Dietary supplementation of dairy cow diets with lipid or starch is a known strategy to mitigate enteric methane (CH4) emission. However, the consequences of these nutritional strategies on N and OM excretion in feces and urine need to be jointly evaluated. This present meta-analysis assessed the impact of dietary lipid supplementation (LS) and increased dietary starch content (IS) on production or excretion (g/d), yield (g/kg DMI), and intensity (g/kg ECM) of CH4; fecal OM (FOM); and fecal, urinary, and total N (FN, UN, and TN) excretions. Records of treatment means were obtained from 172 experiments (678 treatment means) conducted with dairy cows. Univariate and multivariate statistical analyses evaluated the effect of LS or IS (analyzed separately) to quantify the effect of dietary ether extract (EE) or starch (STA) contents on CH4 emissions and FOM N excretions, with or without using additional explanatory parameters obtained from diet composition and animal characteristics. In simple univariate models for LS experiments, an increase in EE content of 1.0 g/kg DM reduced CH4 production, yield, and intensity by 2.03 g/d, 0.077 g/kg DMI, and 0.082 g/kg ECM, respectively; EE content alone did not influence FOM or N yield or intensity, except for a decrease in FN excretion (g/d). With IS experiments, simple univariate models revealed that increasing STA content by 1.0 g/kg DM reduced CH4 yield by 0.0085 g/kg DMI and CH4 intensity by 0.013 g/kg ECM. An increase in STA content increased FOM excretion g/d and decreased TN and FN yield and FN intensity. In complex univariate models, DMI and CP content (g/kg DM) exhibited positive relationships with N excretion (g/d) across both strategies. In the multivariate models, neither EE nor STA content showed trade-offs or synergies in intensity (g/kg ECM). However, an increase in EE decreased CH4 and TN production (g/d) in trivariate models, whereas an increase in either EE or STA content decreased CH4 yield but increased FOM yield (g/kg DMI) in bivariate models for LS diets. A decrease in DMI decreased both CH4 and TN production, as well as decreased both CH4 and FOM production in the bivariate model for both strategies. Similarly, decreased feeding level and NDF content reduced CH4 and FN production (g/d) in the trivariate model for LS diets. Decreased DMI decreased CH4, FN, and UN production, as well as CH4, TN, and FOM production (g/d) in starch-based diets. In trivariate models, for LS experiments, lower NDF content reduced CH4 and FOM yields (g/kg DM), whereas in IS experiments increased percentage of concentrate reduced CH4 and increased FN yields, and increased CP content led to increased TN and reduced FOM yields. In conclusion, reducing DMI and balancing key nutrients such as EE, starch, NDF, and CP can lower CH4, FOM, and N excretion, but trade-offs occur. In lipid diets, higher EE or STA contents reduce CH4 yield while increasing FOM yield, and in starch diets, higher CP content or concentrate proportion reduces FOM or CH4 yield while increasing N yield, highlighting the need for diet- and metric-specific optimization.

La réduction des émissions de méthane entérique associées aux systèmes d’élevage de ruminants est l’un des objectifs de lutte contre le changement climatique pour de nombreux pays. Si les mesures directes des émissions de méthane entérique sont limitées dans leur déploiement à un petit nombre d’animaux, des méthodes indirectes de prédiction via le lait ou les fèces peuvent être appliquées à un grand nombre d’animaux dans des contextes variés. Ces approches permettent de développer de nouveaux modèles de sélection génétique ou de conduites d’élevage des petits et gros ruminants associées à des niveaux d’émissions de méthane entériques plus faibles. Les recherches pour comprendre les relations entre la diversité microbienne du rumen et les phénotypes des ruminants contribuent à l’identification de nouveaux leviers pour moduler les populations microbiennes et les flux d’hydrogène dans le rumen, avec l’ambition de réduire les émissions de méthane tout en préservant la production et la santé de l’animal. L’alimentation et la conduite des troupeaux constituent aussi des leviers d’intérêt, mais leurs effets sur la réduction des émissions de méthane entérique sont variables selon les contextes pédoclimatiques, notamment en régions chaudes. Des stratégies combinant différents leviers doivent être identifiées pour réduire efficacement les émissions de méthane sans compromettre la santé de l’animal et les services écosystémiques rendus par leurs systèmes d’élevage.

Reducing enteric methane emissions from ruminant livestock is a common goal of many countries to limit global warming. While direct measurements of enteric methane emissions are limited to a small number of animals, indirect predictive methods based on milk or feces can now be used to assess emissions from large numbers of animals in a variety of contexts. These approaches allow the development of genetic selection models and management practices associated with lower methane emissions for small and large ruminants. A better understanding of the relationships between rumen microbial diversity and the ruminant (host) will help to identify new solutions for modulating rumen microbial populations and hydrogen fluxes, with the aim of reducing methane emissions while maintaining animal production and health. Ruminant diets and herd management are important ways to reduce enteric methane emissions, but the practices are not always suitable for the different pedoclimatic contexts, particularly in hot regions. Trade-off must be evaluated to identify combinations of levers that can reduce enteric methane emissions without compromising the health of animals and the ecosystem services associated with ruminant livestock systems.

Livestock-forestry (LF) systems enhance the delivery of ecosystem services and sustainability by providing shade, increasing diversity, and improving carbon sequestration. Despite these benefits, more evidence is needed to establish LF systems as a viable alternative for reducing enteric CH4 emissions and improving thermal comfort in beef cattle production. We aimed to evaluate the impact of the forestry component into a forage-based livestock system on animal performance, thermal comfort, and its consequences on enteric CH4 emissions. The experimental design was a randomized complete block with two systems: livestock (L) and LF, each with four replicates. During both seasons, microclimate variables such as relative humidity, photosynthetically active radiation, black globe temperature, and black globe temperature-humidity index were greater in the L system. Plant-part and chemical compositions did not differ between the systems, except for a 10% greater leaf proportion in LF during the rainy season. During the dry season, the LF system showed a 47% greater total gain per ha and 33% greater stocking rate. There was no system effect on CH4 production (g/day). However, in the dry season, LF presented greater CH4 emissions per area. These results indicate that integrating trees into forage-based livestock systems can improve thermal comfort and animal productivity without increasing individual CH4 emission, enhancing long-term productivity and sustainability.

Ruminants plays an important role in global warming by emitting enteric methane (CH4) through the degradation of feeds by the rumen microbiota. To better understand the dynamics fermentation outputs, including methane and volatile fatty acids (VFA) production, mathematical models have been developed. Sensitivity analysis (SA) methods quantify the contribution of model input parameters (IP) to the variation of an output variable of interest. In animal science, SA are usually conducted in static condition. In this work, we hypothesized that including the dynamic aspect of the rumen fermentation to SA can be useful to inform on optimal experimental conditions aimed at quantifying the key mechanisms driving CH4 and VFA production. Accordingly, the objective of this work was to conduct a dynamic SA of a rumen fermentation model under *in vitro* continuous conditions (close to the real *in vivo* conditions). Our model case study integrates the effect of the macroalgae *Asparagopsis taxiformis* (AT) on the fermentation. AT has been identified as a potent CH4 inhibitor via the presence of bromoform, an anti-methanogenic compound. We implemented two SA methods. We computed Shapley effects and full and independent Sobol indices over time for quantifying the contribution of 16 IPs to CH4 (mol/h) and VFA (mol/l) variation. Our approach allows to discriminate the 3 contribution types of an IP to output variable variation (individual, via the interactions and via the dependence/correlation). We studied three diet scenarios accounting for several doses of AT relative to Dry Matter (DM): control (0% DM of AT), low treatment (LT: 0.25% DM of AT) and high treatment (HT: 0.50% DM of AT). Shapley effects revealed that hydrogen (H2) utilizers microbial group via its Monod H2 affinity constant highly contributed (> 50%) to CH4 variation with a constant dynamic over time for control and LT. A shift on the impact of microbial pathways driving CH4 variation was revealed for HT. IPs associated with the kinetic of bromoform utilization and with the factor modeling the direct effect of bromoform on methanogenesis were identified as influential on CH4 variation in the middle of fermentation. Whereas, VFA variation for the 3 diet scenarios was mainly explained by the kinetic of fibers degradation, showing a high constant contribution (> 30%) over time. In addition, the Sobol indices indicated that interactions between IPs played a role on CH4 variation, which was not the case of VFA variation. However, these results are dependent on the way interactions are represented in the model. The simulations computed for the SA were also used to analyze prediction uncertainty. It was related to the dynamic of dry matter intake (DMI, g/h), increasing during the high intake activity periods and decreasing when the intake activity was low. Moreover, CH4 (mol/h) simulations showed a larger variability than VFA simulations, suggesting that the reduction of the uncertainty of IPs describing the activity of the H2 utilizers microbial group is a promising lead to reduce the overall model uncertainty. Our results highlighted the dynamic nature of the influence of metabolic pathways on CH4 productions under an anti-methanogenic treatment. SA tools can be further exploited to design optimal experiments studying rumen fermentation and CH4 mitigation strategies. These optimal experiments would be useful to build robust models that can guide the development of sustainable nutrition strategies.

Over the past decade, there has been considerable attention on mitigating enteric methane (CH4) emissions from ruminants through the utilization of antimethanogenic feed additives (AMFA). Administered in small quantities, these additives demonstrate potential for substantial reductions of methanogenesis. Mathematical models play a crucial role in comprehending and predicting the quantitative impact of AMFA on enteric CH4 emissions across diverse diets and production systems. This study provides a comprehensive overview of methodologies for modeling the impact of AMFA on enteric CH4 emissions in ruminants, culminating in a set of recommendations for modeling approaches to quantify the impact of AMFA on CH4 emissions. Key considerations encompass the type of models employed (i.e., empirical models including meta-analyses, machine learning models, and mechanistic models), the modeling objectives, data availability, modeling synergies and trade-offs associated with using AMFA, and model applications for enhanced understanding, prediction, and integration into higher levels of aggregation. Based on an evaluation of these critical aspects, a set of recommendations is presented concerning modeling approaches for quantifying the impact of AMFA on CH4 emissions and in support of farm-level, national, regional, and global inventories for accounting greenhouse gas emissions in ruminant production systems.

Ruminants play an important role in global food security and nutrition. The rumen microbial community provides ruminants with a unique ability to convert human indigestible plant matter, into high quality edible protein. However, enteric CH4 produced in the rumen is both a potent GHG and a metabolizable energy loss for ruminants. As the rumen microbiome constitutes 15–40% of the inter-animal variation in enteric CH4 emissions, understanding the microbiological mechanisms underpinning ruminal methanogenesis and its interaction with the host animal, is crucial for developing CH4 mitigation strategies. Variation in the relative abundance of different microbial species has been observed in cattle with contrasting residual CH4 emission and CH4 yield with up to 20% of the variation in inter-animal CH4 emissions attributable to the presence of a small number of microbial species. The demonstration of ruminotypes associated with high or low CH4 emissions suggests that interactions within complex microbial consortia and with their host are a major source of variation in CH4 emissions. Consequently, microbiome-assisted genomic approaches are being developed to select low CH4 emitting cattle, with breeding values for enteric CH4 being included as part of national breeding programmes. Generating rumen microbiome data for use in selection programs is expensive, therefore, identifying microbial biomarkers in milk or plasma to develop predictive models which include microbial predictors in equations based on animal related data, is required. A better understanding of the rumen microbiome has also aided the development and refinements of anti-methanogenic feed additives. However, these strategies, which increase the amount of reducing equivalents in the rumen ecosystem, do not generally result in an enrichment of propionate or an improvement in animal performance. Current research aims to provide alternative sinks to reducing equivalents and to stimulate activity of commensal microbes or the supplementation of direct fed microbials to capture lost energy. Furthering our knowledge of the rumen microbiome and its interaction with the host, will aid in the development of CH4 mitigation strategies for ruminant livestock.

Automated measurements of the ratio of concentrations of methane and carbon dioxide, [CH4]:[CO2], in breath from individual animals (the so-called “Sniffer-technique”) and estimated CO2 production can be used to estimate CH4 production, provided that CO2 production can be reliably calculated. This would allow CH4 production from individual cows to be estimated in large cohorts of cows, whereby ranking of cows according to their CH4 production might become possible and their values could be used for breeding of low CH4 emitting animals. Estimates of CO2 production are typically based on predictions of heat production, which can be calculated from body weight (BW), energy-corrected milk yield, and days of pregnancy. The objectives of the present study were to develop predictions of CO2 production directly from milk production, dietary, and animal variables, and furthermore develop different models to be used for different scenarios, depending on available data. An international data set with 2,244 records from individual lactating cows including CO2 production and associated traits, as dry matter intake (DMI), diet composition, BW, milk production and composition, days in milk and days pregnant, was compiled to constitute the training data set. Research location and experiment nested within research location were included as random intercepts. The method of CO2 production measurement (respiration chamber (RC) or GreenFeed (GF)) was confounded with research location, and therefore excluded from the model. In total, 3 models were developed based on the current training data set: Model 1 (“Best Model”), where all significant traits were included, Model 2 (“On-Farm Model”), where DMI was excluded, and Model 3 (“Reduced On-Farm Model”), where both DMI and BW were excluded. Evaluation on test data sets either with RC data (n = 103), GF data without additives (n = 478) or GF data only including observations where nitrate, 3-nitrooxypropanol (3-NOP), or a combination of nitrate and 3-NOP were fed to the cows (GF+: n = 295), showed good precision of the 3 models, illustrated by low slope bias both in absolute values (−0.22 to 0.097) and in percentage (0.049 to 4.89) of mean square error (MSE). However, the mean bias (MB) indicated systematic over-prediction and under-prediction of CO2 production when the models were evaluated on the GF and the RC test data set, respectively. To address this bias, the 3 models were evaluated on a modified test data set, where the CO2 production (g/d) was adjusted by subtracting (where measurements were obtained by RC) or adding absolute MB (where measurements were obtained by GF) from evaluation of the specific model on RC, GF, and GF+ test data sets. By this modification, the absolute values of MB and MB as percentage of MSE became negligible. In conclusion, the 3 models were precise in predicting CO2 production from lactating dairy cows.

Ensuring the sustainability and circularity of mixed crop-ruminant livestock systems is essential if they are to deliver on the enhancement of long-term productivity and profitability with a smaller footprint. The objectives of this study were to select indicators in the environmental, economic and social dimensions of sustainability of crop-livestock systems, to assess if these indicators are relevant in the operational schedule of farmers, and to score the indicators in these farm systems. The scoring system was based on relevance to farmers, data availability, frequency of use, and policy. The study was successful in the assemblage of a suite of indicators comprising three dimensions of sustainability and the development of criteria to assess the usefulness of these indicators in crop-ruminant livestock systems in distinct agro-climatic regions across the globe. Except for ammonia emissions, indicators within the Emissions to air theme obtained high scores, as expected from mixed crop-ruminant systems in countries transitioning towards low emission production systems. Despite the inherent association between nutrient losses and water quality, the sum of scores was numerically greater for the former, attributed to a mix of economic and policy incentives. The sum of indicator scores within the Profitability theme (farm net income, expenditure and revenue) received the highest scores in the economic dimension. The Workforce theme (diversity, education, succession) stood out within the social dimension, reflecting the need for an engaged labor force that requires knowledge and skills in both crop and livestock husbandry. The development of surveys with farmers/stakeholders to assess the relevance of farm-scale indicators and tools is important to support direct actions and policies in support of sustainable mixed crop-ruminant livestock farm systems.

This work aimed to investigate the effect of energy nature (lipids vs. carbohydrates) on enteric methane emission (eCH4) and performance in dairy cows fed grass silage-based diets. Eight multiparous Holstein cows were used in a 4 × 4 Latin square design, with 4 experimental periods of 28 days each. Cows were fed with 4 iso-energetic diets based on grass silage and supplemented with different levels of rapeseed oil (RO) (0, 1.5, 3.0, 4.5% of dietary dry matter (DM), Control, RO-low, RO-medium, RO-high diets, respectively) in substitution of starch from concentrate. Dairy performance, total-tract digestibility, and rumen parameters were measured when animals were in individual respiration chambers for quantifying eCH4. Intake of DM decreased with RO-high compared to other diets. Methane emissions were lowered similarly with all diets containing RO compared to Control (on average, -19% in g/d, −13% in g/kg DMI, −21% in g/kg milk). Ruminal propionate proportion was higher, whereas that of butyrate was lower with RO-high compared to the other diets. Milk yield was higher with RO-low and RO-medium, and was lower for RO-high compared to Control. Milk C16:0 concentration was lower, and C18:1c9, C18:1t11, and other rumen biohydrogenation intermediate concentrations were higher with diets containing RO compared to Control. The shift from the C18:1t11 to C18:1t10 biohydrogenation pathway is in agreement with the milk fat depression observed with RO-medium and RO-high. The inclusion of RO at 1.5% was the best compromise between eCH4, feed efficiency, and milk nutritional quality in cows fed grass silage-based diets.

The production of enteric methane in the gastrointestinal tract of livestock is considered as an energy loss in the equations for estimating energy metabolism in feeding systems. Therefore, the spared energy resulting from specific inhibition of methane emissions should be re-equilibrated with other factors of the equation. And, it is commonly assumed that net energy from feeds increases, thus benefitting production functions, particularly in ruminants due to the important production of methane in the rumen. Notwithstanding, we confirm in this work that inhibition of emissions in ruminants does not transpose into consistent improvements in production. Theoretical calculations of energy flows using experimental data show that the expected improvement in net energy for production is small and difficult to detect under the prevailing, moderate inhibition of methane production (≈25%) obtained using feed additives inhibiting methanogenesis. Importantly, the calculation of energy partitioning using canonical models might not be adequate when methanogenesis is inhibited. There is a lack of information on various parameters that play a role in energy partitioning and that may be affected under provoked abatement of methane. The formula used to calculate heat production based on respiratory exchanges should be validated when methanogenesis is inhibited. Also, a better understanding is needed of the effects of inhibition on fermentation products, fermentation heat, and microbial biomass. Inhibition induces the accumulation of H2, the main substrate used to produce methane, that has no energetic value for the host, and it is not extensively used by the majority of rumen microbes. Currently, the fate of this excess of H2 and its consequences on the microbiota and the host are not well known. All this additional information will provide a better account of energy transactions in ruminants when enteric methanogenesis is inhibited. Based on the available information, it is concluded that the claim that enteric methane inhibition will translate into more feed-efficient animals is not warranted.

Enteric methane (CH4) emissions from sheep contribute to global greenhouse gas emissions from livestock. However, as already available for dairy and beef cattle, empirical models are needed to predict CH4 emissions from sheep for accounting purposes. The objectives of this study were to: 1) collate an intercontinental database of enteric CH4 emissions from individual sheep; 2) identify the key variables for predicting enteric sheep CH4 absolute production (g/d per animal) and yield [g/kg dry matter intake (DMI)] and their respective relationships; and 3) develop and cross-validate global equations as well as the potential need for age-, diet-, or climatic region-specific equations. The refined intercontinental database included 2,135 individual animal data from 13 countries. Linear CH4 prediction models were developed by incrementally adding variables. A universal CH4 production equation using only DMI led to a root mean square prediction error (RMSPE, % of observed mean) of 25.4% and an RMSPE-standard deviation ratio (RSR) of 0.69. Universal equations that, in addition to DMI, also included body weight (DMI + BW), and organic matter digestibility (DMI + OMD + BW) improved the prediction performance further (RSR, 0.62 and 0.60), whereas diet composition variables had negligible effects. These universal equations had lower prediction error than the extant IPCC 2019 equations. Developing age-specific models for adult sheep (>1-year-old) including DMI alone (RSR = 0.66) or in combination with rumen propionate molar proportion (for research of more refined purposes) substantially improved prediction performance (RSR = 0.57) on a smaller dataset. On the contrary, for young sheep (<1-year-old), the universal models could be applied, instead of age-specific models, if DMI and BW were included. Universal models showed similar prediction performances to the diet- and region-specific models. However, optimal prediction equations led to different regression coefficients (i.e. intercepts and slopes) for universal, age-specific, diet-specific, and region-specific models with predictive implications. Equations for CH4 yield led to low prediction performances, with DMI being negatively and BW and OMD positively correlated with CH4 yield. In conclusion, predicting sheep CH4 production requires information on DMI and prediction accuracy will improve national and global inventories if separate equations for young and adult sheep are used with the additional variables BW, OMD and rumen propionate proportion. Appropriate universal equations can be used to predict CH4 production from sheep across different diets and climatic conditions.

Intensification of livestock systems becomes essential to meet the food demand of the growing world population, but it is important to consider the environmental impact of these systems. To assess the potential of forage-based livestock systems to offset greenhouse gas (GHG) emissions, the net carbon (C) balance of four systems in the Brazilian Amazon Biome was estimated: livestock (L) with a monoculture of Marandu palisade grass [Brachiaria brizantha (Hochst. ex A. Rich.) R. D. Webster]; livestock-forestry (LF) with palisade grass intercropped with three rows of eucalyptus at 128 trees/ha; crop-livestock (CL) with soybeans and then corn + palisade grass, rotated with livestock every two years; and crop-livestock-forestry (CLF) with CL + one row of eucalyptus at 72 trees/ha. Over the four years studied, the systems with crops (CL and CLF) produced more human-edible protein than those without them (L and LF) (3010 vs. 755 kg/ha). Methane contributed the most to total GHG emissions: a mean of 85 % for L and LF and 67 % for CL and CLF. Consequently, L and LF had greater total GHG emissions (mean of 30 Mg CO2eq/ha/year). Over the four years, the system with the most negative net C balance (i.e., C storage) was LF when expressed per ha (-53.3 Mg CO2eq/ha), CLF when expressed per kg of carcass (-26 kg CO2eq/kg carcass), and LF when expressed per kg of human-edible protein (-72 kg CO2eq/kg human-edible protein). Even the L system can store C if well managed, leading to benefits such as increased meat as well as improved soil quality. Moreover, including crops and forestry in these livestock systems enhances these benefits, emphasizing the potential of integrated systems to offset GHG emissions.

Stoichiometric models that predict methane (CH4) production of ruminants assume that the amount of hydrogen gas (H2) produced during fermentation in the rumen equals the amount of H2 consumed by electron sinks in the rumen and thus that H2 should not be emitted into the environment. However, some studies have demonstrated that H2 emissions occur under practical conditions. In addition, the H2 emissions increase when nutritional mitigation strategies are used. Hence, this study hypothesized that considering H2 emission would improve the prediction of enteric CH4 emission, especially when using mitigation strategies that induce high variation in the H2 emission. The objective of this study was to develop and evaluate the performance of CH4 emission prediction models using H2 emission as an explanatory variable. A database of CH4 and H2 emission (mean treatment data) from ruminants (dairy cattle, growing cattle and sheeps) was built to develop mixed-effects models at three levels of complexity: level 1 -H2 production/yield, DMI, or both; level 2 -level 1 explanatory variables plus chemical composition of the diet (i.e. CP, EE, NDF, OM, or PCO); and level 3 -level 2 explanatory variables plus animal metabolic weight (BW0.75, kg). When all animal categories were grouped, including H2 production improved the performance of CH4 production prediction only in level 1 models for the electron-receptor mitigation strategy, reducing the root mean square of prediction error (RMSPE) by 18 %. For dairy cattle and sheep, dry matter intake was not a significant explanatory variable in level 1 or 2 models. For growing beef cattle, for all mitigation strategies together and the inhibitor-mitigation strategy alone, including H2 production reduced RMSPE by 13 % and 27 %, respectively. Overall, H2 production was included in 60 % of level 1 models and 100 % of level 2 and 3 models, which best predicted CH4 production. Thus, including H2 emission does improve prediction of enteric CH4 emission, especially when mitigation strategies that induce variation in the H2 are used.

Methane (CH4) emissions from ruminants are of a significant environmental concern, necessitating accurate prediction for emission inventories. Existing models rely solely on dietary and host animal-related data, ignoring the predicting power of rumen microbiota, the source of CH4. To address this limitation, we developed novel CH4 prediction models incorporating rumen microbes as predictors, alongside animal- and feed-related predictors using four statistical/machine learning (ML) methods. These include random forest combined with boosting (RF-B), least absolute shrinkage and selection operator (LASSO), generalized linear mixed model with LASSO (glmmLasso), and smoothly clipped absolute deviation (SCAD) implemented on linear mixed models. With a sheep dataset (218 observations) of both animal data and rumen microbiota data (relative sequence abundance of 330 genera of rumen bacteria, archaea, protozoa, and fungi), we developed linear mixed models to predict CH4 production (g CH4/animal·d, ANIM-B models) and CH4 yield (g CH4/kg of dry matter intake, DMI-B models). We also developed models solely based on animal-related data. Prediction performance was evaluated 200 times with random data splits, while fitting performance was assessed without data splitting. The inclusion of microbial predictors improved the models, as indicated by decreased root mean square prediction error (RMSPE) and mean absolute error (MAE), and increased Lin's concordance correlation coefficient (CCC). Both glmmLasso and SCAD reduced the Akaike information criterion (AIC) and Bayesian information criterion (BIC) for both the ANIM-B and the DMI-B models, while the other two ML methods had mixed outcomes. By balancing prediction performance and fitting performance, we obtained one ANIM-B model (containing 10 genera of bacteria and 3 animal data) fitted using glmmLasso and one DMI-B model (5 genera of bacteria and 1 animal datum) fitted using SCAD. This study highlights the importance of incorporating rumen microbiota data in CH4 prediction models to enhance accuracy and robustness. Additionally, ML methods facilitate the selection of microbial predictors from high-dimensional metataxonomic data of the rumen microbiota without overfitting. Moreover, the identified microbial predictors can serve as biomarkers of CH4 emissions from sheep, providing valuable insights for future research and mitigation strategies.

Significance Agricultural methane emissions must be decreased by 11 to 30% of the 2010 level by 2030 and by 24 to 47% by 2050 to meet the 1.5 °C target. We identified three strategies to decrease product-based methane emissions while increasing animal productivity and five strategies to decrease absolute methane emissions without reducing animal productivity. Globally, 100% adoption of the most effective product-based and absolute methane emission mitigation strategy can meet the 1.5 °C target by 2030 but not 2050, because mitigation effects are offset by projected increases in methane. On a regional level, Europe but not Africa may be able to meet their contribution to the 1.5 °C target, highlighting the different challenges faced by high- and middle- and low-income countries.

CONTEXT: Agricultural systems are generally characterised by many dependent variables that represent their management practices and performances. Parametric approaches are usually used to explore data collected from farms and relations among variables. However, these approaches are generally limited by strong assumptions about the shape of the model that relates variables to each other, which can induce bias in studies.OBJECTIVE: To address these limitations, we investigated the potential of non-parametric kernel density estimators to help explore relations among variables that characterise farms (e.g., forage and milk production, greenhouse gas (GHG) emissions), which have the advantage of requiring no assumptions about the shape of these relations. METHODS: Multivariate kernel density estimation analyses the probability that the values of two or more variables will simultaneously fall within a given range for each variable. The practical utility of this approach was shown by identifying subsets of a population of 96 dairy farms in 2013 in Normandy, France, that had forage production, milk production and GHG emissions that most other farms in the same population were likely to have.RESULTS AND CONCLUSIONS: Several farms outside of the highest density regions, but which lay with the same range of grass or maize production, were able to produce 28% or 27% more milk per cow, respectively (or emit 21% or 9% less GHGs, respectively) each year than farms inside these regions. Characteristics of these farms that increase milk production (e.g., higher maize silage production, more often with majority-Holstein herds) or decrease GHG emissions (e.g., lower maize silage production, more often with majority-Normande herds) were identified. SIGNIFICANCE: Kernel density estimation can be useful for selecting farms with particularly high or low pro-duction or environmental performances in a sample of farms as a function of multiple characteristics.

Manure nitrogen (N) from cattle contributes to nitrous oxide and ammonia emissions and nitrate leaching. Measurement of manure N outputs on dairy farms is laborious, expensive, and impractical at large scales; therefore, models are needed to predict N excreted in urine and feces. Building robust prediction models requires extensive data from animals under different management systems worldwide. Thus, the study objectives were (1) to collate an international database of N excretion in feces and urine based on individual lactating dairy cow data from different continents; (2) to determine the suitability of key variables for predicting fecal, urinary, and total manure N excretion; and (3) to develop robust and reliable N excretion prediction models based on individual data from lactating dairy cows consuming various diets. A raw data set was created based on 5,483 individual cow observations, with 5,420 fecal N excretion and 3,621 urine N excretion measurements collected from 162 in vivo experiments conducted by 22 research institutes mostly located in Europe (n = 14) and North America (n = 5). A sequential approach was taken in developing models with increasing complexity by incrementally adding variables that had a significant individual effect on fecal, urinary, or total manure N excretion. Nitrogen excretion was predicted by fitting linear mixed models including experiment as a random effect. Simple models requiring dry matter intake (DMI) or N intake performed better for predicting fecal N excretion than simple models using diet nutrient composition or milk performance parameters. Simple models based on N intake performed better for urinary and total manure N excretion than those based on DMI, but simple models using milk urea N (MUN) and N intake performed even better for urinary N excretion. The full model predicting fecal N excretion had similar performance to simple models based on DMI but included several independent variables (DMI, diet crude protein content, diet neutral detergent fiber content, milk protein), depending on the location, and had root mean square prediction errors as a fraction of the observed mean values of 19.1% for intercontinental, 19.8% for European, and 17.7% for North American data sets. Complex total manure N excretion models based on N intake and MUN led to prediction errors of about 13.0% to 14.0%, which were comparable to models based on N intake alone. Intercepts and slopes of variables in optimal prediction equations developed on intercontinental, European, and North American bases differed from each other, and therefore region-specific models are preferred to predict N excretion. In conclusion, region-specific models that include information on DMI or N intake and MUN are required for good prediction of fecal, urinary, and total manure N excretion. In absence of intake data, region-specific complex equations using easily and routinely measured variables to predict fecal, urinary, or total manure N excretion may be used, but these equations have lower performance than equations based on intake.

à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.