A study by Assouma et al. (2019a), Assouma et al. (2019b) was set in the pastoral area of Ferlo, Senegal, with a “landscape” ecosystem approach measuring the overall carbon balance of the Widou Thiengoly service area borehole, a 706 km² circular area that hosts 354 pastoral camps with a population of 4,800 inhabitants. The total livestock population of the area is about 25,000 free grazing animals, including cattle (54% of total livestock units), sheep (27%), goats (8%), donkeys (8%) and horses (3%). The area under study was subdivided into different units based on land use, the type of vegetation and soils (Assouma et al., 2017), with grazing lands occupying the greater area (90% of the total area) with sparsely wooded savannas, and the remaining area shared between settlements, forest plantations, natural ponds, boreholes and enclosures, set up for experimental plots. In this area, a formal research project was set up to assess all main sources of GHG emissions and C sinks of the whole pastoral ecosystem, taking into account not only anthropogenic sources.

Two territorial-scale LCA-based studies were carried out in Qinghai Province of China on GHG emission of pastoral systems in transition. In the first study (Zhuang and Li, 2017), Chanaihai village and Guinan Grassland Development Limited Company offered the opportunity to compare a traditional pastoral system with a livestock husbandry system, combining extensive and semi-intensive management systems, respectively. The Chanaihai pastoral system is characterized by 431 households practicing transhumant seasonal pastoralism on 18,779 ha of rangelands, where 63,701 sheep and 4,000 yaks are reared. The different seasonal pastures provide for the needs of livestock production throughout the year. The Grassland Development Limited Company operates an intensive livestock husbandry system, with a combination of rangelands, feedlots, artificial grasslands and stations for forage and fertilizer processing. This demonstration area is characterized by 15,800 ha of rangelands and 2000 ha of artificial grassland supporting 80,000 sheep. The rangeland is utilized for grazing, similarly to the extensive system. Feedlots are only used for rams feeding, while the rest sheep are still grazing on the rangeland. The study assessed and compared the GHG emissions under these two production management systems, and explored the major causes for the differences, comparing GHG emissions from a pastoral system with a more industrialized intensive system, involving feedlots, seeded pastures and imported feed (Zhuang and Li, 2017).

On the other side, in Sichuan Province of China, GHG emissions under two rangeland management systems were estimated in the two villages Axi and Xiareer (Zhuang et al., 2019). The two villages have similar socio-economic and ecological environments but different forms of rangeland management systems. Axi village allocated its rangelands to individual households, with all-year continuous grazing under the individual use of rangeland with fences demarcating boundaries. In the Axi territory, 20,000 ha of grassland provide pastures to 16,000 sheep and 15,000 yaks, reared by 203 households with around 1,000 herders. Xiareer village has a community-based system for the common use of the whole rangeland. Xiareer village has 12,000 ha of grassland grazed by 20,180 sheep and 6,000 yaks, reared by 140 households with 802 herders. This study compared the Axi household-based system under continuous grazing in fenced, individualised plots with the Xiareer more traditional community-based system, involving movement across four seasonal pastures.

The Sardinian case study is based on the analysis of a transition from a semi-intensive to a semi – extensive dairy sheep farming system (Vagnoni and Franca, 2018; Arca et al., 2021). A farm located in North Sardinia, Italy (Osilo municipality), shifted from an intensive feeding system based on grain-cereal crops, annual forage crops and irrigated crops, with cheese industry-oriented production (milk sold to the cheese industry), towards a semi-extensive system switched to valorisation of natural feed resources. Extensification of the production system was part of a wider new farm management strategy oriented to a direct on-farm cheese manufacturing, and to reducing production costs. Extensification slightly reduced flock size (from 340 to 320 sheep) and animal diet switched from hay/silage consumption from temporary grasslands to the grazing of legume-rich permanent grasslands that increased from 10% to 90% of 70 ha farm agricultural area, with a strong reduction of concentrates, fertilizers and irrigation use. This diversification process has certainly increased the complexity of the work on the farm, requiring know-how on cheese making and marketing, but valorising the role of young family members employed into the renewed farming system (Dumont et al., 2022).

The conceptual model based on an ecosystem approach was aimed at estimating the main C and N stocks and flows at Sahelian pastoral ecosystem level. In detail, GHG emissions to the atmosphere (CO₂, CH₄, N₂O), were analysed accounting for enteric fermentation (CH₄) emitted by both livestock and termites, anaerobic degradation (CH₄) in standing water of ponds closed to the boreholes, organic matter decomposition (N₂O and CH₄), combustion (CO₂) and bush fires from grasses and trees (N₂O and CH₄). On the other hand, C sinks are represented by woody plants, deposited excreta and bodies of the animals. In the study, Assouma et al. (2019a) have found that the emission estimated by IPCC Tier I protocol for Africa (25 g dry matter/kg live weight) is probably too high for the pastoral landscape they observed. The authors proposed a new standard of either 18 g dry matter/kg live weight for cattle and 34 g/kg live weight for small ruminants, or 73 g/kg metabolic weight for all ruminants, as a result of this ecosystem approach, revealing a level of dry matter intake/kg live weight for cattle and for small ruminants lower than that estimated by the current standard of IPCC used in emissions calculations. This approach application permitted finding a negative C balance, which clashes with the negative image of African livestock consolidated by previous sectorial and LCA studies (Gerber et al., 2013), based only on anthropogenic sources of GHG.

Two studies were carried out, both applying an LCA approach. In the first one (China 1), the study was based on the GHG emission inventory and a calculation method applied to two different livestock production systems (traditional pastoral vs. combined extensive/intensive). The GHG emission inventory included enteric CH₄ emission, direct and indirect N₂O emission from manure, CH₄ soil emission, N₂O and carbon emission/uptake, and emission from manure combustion. In addition, carbon emission/uptake from artificial grassland, and feedlot, and other external inputs (seed, diesel, electricity, organic fertilizer, etc.) were accounted for in the estimation of GHG emission in the combined extensive/intensive system. GHG emissions were higher in the intensive system, both per ha of area and per kg of carcass weight. The intensive system had a slightly lower production of methane, but with a reduction of methane emissions significantly lower than is suggested by the literature. The reduction of methane emissions in combined extensive/intensive system was significantly lower than often suggested by the literature, at 6.95% rather than between 22% and 62%. Moreover, the higher emissions in the traditional pastoral system were offset by lower emissions from external inputs and higher levels of C sequestration. Overall, the intensive system had 40% higher emissions per carcass weight.

In the second study (China 2) two livestock production systems (household-based under continuous grazing vs. traditional community-based system) were compared. LCA was used to evaluate the GHG emissions intensity of both systems. Enteric emissions and manure CH₄, and N₂O emissions were calculated using the S-GAF model (Zhuang and Li, 2017). The soil CH₄ and N₂O emission data were measured on-site by using the static chamber method. On the other hand, soil organic carbon was measured by sampling the topsoil (0–30 cm) in 2011 and 2014 on-site. As a result, on-farm emission patterns of the two systems were broadly similar (around 9 kg CO₂-eq per kg of meat), but when the C sequestration levels were added, the contrasts were striking. The traditional community-based system showed a positive C balance (0.62 kg CO₂-eq per kg meat), while the household-based system had a relatively high net emission level (10.51 kg CO₂-eq per kg meat). This huge difference is explained by the authors through differences in soil C sequestration, due to the predominance of perennials, incorporation of litter and the spreading and trampling of manure in the traditional system (Chen et al., 2015). Overall, the study reported that light grazing contributes to soil C and N sequestration, while enclosed areas have compacted soil, less mineralization, lower root biomass and reduced quality of forage biomass from artificial grassland.

The LCA was conducted using two different impact assessment methods: Carbon Footprint-IPCC (2013) and ReCiPe endpoint (H) V1.12. From an environmental point of view, extensification of this dairy sheep system increased grassland biodiversity (+53% of surface covered by natural grasslands), provided a reduction of environmental impact assessed through a land use-based FU, with −43% kg CO₂-eq per ha of utilized agricultural area-UAA (Vagnoni and Franca, 2018) and a further environmental benefit in terms of soil C sequestration (+63% of sequestered C in the soil) (Arca et al., 2021). On the other hand, carbon footprint assessed using a mass-based FU was similar in both systems at farm level, being respectively 2.99 CO₂-eq and 3.25 CO₂-eq per kg of normalized milk. In both systems, enteric methane resulted in around half of all GHG emissions. Similar results have been found in other Mediterranean systems, including in northern Spain (Batalla et al., 2015), as well as elsewhere in Sardinia (Atzori et al., 2017b). Two methodological issues arise from this study:

1) When emission intensity was expressed per ha of UAA, semi-extensive dairy sheep farming led to better environmental performance even without C sequestration accounting. On the other hand, semi-intensive systems resulted in lower impacts when GHG emission intensity was referred to 1 kg of normalized milk.

This case study estimated the main C and N stocks and flows and GHG fluxes between the pastoral system and the atmosphere, measuring the overall emissions balance at landscape scale. For doing this, Assouma et al. (2019b) suggested an ecosystem approach, considering all the ecosystem components (livestock, soil and plants) and the interactions between themselves and with the atmosphere. Indeed, this ecosystem angle revealed some key points that should be taken into account when allocating impacts between livestock and natural processes: firstly, seasonal patterns of rainfall may drive the entire variability of gas flows, which should be taken into consideration when approaching studies on the environmental performance of systems with a strong territorial and ecosystemic vocation such as extensive pastoral ones. Thus, wet and dry seasons can completely change the outcomes of the environmental assessment of these systems. Focusing on specific emissions for each season could enable addressing seasonally hotspots in emission and designing more relevant mitigation options. Another source of variation was referred by the authors to emissions from manure in pastoral systems with such a wide territorial basis. In fact, in such conditions manure cannot be assumed to be distributed uniformly, as pointed out in most environmental assessments (Houzer and Scoones, 2021). Thus, direct field-level data and not constant livestock excretion rates should be established and used in the emission analysis. Moreover, further uncertainties arise in relation to the proportion of the manure that is managed and the emission factors for nitrogen used may not reflect the context of most extensive livestock production systems (Rufino et al., 2014). In general terms, when only anthropogenic sources of GHG are taken into account, instead of considering the whole ecosystem angle, studies about the environmental impact of pastoral products might be a source of uncorrected interpretation of GHG flows. Starting from this assumption, Assouma et al. (2019a) suggested that, in the case of transhumant livestock that grazes seasonally in different ecosystems during a year, estimating C balances for diverse ecosystems along agro-climatic gradients might better integrate mobile livestock impacts in studies on the environmental impacts of pastoral products, with a focus on the dynamic relationship between pastoral practices and land use. By recognizing the importance of landscape variability and seasonal pasture availability, pastoralists can develop land use strategies that support sustainable pastoral systems, enhance biodiversity, and improve the resilience of both pastoralists and ecosystems in the face of climate change.

Both Chinese case studies investigated GHG emissions of pastoral systems at territorial scale, by using an LCA “from cradle to gate” approach. The two studies underlined some recommendations for better address the studies on GHG emissions from pastoral systems, albeit with different nuances. The Case Study China 1 Zhuang and Li (2017), revealed that enteric CH₄ emissions calculations based on a model rather than field observations could affect the accuracy of results to some extent. Indeed, direct measurements and estimations of emissions allow a context-based analysis, bounding the system not just by focusing on animals and feed, but using a wider system approach. More precisely, the context-based approach should include observations on the vegetation (e.g., plant biodiversity) and grazing managements, in order to have direct and more detailed data regarding, for example, the effect of vegetation and grazing patterns on C sequestration. It means that the approach to be used for estimating GHG emission should not exclude a complex framework of data collection, with direct measurements on plant biodiversity at territorial scale.

In the China 2 case study, recommendations by authors Zhuang et al. (2019) mainly concerned issues relating to the accounting for soil C sequestration, in the case in which different grazing techniques and regimes are taken into consideration, frequent when pastoral systems are linked to the grazing of large portions of territories with mobile and transhumant animals. Land use practices directly influence soil carbon sequestration potential. As highlighted in this case study, understanding the dynamics of soil C sequestration requires accurate assessments of land use changes and grazing management. Zhuang et al. (2019) suggested how an accurate study on soil C sequestration dynamics should be included in the LCA analysis, when it is addressed to define the environmental impact of a changing scenario in terms of a transition from mobile seasonal to sedentarized continuous grazing. Also, more direct field data is needed to better describe this transition. For example, exactly as in the China 2 case study, they suggested that enteric CH₄ emissions should be based on field observations for having more accurate results to some extent. Finally, when grazing management affects vegetation, Zhuang et al. (2019) suggested the collecting of more detailed vegetation measurements, which might help in making observation on the indirect effect of changing vegetation on, for instance, soil C sequestration potential of the changed plant environment.

In both Chinese case studies, the interplay between land use and pastoral sustainability is clear. By adopting practices that enhance soil health, biodiversity and resource management, pastoralists can create systems that are not only productive but also resilient to environmental changes (Teague and Wright, 2018; Díaz et al., 2018). Understanding and integrating land use considerations into GHG emissions assessments and lifecycle analyses will be essential for developing sustainable pastoral systems in the Qinghai-Tibetan Plateau and beyond.

When estimating GHG emissions from a Sardinian dairy sheep farm, after a land use change due to a transition scenario of extensification, a clear environmental benefit was revealed, accounting for soil C sequestration in emission intensity estimation, whatever the FU considered, product-based or area-based. It suggests that LCA of pastoral products, systems and scenarios should be based on a more complete and wider estimation of the environmental performances, reclaiming the use of different FUs in order to allow a more comprehensible and suitable assessment of the environmental impacts. Moreover, in a land use change scenario towards pasture-based systems, soil C sequestration is favoured by the presence of large grazing areas covered by permanent grasslands, whether natural (natural pastures) or semi-natural (artificial pastures). The improvement of soil organic C stock associated with permanent grasslands would contribute effectively to mitigating GHG emissions in Mediterranean dairy sheep farms, highlighting the positive role of ecosystem services provided by extensive farming systems.

Considering the data scarcity on pastoral systems, and high time, energy and economic costs to complete them, the majority of LCA studies in literature make use of data from high-income countries and industrial systems. These global and standardized assessments do not permit catching the complexity and variability of pastoral systems and therefore risk being highly partial and poorly adequate (Houzer and Scoones, 2021). Some authors (Glatzle, 2014; Manzano et al., 2023a; Vagnoni et al., 2024), highlighted that it a new front of investigation and research is urgent, aimed at deepening the assessments relating to the environmental impacts of extensive livestock systems, particularly in relation to the emissions of climate-altering gases, in such a way as to establish, more and more accurately, the effective impact of pastoral farming on climate change. This would allow studies of extensive systems to be freed from standard parameters relating to industrial, intensive and product-oriented farming systems. Direct data and surveys at territorial and ecosystem level are necessary for better describing the complexity of emissions from pastoral systems, and it is even more necessary as these systems are characterized by mobility, nomadism and the use of large portions of grazing lands. The use of default emissions factors (Tier 1), can induce the risk of not reflecting the variability inherent in the extensive production conditions, which is the very basis of production in a non-equilibrium system, especially in pastoral settings (Krätli, 2015). Also in semi-extensive systems, with flock partial sedentarization and where pasture utilization is the principal feed resource, direct data collection is needed to better express the variability of quali-quantitative pasture traits. Models that continue to assume stability and uniformity in livestock systems, ignoring the pivotal importance of this variability, cannot be applied in region-specific contexts where ecosystem and livestock differences emerge. New models, and LCA approaches, should be able to capture this contextual variability which translates into a range of operational options, in particular mobility, as a crucial tool for the adaptive capacity and resilience of pastoralists (Krätli et al., 2023).

The mobility of livestock on large portions of grazing territories and their use characterizes pastoral systems in all their forms (Rugani et al., 2019; Gac et al., 2020). Considering impacts from pastoral systems simply as an aggregate assessment of impacts could miss important patterns of spatial heterogeneity and temporal variability (Houzer and Scoones, 2021). Particularly, GHG emissions are highly variable over space and time in extensive systems, being potentially positive and negative in the same area at different times, requiring much more focused mitigation measures compatible with livestock-keepers’ practices (Charteris et al., 2021). This spatial and temporal heterogeneity should be reflected when analysing GHG emissions from pastoral systems. As an example, permanent grasslands in extensive systems are the most important, and in many pastoral settings the unique, source of feed for livestock. Numerous studies have established that in permanent grasslands the relationship between soil organic carbon, nitrogen, and site factors varies according to the land use types, vegetation types, environment, and soil types (Busman et al., 2023). But, on the contrary, when LCA analysis is performed on productive basis, pasture-based diets, that valorise the availability of permanent grasslands, are often associated with higher GHG emissions due to higher methane emissions from low-quality diets and to the low production efficiency of extensive livestock (Steinfeld et al., 2006; Gerber et al., 2013). Efficiency is often measured in terms of emissions per unit of output (milk or meat), but this does not take account of the multi-functional use of livestock and land into pastoral systems. This standardised measure of efficiency reveals an anthropogenic perspective, for which this mass-based FU results in an indicator aimed at minimizing the input use and the impacts per unit of marketable product (Molle et al., 2017). The utilization of these units derives from the challenge of increasing productivity and production efficiency, typical of industrialised sectors. Extensive systems are considered the least efficient, although they actually make productive use of areas that have limited alternative uses and allow the production of valuable proteins for food use starting from herbaceous biomasses of low nutritional value (Houzer and Scoones, 2021).

Land use is defined as “total arrangements, activities, and inputs undertaken in a certain land cover type (a set of human actions)” and it refers to the social and economic purposes for which land is managed (Verones et al., 2017). Variability of land use and land use changes may affect not only productivity and resilience of land and rural territories and related communities, but also the wider ecosystem benefits provided by their pastoral use (Davies et al., 2012). In many marginal areas of the world, where livestock have long played an important role in maintaining silvopastoral ecosystems, land abandonment and intensification of agricultural production systems via land use intensification is affecting farmland biodiversity (Houzer and Scoones, 2021). The positive impacts of extensively grazing livestock for reducing GHG emissions in open forest-mosaic and savanna landscapes is clear, when compared with land abandonment scenarios, moreover when practices of sustainable grazing are applied (Manzano and White, 2019). Land abandonment leads to woody vegetation encroachment, and related increased fire risk. Pastoral land use can greatly reduce the potential fuel load for forest fires and vegetation structure can be maintained in a way that promotes greater species diversity (Franca et al., 2016; Bernués et al., 2011). On the other hand, environmental benefits of grazing activity may be hindered by inadequate grazing management, and can be obtained only if the pastures are managed in a rational and sustainable way. In practice, avoiding phenomena of overgrazing and undergrazing, which vice versa can create environmental damage, such as soil compaction and loss of biodiversity (Schrama et al., 2023). For these reasons, the proper characterization of Land Use and Land Use Changes (LULUC) within extensive small ruminant systems is of special concern and it’s a matter of yet greater complexity, as no consistent methodology exists for calculating and characterizing the impacts of LULUC over time. Regarding the inclusion of LULUC impacts within LCA of livestock systems, several organizations have given guidelines for their incorporation. In case of an attributional LCA of a small ruminant supply chain, for instance, different standards may be helpful for considering the inclusion of LULUC, including recommendations from the International Life Cycle Data System (ILCD) Handbook (ILCD, 2010; ILCD, 2011) and Product Environmental Footprint (PEF) guidelines (PEF, 2013), the International Dairy Federation (IDF) (IDF, 2010), and FAO-LEAP guidelines (FAO, 2016a; FAO, 2016b). In the framework of SheepToShip LIFE project (www.sheeptoship.eu), recommendations of the EU-PEF, IDF, and LEAP guidelines were followed, utilizing the Ecoinvent database as a source of secondary data, as the LULUC modelling tool utilized by this database considers the PAS 2050 calculation scenarios, and is therefore already compliant with both PEF and FAO LEAP guidelines (Reinhard et al., 2017).

When considering LULUC, great attention needs to be taken on the baseline scenarios. Extensive livestock systems, making use of permanent grasslands (often as patches within wooded ecosystems), can be considered as replications of ‘natural’ or ‘wild’ systems. This makes assessing baselines crucial, as livestock may not add to ‘natural’ emissions (as assumed in standard LCA measures and climate scenarios). As a result, the complete abandonment of livestock production in certain extensive contexts could therefore have negative impacts on landscape-level emissions (Manzano and White, 2019). LCAs current practice used to derive emission estimates usually do not consider such baseline conditions, whether shifts following advocated removal of livestock resulting in a return to ‘wild’ ecosystems or shifts to intensified agriculture. IPCC (2010) uses managed land as a proxy for anthropogenic emissions and removals, thereby not considering such baseline emissions either. This results in large distortions in the interpretation of results, with major implications for policy (Manzano and White, 2019). A new perspective in accounting GHG emissions from extensive livestock systems requires the accounting of natural baseline emissions, the fluxes from a livestock abandoned scenario, where the increase of wild herbivores and fires certainly does not paint an “emissions-zero” scenario (Pardo et al., 2024).

The LCA food scientific community has been wondering for some time about the criteria to be used to identify the most appropriate FU to express the environmental impacts of livestock systems. In fact, the effects of intensification on emission intensity can significantly depend on the FU adopted; some authors have observed that intensification does not produce variations in emission intensity per kg of normalized milk (mass-based FU), while higher values of emission intensity per ha (area-based FU) have been observed in more intensive systems (Salou et al., 2017). Furthermore, the choice of FU can produce controversial results in terms of environmental output, depending on whether the systems are intensive or extensive (Baldini et al., 2017). The use of the FU based on area and the inclusion of the soil Cseq in the estimation of emission intensity are considered by Escribano et al. (2020) as more appropriate for the assessment of environmental impacts of extensive agricultural systems based on permanent grasslands. This conclusion was confirmed by an LCA study on a semi-extensive Sardinian dairy sheep farm, conducted by Arca et al. (2021). In addition, FU based on area avoids large debate about mode of allocation between products (Kyttä et al., 2022). Finally, the attribution of emissions that have a fossil origin vs. a biogenic origin, and the distinction in the latter between their natural and anthropogenic character, can also be very relevant to measure impacts (Manzano et al., 2025).

Pastoral systems can contribute to ES in a broader and more complex way and the extent of this contribution can vary widely. Indeed, they are mainly based on land use and in particular on pastures, often of low productivity, and with extensive use of grasslands, associated not only with the ES mentioned above, but also with a wider series, including, for example, biodiversity conservation (which contributes, for example, to pollination), fire prevention and many others (Vagnoni et al., 2015; Bengtsson et al., 2019; Zhao et al., 2020; Arca et al., 2021; Von Greyerz et al., 2023).

Given the complexity and multifunctionality of pastoral systems and the wide variability of climatic and economic conditions that they have to face, a large scale assessment of all ecosystem services for all general land use types would be necessary for the application in LCA. Yet, as underlined by Barnaud and Couix (2020), ecosystem services as a concept appear as a legitimate analytical tool to investigate multifunctionality. In this perspective, in order to assess the impacts on ecosystem services, a set of indicators would be necessary (van Oudenhoven et al., 2012). Several authors include suggestions for indicators of ecosystem services for land management, both capacity indicators (representing the potential to provide services, e.g., carbon storage) and flow indicators (representing the actual provision of a service experienced by people, e.g., erosion prevention) (Tsonkova et al., 2014). Some ecosystem services can instead be measured directly, mainly provisioning services: e.g., crops, wood or fibre, number of animals and freshwater withdrawal (Kandziora et al., 2013). Cultural services depend on human values and can only be estimated indirectly, e.g., from the number of visitors or facilities, preference surveys and the number of protected species or habitats (Kandziora et al., 2013; de Groot et al., 2010). Regulatory and supporting services are the most difficult to measure, because they often arise from complex interactions of ecosystem properties or occur at spatial scales other than the observed landscape scale (e.g., climate regulation or gene pool protection) (Zhang et al., 2022).

Having pastoralists a functional role for the use and preservation of permanent grasslands, their traditional pasture practices and grazing management help in providing various ecosystem services. But, in general, the multiple functions of pastoral systems, are not considered in LCA calculations. LCAs, for example, do not weigh production against any upper limit on demand, thus distorting the results in favour of industrial systems. Failing to distinguish between different types of “proteins” and specific dietary requirements, these biases are further reinforced (García-Dory et al., 2022; Moughan, 2021). Conversely, an adequate LCA study should be able to highlight whether the negative impact of the low production efficiency of extensive grazing systems can be offset by the positive effect of the ES they provide to the territories in which they operate (Vagnoni et al., 2018; Manzano et al., 2023b). LCA for ruminant systems commonly include only beef and milk as valuable outputs, and no other ES potentially provided (Baldini et al., 2017; Clune et al., 2017). The risk is to overlook the positive contributions to ES, other than direct food provisioning, and this can be misleading when making decisions about future livestock systems, such as promoting the idea that meat produced in intensive systems generally has lower climate impacts than meat from extensive systems (Von Greyerz et al., 2023). The risk is even greater in the case of traditional, mobile pastoralism (Manzano et al., 2023b).

So, including non-provisioning ES in LCA studies might represent a new strategy for more adequately accounting for environmental impacts of multifunctional pastoral productions. However, while food products can be valued using market prices, attributing value to other ES is less straightforward. Some authors (Ripoll-Bosch et al., 2013; Kiefer et al., 2015; Bragaglio et al., 2020) have attempted to include foods and other ES into the LCA study, using economic allocation, where the environmental impact is distributed proportionally to outputs (foods and other ES) based on their economic value. In order to allocate emissions to ecosystem services, Ripoll-Bosch et al. (2013) proposed for grazing systems an approach that uses agri-environmental payments from CAP (EU Common Agricultural Policy). This method is a viable solution to consider for future LCA application to Mediterranean sheep farming system. Of course, this method might be taken into consideration only in European Union countries where these payments are applied, where they are interpreted as subsidies supporting farmers because of recognizing the environmental role of their sustainable productive activities into marginal and rural territories (Tchakerian, 2008). In these countries, agri-environmental payments are directly associated with support for both organic farms and farms in areas of “natural constraints,” where agricultural production is more difficult due to unfavourable natural conditions. There are also a range of payment schemes more directly connected to livestock, including support for biodiversity conservation in semi-natural pastures or preservation of local livestock breeds. It can therefore be difficult to decide which payment scheme/s to include when using these as a base for economic allocation in LCA to account for non-provisioning ES, especially as payments are sometimes only vaguely reflecting the ES provided (Simoncini et al., 2019), which can give variable results.

More recently, von Greyerz et al. (2023) studied the environmental impacts of varying cattle production systems in Sweden, with different levels of extensification, and attributed the climate impact not only to beef and milk, but also to a set of provided ES, applying economic allocation. Results were influenced by the differences between farms in coupling the payment schemes. Particularly, when payments were directly related to livestock rearing (e.g., as mentioned before, related with biodiversity conservation or preservation of local breeds) the differences in the climate impact were still large between farm types, while the difference decreased considerably when all environmental schemes were included. Including ES in the LCA analysis, does not reduce or cancel the emissions, but gives a more realistic and adequate estimate of the environmental impact of a livestock system, moreover when it is related with a large pasture utilization and land occupation, like pastoral systems are. While emissions do not disappear, ES-corrected climate impact can potentially be useful as part of consumer communication or in decision-making, reducing the risk of overlooking ES provided by ruminant production in a simpler way than using separate indicators. Otherwise, including ecosystem services in LCA shall lead to a more complete comparison among the environmental profile of extensive livestock systems. Vagnoni et al. (2018) consider this method as a viable solution to consider for future LCA application to Mediterranean sheep farming systems, mainly if further CAP measures will more strictly recognize the environmental role of Mediterranean pastoralism. In the long term, Atzori et al. (2022) highlighted that considering the LCA method from a systemic perspective, which may enable a transition in the small ruminant sector toward a more sustainable bio-economy based society, prioritizing rural development and capacity building for sustainable production of dairy sheep sector is likely to stimulate both productivity and ecosystem services maintenance.

The majority of LCA studies prioritize the assessment of a limited number of environmental indicators, i.e. GHG emissions and land use, reflecting the limited data available for other factors such as biodiversity and ecotoxicity (Clark and Tilman, 2017; Sahlin et al., 2020). LCAs therefore often fail to integrate the diverse interactions (both positive and negative) between livestock and the ecosystems they encounter into their analyses, and they often do not account for the various ecosystem services that well-managed livestock can offer in a pastoral context (Houzer and Scoones, 2021). Besides its primary function of producing milk and meat (and wool), most extensive farming systems provide other functions to society. Some examples of ecosystem services attributed to pastoral grazing include: maintaining both the vitality and the traditions of rural communities (specially, in marginal areas), as well as preventing environmental issues (i.e., soil erosion and desertification, wildfire, biodiversity depletion, etc.); conserving cultural landscapes; maintaining rangeland (which maintains sequestration potential in soils and vegetation), nutrient and water cycle regulation and soil health (Schils et al., 2022). The actual contribution of pastoral systems to such services has rarely been quantified (Glimskär et al., 2023).

In Europe, farming systems such as pasture-based systems - evolved towards complex socio-ecological systems in situations of natural resource scarcity, where production is integrated with biodiversity conservation and the delivery of ecosystem services to society - are considered farming systems with High Nature Value (HNV) that require quantifying and paying ecosystem services as a tool for developing strategies to simultaneously meet future food security and sustainable agriculture (Pinto-Correia et al., 2018; Pinto-Correia et al., 2022). However, this approach constitutes a formidable challenge and a requires continuous renewal of policy mechanisms to compensate management models that ensure the delivery of societally desired public goods, gaining new practices and business models that allow for the preservation or even regeneration of natural resources, biodiversity and landscapes (Bouma, 2021; Schröder et al., 2020). After performing as literature review about the inclusion of ecosystem services into LCA analyses on food production systems, Vanderwilde and Newell (2021) clustered the LCA-ES literature in the main categories of “biodiversity and ES clusters,” “land use cluster” and “dynamic modelling,” highlighting how reconsidering LCA with an ecosystem services extension can provide a complementary evaluation both of beneficial contribution of ecosystem services to wellbeing and, also, their detrimental impacts. Nevertheless, despite the recent progress in explicit consideration of ES in LCA (Zhang et al., 2010), today there is still a general lack of a methodological consensus on how to comprehensively integrate ES accounting in LCA and the first steps are being taken to unite the theory and practice of accounting for ecosystem services in LCA, albeit with difficulties linked to a methodological and procedural consensus (McLaren et al., 2021). More recently, Damiani et al. (2023) stated that biodiversity loss can be assessed in different ways into LCA environment but will presumably include a broader set of information regarding, for example, taxonomic coverage and other ecosystem services.