Breeding a sustainable future for milk production

Selective breeding to improve the productivity of livestock systems has been carried out for centuries as a cost-effective approach to bring about rapid and permanent change. Genetic improvement can enhance food production and help mitigate greenhouse gas emissions by improving production potential, reproductive efficiency and lifespan, health, and lifetime productivity, thereby reducing emissions per unit of product. Historically, focusing selection solely on production traits (e.g., milk and meat yield) had detrimental effects on health, fertility and lifespan, prompting a shift in the last 20 years toward balancing more heritable production traits with less heritable fitness traits due to economic, environmental, and social imperatives. Poor health and fertility lead to inefficiencies, higher feed use (often 50–70% of production costs), and nutrient losses as enteric and manure GHGs. In dairy cows, genetic selection on health, fertility and lifespan reduces the carbon footprint of milk largely via dilution of maintenance as fewer animals and resources are required for a given milk output. Various national economic and environmental breeding indices (e.g., Profitable Lifetime, Envirocow, Healthy Cow in the UK) balance production and fitness traits. Prior research indicates that a carbon-focused index would place relatively higher weight on fitness traits than a profit-focused index, aligning with environmental and social priorities. While genetic selection has already reduced GHG emissions per unit milk over decades, targeting traits associated with carbon emissions can further reduce emissions per cow as well as per unit product. This study compared a traditional economic index with a carbon index using detailed historical lifetime records for dairy cows to assess the balance of traits for sustainable milk production. Economic and carbon coefficients derived previously were applied to predicted transmitting abilities (PTAs) for key traits in 1042 cows. The PTA represents the genetic potential a cow transmits to offspring relative to a zero-average reference. The study tests whether greater emphasis on non-milk production traits (fertility, lifespan, efficiency) can enhance resource use and reduce the absolute carbon footprint of milk production. Cows with a positive economic index and a negative carbon index are considered more sustainable.

The paper situates its work within decades of research on selection indices and sustainability. National breeding indices increasingly include both production and fitness/functional traits to reflect economic and societal priorities. Evidence shows that selection solely for production can harm health, fertility, and longevity, motivating balanced multi-trait indices. Studies have documented sustained reductions (about 0.6–1% per year) in emissions intensity per unit milk in countries implementing national economic indices. Research indicates that a carbon-focused index would assign higher weight to fitness traits than a profit-focused index, potentially enabling permanent, cumulative reductions in emissions per cow. Internationally, indices such as the Nordic Total Merit maintain balanced weighting across production, fitness, and functional traits, and other countries are adopting similar approaches. Emerging genomic PTAs for feed intake and enteric methane are increasingly available and important for sustainability, with modeling suggesting much higher relative weight for methane in a carbon index than in an economic index, and substantial weight for feed intake in both.

Design and data: Genetic and performance records for 1042 Holstein Friesian dairy cows from the Hartpury University commercial herd (UK) that had completed their lifespan between 2013 and 2022 were analyzed. The herd is autumn-calving, pasture- and silage-fed with concentrates, milked twice daily. PTAs for production traits (milk volume, milk fat, milk protein) and fitness traits (somatic cell count, lifespan, fertility via calving interval) were obtained from April 2023 genetic evaluations. PTA reliabilities for production traits averaged 56% (±11; range 25–81%). Traits: Milk volume, fat and protein represent yield and composition; somatic cell count (SCC) is a milk quality and udder health indicator; lifespan is days from birth to end of productive life; fertility is days between calvings (calving interval). Coefficients and modeling: Economic (£ per unit trait per cow) and carbon (kg CO₂-eq per unit trait per cow) coefficients for each trait were sourced from a prior study on the same herd, derived using a bioeconomic model. The model integrates herd production, financial, and nutritional data (averaged 2017–2022) to describe nutrient partitioning over a cow’s productive life, assuming diets meeting energy requirements for maintenance, growth, pregnancy, activity, and lactation. Emissions estimated include enteric and manure methane, and direct/indirect nitrous oxide from excreta and stored/applied manure, aligned with UK National Inventory methodology. The model employs a Markov chain stochastic framework describing herd structure across 11 age groups (pre-entry and lactations 1–10), with survival-dependent transition probabilities. Index construction: For each cow, the overall index value equals the sum over traits of (trait PTA × coefficient).

• Economic index (£ per cow) = -0.08 × milk volume PTA + 3.24 × milk fat PTA + 3.91 × milk protein PTA – 1.50 × SCC PTA – 3.61 × calving interval PTA + 1.48 × lifespan PTA.
• Carbon index (kg CO₂-eq per cow) = 0.12 × milk volume PTA + 5.55 × milk fat PTA + 1.04 × milk protein PTA + 0.84 × SCC PTA + 16.83 × calving interval PTA – 3.65 × lifespan PTA. Trait weightings in this herd’s indices were approximately 49% production and 51% fitness for the economic index, and 29% production and 71% fitness for the carbon index, with greater emphasis on fertility and lifespan in the carbon index. Statistical analysis: Associations between economic and carbon index values were tested using Spearman’s rank correlation (rho). A “sustainable” animal was defined as having a positive economic index and a negative carbon index value.

• Overall, there was a high negative rank correlation between economic and carbon index values among cows: rho = -0.72 (P < 0.001).
• Rank correlations by parity were consistent: first parity -0.72, second -0.80, third -0.66, fourth -0.70, fifth or greater -0.60.
• Average per-cow per-lactation effects over the 10-year period: herd profit change -£27 ± 105 (range £381 to £253 per cow), and carbon footprint change -20 ± 293 kg CO₂-eq (range -784 to 790 kg CO₂-eq per cow).
• Increasing economic and carbon coefficients for fitness traits (SCC, lifespan, calving interval) by 10% widened ranges: economic index from -£409 to £270 per cow; carbon index from -838 to 864 kg CO₂-eq per cow per lactation.
• 31% of cows were “sustainable” (positive economic, negative carbon); 41% had positive economic values; 45% had negative carbon values.
• Sustainable cows tended to have lower average PTAs for milk volume, fat, protein, SCC, and calving interval, and higher average PTA for lifespan than non-sustainable cows.
• Trait–index rank correlations (Table 2) highlighted lifespan as strongly associated with both indices (economic 0.910; carbon -0.914) and fertility (calving interval) as strongly associated with the carbon index (rho = 0.759).

Findings indicate that selecting on a carbon index per cow complements economic selection by placing greater emphasis on non-milk production traits—particularly fertility and lifespan—linked to more efficient resource use and reduced absolute emissions per cow. Historically, economic indices focused on output traits have driven consistent reductions in emissions intensity per unit milk, but a combined approach targeting both profitability and per-cow emissions can achieve further sustainability gains, potentially delivering annual reductions of 0.5–1% in emissions per cow while maintaining productivity. The strong association of lifespan with both indices supports its central role in lifetime productivity and sustainability. National indices are increasingly rebalancing weights from milk yields toward health, fertility, survival, and feed efficiency, with examples such as the Nordic Total Merit maintaining balanced weights across production and functional traits. In this herd, relatively higher weights were placed on lifespan than in national averages, aligning with the observed increase in the proportion of sustainable cows over time. Incorporating emerging genomic PTAs for enteric methane and feed intake into multi-trait indices is expected to further reduce emissions with minimal impact on profitability, given the higher predicted weighting of methane in carbon indices and substantial weighting of feed intake in both carbon and economic indices.

Ranking cows using both economic and carbon indices is feasible and informative for selecting more sustainable dairy animals. A strong negative rank correlation between the indices was observed, and the proportion of cows classified as sustainable (positive economic and negative carbon) increased across birth years studied. Benchmarking on a carbon index per cow alongside an economic index enables reductions in the carbon footprint of milk production when using balanced weightings that emphasize fertility, lifespan, and efficiency traits in addition to milk production.

The analysis is based on a single commercial herd with coefficients derived from prior modeling on the same herd, which may limit generalizability to other populations or systems. The cows were not genotyped; incorporating genomic PTAs—particularly for efficiency traits like feed intake and enteric methane—could refine selection and sustainability outcomes. Results rely on a bioeconomic and emissions modeling framework with assumed diets and herd structure, which, while aligned with national inventory methods, introduces model-based uncertainties.