal of feeding 9.5 billion people in 2050, the efficiency of human-edible protein generation from pasture versus TMR should also be considered.

Introduction

Despite a general acceptance that a diet with a lower forage proportion (FP, %) will reduce CH4 emissions (CH4, g/kg DMI or milk) in cattle, relatively few direct comparisons of pasture vs. TMR feeding regimes that examine environmental impact relative to production levels actually exist. With increasing concerns about the contribution of livestock to climate change (IPCC, 2006), such comparisons are valuable in identifying ‘best practice’ approaches, but considerations must go beyond CH4 emission intensity. Such a seemingly simple comparison is complicated by the fact that both pasture and TMR composition vary widely, and TMR feeding does not simply represent the exchange of pasture in the diet for grain. TMRs can be based on corn, barley or grass silage/hay and cereal grains, but may also include substantial amounts of crop residues (e.g. cereal straws, sorghum stalks, sugarcane tops/bagasse or maize stover) and grain by-products (e.g. dried distillers grains, soy hulls, wheat bran/middlings or corn gluten meal/feed), the latter categories typically being higher in fiber and lower in starch compared to cereal grains. Pasture may be composed of grasses, legumes and/or shrubs and tree foliage, depending on geographical location. Further, the quality and digestibility of both pasture and TMR feeds will vary based on environment, management, fertilization regime and season or weather conditions, all of which will influence the relative CH4 emission intensity of animals on these feeding programs. Resulting changes in CH4 emission intensity may be direct via changes in chemical composition of the diet or diet digestibility, or indirect via changes in DMI (Dry Matter Intake). Relatively little has been published in the peer-reviewed literature on broadly comparing pasture vs. TMR feeding strategies, likely because of the complexities outlined above. This paper will therefore review some of the sources of variation in CH4 emission intensity within pasture and TMR feeding systems, as well as touch on the broader considerations relevant when attempting to compare these systems.

Pasture

Pasture Quality

Pasture quality and digestibility are influenced by numerous management practices (e.g. fertilization regime, regrowth period, maturity and stocking density) which have the potential to significantly alter CH4 production and CH4 emission intensity (per unit intake or product). In general, CH4 reductions on forage diets are correlated with increases in nutrient quality and digestibility. However, these are also often accompanied with increases in DMI. In a classic study, Blaxter and Clapperton (1965) noted that increased intake of poor quality forage has little effect on CH4 production relative to DMI. However, for intake of high quality forages, increased DMI depresses CH4 relative to DMI and per unit of product (Hammond et al., 2009, 2013). The magnitude of CH4 mitigation, due to DMI, when scaled per unit of production, is generally higher when cattle consume higher quality forages.

Grazing management may be a CH4 mitigation strategy. By grazing forages at optimal maturity, allowing for adequate accumulation of pre-grazing herbage mass, or by using intensive grazing practices, forage quality and/or forage DMI may increase. Increasing quality of grazed forages will increase production efficiency, resulting in a reduction in CH4 emission intensity.

DeRamus et al. (2003) showed that management-intensive grazing resulted in more efficient use of grazed forage and conversion of forage into meat and milk, resulting in a 22% reduction in projected annual CH4 emissions from beef cattle. In Warner et al. (2015), fat and protein corrected milk yield (FPCM) and feed digestibility increased with increased N fertilization (20 vs. 90 kg/ha) and a shorter regrowth interval (3 vs. 5 weeks), when DMI was held constant in a zero grazing study. The high N fertilization rate increased CH4 relative to intake (22.0 vs. 20.2 g/kg DMI) due to higher digestibility, but not relative to FPCM (16.2 vs. 15.4 g/kg FPCM) due to increased productivity. Although the shorter regrowth period did not alter CH4 relative to DMI, it did significantly reduce CH4 relative to FPCM (14.8 vs. 16.8 g/kg FPCM). Warner et al. (2015) did not consider the interaction between increased N fertilization and regrowth period, which would likely be negatively correlated in practice.

Overall these results indicate that pasture management, as well as N fertilization, can significantly alter CH4 emission intensity.

Pasture Type

Between forage types, considerable variation in chemical composition, quality and presence of non-nutritional factors exists. Generally, the C4 metabolic pathway for carbon fixation in grasses leads to a higher rate and degree of lignification in plant tissue compared to C3 pathways, which when consumed by ruminants, may affect DMI and digestibility. In the database used by Hristov et al. (2014), C4 grasses on average had 16% higher NDF content than C3 grasses, suggesting a higher CH4 emission intensity potential, given the greater methanogenic potential of structural vs. non-structural carbohydrates (Moe and Tyrrell, 1979). Some tropical forage such as legumes, shrubs and tree foliage also contain secondary metabolites such as condensed tannins and saponins, which have been shown to inhibit methanogensis in the rumen (Jouany and Morgavi, 2007, Martin et al., 2010).

In a meta-analysis, Archimede et al. (2011) found significantly lower CH4 emissions in C3 vs. C4 grasses (30.0 vs. 33.7 L/kg DMI, and 52.1 vs. 57.7 L/kg digested OM, respectively), and in warm vs. cold-climate legumes (25.9 vs. 30.1 L/kg DMI, and 40.7 vs. 52.4 L/kg digested OM, respectively), after equalizing for fiber content, digestibility and level of intake. Differences in CH4 emission after these factors were equalized implies that the forages differ in other respects. Wilson (1994) indicated that the nature of fiber differs between tropical and temperate forages, with C4 grasses being more lignified. Assoumaya (2007) found that at the same level of digestibility, tropical C4 grasses have a longer retention time in the rumen. Both of these differences may contribute to higher CH4 emissions for C4 compared with C3 grasses.

There has also been interest in high water-soluble carbohydrate (WSC) grasses for their potential as an N mitigation strategy, but their effect on CH4 emissions is unclear (Parsons et al., 2011). Ellis et al. (2012) suggested this was because defining ‘high WSC grass’ by WSC concentration alone is not adequate, and the total chemical composition and digestibility of the grass species must be considered to impact CH4 emissions as well. Staerfl et al. (2012) found no effect of high WSC grass on CH4 emissions in dairy cows.

Comparisons between forages are often contradictory (e.g. Hart et al., 2009, Nishida et al., 2007), likely because unlike other feed ingredients with a narrow range of variation in chemical composition and digestibility, forages – even the same type of forage – can vary widely based on management conditions (as can the level of DMI). For example, Hammond et al. (2011) found no differences in CH4 production (22.5 vs. 23.4 g/kg DMI) in sheep fed either ryegrass or white clover, over a range of intakes, despite a > 2 fold range in readily fermentable CHO:NDF ratio. Sun et al. (2011) also reported similar CH4 production (g/kg DMI) from sheep fed fresh chicory or ryegrass that differed widely in chemical composition.

Additional challenges of pasture studies which may contribute to conflicting results include differences in measurement method accuracy for DMI (zero-grazing vs. grazing) and CH4 (calorimetry vs. SF6 tracer technique), as well as additional challenges controlling variation in grass herbage sampling for nutritional analysis. Interestingly, as a result of the variation in forage composition and quality, and the resulting direct (quantity and type of nutrients digested) and indirect (DMI) effects, Hristov et al. (2014) identified improving forage quality as one of the most promising CH4 mitigation strategies available, upon review of the literature.

TMR

Forage Type & Quality

Methane emissions from cattle on TMR diets will be dependent on the composition of the TMR, which varies widely across the world. As conserved forage (silage, hay) typically comprises 40-60% of a TMR diet, many of the forage type and quality considerations relevant for pasture systems are relevant within a TMR feed. For example, Keady et al. (2012) recently produced a review of the effects of silage quality on animal performance in various production systems in Ireland. They concluded that a 10 g/kg DM increase in digestible OM of grass silage could increase milk yield by 0.37 kg/d, and found that each further week of delay in grass harvesting date reduced digestibility by 3-3.5%. Similar to pasture, positive effects on digestibility will likely reduce CH4 per kg of milk produced, through increases in digestible DMI. Numerous studies have also indicated reduced CH4 production with corn vs. grass silages (e.g. Doreau et al., 2012), and that (comparatively) the corn silage diets tended to increase DMI and milk yield (Dewhurst, 2012). In the review by Dewhurst (2012), lower NDF content and higher DMI of legume vs. grass silages are suggested to decrease CH4 production, but these relationships (legume vs grass silage) are less conclusive, due to similar variation as pasture.

Unlike pasture, silage for use in TMR rations may additionally be treated to improve quality with formic acid or lactic acid bacteria inoculants, for example. These additives are thought to improve animal performance (Keady et al, 2012), however, effects on CH4 emissions are not conclusive (Ellis et al., 2016a). In Ellis et al. (2016b) no effect of lactic acid bacterial inoculation of grass silage was found on DMI, milk yield or CH4 emissions. The potential to increase silage quality and reduce CH4 produced per unit product exists, but seems dependent on the specific inoculant and strain, silage type, harvesting conditions and presence/absence of spoilage.

Cereal Grains

It is well established that increasing the proportion of starchy grain in the diet will lower CH4 emissions per kg DMI and per kg milk yield, as long as production stays the same or increases (e.g. Tyrrell and Moe, 1972, Yan et al., 2000). Linear decreases in CH4 emissions with an increasing proportion of grain in the diet have been observed in lactating dairy cows as well as beef cattle (Aguerre et al., 2011, McGeough et al., 2010), but small and moderate changes at low levels of inclusion are unlikely to greatly affect CH4 emissions. Sauvant and Giger-Reverdin (2009) concluded that improvements may be expected beyond 35 to 40% inclusion, dependent on level of intake, but that beyond certain levels additional grain may also have a negative effect on fiber digestibility and productivity (e.g. Firkins, 1997).

In addition to the total inclusion level, the type of grain and processing method can have significant effects on rumen starch degradation and hence on CH4 emissions. Beauchemin and McGinn (2005) found that CH4 emissions from beef cattle per kg DMI was not affected by grain source during the backgrounding period, but was less for corn vs. barley during the finishing phase (9.2 vs. 13.1 g/kg DMI, respectively). Differences are likely related to degradation characteristics, total inclusion level and intake differences between the backgrounding and finishing phases. Processing of grain may affect digestibility, feed efficiency, DMI and passage rate of particles and thus CH4 emissions from the animal. Firkins et al. (2001) summarized the effects of corn processing on starch digestibility, and generally saw improvements with steam-flaked corn vs. steam rolled corn grain, resulting in higher milk production. Hales et al. (2012) found 17% lower CH4 emissions (per kg DMI) with steers fed steam-flaked corn vs. dry-rolled corn, accompanied by increases in digestibility. Potential negative effects on NDF digestibility must also be considered at high inclusion levels though (Firkins et al., 1997).

By-products & Crop-residues

Globally, while cereal grains and silages are the cornerstones of a TMR ration, the use of by-products and crop-residues are increasingly found to be economical alternatives. Crop residues can include cereal straws, sorghum stalks, sugarcane tops/bagasse or maize stover, and common by-products may include dried distillers grains (DDG), soy hulls, pulps, wheat bran/middlings and corn gluten meal/feed. Use of by-products from the biofuel industry, such as dry or wet distillers grains, may naturally serve as CH4 mitigating feeds due to their high oil content. McGinn et al. (2009), for e.g. reported 24% less CH4 emissions when DDG replaced barely grain in the backgrounding diet of beef cattle (resulting in a 3% increase in diet EE). In contrast, Hales et al. (2013) found a linear increase in CH4 (/kg DMI) when wet distillers grain with solubles (WDGS) was increased from 0 to 45% of the diet in exchange for steam-flaked corn. This increase was mainly attributed to increases in NDF content of the diet, despite an increase in EE from 5.9 to 8.3%. Beneficial effects can be seen in some experimental studies and not others, depending on the by-product or crop-residue being considered, what it is being compared to (i.e. relative differences in chemical composition), as well as presence/absence of effects on DMI or diet digestibility.

Pasture versus tmr diets

Based on the above discussion, it is evident that a direct comparison of pasture and TMR systems in terms of CH4 emission intensity is complicated. Several specific comparisons are available in the literature, but specific comparisons cannot be generalized to the system level due to significant variation within TMR and pasture systems, and among forages, silages and by-products.

Fredeen et al. (2013) compared CH4 emissions from dairy cows on a TMR diet (grass/alfalfa silage, concentrate and hay) to those on pasture harvested from a management intensive grazing (MIG) regime (which includes supplemental grain), in both spring and fall. In spring, pasture fiber quality was higher than that of the silage used in the TMR, resulting in less grain feeding with MIG compared to the TMR (0.24 vs. 0.32 kg DM/kg milk). In the fall, this difference in quality (and thus grain intake) was not as pronounced. Methane emissions (g/kg DMI and g/kg FCM), DMI (kg/d) and FCM (kg/d) were not affected by diet in the spring. In the fall, fat corrected milk yield was significantly lower for the MIG treatment compared to the TMR treatment, but did not result in any significant differences in CH4 (g/kg DMI and g/kg FCM) or DMI.

O’Neill et al. (2011) compared perennial ryegrass to TMR diets offered in early lactation in spring, and found that CH4 emissions relative to DMI and to milk fat and protein yield were significantly higher for the TMR diet (20.28 vs. 18.06 g/kg DMI and 200 vs. 174 g/kg FPCM, respectively). The TMR was based on corn silage, a concentrate blend and barley straw, and had a lower OM digestibility compared to the spring ryegrass (768 vs. 830 g/kg DM, respectively). Cows on the TMR diet had a significantly higher DMI and milk yield (though DMI was measured with less accuracy for grazing animals).

Dall-Orsoletta et al. (2016) also found that CH4 (g/kg DMI and g/kg milk yield) was significantly higher on a TMR diet compared to cows that received a partial TMR with ryegrass strip-grazing during two time periods throughout the day. Total DMI and milk yield did not differ between the treatment groups.

Bannink et al. (2010) compared grass silage directly to grass herbage (fresh cut grass) in a series of model simulations, and predicted lower or equal CH4 emissions for grass herbage compared to early cut grass silage with units g/kg FCM, but higher emissions for grass herbage when CH4 was expressed as g/kg DM or as a % of GEI. This was mainly due to improved digestibility of the grass herbage relative to the silage.

From these studies, it is evident that pasture and silage quality (chemical composition, digestibility) as well as DMI largely dictate the comparison, and that results may differ based on the season (spring vs. fall, first vs. second/third cut silage/hay). While a TMR diet may contain more cereal grain, compared to a high quality pasture diet, it can have a lower forage fiber or OM digestibility. The realized DMI level can also differ, influencing these comparisons and magnitude of response.

Whole farm basis

Beyond CH4 emissions, comparison of farming systems must include consideration of other GHG sources and environmental pollutants such as nitrogen (N) and phosphorous (P), not only at the cow level but at the whole farm level. If MIG and TMR diets result in similar CH4 emissions (g/kg milk), one must consider the carbon footprint and environmental ‘cost’ of grain growing, processing and transport of this input. This would make the environmental cost of the TMR higher. In turn, however, N emissions from pasture may be significantly higher than from confined feeding systems, depending on the level of pasture N fertilization, grazing management, and the CP content of TMR feeds (e.g. Burke et al., 2008). Finally, pollution ‘swapping’ may occur, for e.g., between CH4 and N excretion (Dijkstra et al., 2011). Therefore, using only CH4 (g/kg milk) as a metric to guide the dairy industry towards lower total GHG emissions ignores GHGs created by other inputs and outputs to the system. A farm level comparison is largely beyond the scope of the current paper, mainly due to a lack of available appropriate literature at the farm system level. While many farm level comparisons exist, they fail to capture complexities of variation at the animal level by using standard emission estimates to describe emissions from the cow (e.g. IPCC, 2006 Tier 1 and 2). Whole farm models which incorporate more detailed models at the animal level are needed to determine the relative GHG impact of a pasture vs. TMR feeding system.

Generation of human-edible protein

In addition to variation in composition and quality of pasture and TMR feeds, as well as other GHG emissions from the farm system level, additional factors must still be considered. With the prediction that we need to feed 9.5 billion people in 2050, the ability of a farming system to generate human-edible protein from human-non-edible protein or nitrogen sources is increasingly relevant (Wilkinson, 2011). Although reducing the forage proportion of the diet is generally proposed as a CH4 mitigation strategy, many livestock feeds (cereal grains) could be eaten directly by humans, which leads to debate about the competition between land use for livestock feed and human food production, and questions the appropriateness of suggesting a lower FP and higher grain diet.

Conclusions

Comparison between pasture and TMR feeding systems for cattle reveals similar CH4 emission (g/kg DMI or g/kg milk) at times while differences are evident at other times. Variability is largely due to extreme variation in ingredients, chemical composition, diet digestibility and DMI. While insightful, direct comparison of CH4 emissions from pasture and TMR systems will not yield full information on the environmental impact of a change in the farming system, as it excludes consideration of N and P at the whole farm level, effects of land use change, as well as differences in the potential to generate human-edible protein. Such a comparison can only be performed with advanced whole-farm simulation models which also consider the intricacies of animal-level responses to nutritional changes, and which are demonstrably able to take into account the trade-offs between C- and N-related GHG emissions, on- and off-farm. Across the feeding system, improving forage quality may be the most promising CH4 mitigation approach which may also benefit other GHG emissions and generation of human-edible protein.