1 BACKGROUND

The concept of Sustainable Intensification (SI) in agriculture emerged as a proposed solution for navigating the complicated network of competing food system priorities while adhering to the defined planetary boundaries for agricultural production (Garnett & Godfray, 2012). This concept emphasises the imperative to enhance food production on existing agricultural land without further interfering with the natural ecosystem. Nevertheless, there has been a recent misinterpretation of the concept, leading to a global increase in agricultural production intensity and resource utilization. This trend has culminated in over 80% of the world's arable lands being allocated to monocultures for grain production, resulting in a decreasing crop diversity per unit of land (Altier et al., 2015). This phenomenon is also attributed to the growing expansion and intensification of the livestock sector in various regions (Godfray et al., 2010), leading to shifts in land use and a decline in agro-diversity within landscapes.

In high-income nations, specialised dairy farms are characteristic of the livestock sector, where highly productive dairy breeds are housed indoors year-round. Commonly seen in intensively managed farms in the US and central Europe, the animals are fed energy- and protein-rich diets, relying mainly on maize and imported concentrate feed (EC, 2021) to maximise milk performance per cow, while the share of grass products in these diets drops down to levels lower than 30% in terms of total energy demand. While this approach has elevated crop productivity, milk production and the dairy industry's efficiency and profitability, it has also introduced challenges such as nutrient accumulation and leaching, as well as surplus manure accumulation, reduction in drinking water and air quality, biodiversity and climate stability (Taube et al., 2014). Decoupling crop and livestock farming has resulted in a high N surplus and social cost. In northern Germany, for instance, about 75% of the surplus N resulting from N application (+100 kg N ha⁻¹) is directly linked to animal husbandry, inducing an estimated social cost of nearly 1000 € ha⁻¹ (Table 1). In some regions, former grasslands have been repurposed for cultivating cereal crops and silage maize for ruminant feeding, sparking competition between arable land for animal feed and direct human food production (Karlsson & Röös, 2019).

Consequently, an urgent need is to introduce innovative paradigms that embrace solutions advantageous to farmers, the environment and society—an actual win-win-win scenario. A shift should occur in the EU towards multi-functional grass-based dairy farming to produce milk and better protect the drinking water quality, contribute to climate change mitigation, enhance biodiversity, improve animal welfare, and create attractive landscapes. Thus, dairy farms should provide essential ecosystem services (ES) beyond milk production. However, the challenge is to integrate these economic and ecological benefits into a farming system.

2 ECOLOGICAL INTENSIFICATION

The concept of Ecological Intensification (EI), as introduced by (Tittonell 2014), aims to improve ES and strengthen biodiversity within agricultural frameworks. Relatedly, ecological efficiency encompasses the adeptness with which ecosystems convert inputs (like sunlight, water and nutrients) into valuable outputs (such as biomass, food and diverse ES). In agriculture, attaining ecological efficiency involves realising elevated productivity with restrained resource inputs while mitigating adverse environmental repercussions. With this, agricultural systems curtail their environmental impact by optimising ecological efficiency while fulfilling the imperative for sustenance and goods. These concepts synergize closely with the principles of resource-use efficiency and waste reduction. Achieving high eco-efficiency necessitates strategic adjustments in managing individual crop and livestock operations or the broader land-use paradigm, as underscored by (Wilkins, 2008). This might encompass judicious manipulation of input factors to fine-tune the equilibrium between economic outcomes (products, services, activities) and the environmental consequences embedded in production, consumption and disposal cycles.

The core principle of achieving high eco-efficiency is illustrated in Figure 1, demonstrating that as resource inputs increase, the output level typically exhibits diminishing marginal productivity; however, environmental load escalates disproportionately. The pinnacle of eco-efficiency lies where the quotient of ecological load and productivity is minimised, necessitating the acceptance of slightly reduced production levels (de-intensification) for agricultural commodities when necessary. Accordingly, to ensure overall high eco-efficiency, high-input farming systems (Point ‘a’ in Figure 1), characterised by substantial environmental load, should consider de-intensifying agricultural commodity production to enhance ecosystem service provision (ecological intensification). This involves a nuanced approach, which does not necessarily entail a complete shift to organic practices but rather a combination of strategies, so-called ‘hybrid agriculture’ strategies (Taube, 2022), to maintain a high level of land use efficiency (LUE) and to avoid leakage effects (Lambin and Meyfroidt, 2011). For example, optimising eco-efficiency might involve relocating the production of specific agricultural commodities to regions with the highest ecological efficiencies. Conversely, in regions where farms (for example, smallholder farms) are positioned at a lower point on the production function (Point ‘b’; Figure 1), characterised by low productivity, efficiency and environmental load due to minimal inputs, a strategy of sustainable intensification is required to boost food production and land use efficiency (LUE) without compromising ES (Brandt et al., 2020).

A prerequisite for high eco-efficiency in agricultural systems is circularity. The circularity concept posits the strategic harnessing of the potential of animal-sourced foods within the global food paradigm, capitalising on the intrinsic trade-offs embedded within the food system and explicitly mitigating feed-food competition. A perspective gaining traction suggests that the presence of livestock in future food systems should be contingent upon minimal feed-food competition (Mottet et al., 2017; Springmann et al., 2018; van Zanten et al., 2018). With this, livestock emerges as a transformative agent if it can convert otherwise disregarded by-products—resources that remain unconsumed or undesirable for human consumption, such as crop residues, co-products, and food chain waste—into valuable commodities like meat, milk, eggs and manure, while simultaneously rendering various ES. A study by van Zanten et al. (2018) illuminates the potential of livestock in recycling these remnants and harnessing existing grassland resources for animal production. According to these authors, such ‘low-cost livestock’ could meet an optimum of 14%–23% of the daily per capita protein demand derived from animal-sourced foods (ASF), while current figures in Europe indicate a two to three times overconsumption of protein from ASF. This strategic approach presents an effective and efficient avenue for leveraging arable land for food production. Such a trajectory is projected to curtail nitrogen (N) and phosphorus (P) losses by 40% and 46%, respectively, and greenhouse gas (GHG) emissions by 19%–50%, vis-à-vis a business-as-usual scenario (Röös et al., 2017).

In a concerted effort to tackle the challenges mentioned earlier, which are explicitly linked to animal agriculture, the European Union (EU) has embraced the ‘Farm to Fork’ strategy to address multifaceted issues within European agriculture (EC, 2020). This strategy, underpinned by agricultural circularity principles, envisions a future where livestock's role is carefully calibrated to minimise feed-food competition, leveraging their unique ability to transform discarded resources into valuable sustenance and services. In the ongoing discourse, therefore, the efficacy of ruminant livestock production in converting leftovers into animal-derived sustenance comes to the fore, albeit with relatively greater associated GHG emissions than their non-ruminant counterparts. Nevertheless, the latter escalates the challenge of feed-food competition on agricultural land (van Selm et al., 2022).

3 A CASE FOR LEYS IN CROPPING SYSTEMS

Acknowledging that “low-cost livestock’ could play a beneficial role in future food systems, the need to implement ecological intensification and agricultural circularity concepts for diverse ecosystem benefits through a careful localization of a context of sustainable practices comes to the fore. One possibility to reduce the high environmental loads associated with the highly specialised farming practices of our modern production system (separated arable crops and dairy/pig farms) is to promote cooperation by developing joint crop rotations to facilitate nutrient cycling. Accordingly, a concrete illustration of translating the principles of ecological intensification into practice emerges through the implementation of integrated crop-livestock systems (ICLS) that includes leys. This cooperation could achieve the highest LUE and ensure social acceptance of low-level ASF in Europe when forage for dairy cows, for instance, comes from absolute grasslands (but not from peat soils due to high GHG emissions) and/or from arable land if dairy inclusion makes arable cropping systems better in terms of soil fertility and additional ES. ICLSs maximise the interaction between system components, that is, soils, plants, and animals, and, therefore, have the potential to decrease environmental impacts while maintaining production levels (Ryschawy et al., 2017). Such systems were essential in recycling nutrients in the past, but the advent of synthetic fertilisers has made them less critical (Schut et al., 2021).

The role of leys in integrated crop-livestock systems and their positive impact on the livestock of South Australia and Mediterranean states of Africa pre-and post-World War II eras, respectively, have been well chronicled by Byerlee (2023). In their review, Taube et al. (2014) expounded on the positive effect permanent grassland-based dairy farming could have on the ecological footprint of milk in our world today. Production grasslands, consisting of permanent and temporary grasslands, provide feed for herbivores and ruminants but can offer additional benefits. Perennial grasses have a more extended growing season, a dense, fibrous root system, and substantially larger below-ground biomass productivity than annual cash crops (Loges et al., 2018). Where grass-legume leys are a component of the crop rotation within the context of ICLS, nitrogen fertiliser and plant protection inputs can be lowered while additional carbon is sequestered in the soil (Goudriaan, 1992). Moreover, mixed grass-clover ley-arable systems (temporary grass of 2–5 years) can capture the benefits of grass swards temporally in combination with periods of intensive cropping and ensure greater protein self-sufficiency of the livestock farm (Peyraud et al., 2014). The termination of grasslands results in a large quantity of mineral N available to subsequent crops, thus reducing N fertiliser requirements in the arable phase of the rotation with greater yields of the following arable crop (Alderkamp et al., 2022; Lemaire et al., 2015), thus minimising nutrient leaching. Accordingly, ley systems potentially yield positive pre-cropping effects, adding biologically fixed N when legumes are included, thus reducing external N input and pesticide use (Martin et al., 2020).

The use of mixed grass-clover ley-arable systems within ICLS arrangements could improve the eco-efficiency of agricultural systems. Ley-crop rotations can provide indirect ES, such as improving soil fertility, controlling weed and pest populations, and saving energy due to the economy of avoiding tillage and fertiliser applications (Lemaire et al., 2014). Multispecies swards in leys, including tannin-rich species, may have additional benefits like reducing N₂O emission from soils via biological nitrification inhibition (de Klein et al., 2019) and mitigating enteric methane emissions when fed to livestock (Beauchemin et al., 2020). Multispecies leys could conserve and improve biodiversity by supporting various insect species (Badenhausser et al., 2009; Beye et al., 2022), thus contributing to the food chain for birds and other species (Lemaire et al., 2015). Compared to continuous arable cropping, leys improve soil structure through several biological processes, including more significant litter inputs, as well as high C and N returns from grazing animals, supplying constant food or C to soil organisms (Yeates et al., 1998). The reduced tillage disturbance characteristic of leys promotes a larger population of earthworms (Yeates et al., 1998) and herbivorous nematodes (Bouwman & Arts, 2000).

4 ECO-EFFICIENT PASTURE-BASED MILK PRODUCTION AT LINDHOF EXPERIMENTAL FARM

The Lindhof research farm of Kiel University is an ICLS designed to demonstrate the high eco-efficiency of pasture-based milk production, where at least 75% of milk came from grassland of multispecies forage mixtures (of functional diversity). The system promotes homegrown protein provision, ensures yield stability and maintains high nutritive value. In addition, the arrangement was aimed to provide multiple ES (clean water, climate change mitigation, biodiversity, animal welfare and attractive landscapes) while maximising milk production per hectare. It was hypothesised that combining N self-sufficient high-yielding grass-clover leys with low N₂O emissions and significant carbon sequestration, along with the benefits of low input pasture-based milk production, would lead to high LUE and low environmental footprints per kg of energy-corrected milk (ECM) (product carbon footprint [PCF] and product nitrogen footprint [PNF]). Moreover, it could provide additional benefits for biodiversity and significantly reduce the social costs per kg of ECM produced. The study aimed to provide compelling data to inform policy transitions of EU agriculture towards ICLS, creating farming schemes that maximise milk yield from grass with low concentrate supplementation in a mixed farming setting based on leys (here, organic farming). The research methods adopted at the farm comprised assessing the nutritive value and forage accumulation rate, optimising pasture management, partitioning C, measuring N fluxes and enteric methane and modelling.

4.3 Ecological metrics

4.3.1 Nitrate leaching

Dairy production significantly contributes to the pollution of surface and groundwater with nitrate (NO₃). This mounting concern has spurred the implementation of various regulations, including the European Union Nitrate Directive. The findings from 3 years of field measurements at Lindhof (Smit et al., 2021) offer valuable insights into nitrate leaching across different grassland systems. The outcomes reveal that fertilised permanent grasslands and grazed grass-clover swards exhibited relatively low (<10 kg N ha⁻¹) and medium (10–28 kg N ha⁻¹) levels of nitrate leaching, respectively (Figure 2). As anticipated, grazing led to greater NO₃N losses than cutting. Within grazed leys, the extent of leaching losses grew with the age of the swards. On average, significant differences between permanent grasslands and leys were not observed. The relatively modest nitrogen losses from grazed swards at Lindhof can be attributed to two factors: (i) diminished sward clover proportion in grazed versus cut swards, and (ii) the extraction of nitrogen through at least one summer silage cut, even from all grazed paddocks.

In the above study, NO₃N leaching was contingent on the availability of mineral nitrogen, land usage, soil type, prevalent weather conditions, and the duration of the growing period. The mean cumulative nitrate leachate over the experimental period was 8.5 ± 1.8 kg NO₃N ha⁻¹. Cover crops under grazing and fertilisation exhibited the most substantial N-leaching losses (~20 kg NO₃N ha⁻¹). However, the values documented at Lindhof fell below those reported in the literature range (10–80 kg NO₃N ha⁻¹; Biernat et al., 2020; Eriksen et al., 2015; Askegaard et al., 2011), positioning Lindhof at the lower spectrum under comparable environmental conditions. Overall, the examined treatments consistently maintained nitrate concentrations well below 25 ppm, thus promoting a dilution effect within the crop rotations. This might be due to a highly effective rooting system, with root length density exceeding 100 km per m² per year measured on grasslands at Lindhof (Chen et al., 2016). Moreover, a pronounced N carry-over effect was observed from grass-clover swards to the ensuing cash crop unit (Smit et al., 2021), consequently diminishing the risk of groundwater contamination stemming from grazed leys. This carry-over effect also led to a decreased nitrogen fertiliser requirement for the subsequent crop following the grazing of the leys.

4.3.2 Biodiversity

Flower-visiting insects have an essential function in plant pollination, and pollination by bumblebees, for instance, increases the yield of many crops (Orford, 2016). Regrettably, pollinator populations are facing a decline, raising concerns (Bommarco et al., 2012). A study conducted at Lindhof (Beye et al., 2022) shows the potential of legume mixtures, particularly those incorporating Trifolium spp., in augmenting bumblebee abundance. Among the diverse mixtures explored, a noticeable effect was observed in the grazed binary and tertiary mixtures of grass-clover leys (consisting of perennial ryegrass (Lolium perenne L.), white clover (Trifolium repens L.) and red clover (Trifolium pratense L.). The study underscores that augmenting species richness and incorporating legumes in grass-clover leys promotes agro-biodiversity compared to conventionally managed grasslands without flowering legumes. Clover incorporation in the grassland (between 20 and 40%) elevated wild bee abundance (Figure 3).

The study recorded a remarkable count of 541 wild bees, encompassing bumblebees and a solitary bee species, spanning two genera and 10 species within plant-species-enriched grass-clover swards. In contrast, no wild bees were observed within nearby conventionally managed grasslands for silage production (with 4–5 cuts annually). Nevertheless, the potential of these species-enriched mixtures to strengthen wild bee populations faced limitations under intensive stocking. Notably, areas excluded from stocking harboured a greater number of long-tongued bumblebees in binary mixtures (Lolium perenne and Trifolium repens) and multispecies mixtures (L. perenne, T. repens, T. pratense; Lotus corniculatus, Cichorium intybus, Plantago lanceolata, Carum carvi, and Sanguisorba minor) than grazed areas. Although the outcomes of this study displayed modest results, the authors indicated that by introducing a more diverse range of forage species with open flowers across successive phenological phases and adjusting stocking practices accordingly, a continuous and varied floral resource for insects could be established.

4.3.3 Greenhouse gas emissions

In an experiment at Lindhof, adding herbs to the grass-clover swards did not provide additional benefits in terms of enteric methane reduction (Loza et al., 2021), nitrification and denitrification losses (Nyameasem et al., 2021) or forage accumulation (Lorenz et al., 2020) as the herbs were maintained at low-yield proportions throughout the stocking season. Nevertheless, enteric methane emissions per unit of milk were remarkably low (averaging 8.8 g CH₄ kg ECM⁻¹ and 9.8 g CH₄ cow⁻¹ day⁻¹ for spring and summer measurements, respectively) compared with literature values (Figure 4). Notably, these values stood significantly lower than the previously documented 17.9 and 17.4 g CH₄ kg ECM⁻¹ for Jersey cows fed 0 and 4 kg of concentrate, respectively (van Wyngaard et al., 2018), as well as the 13.4 g CH₄ kg ECM⁻¹ reported for Jersey cows fed 61% concentrate (Olijhoek et al., 2018). The lower CH₄ emissions and enhanced milk output were attributed to the superior quality of the feed (Loza et al., 2021).

In a related study at Lindhof, Nyameasem et al. (2021) observed that the greatest emission factor for N₂ON (stemming from cow urine and dung) was less than half of the IPCC default value (0.30% compared to 0.77%). The capacity of the pastures to effectively harness nitrogen inputs from excreta patches was linked to their heightened nitrogen uptake ability. Although differences were not statistically significant (p > .05), there was a notable trend of increasing pasture diversity associating with decreasing N₂O emission intensity in the above study. Also, the long-term use of grass-clover in crop rotations was found to accumulate soil carbon substantially (De Los Rios et al., 2022; Reinsch et al., 2018), which significantly reduced the PCF of milk production at Lindhof. The PCF linked to milk from the Lindhof grazing system stands at approximately 0.6 kg CO₂eq kg⁻¹ ECM, in stark contrast to the more than 1 kg CO₂eq kg⁻¹ ECM attributed to conventional milk from year-round indoor systems (Figure 5).

Although the additional benefits offered by the ley system are yet to be translated into monetary terms, it is anticipated that with milk yields ranging from 750 to 900 Mg ECM per farm annually and a projected CO₂ price of 100 € t⁻¹ of milk set for 2030 (Statista, 2023), the advantageous position of the grazing system in terms of mitigated greenhouse gas emissions amounts to around 5 cents kg ECM⁻¹, while also maintaining comparable LUE. In their study, Reinsch et al. (2021) underscored the challenges faced by dairy-focused farm systems in achieving farm nitrogen (N) surpluses below 100 kg N ha⁻¹ (refer to Figure 5). A clear relationship between N balance and GHG emissions showed that the farms with greater N surpluses also had greater GHG emissions per unit of milk and unit area. A decrease from an average farm N balance of +150 kg N ha⁻¹ down to about +50 kg N ha⁻¹, as realised in the Lindhof system, would result in a substantial mitigation potential for social costs caused by lower environmental N pollution. We estimated the avoided environmental costs (Table 4; ct kg⁻¹ ECM) using available prices from the literature to provide a rough estimate regarding social cost avoided by the Lindhof system (IFG) relative to the indoor system (IC), according to Reinsch et al. (2021). The socially spared cost attributed to adopting the Lindhof system was computed at 25 cents kg ECM⁻¹, with the most significant cost component being 13 cents kg⁻¹ ECM arising from phosphorus (P) pollution. We envisage that comparable calculation methods will become more critical regarding greenhouse gases and nitrogen loads going forward.

TABLE 4. Avoided abiotic environmental costs per kg ECM compared between the Lindhof system and intensive indoor system.

	Indoor dairy	Lindhof	Difference	Unit cost	Social cost avoided by Lindhof system (€ kg−1 ECM)
GHG, kg CO2eq kg−1 ECM	1.10	0.60	0.50	100 € t−1 CO2a	0.05
Surplus N, g N kg−1 ECM	12.00	5.00	7.00	10 € kg−1 Nb	0.07
Surplus Pd, g P kg−1 ECM	1.20	0.01	1.10	120 € kg−1 Pc	0.13
Total					0.25

• Abbreviation: ECM, energy-corrected milk.
• ^a Statista (2023). https://www.statista.com/statistics/1334906/average-carbon-price-projections-worldwide-by-region/.
• ^b Brink and van Grinsven (2011).
• ^c UBA (2021). https://www.umweltbundesamt.de/tags/phosphor .
• ^d Phosphorus was calculated using data on concentrate feeding and mineral fertiliser import.

The 2.5-year grass-clover ley phase, integral to milk production at Lindhof, yielded supplementary benefits for the downstream cash crop segment of the farm. Within the soil, nitrogen-rich grass-clover roots and stubble are stored, safeguarding them against leaching, and subsequently serve as a vital N source, furnishing nearly 150 kg N ha⁻¹ to the directly succeeding crop (in this case, edible oats). The farm's manure resources were applied to the subsequent winter cereal crops, leading to heightened yields. This approach effectively ensured that the cereal crops functioned as a practical counterbalance, compensating for the excess nitrogen from milk production (approximately 88 kg N ha⁻¹ of grass-clover forage; Table 2). The cumulative areas allocated to crops and pastures culminate in an N balance of +18 kg N ha⁻¹. This approach concurrently facilitates the reduction of inevitable N losses while contributing to a positive humus balance. Accordingly, ley-based milk production can be associated with a very low carbon and nitrogen footprint as well as a high LUE, contributing significantly to more functional diversity in agriculture. On the other hand, specialised all-indoor high-input/high-output milk production systems can be costly for society.

5 CONCLUSION

Specialised dairy systems within Western Europe exhibit robust milk output per hectare, deriving considerable economic advantage. These systems possess the potential to yield substantial incomes and effectively compete within the milk market, provided social costs remain unaccounted for. This economic competition, however, depends on milk price and the extent of the competitive advantage of low feed costs generated from grazed systems. The ongoing trend of specialisation and intensification in dairy production must align with the imperative to mitigate regional-level greenhouse gas and nitrogen emissions. Currently, these systems deviate from the principles of ecological intensification. Our case study confirmed that the ongoing specialisation and the intensification of dairy production do not adhere to the need for mitigating GHG and N emissions at the regional level.

Contrastingly, the Lindhof case illustrates the advantages conferred by a grass-clover ley-based ICLS, particularly when compared to specialised systems in terms of a spectrum of ES. Even when considering equivalent producer prices observed in conventional systems, the Lindhof approach emerges as economically competitive. Remarkably, the Lindhof approach reaches these outcomes while maintaining a commendable level of LUE. As such, this system embodies a ‘resilient ICLS narrative’ (comprising ley systems and pasture-based dairy) that markedly enhances functional diversity within agriculture, considerably curtailing social costs, and aligns with the goals outlined in the Farm to Fork Strategy. Linking high milk production with grazing in mixed farming systems is one strategy for the economic resilience of (dairy) farming and providing long-term ES for society. However, integrating ley-based ICLS into the dairy production sector necessitates further research, development, advisory services, and political support to realise its full potential.

6 CONSIDERATIONS FOR THE FUTURE

Enhancing the eco-efficiency of agricultural land-use systems by including ruminants as recyclers (consuming residues and products from permanent grasslands and leys) holds significant promise. This approach not only presents the potential to decrease environmental impact but also offers the advantage of remarkably low-cost food production. By capitalising on these aspects, agricultural systems can take steps towards sustainability while maintaining economic viability. Nevertheless, ecological intensification might only work under some conditions, for instance, due to crop-specific environmental constraints, reducing crop productivity. For such cases, Taube (2022) suggested the concept of hybrid agriculture, which combines ecological and conventional elements to produce optimum economic and environmental goods and services. Such a practice would require further research to establish its workings and quantify its effects. Additional specific questions arise as we consider adopting and implementing such strategies. First of all, are more stringent certification measures necessary to ensure the integrity of these practices? For instance, could a ‘grass milk’ certification, which requires that a substantial portion of feed protein and energy, as mentioned above, originate from grass (e.g., >75%), give consumers greater confidence in sustainability claims? Secondly, while life cycle assessment methodology offers a comprehensive view of the environmental implications of various systems, does it include the complete spectrum of benefits associated with ley systems, potentially extending beyond metrics like the carbon footprint? Could further research show the often-hidden social costs of high input and high output systems, aiding in quantifying their overall impact on society? Finally, considering policy implementation, a critical question regarding how the benefits of ley systems can be effectively integrated into the Common Agricultural Policy (CAP) arises. Could mechanisms such as a ‘Public Goods bonus’ (DVL, 2020) incentivise farmers to transition towards these more sustainable practices, acknowledging their positive impacts on the environment and the broader society? As we move forward in shaping the future of agriculture, these questions need further exploration and consideration.

FUNDING INFORMATION

The EU-Horizon-2020 Project financed this project: R4D-Resilience for Dairy/Grant agreement ID: 101000770.

CONFLICT OF INTEREST STATEMENT

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.