An assessment of global ruminant methane-emission measurements shows bias relative to contributions of farmed species, populations and among continents

Keywords: agriculture, diets, farmed ruminant species, global distribution of studies, methane measurement techniques, methane mitigation, ruminant methane emissions, trends.

Enteric methane (CH₄) from domestic ruminants is a major source of anthropogenic greenhouse gas (GHG; Myhre et al. 2013). Methane emissions from livestock and mitigation strategies are reported biennially in a national GHG inventory report from countries associated with the ‘Paris Agreement’ (United Nations of Climate Change (UNCC) 2015). National CH₄ estimates can be calculated using the Intergovernmental Panel on Climate Change (IPCC) Guidelines (Intergovernmental Panel on Climate Change (IPCC) 2006, 2019), and the level of uncertainty is dependent on the tier level (1–3) used for calculations. The lower and upper uncertainty (95% confidence interval) for global enteric-CH₄ emissions is −11% to 18% (United States Environmental Protection Agency (US EPA) 2017) and this is a consequence of limited local information about livestock numbers, dry-matter intake (DMI), diet chemical composition and CH₄ emissions per unit of DMI (i.e. CH₄ yield; Hristov et al. 2018).

Globally, ruminant livestock numbers and current estimates of ruminant enteric-CH₄ emissions (Statistical Database Food and Agriculture Organization of the United Nations (FAOSTAT) 2020) show that emissions are highest from Asia (37%), followed by South America (SA, 23%), Africa (17%), Europe (10%), North America (NA, 9%) and Oceania (3%). Non-dairy and dairy cattle account for 56% and 19% of estimated enteric emissions respectively, followed by buffalo (12%) sheep (7%), goats (5%) and camelids (1%). Statistical Database Food and Agriculture Organization of the United Nations (FAOSTAT) (2020) data were available only for South American camelids (llama and alpaca) but not for other minor categories such as farmed deer and bison.

Refining and implementing methodology to estimate livestock CH₄ emissions is ongoing, and the aim is to reduce uncertainty of national CH₄ estimates by validating estimates of animal DMI and CH₄ yield (summarised by Intergovernmental Panel on Climate Change (IPCC) 2006, 2019) with measured values. Ruminant CH₄ yield depends on several interacting feed and animal factors, which should be captured by in vivo CH₄-emission measurements. Although the United Nations Framework Convention on Climate Change (UNFCCC) (1997) stated that ‘comparable methodologies’ should be used for compiling the GHG inventory to make national results comparable in a consistent manner, several methods have been developed to determine CH₄ emissions from individual animals. Each method affects the variability of the enteric-CH₄ estimate (Hammond et al. 2016a; Jonker et al. 2020; Della Rosa et al. 2021), and all need to be validated against the ‘gold standard’, respiration chambers (RC). Methodology may contribute variability to CH₄ emission measurements and national inventories.

Global efforts to reduce the rise in anthropogenic GHG emissions require robust data to establish current emissions, develop projections of future emissions as well as identify and evaluate emission-mitigation strategies. The global statistics from Statistical Database Food and Agriculture Organization of the United Nations (FAOSTAT) (2020) give a macro view of ruminant populations and its contribution to CH₄ from ruminant enteric fermentation. These statistics and expert advice guide global and regionalised policies to identify new research areas appropriate for local conditions.

Every published study has its own objective that adds information to the scientific community, but a macro analysis of where, how and what studies were conducted will enable comparisons among ruminant populations and production systems. The objective of this literature survey was to summarise published information on CH₄ measurements performed on farmed ruminant species, categories of feeds, methodology and research aims on six continents, in relation to ruminant populations and the contributions of each species to estimated global CH₄ emissions for each continent.

A literature search was conducted using Google Scholar, Scopus, and Ovid to retrieve relevant scientific peer-reviewed papers. The keywords used in the literature search were livestock, cattle, beef, dairy, goat, sheep, deer, buffalo, llama, alpaca, methane, chambers, sulfur hexafluoride (SF₆), GreenFeed (GF), laser CH₄ detector, sniffer and words related to continents, such Africa, Asia, etc. Only peer-reviewed papers written in English were considered, not technical notes, conference abstracts or technical reports. Articles published between 1994 and 2018 that reported CH₄ emissions measured using RC, SF₆, GF and ‘Other’ techniques (face mask (FM), portable accumulation chambers (PAC); hand-held laser CH₄ detector (LMD) and ‘sniffer’ (SNF)) were considered (n = 415 treatment means; listed in Supplementary material Table S1). Articles that were based on simulated CH₄-emission data were not considered.

Information extracted from each publication included year of publication (1994–2018), country where the trial was conducted, animal species/type, the technique used to determine CH₄ (RC, SF₆, GF and Other techniques), as well as the aims of each study and animal diets. Animals were categorised as cattle, buffalo, goat, sheep and other species (deer, alpaca, llama, etc.). Cattle were further categorised as either dairy (growing, lactating and non-lactating) or beef cattle (growing and mature) and either Bos indicus or Bos taurus. If the studies reported a cross between Bos taurus and Bos indicus, they were included as Bos indicus.

The countries where the studies were conducted were grouped into continents as follows: Asia, Africa, Europe, North America (including Mexico; NA), South America (including Caribbean, except Mexico, SA) and Oceania. The aims of the research studies were classified into the following four groups: diet evaluation (Diet); technique/s (Tech), involving comparisons among, or modifications within a technique to measure CH₄; animal efficiency (Efficiency) for animals with divergent feed efficiency (residual feed intake), CH₄ emissions or breeds and ‘Other objectives’, such as vaccines, defaunation, etc.

Information on composition of diets fed during CH₄ measurements was used to identify the following three subgroups: ‘High-forage diets’ when fresh or conserved forage, i.e. hay, haylage and silage, represented 65% or more of the diet; ‘High-concentrate diets’, when energy and/or protein concentrates represented 65% or more of the total diet; and ‘mixed diets’, when there was no-clear dominance of either forages or concentrates. If a comparison was made between two contrasting diets such as forage versus concentrate, or if diets changed from high forage to high concentrate (or mixed) or vice versa, then the two extreme diet types were recorded within the same study.

Within each diet subgroup, further classification was performed. Diets were grouped as ‘Supplement’ when any type of forage/concentrate (e.g. different plant species, oils, seeds, lipids or different combinations of dietary components) was added or substituted in the control diet. Diets were grouped as ‘Additive’ if bio-active extracts or compounds such as ionophores, urea, tannins, mix of herbs, saponins, nitrate, CH₄ inhibitors, and/or essential oils were tested. If the aim of a study was to test different proportions of ingredients in a diet (e.g. compare an unbalanced diet with a nutrient-balanced diet), or forage qualities/types (e.g. C3 vs C4 species, ryegrass vs clover) or feeding frequencies, they were classified as ‘Nothing added’. If the combination of components classified as ‘Supplement’ or ‘Additive’ were tested in the same study, it was stated as ‘Both’.

Because of the nature of the data and the scope of this survey, statistical comparisons were neither possible nor required. Data were collected and organised in Microsoft Excel. Each Excel row entry (unit of evaluation) contained the following information: authors, publication year, journal, CH₄ measurement technique, continent, livestock species, productive state and subspecies of beef and dairy cattle, study aims and type of diet. If one study included more than one CH₄ measurement technique, livestock species, study aim or diet, the same study was counted as many times as variations showed for that classification criteria. For example, a study using three CH₄ measurement techniques in cattle was counted three times for technique classification and one time for livestock species. Although 415 publications were collected for the current evaluation, there were 450 entries for technique by continent, because 28 publications used two or more techniques to measure CH₄. There were 435 entries for the evaluation of species by continent, 467 entries for the evaluation of techniques by species and 446 entries for study aim combined with continent. Only studies that were classified as ‘Diet’ by their aim were used to analyse diet by continent and animal species, resulting in 348 entries for diet by continent and 360 entries for diet by animal species.

Only data from cattle were further classified by subspecies (Bos taurus, Bos indicus), production type and physiological state, i.e. dairy (growing, lactating, non-lactating) and beef (growing, mature). For cattle, 292 entries were available for the evaluation. Twenty-three publications with cattle did not report subspecies and one did not report the physiological state of the animals and these were not included in the cattle subclassification.

So as to have a reference point and to aid interpretation, data from Statistical Database Food and Agriculture Organization of the United Nations (FAOSTAT) (2020) were included. Classification of countries from Statistical Database Food and Agriculture Organization of the United Nations (FAOSTAT) (2020) was summarised as continents. ‘Africa’ included near east as well as northern and sub-Saharan Africa; ‘Asia’ included eastern Asia, Russian Federation and southern Asia regions; western Europe, eastern Europe were grouped as ‘Europe’; North America, plus Mexico as NA; Latin America and the Caribbean (Central America), excluding Mexico as ‘SA’ and Oceania as ‘Oceania’. Then, the number of animals per species and emissions per species per continent were summarised from Statistical Database Food and Agriculture Organization of the United Nations (FAOSTAT) (2020).

This survey defined disparities between continents and animal species in trials involving CH₄ measurements, as well as the focus of research across regions and the methodology used. This survey did not evaluate CH₄-emission values, and the accuracy of published data was not judged. However, the survey did provide details concerning the scarcity of measurements in some continents, bias regarding some ruminant species, production systems and feed type. Identifying these shortcomings provides an insight for future research measurements, to address the current imbalance of global information, enabling progress towards an integrated CH₄-emission mitigation program. The current study considered only papers published in English language in peer-reviewed journals and therefore excluded information from publications published in local languages and published in reports and conference proceedings.

Between 1994 and 2018, approximately half (51%) of the published studies used RC to measure CH₄ emissions, followed by 36% with SF₆ and 7% with GF. The remaining 6%, mostly from Europe, comprised 2% FM, 2% SNF, 1% PAC (from Oceania), and 1% LMD (9, 9, 4 and 6 studies respectively) and have been categorised as ‘Other tech’ in Table 1. From 2006, there was an increase in annual studies of ruminant CH₄ emissions and numbers of studies using RC and SF₆ were similar from 1994 until 2011, after which the number of studies with RC increased at a greater rate (Fig. 1).

Table 1.  Continental summary of studies published from 1994 to 2018, classified by methane measurement techniques, study aims and ruminant species.

Fig. 1.  Cumulative number of published studies from 1994 to 2018, with seven in vivo methane measurement techniques. RC, respiration chamber; SF₆, sulfur hexafluoride; GF, GreenFeed; FM, face mask; PAC, portable accumulation chamber; LMD, laser methane detector; SNF, sniffer.

Respiration chambers provide accurate and precise measurements of CH₄ emitted by individual animals (Pinares Patiño and Waghorn 2014), but RC facilities are not available in many countries. The need to measure CH₄ emissions under conditions more representative of livestock farming led to the development of methods including SF₆ tracer technique in the early 1990s (Johnson and Johnson 1995) and GF spot-sampling in the past decade. These methods enable CH₄ emissions to be estimated on-farm or while animals are grazing (Della Rosa et al. 2021), but intakes of grazing animals cannot be determined accurately (Undi et al. 2008).

The SF₆ and GF methods have an acceptable accuracy for CH₄ measurement (Grainger et al. 2007; Hammond et al. 2016a; Jonker et al. 2020), and PAC has an acceptable precision for ranking animals, when compared with results from RC (Jonker et al. 2018). Other spot-sampling methods to measure CH₄ emission from individual animals provide less accurate or precise estimates of daily CH₄ production and have been developed with a view to screening and selection of individual animals, these are FM (Hammond et al. 2016b), LMD and SNF (Garnsworthy et al. 2012; Chagunda 2013).

Prior to 1990, most research involving RC for CH₄ determinations was to evaluate energetic metabolism in ruminants (Lassey 2007). The increasing number of studies addressing CH₄ emissions since 1990, and especially from 2006 onward, aligns with international efforts and interests to quantify and reduce GHG emissions from livestock in relation to global warming. These include the international environmental treaty under the United Nations Framework Convention on Climate Change (UNFCCC) which was adopted in 1992 at the Rio de Janeiro Earth Summit, followed by the Kyoto Protocol agreements that came into force in 2005, with a second commitment period after 2012.

Most animal-CH₄ measurement studies were from Europe (32%), followed by Oceania (23%), North America (20%), Asia (15%), South America (9%), and Africa (1%). When considering cattle data only (Europe, 35%; North America 27%, Oceania 18%, South America 10%, Asia 9%, Africa 1%), the distribution of the number of studies by region was similar to the number of studies providing CH₄ data from dairy and beef cattle to the ‘GLOBAL NETWORK’ database (Europe 49%, North America 38%, South America 7%, Oceania 6%, Asia <1%; Niu et al. 2018; van Lingen et al. 2019), although with relatively more studies from Oceania, South America and Asia in the current study. In contrast, the largest total emissions of enteric CH₄ from ruminant livestock are produced in Asia, SA and Africa, which also have the largest ruminant livestock populations (Table 2). The increase in total CH₄ emissions from the ruminant livestock sector from 1890 to 2014, estimated using Tier 2 IPCC guidelines, was mainly due to increased emissions from livestock in Africa, Asia and SA (Dangal et al. 2017). In addition to the numbers of ruminants, the high proportion of cattle in SA, relative to Asia and Africa, accounts for the high emissions relative to total ruminant numbers in SA.

Table 2.  Continental summary of animal numbers (millions), annual enteric-methane emissions (CO₂-equiv; million tons) by species (Statistical Database Food and Agriculture Organization of the United Nations (FAOSTAT) 2020).

With regard to measurement technique (Table 1), RC has been the most used (percentage of total measurements) in Asia (54%), Europe (57%), NA (63%) and also Oceania (48%) but has been used only in 17% of SA studies. The SF₆ tracer technique has been used in all continents (Table 1) and has the advantage over RC in that it can be used with free-ranging and grazing animals (Johnson and Johnson 1995). The GF system has mainly been used in Oceania, Europe and NA and the other spot-sampling techniques have mainly been reported in Europe (Table 1).

The development and increased use of spot-sampling methods reported in the literature are a response to the need to measure larger number of animals for genetic selection or to measure emissions on commercial farms, and most of these methods require less technical expertise by the user (Hammond et al. 2016a; Garnsworthy et al. 2019). However, some of these spot-sampling methods (i.e. FM, LMD and SNF) require further investigation due to high variance and uncertainty regarding accuracy (Oss et al. 2016; Hristov et al. 2018). Differences among continents in the techniques used for determination of CH₄ emissions are indicative of available infrastructure (RC especially), but also ruminant species and production systems. Sharing of capability among countries could enable development of strategic research alliances, enabling research gaps to be filled, especially in Africa, Asia and SA. Such initiatives are in progress with projects through, for example, the Global Research Alliance on Agricultural Greenhouse Gasses (globalresearchalliance.org) and Research Program on Climate Change, Agriculture and Food Security (ccafs.cgiar.org).

The distribution of published studies concerning CH₄ measurement from animal species and across continents is summarised in Table 1. Most studies have been conducted with cattle (64%) and sheep (22%), with only 7% from goats and 5% from buffalo and 2% from other ruminants (alpaca, bison, llama, deer). The large number of studies with cattle is likely to reflect their dominance in numbers, importance in global agriculture and contribution to the total enteric-CH₄ emissions from ruminant livestock (76%, Table 2).

The relatively large number of studies with sheep, compared with other livestock, is likely to reflect their productive value for some continents (e.g. Europe and Oceania) and the ease of measurement, although sheep contribute only 7% of livestock enteric-CH₄ emissions (Table 2). Sheep are often used to explore mitigation options for ruminants in general, possibly because of lower costs than with cattle studies. However, goats are not studied as much as sheep, despite large global numbers (1.06 × 10⁹). The percentage of studies involving goats (7%) was similar to their contribution to ruminant enteric-CH₄ emissions (5%; Table 2). The imbalance between studies and emissions for sheep versus goats is also evident for cattle versus buffalo. Buffalo is responsible for approximately 12% of ruminant enteric-CH₄ emissions, yet only 21 studies have been published (5% of the total). This assessment shows little affiliation between the number of studies on individual species and either animal numbers or their contribution to ruminant enteric-CH₄ emissions.

The species from which CH₄ measurements have been made, within continents, do not relate to their contribution to emissions. This is illustrated in Fig. 2, where both the quantity of emissions and the number of studies have been expressed on a percentage basis for each continent for cattle, sheep and goats. Most apparent is the low number of studies undertaken in Africa over the assessment period. Asia is responsible for 37% of enteric emissions and, although all measurements from buffalo have been undertaken in that continent, studies with sheep, goats and cattle are substantially under-represented in a global context. In SA, cattle are by far the dominant emitters and, although most studies have been undertaken with cattle (Table 1), numbers have been low but are most likely to grow in the near future (Congio et al. 2021). In NA, cattle are also the dominant ruminant species and, although the numbers are less than half of those for SA, the number of studies undertaken are more than twice that for SA (Fig. 2). In contrast, the percentage of studies with goats and cattle in Europe far exceeds their contribution to global emissions for those species. Similarly, for sheep and cattle in Oceania, where the number of studies accounted for 52% and 18% of those undertaken globally, corresponding emissions accounted for 12% and 3% of global emissions.

Fig. 2.  Relationship between ruminant enteric-methane emissions and number of studies in six continents for cattle (♦), sheep (▴) and goats (▪). Continents are as follows: AF, (yellow) Africa; AS, (red) Asia; EU, (green) Europe; NA, (violet) North America; SA, (black) South America; and OC, (orange) Oceania.

Although global populations of sheep and goats exceed those of cattle (Table 2) the percentage of total emissions from cattle are highest across all continents (Africa 74%, Asia 53%, Europe 88%, NA 98%, SA 98% and Oceania 75%) and this would suggest that most evaluations should be from cattle. This was the case in Europe, NA and SA where 73%, 88% and 73% of studies were with cattle, and although in Asia 37% of studies were with cattle, another 30% were with buffalo. In Oceania, 48% of studies were with cattle; so, globally and also regionally, most studies have been undertaken with the highest-emitting species.

Only 13% of global enteric-CH₄ emissions come from sheep, goats and camelids, with 44% of this from Asia, yet studies with sheep and goats from this continent are only 19% of global studies with these species (Table 1). Sheep and goats from Africa account for 34% of global emissions from these species, and only one study has been published. Given differences in breeds and diets available to sheep and goats in Africa and Asia, compared with Europe and Oceania, it is important that more studies with these species are undertaken to achieve defensible estimates of emissions.

Overall, number of studies appear to be related to infrastructure and wealth of continents and there are limited associations between the number of studies and either species CH₄ emissions (Fig. 2) or species numbers. Although it may be argued that CH₄ yields are similar for all ruminants when given similar diets, ‘typical diets’ used in each continent differ among species, as do responses in CH₄ yield to the level of intake. Differences in diet, digestive physiology, animal management and levels of production justify the requirement for species-specific data (van Gastelen et al. 2019), so as to achieve defensible global estimates of CH₄ emissions and increase accuracy to reduce uncertainty. It is also clear that small populations of large ruminants (e.g. buffalo) can contribute with higher amounts of CH₄ than do large populations of small ruminants (i.e. goat and sheep), suggesting that measurements from large emitters (cattle and buffalo) could be a priority. Achieving accurate estimates of CH₄ emissions across all species, diets and continents will guide future global research.

Bos indicus cattle represent more than half the global cattle population and are dominant in tropical and subtropical regions (Utsunomiya et al. 2019), yet only 12% of CH₄ measurement studies with cattle have been undertaken with Bos indicus, compared with Bos taurus (79%). Twenty-four studies (9%) did not specify breed or physiological state, so they were not included in the classification. Most studies with Bos indicus were undertaken in growing beef cattle (87% of studies), while dairy cattle were studied only in 13% of the publications (3% growing, 5% lactating and 5% non-lactating). In contrast, most studies with Bos taurus were with dairy cattle (62% of studies; 4% growing, 50% lactating, 8% non-lactating) and 38% of studies were with beef cattle (35% growing and 3% mature).

Given the genomic and phenotypic differences, such as heat tolerance, a faster rumen fermentation rate, shorter rumen retention time and higher protozoa populations, as well as diet, there may be differences in CH₄ emissions between Bos indicus and Bos taurus (Hegarty 2004). Estimates (Table 2) indicated that non-dairy cattle were the largest contributor (75%) to global enteric-CH₄ emissions, followed by dairy cattle (25%). However, beef cows have been poorly addressed in published studies. Therefore, to improve the accuracy of CH₄ estimates, the database of beef cattle needs to be increased, including data for mature (breeding) cows. Beef cows are an essential category for beef production, but their diet quality is often poor, compared with dairy cows, and this is likely to affect CH₄ yields. Addressing these imbalances will improve the accuracy of global emission data from ruminants, as indicated in the assessment by Cottle and Eckard (2018).

The SF₆ and RC techniques were used to determine CH₄ emissions with most ruminant species. RC was the most commonly used technique with cattle (45%), goats (70%) and sheep (64%), while the SF₆ tracer technique was the most commonly used technique in buffalo (67%; Table 3). The less frequent use of RC in large ruminants than in small ruminants is probably associated with experiment costs, limited number of RC units for large animals per laboratory and greater development and use of alternative methods in cattle. Although measurements from grazing cattle may be better suited to the SF₆ tracer technique or GF spot-sampling, there remains the challenge of determining feed intake. The increasing use of RC to measure CH₄ emissions for all species maybe explained by their accuracy and precision in measurements of both CH₄ and feed intake (Jonker et al. 2020).

Table 3.   Number of studies from 1994 to 2018, classified by ruminant methane-measurement techniques and ruminant species evaluated using each technique.

Within the spot-sampling methods, GF, LMD and FM were used with large and small ruminants, while SNF was used only with cattle and PAC was used only with sheep. The SNF technique was developed for dairy cattle (Garnsworthy et al. 2012), whereas the PAC technique was developed for sheep (Goopy et al. 2011). The SNF and LMD methods record only CH₄ concentrations, which limits their wider application to generate accurate CH₄-emission estimates. Although some spot-sampling methods are still under evaluation to assess their capability to provide accurate emission-factor estimates (FM and PAC), the accuracy of GF has been demonstrated for cattle (Jonker et al. 2020) and may enable more measurements from less studied species, such as, for example, camelids, and buffaloes, so as to provide reliable data in continents where these species are common.

In terms of study objectives, the majority of CH₄ measurement studies evaluated diet as the main study aim (Africa 100%, Asia 92%, Europe 71%, NA 84%, SA 85%, Oceania 53%) whereas animal efficiency was the second-most studied strategy (Asia 5%, Europe 17%, NA 7%, SA 12%, Oceania 24%; Table 1). Oceania and Europe also reported a higher number of studies that evaluated CH₄ measurement techniques, than did other continents. The focus of studying diets as a mitigation strategy may have been a response to obtaining emission factors for inventories, especially as dietary strategies can be a method ready to be applied to reduce CH₄ emissions (Martin et al. 2010) through the use of supplements and additives. In the past decade, studies have also attempted to identify animals that vary in efficiency for production and also low CH₄ emitters, both of which could provide options to reduce CH₄ emissions, provided these characteristics are heritable (Thompson and Rowntree 2020), as shown in sheep and cattle (Lassen and Løvendahl 2016; Hayes et al. 2016; Jonker et al. 2018).

Diets used for animal production are dependent on feed sources available, animal species and affordability. Globally, CH₄ measurement studies reporting diet composition comprised 197 high-forage (fibrous), 35 high-concentrate (grain) and 116 mixed diets (Table 4). The high-forage diets were the most common in evaluations undertaken in Africa (100%), Europe (55%), NA (44%), SA (72%) and Oceania (86%), while mixed diets were the most prevalent in Asia (56%), and although few in number, most high-concentrate diets were evaluated in Europe and NA. Forage is less expensive than grain and it is the main dietary component (86%) of ruminant diets (Mottet et al. 2017), especially in Africa, SA and Oceania (Seré and Steinfeld 1995). Forages are important in less intensive farming systems in all continents; however, it is expected that intensification of production systems continues and changes in diets offered to ruminants will be most evident in continents with extensive systems (Thornton et al. 2009; McAllister et al. 2020).

Table 4.  Continental distribution of studies published from 1994 to 2018 of ruminant feeding, expressed in terms of the basal diets with supplements and additives fed during methane measurements.

Most (72%) diets fed during studies of CH₄ emissions included supplements that provided energy or protein, but which differed from the main components and/or additives, such as ionophores, nitrate, CH₄ inhibitors, etc. (Table 4). Their use was widespread in CH₄ studies and only 73 of 197 studies with high-forage diets did not include supplements or additives (these were 100% forage). Half or more of studies with sheep, goats and cattle fed high-forage diets included addition of supplements or other additives (Table 5) as did most studies with high-concentrate diets (most of these were with cattle) and mixed diets. The supplements provide nutrients to meet animal requirements and the choice depends on the main dietary component, level of production and are species-specific, whereas many additives may have been included in an effort to reduce methanogenesis.

Table 5.  Number of studies published from 1994 to 2018 classified by animal species and diets fed during methane measurements, expressed in terms of the basal diets with supplements and additives.

The analysis of the diet type per species shows that a wide range of feedstuffs has been evaluated, perhaps reflecting the diversity in farming systems, or in anticipation of changes to future feeding. The ‘high-forage’ diets were dominant in studies with sheep (73%) and cattle (56%), while the ‘mixed’ diet was most prevalent with buffalo (78%) and both ‘high-forage’ and ‘mixed’ diets dominated goat studies (Table 5). Although ‘high-concentrate’ diets were the least studied diet types in all species, their use was most prevalent with cattle. Mixed diets dominated buffalo studies because they are fed agricultural crop-residues and industrial by-products to ensure low-cost production (Deb et al. 2016). In summary, a wide spectrum of diets has been evaluated (Table 5), providing a broad base of emission data and supplements and additives that were in common use.

Overall, the number of studies that have measured CH₄ from cattle align with their contribution to enteric emissions. However, buffalo, Bos indicus cattle and mature beef cows were under-represented relative to their global populations and contribution to global emissions. The second-largest contributor to global CH₄ emissions are buffaloes, which globally produce a similar quantity of CH₄ as do sheep and goats combined, yet only 14% of studies from these three species were from buffaloes.

Most published research has focused on evaluating the effect of diets on CH₄ emissions. High-forage diets were the most evaluated across species. Current trends include studies of individual variation within species and suggest a change in focus, along with increased implementation of spot-sampling methods for estimation of CH₄ emissions.

Finally, it is well known that there have been fewer studies in continents with lower per capita incomes, but, in general, their research follows global trends regarding measurements from animal species. The use of spot-sampling methods to measure CH₄ emissions may accelerate the research in geographical areas of data scarcity. Additionally, identifying the typical diets used to feed the predominant CH₄-emitting species within each macro region will contribute to more accurate CH₄ estimates on a global and regional scales.

Supplementary material is available online.

The data that support this study are available in the article and the list of studies retrieved are included as supplementary material.

The authors declare that they do not have conflicts of interest.

Maria M. Della Rosa was financially supported by the New Zealand Government through the Global Research Alliance Livestock Emissions and Abatement Research Network (LEARN) awards program (https://livestockemissions.net/). The information contained within this paper should not be taken to represent the views of the Alliance as a whole or its partners.