Agronomy for Sustainable Development (2022) 42:47 https://doi.org/10.1007/s13593-021-00751-7 RESEARCH ARTICLE Mountain sheep grazing systems provide multiple ecological, socio-economic, and food quality benefits Eneko Garmendia 1,2 & Arantza Aldezabal 3 & Elena Galan 2 & Aitor Andonegi 2,3 & Agustín del Prado 2 & Gonzalo Gamboa 4 & Oihana Garcia 1 & Guillermo Pardo 2 & Noelia Aldai 5 & Luis Javier R. Barron 5 Accepted: 25 December 2021 # The Author(s) 2022 Abstract Pastoral systems face increasing pressure from competing global markets, food sector industrialization, and new policies such as Europe’s post-2020 Common Agriculture Policy. This pressure threatens the use of extensive sheep-grazing systems in mountain areas of low productivity but high natural value. Using information gathered at a long-term research setting in a mountainous area of the Basque Country (northern Spain), we assessed the multiple benefits of extensive dairy sheep grazing systems from multiple perspectives using indicators pertaining to ecological, socio-economic, and food quality domains. In this way, we address the benefits that would be lost if sheep grazing abandonment persists in mountain regions. Our results show that the benefits of extensive dairy sheep grazing in the research area include the production of healthy and high-quality foods and multiple ecological benefits including biodiversity conservation. Extensive dairy sheep grazing also contributes to rural development by generating employment and income in marginal, low-productivity lands that can support few economic alternatives. In particular, we found that sheep farmers who produce high-value products, such as cheese, have enhanced their economic profitability and are less dependent on public subsidies. However, careful attention to sustainable practices, support for new generations of farmers, and streamlined supply chains are required. These would contribute to ensure socio-economic benefits for farmers, avoid the ecological costs associated with grazing abandonment, and enhance ecosystem services for the whole society. Keywords Sustainability . Multi-criteria evaluation . Dairy sheep . Mountain areas . Rural development . Biodiversity . Nutritional and sensory food quality 1 Introduction people around the world (Herrero et al. 2013). Small ruminants alone provide vital resources to 300 million smallholders world- Rangelands make up 30–40% of the Earth’s land surface (Sayre wide using mainly rangelands unsuitable for crops (FAO 2018). et al. 2013) and are intertwined with extensive livestock systems. Grassland-based systems include a variety of different These systems depend on grazing of marginal lands, providing management approaches in diverse habitats. This heterogene- multiple services, income, and employment for 1 to 2 billion ity of contexts makes the management of grasslands a key driver of land use changes with multiple, often contradictory, consequences. For example, in some parts of the world, graz- * Eneko Garmendia ing is a main cause of deforestation, while in other areas, eneko.garmendia@ehu.eus overgrazing of grasslands leads to land degradation (Herrero et al. 2013). Elsewhere, grazing abandonment threatens the 1 Department of Applied Economy, University of the Basque Country existence of species-rich semi-natural grasslands (UPV/EHU), Comandante Izarduy Kalea, 23, (MacDonald et al. 2000). In general, grassland-based systems 01006 Vitoria-Gasteiz, Spain in temperate mountainous areas are used in marginal locations 2 Basque Centre for Climate Change (BC3), Leioa, Spain of low economic profitability but high natural and cultural 3 Department of Plant Biology and Ecology, UPV/EHU, Leioa, Spain value (Lomba et al. 2017; Navarro & López-Bao 2018). 4 Institute for Environmental Sciences and Technologies (ICTA), Extensive grazing is threatened by intensification and Universitat Autonoma de Barcelona, Cerdanyola del Vallès, Spain abandonment, especially in low-productivity zones such as 5 Lactiker Research Group, UPV/EHU, Vitoria-Gasteiz, Spain mountainous areas, which are characterized by ecological, 47 Page 2 of 13 Agron. Sustain. Dev. (2022) 42:47 economic, and political marginality (Sayre et al. 2013). In biological diversity and cultural heritage (Fig. 1 illustrates Europe, where traditional farming systems are in decline, key features of the study area) within the Aralar Natural grazing abandonment has been observed in various mountain Park. Native grasslands included in the Habitat Directive are regions (MacDonald et al. 2000). Factors leading to grazing dominant: the most relevant is the Jasiono-Danthonietum abandonment include meager incomes, tedious working con- grassland (code 6230, subtype a, 92/43/EEC, European ditions, competing land uses, integration into global markets, Commission 2013). This consists primarily of perennial and changes in Europe’s Common Agriculture Policy; these graminoids such as Festuca nigrescens subsp. microphylla factors have been discussed elsewhere in detail (e.g., (St-Yves) Markgr.-Dannenb. (chewings fescue), Agrostis MacDonald et al. 2000). The consequences of grazing aban- capillaris L. (common bent), Luzula campestris (L.) DC. donment are also diverse, although most studies have focused (field woodrush), and herbaceous dicotyledons such as on associated costs and dynamics from a single perspective. Galium saxatile L. (heath bedstraw), Trifolium repens L. Recent efforts have integrated multiple fields of knowledge (white clover), and Cerastium fontanum Baumg. (starweed) into the study of extensive grazing systems (e.g., Bernués (Aldezabal et al. 2015; Odriozola et al. 2017a). The mean et al. 2011; Ripoll-Bosch et al., 2012; Benoit et al. 2019). aboveground net annual primary production of Jasiono- Our aim is to address the full range of socio-ecological costs Danthonietum grassland is 1.97 ± 0.39 tons of dry matter associated with mountain sheep grazing abandonment by per hectare (Aldezabal et al. 2019). assessing extensive grazing systems benefits from multiple Livestock herds are managed extensively in transhumance, perspectives that integrate all available knowledge. using lowland farms in winter and upland grasslands from With this aim, this study applies multiple ecological, socio- May to November. The number of livestock in Aralar economic, and food quality indicators to a mountain sheep Natural Park has remained almost stable for the last 25 years grazing system at a long-term research setting in the Aralar (17,000 Latxa ewes; average weight of a female Latxa ewe Natural Park (Basque Country, northern Spain). This area, 55 kg); however, as in other Spanish regions, there has been a which is representative of Atlantic European temperate mar- shift towards fewer sheep farm holdings with higher livestock ginal lands, provides a unique laboratory for the study of loads. Most transhumance flocks range between 200 and 500 mountain pastoral systems. In addition to a long tradition of lactating ewes. Latxa sheep produce one lamb per year after a sheepherding that goes back to Neolithic times, the area has a gestation period of 5 months. During the 140-day lactation strong and dynamic shepherd’s community, includes species period, a ewe produces an average of about 1.3 kg of milk and habitats of high conservation value, and produces highly per day. The milk is used primarily for making cheese: the valued sheep products (Andonegi et al., 2021). Idiazabal Protected Denomination of Origin (PDO) is recog- nized worldwide as a high-quality cheese, but most of the production is sold in local markets and/or directly from farms. 2 Material and methods 2.2 Methodology: multiple indicators, data collection, 2.1 Study area and analysis Aralar Natural Park is an 11,000-hectare Special Area of The application of multiple indicators in agri-food systems in Conservation (in the European Natura 2000 network) located general and pasture-based livestock systems in particular has a in the Atlantic Basque Country of northern Spain. The area long trajectory (e.g., Bernués et al. 2011; Ripoll-Bosch et al., has an oceanic climate, with a mean annual temperature of 2012; Botreau et al. 2014). These indicators can be used for 7 °C and annual precipitation of about 1,330 mm. Although assessments (e.g., Singh et al. 2012) or as inputs for multi- calcareous substrates predominate, the oceanic climate and criteria modeling approaches to choose, sort, or rank alterna- abundant precipitation throughout the growing season have tive options (see Gésan-Guiziou et al. (2020) for a recent re- led to the development of acidic soils. The soils contain an view on the diversity and potentiality of multi-criteria decision average of 6,534 ppm Kjeldahl total nitrogen (SD 607.3; min. analysis for agri-food research). In this study, we adopt the 4,500; max. 9,700) and 8.3 ppm available phosphorus (SD 2.7 first approach and use multiple indicators pertaining to the ppm; min. 3 ppm; max. 23 ppm) and have a mean soil water ecological, socio-economic, and food quality domains to as- content of 46.3 % (SD 4.7 %; min. 27 %; max. 65%) at a 10- sess the benefits of a mountain sheep grazing systems from cm depth and a mean pH of 4.8 (SD 0.3; min. 4.2; max. 7.4) multiple perspectives. In this way, we address, with input from (Odriozola et al. 2017a). The vegetation in the park comprises multiple disciplines, the benefits that would be lost if moun- a mosaic of gorse-heather shrublands and semi-natural tain sheep grazing abandonment takes place, a trend that has grasslands. been observed in many European regions (e.g., MacDonald The study area is called the Community (or communal et al. 2000). To cover the wide range of benefits associated land) of Enirio-Aralar and occupies 3387 ha of remarkable with these systems in a balanced way, we assess an equal Agron. Sustain. Dev. (2022) 42:47 Page 3 of 13 47 Figure 1 Key components associated with mountain sheep grazing in Aralar. L.J.R. Barron), (c) Idiazabal PDO cheese ripening in a mountain dairy farm (a) A mosaic of gorse-heather shrublands and semi-natural grasslands (photograph by L.J.R. Barron), (d) vegetation survey in the species-rich (photograph by A. Aldezabal), (b) flock of Latxa breed (photograph by Jasiono-Danthonietum grassland (photograph A. Aldezabal). number of indicators for each of the dimensions. Unlike total production of dairy products and consumption of re- modeling approaches, our assessment uses actual data collect- sources. By moving down one level of analysis (Level n-1 ed over a long period from a research setting. The incorpora- in Fig. 2), we differentiate the relationship between a graz- tion of multiple disciplines in the data collected at the research ing system and the ecological context (e.g., by comparing setting has occurred progressively. While the ecological as- grazing systems with non-grazing systems, as we have sessment is based on more than 10 years of data, food quality done, or with partial-grazing systems). To determine the and socio-economic data were incorporated into the analysis nutritional benefits of dairy products associated with dif- more recently. Data collection methods, including sample size ferent grazing systems in this research area, we analyze the and sampling period for each indicator, are provided in the effects of mountain versus valley grazing during the graz- following subsections. ing season (Level n-2 in Fig. 2). Finally, to understand the The analytical strategy of this study, for identifying and socio-economic effects of different production systems on evaluating the benefits of extensive grazing systems (see mountain herders, we examine the two main typologies of Fig. 2), considers the existence of emerging properties of mountain grazing production systems in the area: cheese complex systems at multiple scales, which provide non- and milk producers (Level n-3 in Fig. 2). Our study focuses equivalent information to understand the performance of solely on grazing systems through an analysis of Levels the analyzed systems (Giampietro 2004). That is, when n-1, n-2, and n-3. A broader Level n analysis, which would analyzing the performance of sheep dairy systems at the assess the contribution of both grazing and non-grazing level of the whole (Level n in Fig. 2), we would analyze the systems to the whole sheep dairy system, requires aggre- gated socio-economic and biophysical data related to the entire dairy system and is beyond the scope of this article. 2.2.1 Ecological indicators Over a 10-year period, we measured four essential eco- logical indicators that are sensitive to grazing in Jasiono- Danthonietum grasslands. These indicators respond (pos- itively or negatively) to grazing management or can be affected significantly by grazing disturbance (Odriozola Figure 2: Multi-scale analytical framework of this study. et al. 2014, 2017a). 47 Page 4 of 13 Agron. Sustain. Dev. (2022) 42:47 & Species richness (S) is the number of species in a given combination with the soil microbial metabolic quotient, surface unit. The effects of grazing on the biodiversity and provides a proxy to assess soil health. species composition of plant communities depend on pro- & Carbon/nitrogen ratio (CN ratio ), a quotient (non- ductivity, evolutionary history, and grazing pressure. dimensional ratio) of leaf carbon content and leaf nitrogen According to the generalized model, productive grass- content per leaf dry matter, is used to estimate forage nu- lands with a long evolutionary history of grazing have tritional value (a proxy of forage quality). Leaf nitrogen experienced divergent selection for species traits that in- content is positively correlated to forage nutritional value crease resistance to grazing and the capacity to compete and livestock productivity, while leaf carbon content is for light, resulting in a species pool of both short and tall related to fiber content and is negatively correlated with species. Under such conditions, grazing intensity strongly livestock productivity. This means that the lower the affects both species composition and community struc- carbon/nitrogen ratio, the higher the forage digestibility. ture, as rapid competitive exclusion of short species for Livestock grazing enhances the forage quality at moder- light would occur with a hypothetical cessation of grazing. ately high stocking rates (i.e., 3.2 livestock units ha−1 d−1: Grazing mediates spatial heterogeneity in grasslands by 13% beef cattle, 52% dairy sheep, and 35% horses; see modulating plant inter- and intra-specific interactions; Odriozola et al. 2014) by increasing white clover abun- the patterns generated by species interactions are directly dance in grasslands (Aldezabal et al. 2019). As a conse- related to diversity (Deléglise et al. 2011). To determine quence, the production of enteric methane per unit intake the contribution of grazing to biodiversity, we assessed diminishes, especially in ruminant livestock (Waghorn plant diversity by measuring species richness. This is et al. 2002, Lee et al. 2004). one of the most frequent and recommended quantitative measures of diversity (Wang et al. 2019). The analysis We estimated the ecological indicators described above was developed at a 0.25 m2-quadrat level, the recom- using data collected in a manipulative experiment conducted mended sampling size in grassland studies. in the study area. For this study, permanently fenced plots & Soil microbial metabolic quotient (qCO2) is the soil mi- (50 m × 50 m each) were installed in May 2005 at four exper- crobial CO 2 emissions (CO 2 flux, expressed in imental sites to ensure grazing exclusion. Next to each of these g CO2 m−2 h−1) per microbial biomass (a measure of the exclusion plots, there was a grazed plot, where sheep, cattle, mass of bacteria and fungi living in soil organic matter, and horses were allowed to graze continuously from May to expressed in mg microbial C kg−1 soil). This soil eco- November (Odriozola et al. 2014) (see Fig. 3). physiological trait indicates the ecological well-being of To estimate species richness, we located 100 sampling the soil ecosystems and is used to monitor ecosystem points systematically in both exclusion (E) and grazing (G) health. As microbial biomass responds quickly to soil plots at three experimental sites (sites 1, 2, and 3). We took management changes, the soil microbial metabolic quo- samples during the growing season of 2014, 9 years following tient provides information about the efficiency of micro- livestock exclusion. Sampling points were spaced 2 m apart bial metabolism. Lower values reflect a more efficient using the triangulation method, as a lag of 2 m proved suitable microbial metabolism typical of mature, stable ecosys- for pattern analysis in a pilot study of these grassland plots tems, while higher values may indicate more immature (unpublished data). At each sampling point, we measured flo- or stressed communities (Aldezabal et al. 2015; Epelde ra composition and structure using two types of abundance et al. 2017). Since this quotient demonstrates the efficien- metrics in quadrats (overlaid on the sampling points) of cy of microbial communities in utilizing C substrate in the 0.5 m × 0.5 m: (i) species frequencies in 49 sub quadrats of soil, it also assesses indirectly the soil potential for C 0.07 m × 0.07 m each and (ii) visual estimation of species sequestration. percentage cover (Odriozola et al. 2017b). From these data, & Soil compaction (resistance to penetration, Mpa) reflects we calculated species richness for each experimental plot. the increase in bulk density or decrease in porosity of soil We measured soil health and ecological indicators (soil due to externally or internally applied loads. This indicator microbial metabolic quotient and soil compaction) in E and provides a useful measure of the soil’s resistance to root G plots at two experimental sites (sites 1 and 4) in 2010. After growth (Correa et al. 2019). Soil compaction, which may cutting the grass to ground level, we collected a total of 48 be caused by trampling under high grazing intensity, can random soil samples (12 samples per site and plot) from a adversely affect nearly all physical, chemical, and biolog- depth of 0–10 cm using a 3-cm diameter core. Each soil sam- ical properties and functions of soil. For example, ple comprises a pool of 10 cores (subsamples). All laboratory compacted soils may hinder root penetration and conse- analyses were carried out in duplicate within the following 2 quently the development of rhizosphere and edaphic in- months. We determined microbial biomass C, following the vertebrates, decreasing soil health. This indicator, in original method of Vance et al. (1987), and soil CO2 emis- sions (CO2 flux), using a PP-Systems® EGM-4 IRGA linked Agron. Sustain. Dev. (2022) 42:47 Page 5 of 13 47 Figure 3 The four experimental sites, where a fenced plot (exclusion or non-grazed area, E plot) and contiguous grazed area (outside the exclusion, G plot) could be observed at each site. (Photographs by A. Aldezabal). to a cylindrical soil respiration chamber SRC-1 (in triplicate content per leaf dry mass. Finally, we calculated the com- around each sampling point, to give an average value). The munity weighted mean, obtaining indicator estimates per soil microbial metabolic quotient (qCO2) was inferred directly site × treatment combination (for detailed sampling pro- by dividing the CO 2 flux by the microbial biomass C tocol, see Odriozola et al. (2014)). (Aldezabal et al. 2015). To determine soil compaction (resis- We used the statistical package SPSS (Version 25.0, IBM tance to penetration, measured with Rimik CP40® cone pen- Corp., Armonk, New York) for our statistical analyses. To etrometer), we carried out five in situ cone index measure- investigate the effect of treatment (exclusion versus grazing) ments around each sampling point and obtained an average on ecological indicators (site was included as random factor), value. we used the generalized linear mixed model. For species rich- Additionally, we defined structural plant species as ness, we used data available from Odriozola et al. (2017b). those which account for at least the 80% of the commu- The raw data we used for calculating soil microbial metabolic nity cover and measured leaf carbon content (C), leaf quotient (qCO2) and soil compaction are presented in nitrogen content (N), and CNratio in seven randomly se- Supplementary Material (Table S1 and S2). For leaf nitrogen lected (7 in G and 7 in E plots), fully grown, undamaged content (LNC), leaf carbon content (LCC), and LCC to LNC individuals from each species (common bent Agrostis ratio (CNratio), we used raw data available from Appendix C capillaris L., chewing fescue Festuca nigrescens subsp. (trait.csv) of Aldezabal et al. (2019). Statistical significance microphylla [St.-Yves] Markgr.-Dannenb., heath bedstraw was declared at P ≤ 0.05. Galium saxatile L., field wood-rush Luzula campestris [L.] DC., and white clover Trifolium repens L.), following the procedure described in Cornelissen et al. (2003). After 2.2.2 Food quality indicators removing the petiole or rachis, we prepared powdered leaf samples and used these to determine the C and N contents To assess the contribution of an extensive mountain dairy using combustion elemental analysis (AOAC 1990). We sheep farming system in terms of healthy and high-quality obtained the CNratio by dividing the C content by the N food production, we focused on Idiazabal PDO cheese, the 47 Page 6 of 13 Agron. Sustain. Dev. (2022) 42:47 reference product of the area and main destination of the milk et al. 2015). Moreover, as flock levels of concentrate and produced by shepherds in lowland and highland farms. oil supplementation increase, the content of non-healthy trans-10 octadecenoic acid (trans10-18:1) in ruminant- & Grass-related fatty acid accumulation (GRFA). derived products increases (Aldai et al. 2013). This indi- Cheeses made with milk from extensively managed flocks cator assesses the cheese fat nutritional quality in relation provide a higher content of healthy fatty acids (e.g., to the forage intake of animals. The grass-related fatty acid vaccenic, rumenic, and α-linolenic acids) as compared to accumulation (non-dimensional ratio) is calculated as the flocks managed indoors and/or under part-time grazing content ratio between certain unsaturated fatty acids in that are supplemented with concentrate (Valdivielso cheese, as follows: GRFA ¼ ½vaccenic acidðtrans11−18 : 1Þ þ rumenic acidðcis9; trans11−18 : 2Þ þ α−linolenic acidð18 : 3n−3Þ=trans10 octadecenoic acid & Fat healthiness (FH): The fat healthiness indicator, a non- grazing (Valdivielso et al. 2015, 2016b). Likewise, the dimensional ratio that assesses the healthiness of the ratio between polyunsaturated and saturated fatty acids is cheese, is adapted from the atherogenicity index defined higher in milk from extensive mountain grazing flocks by Ulbrich and Southgate (1991). As previously reported, compared to that from valley grassland flocks (Bravo- sheep milk fat from flocks managed by mountain exten- Lamas et al. 2018). The fat healthiness indicator is calcu- sive grazing shows lower atherogenicity values than that lated as the content ratio between unsaturated and saturat- from flocks managed by indoor feeding and/or part-time ed fatty acids in cheese, as follows: FH ¼ ½total unsaturated fatty acids½½total trans octadecenoic acidsvaccenic acidðtrans1118 : 1Þ½total conjugated linoleic acidsrumenic acidðcis9; trans1118 : 2Þ= ½lauric acidð12 : 0Þ þ ½4 x myristic acidsð14 : 0Þ þ palmitic acidð16 : 0Þ & Antioxidant capacity (AC): The antioxidant capacity in- flavor and odor than valley cheese (Valdivielso et al. dex (arbitrary units) assesses the presence in cheese of 2016a; Amores et al. 2021). Sensory typing is calculated antioxidant compounds beneficial for human health that as the ratio between specific sensory attributes in cheese, come directly from the animal diet. Previous studies report as follows: significantly higher levels of tocopherols (vitamin E) and terpenoids in cheese from mountain-grazed flocks com- ST ¼ ½acid flavor þ ½4 x rennet odor þ ½4 x piquant flavor=½milky odor þ toasty odor pared to those from flocks managed indoors and/or We based the food quality indicators on chemical and sen- through part-time grazing, although the retinoid content sory data collected at 22 small rural dairies in the study area was similar for each (Valdivielso et al. 2015, 2017). during the Idiazabal PDO cheese making seasons of 2009– Antioxidant capacity is calculated as the content ratio be- 2015. During 2009 and 2010, 10 farms located in valley grass- tween fat-soluble antioxidants in cheese, as follows: lands were monitored monthly from February (indoor feed- ing) to July (extensive grazing) (Virto et al. 2012). In 2011, 6 flocks were monitored in February (indoor feeding), April AC ¼ ½½4 x total tocopherols þ total terpenoids=total retinoids (part-time valley grazing), and June (extensive mountain graz- ing), after having been moved from valley to mountain farms in May (Valdivielso et al. 2015). In 2015, a comparative study during May and June was carried out with 3 flocks located in & Sensory typing (ST): The sensory typing indicator (non- valley farms and 3 other flocks that were moved to mountain dimensional ratio) assesses the typical cheese flavor in farms. Both valley and mountain flocks were managed under terms of sensory quality. The typical sensory characteris- extensive grazing, with valley flocks grazed on semi-natural tics of cheeses are strongly linked to the type of forage and pastures (Bravo-Lamas et al. 2018). The Jasiono- to specific botanical species. Some studies have found Danthonietum grassland dominates the grazing areas of the significant differences in the sensory profile of the moun- mountain flocks (Valdivielso et al. 2016b). In all studies, rep- tain cheese compared to valley cheese. Overall, Idiazabal licate raw bulk milk samples and cheeses ripened for 3 to 6 PDO cheese made on mountain farms has more intense months were sampled from each dairy. Fatty acids, Agron. Sustain. Dev. (2022) 42:47 Page 7 of 13 47 tocopherols, retinoids, terpenoids, and sensory attributes were the percentage of subsidy over the net margin and com- determined in the samples according to previously described pared it with other agricultural holdings in the same methods (Valdivielso et al. 2016a, 2016b). More details on region. analytical data from the different samplings used for estima- tion of food quality indicators are presented in the Two types of dairy producers were identified in the study Supplementary Material (Tables S3–S6). area: those who transform milk into artisanal Idiazabal PDO We performed a statistical analysis using the statistical cheese and sell it directly to consumers (cheese makers) and package SPSS and used one-way ANOVA to investigate the those who sell milk to cheese factories (milk sellers). To cap- effect of the livestock system (indoor feeding in valley, part- ture the diversity of shepherds operating in the area, we sur- time grazing in valley, extensive grazing in valley, and exten- veyed 40% of them, 8 cheese makers and 12 milk sellers. Of sive grazing in mountain) on food quality indicators. We in- these, 7 owned 150 to 300 sheep, 11 owned 300 to 500 sheep, cluded the livestock system as a fixed effect in the model and and 2 owned more than 500 sheep. On average, cheese makers used the Tukey test for pairwise comparisons. Numerical use approximately 55.9 ha of mainly communal land and pro- values of all dependent variables were log transformed for duce an annual average of 34,611 L of milk, while milk sellers statistical data treatment, and we checked normality and ho- use 57.3 ha of land and produce 37,375 L of milk annually per moscedasticity of the data within each group according to the farm. different levels of the main factors. Statistical significance was We used the statistical package SPSS for our statistical declared at P ≤ 0.05. analysis. To investigate the effect of farm typology (cheese maker versus milk seller) on socio-economic indicators, we 2.2.3 Socio-economic indicators used the General Linear Model of Analysis of Variance (GLM-ANOVA) as follows: We measured four socio-economic indicators using data col- Y ijkl ¼ μ þ Di þ L j þ M k þ εijk lected from 20 Aralar farms during 2016. The representative- ness of this year was checked against data available for 7 where Yijk = dependent variables, μ = intercept, Di = farm farms since 2012. typology as fixed effect, Lj = livestock (flock size) and Mk = total milk production as covariates, and εijk = residual random & Net margin. Livestock grazing generates economic reve- effects. Numerical values of all dependent variables were log nues in marginal lands of low productivity (Sayre et al. transformed for statistical data treatment. We checked normal- 2013). To assess such benefits, we estimated the net mar- ity and homoscedasticity of the data within each group accord- gin (in euros) generated by cheese, milk, meat, and live- ing to the different levels of the main factors. Statistical sig- stock production at each local farm. We calculated net nificance was declared at P ≤ 0.05. margin for a farm as follows: income (including revenues from sales plus other incomes such as subsidies) – variable costs (feed, herd replacement, commercialization, veteri- nary) – fixed costs (land rental, salaries, maintenance). 3 Results and discussion & Employment. Small ruminants provide vital resources and employment opportunities in remote rural landscapes 3.1 Ecological assessment with limited alternatives (FAO 2018). For this study, we measured the employment generated directly by sheep All ecological indicators measured at our experimental sites grazing in annual work units (AWU). were affected significantly (P ≤ 0.05) by grazing abandon- & Economic profitability. The long-term viability of moun- ment (Table 1). Species richness was higher in the grazing tain grazing systems depends on its ability to provide re- plots, with an average of seven more species than in the ex- turn to labor and total family income. To assess economic clusion plots. This increase in species diversity was due to the viability, we divided the net margin of each farm by the presence of more forbs and legumes in the grazing plots. This number of AWUs it uses. was especially true for Trifolium repens, which benefits from & Dependency on premiums. Another important socio- reduced competition for light in grazed areas. economic feature of agricultural sector activities is their The soil microbial metabolic quotient was around 0.27 units dependency on public subsidies. This is of particular rel- lower in the grazing plots than in the exclusion plots. Higher evance in European mountain grazing systems, where basal soil respiration (CO2 flux: 1.55 mg CO2 m−2 h−1in ex- Common Agricultural Policy support income schemes clusion plots vs 1.32 mg CO2 m−2 h−1 in grazing plots) and are often employed to compensate activities that have lower microbial carbon biomass (1315.0 mg microbial C kg−1 multiple socio-ecological benefits but low productivity. soil in exclusion plots vs. 1451.6 mg microbial C kg−1 soil in To assess such dependency in the study area, we estimated 47 Page 8 of 13 Agron. Sustain. Dev. (2022) 42:47 a,b Table 1 Mean values and standard deviations for ecological indicators. Means with different superscripts in the same row indicate statistically significant differences (P ≤ 0.05) between treatments (exclusion, grazing). Indicators Grazing treatment Exclusion treatment Species richness (number of species) 32.33a ± 2.33 25.00b ± 2.33 Soil microbial metabolic quotient (mg CO2 m−2 h−1/mg microbial C kg−1 soil) 0.93a ± 0.28 1.20b ± 0.26 Soil compaction (Mpa) 3.31a ± 0.86 1.28b ± 0.21 CNratio 24.51a ± 0.35 29.06b ± 0.88 grazing plots) were responsible for the higher soil microbial As shown in Table 1, medium- to long-term grazing aban- metabolic quotient observed in exclusion plots. donment decreased soil compaction at the 0–10 cm soil depth, Soil compaction was higher in grazing plots, although no- improving conditions for root development. This is consistent ticeable variation was found among sites, probably due to with previous studies (Epelde et al. 2017). Decreased soil differences in stocking rates or grazing intensities. In our compaction in exclusion plots could promote water infiltration study, soils under grazing regimes exceeded 3 MPa. The high capacity, gas circulation through the soil, and the accumula- grazing pressure in these grasslands clearly contributes to soil tion of plant necromass due to the absence of herbivores. compaction, especially at site 4, where horses and cattle are However, grazing abandonment also lowers the amounts of present in higher densities because of the lower altitude. After available nutrients (e.g., nitrogen, phosphorus) from animal 5 years, grazing exclusion allowed soil to recover to compac- excreta and urine, which changes the quantity and quality of tion values close to 1–1.5 MPa. root mass. Changes in organic matter dynamics and nutrient Finally, the carbon/nitrogen ratio was lower in grazing supply are likely to have a profound influence on microbial plots because of the higher nitrogen content of plants under community structure as well as on plant physiological re- a grazing regime (there were no significant differences in car- sponses to grazing. Differences in soil microbial metabolic bon content). values for grazed and excluded plots suggest that soil micro- All indicators, except soil compaction, benefit from bial community functions are sensitive to the impacts of live- grazing (Table 1). Similar trends for the Jasiono- stock grazing exclusion. According to Aldezabal et al. (2015), Danthonietum community are reported by Epelde et al. soil microbial metabolic quotients are lower in grazed plots (2017), who observed that species richness decreased due to reduced CO2 emissions and increased microbial bio- from 16 to 14 after short-term grazing exclusion. mass. This effect could be related to higher soil temperatures Likewise, Odriozola et al. (2017a) detected a negative observed in grazed plots (Odriozola et al. 2014), which also linear relationship between species richness and abun- positively affect microbial biomass. In the absence of herbi- dance of competitive species, suggesting that strong vores, a rise in microbial oxidative activity implies an unde- aboveground competitors outcompeted other species sirable increase in CO2 emission into the atmosphere from (mainly forbs) when herbivores were excluded. This microbial respiration. means that one important benefit of livestock grazing is At this point, it is worth noting that a complete climate the maintenance of plant diversity, which, in turn, en- impact assessment would consider other sources of GHG hances the forage quality of Jasiono-Danthonietum grass- fluxes (e.g., sheep enteric fermentation). For instance, in the lands. Regarding forage quality, similar trends were ob- study area, Batalla et al. (2015) found that the C-footprint of served by Aldezabal et al. (2019), who noted that an in- sheep milk in Latxa breed farms grazing in mountain and crease in carbon/nitrogen ratio after grazing abandonment valley areas was greater (2.8–15 kg CO2-e/kg fat- and was related to a reduction in white clover. Odriozola et al. protein-corrected milk (FPCM)) than that from dairy sheep (2017a) demonstrated that strong aboveground competi- farms under confined management (2.3 kg CO2-e/kg fat- tors (mainly tall grasses and long-spreading stoloniferous and protein-corrected milk (FPCM)). Nevertheless, if soil or- species) outcompeted small species such as white clover ganic carbon (SOC) changes are considered, the C-footprint of via a light-competition mechanism (by creating large milk produced by Latxa breed farming systems can be signif- intra-specific aggregations when herbivores were exclud- icantly reduced, offseting GHG emissions, on average, about ed). Sheep grazing exerts an equalizing effect on compet- 55% (Batalla et al. 2015), of the total C-footprint. Last, but not ing relationships among species because sheep preferen- least, in the light of pledges calling for afforestation of pastoral tially consume grasses (more than 60% of their diet) and lands (e.g., mountain areas) as a means of mitigating climate nitrogen-rich white clover to a lesser extent (Valdivielso change (e.g., Harwatt et al. 2020; Hayek et al. 2021), it is et al. 2016b). important to note that the resultant changes in C stocks or Agron. Sustain. Dev. (2022) 42:47 Page 9 of 13 47 albedo are highly context-dependent (Bond et al. 2019), po- The antioxidant capacity index is related to the presence of tentially counterproductive (Nuñez et al. 2021), and subject to fat-soluble antioxidant compounds, which are beneficial for large uncertainties as landscapes will be colonized by wild human health. Several studies have reported that plant biodi- ruminants and other methane emitters (Manzano and White versity and botanical composition of mountain grasslands 2019). contribute to a high-quality animal diet (Falchero et al. 2010; In terms of biodiversity, our results corroborate that live- Revello Chion et al. 2010; Valdivielso et al. 2016b). We found stock grazing exerts important and direct benefits for plant significantly (P ≤ 0.05) higher AC values in milk and cheese species richness, which also benefits the diversity of other from extensively grazed flocks as compared to those from aboveground and belowground organisms (e.g., insects, flocks managed under part-time grazing and indoor feeding worms, microorganisms) (see also Wang et al. 2019). In this (Table 2). These results confirm the beneficial effect of exten- regard, enhancing shepherds’ skills on guided grazing sive grazing management on the nutritional quality of milk (Meuret and Provenza 2015) can help ensure these benefits and cheese, particularly when produced from mountain farms. across the entire landscape. Active livestock management The higher nutritional quality and healthiness of mountain- plans, such as the new payments for ecosystem services produced dairy foods are related to increased content of schemes being promoted to increase grazing on abandoned healthy mono- and poly-unsaturated fatty acids and fat- (under-grazed) areas in the Navarran side of the park, support soluble antioxidants and decreased specific unhealthy trans the conservation of biodiversity-rich Jasiono-Danthonietum and saturated fatty acids (Aldai et al. 2013; Valdivielso et al. grasslands. 2015; Bravo-Lamas et al. 2018). Therefore, the abundance of specific botanical species in pastures and their fatty acid, to- copherol, retinoid, and terpenoid contents strongly influence 3.2 Food quality assessment the sheep milk and cheese composition (Valdivielso et al. 2016b). Table 2 shows mean values of food quality indicators, mea- With respect to sensory typing, the typical odor and flavor sured in milk and cheese samples from commercial Latxa of Idiazabal PDO cheese have been described and defined. flocks managed under different livestock systems on valley Our study is the first, however, to report scientific information and mountain farms. Feeding management (indoor, part-time on the specific sensory properties of Idiazabal PDO mountain grazing, extensive grazing) and/or farm location (valley or cheeses. Table 2 compares the mean sensory typing values for mountain) influenced all indicators significantly (P ≤ 0.05). ripened cheeses made on valley and mountain farms during As expected, milk and cheese from commercial flocks fed in- the extensive grazing period. Cheeses produced on mountain doors on valley farms yielded the lowest values for grass- farms show significantly (P ≤ 0.05) higher values than those related fatty acid accumulation and fat healthiness indices. produced on valley farms. The fact that Idiazabal PDO These indices increased in flocks managed with extensive graz- cheeses from extensive mountain grazing can be sensory dif- ing systems, indicating a healthier fatty acid profile for milk and ferentiated from those made on valley farms may be an im- cheese from the latter. Furthermore, mean values for the grass- portant factor for increasing consumers’ willingness to pay related fatty acid accumulation index were significantly (P ≤ more for mountain cheeses. Compared to valley farms, the 0.05) higher for milk and cheese produced from extensive incorporation of automatic milking machines, small cooling mountain grazing as compared to extensive valley grazing. In tanks, automatic or semi-automatic vats, and small ripening both types of extensive grazing, sheep were allowed to graze chambers with controlled conditions (temperature and humid- freely outside for the entire day, except during milking (twice a ity) in mountain farms could also contribute to improved day, 12 h apart) (Bravo-Lamas et al. 2018). Table 2 Mean values and standard deviations for food quality the same row indicate statistically significant differences (P ≤ 0.05) indicators measured in milk and cheese samples from commercial Latxa among different livestock systems. 1Index not calculated due to lack of flocks managed under different livestock systems on valley and mountain tocopherol and retinoid data in milk and cheese samples. 2Index not farms. The number of farms (n) sampled in each livestock system was calculated due to lack of sensory attribute data in cheese samples. different depending on the year. a,b,c,dMeans with different superscripts in Indicators Indoor feeding in Part-time grazing in Extensive grazing in Extensive grazing in valley (n = 16) valley (n = 16) valley (n = 13) mountain (n = 9) Grass-related fatty acid accumulation 11.8 ± 2.1d 18.3 ± 2.7c 25.2 ± 3.7b 31.5 ± 2.1a (GRFA, non-dimensional units) Fat healthiness (FH, non-dimensional units) 0.29 ± 0.04c 0.32 ± 0.03b 0.47 ± 0.07a 0.46 ± 0.03a Antioxidant capacity (AC, ratio in arbitrary units) 14.4 ± 3.6b 18.6 ± 2.5b ---1 46.9 ± 5.4a Sensory typing (ST, non-dimensional units) ---2 ---2 7.8 ± 5.3b 12.1 ± 7.6a 47 Page 10 of 13 Agron. Sustain. Dev. (2022) 42:47 production of high-quality products that are better valued in Contable Agraria Nacional), 2015). Because of their higher the market. These technological improvements will undoubt- workload, Aralar cheese makers employ 1.8 AWU per edly improve the cheese making process in mountain facilities farm, while milk sellers employ only 1.3 AWU (Table 2). through increased control of the manufacturing process, better These numbers are similar to those reported in the Basque working conditions, and time savings for the farmer although Country for cheese makers (1.9 AWU) and milk sellers they will also imply an additional economic cost for the shep- (1.5 AWU). On average, cheese makers require 216 ewes herds. Finally, considering the intensification trend that has to generate 1 AWU, while milk sellers require 257. The 50 been observed in some European regions, it would be worth- farms registered in the area, both full and part-time graz- while to assess the benefits of extensive mountain sheep graz- ing, account for in total 99.5 AWU. These results show ing in comparison to more intensive types of lowland grazing that, although mountain dairy sheep farming systems may systems than those addressed in this study. A comparison have little impact on national economies, they can create could combine multiple indicators, such as the ones consid- jobs and increase income in remote mountainous areas ered in this study as well as other indicators defined in collab- such as Aralar. This study considered only the direct crea- oration with relevant stakeholders. Multi-criteria aggregation tion of jobs, but there are multiplier effects in local income algorithms could be used to rank and compare different inten- and employment derived from both backward linkages sification levels of farming systems. In addition, further re- with local suppliers and forward linkages stemming from search is needed to assess the impacts of different degrees of successive expenditures created by increased sales in the abandonment in order to identify critical socio-ecological region. In terms of employment, these linkages imply that thresholds and tipping points at which the multiple benefits for every job created in dairy sheep farming, indirect and noted in this study might be lost irreversibly. induced jobs are created in other sectors. A multiplier anal- ysis, based on Social Accounting Matrices, found that 3.3 Socio-economic assessment backward linkages are greater than forward ones for the dairy sector in Spain (Philippidis et al. 2014). The analysis In socio-economic terms, the net margin and employment of estimated an employment multiplier of 13 in Spain’s dairy the farms we analyzed show that sheep grazing systems have sector. The additional jobs induced by farming vary at lo- the potential to generate economic benefits in marginal moun- cal and national levels but are usually low compared with tain regions with harsh working conditions and few economic other activity sectors (tourism, industry) (Hostiou et al. alternatives. In this regard, we also identify significant differ- 2020). However, farms that sell directly in local markets, ences (P ≤ 0.05) among cheese makers and milk sellers for the as those of Aralar, tend to have higher forward multipliers selected indicators (see Table 3). and consequently generate more income and employment Cheese makers show a superior net margin, due to the high in the region (Ekanem et al. 2016; Malagon-Zaldua et al. added value of the cheese. This is reflected in a higher income 2018). from sales that compensate higher variable and fixed cost of The average economic profitability of cheese makers is 1.7 cheese makers in comparison to milk sellers. The net margin times higher than that of milk sellers and is higher than that of of cheese makers is 2.5 times higher than of milk sellers. Basque (28,938 €/AWU) and Spanish (28,768 €/AWU) small On average, each farm employs 1.49 shepherds (AWU), ruminant farms (RECAN (Red Contable Agraria Nacional), which is close to the averages for Basque (1.7 AWU) and 2015). The dependency of cheese makers on subsidies is in Spanish (1.6 AWU) small ruminant farms (RECAN (Red line with the average for Basque (30%) and Spanish (34%) Table 3 Mean values and standard deviations for socio-economic indicators by farm typology (n = 20). a,b Means with different superscripts in the same row indicate statistically significant differences (P ≤ 0.05) between farm typologies. Indicators Cheese makers (n = 8) Milk sellers (n = 12) Income from sales (€) 85,171 ± 35,632a 36,099 ± 9,145b Other income (i.e., subsidies) (€) 21,615 ± 9,370 15,010 ± 3,400 Variable costs (€) 32,927 ± 18,367 19,786 ± 4,123 Fixed costs (€) 13,611 ± 6,469 7,571 ± 4,429 Net margin (€) 60,248 ± 23,546a 23,751 ± 7,854b Employment (annual work units) 1.8 ± 0.7a 1.3 ± 0.4b Economic profitability (€/annual work unit) 33,647 ± 12,849a 19,287 ± 4,025b Dependency (%) 36 ± 10b 63 ± 30a Agron. Sustain. Dev. (2022) 42:47 Page 11 of 13 47 small ruminant farms (RECAN (Red Contable Agraria 4 Conclusion Nacional), 2015). The dependency of milk sellers is much higher, almost double that of cheese makers, due to lower This study confirms that extensive mountain dairy sheep graz- income from sales. The topography, climate, and remoteness ing systems provide multiple ecological, socio-economic, and of mountainous areas detract from the economic competitive- nutritional benefits. Interestingly, in contrast to the trade-offs ness of extensive grazing systems unless these are linked to between ecological and economic dimensions found in other food products with high added value (Madelrieux et al. 2018), studies (e.g., Ripoll-Bosch et al. 2012), we found significant such as Idiazabal PDO label cheeses. In the study area, the net synergies among ecological, socio-economic, and nutritional margin per AWU of cheese makers is almost double that of benefits. These results demonstrate that it is possible to en- milk sellers, and, the per liter price for milk used to manufac- hance the economic profitability of farms based on sustainable ture cheese sold directly by farms is almost double the price of extensive grazing practices that provide socio-ecological ben- milk sold to dairy factories. This is in line with other studies efits for the whole society (e.g., high-quality food and biodi- (e.g., Renting et al. 2003) that address farmers’ benefits asso- versity conservation). Indeed, the economic profitability is ciated with short supply chains and direct commercialization. most remarkable for on-farm cheese makers, who produce a Fresh milk sales direct to consumers could also increase farm high-quality food product (Idiazabal PDO cheese) that de- incomes, but at present, this type of sales has no market niche. pends on the maintenance and grazing of ecologically rich Despite its advantages, cheese making is not a silver mountain pastures (see also Valdivielso et al. 2016b). bullet for all shepherds. Cheese makers assume higher Even so, European extensive mountain sheep grazing faces variable costs (purchased feed), higher fixed costs (invest- significant challenges: a high risk of economic failure, an ag- ment in equipment), and higher dedication in terms of ing workforce, low social regard, competing land use, and working hours for manufacturing and marketing cheese. integration into global markets (MacDonald et al. 2000). In Note that the additional income is often not enough to addition to subsidies that will help sustain the socio-ecological hire extra personnel and/or to provide attractive salaries benefits of extensive sheep grazing systems in mountain areas, for qualified people. These requirements may discourage measures are needed to attract new generations of sheep new generations from livestock farming (Benoit et al. farmers (e.g., improved quality of rural life and working con- 2019). As in other Spanish regions, there are few young ditions), enhance short supply chains, and improve the mar- farmers in Aralar. The average age of Aralar farmers is 51 keting and recognition of value-added mountain products years, and only 11% of the shepherds aged 60 years and (e.g., through specific labeling). older have ensured generational replacement. The high dependency on subsidies, especially in the case of milk sellers, makes the economic viability of Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s13593-021-00751-7. mountain sheep grazing systems vulnerable to external changes. Without subsidies, the number of flocks might Authors’ contributions Conceptualization, E.G, EL.G, A.A, A.P, G.G, be reduced. Even so, the dependency of sheep grazing O.G, and G.P; methodology, E.G, A.A, A.P, G.G, and G.P; validation, systems on subsidies is much lower than that for horse E.G, A.A, A.P, G.G, G.P, N.A, and LJR.B; formal analysis, E.G, A.A, (mares) or dairy/meat production systems, whose subsi- AT.A, G.G, O.G, N.A, and LJR.B; investigation, E.G, A.A, AT.A, LJR.B, and O.G; writing (original draft), E.G, EL.G, A.A, and LJR.B; dies exceed in both cases their net income. Considering writing (review and editing), E.G, A.A, and LJR.B; funding acquisition, social preferences for multifunctional agricultural land- E.G, A.A, and LJR.B; project administration, E.G, A.A, and LJR.B. scapes and the multiple benefits of mountain sheep graz- ing systems documented in this and other studies (e.g., Funding Open Access funding provided thanks to the CRUE-CSIC Dumont et al. 2019), subsidies may be justified. At the agreement with Springer Nature. Financial support was provided from the Spanish Ministry of Economy, Industry and Competitiveness European level, payments for ecosystem services are (AGL2013-48361-C2-1-R), Spanish Ministry of Science and gaining popularity as a way to recognize good livestock Innovation (PID2020-113395RB-C21), and Basque Government to E.G management practices that support high natural value (IT1022-16), A.A (IT1365-19), and L.J.R.B. (IT944-16). E.G, A.P, and systems benefiting the entire society. To promote socio- G.P research was supported by María de Maeztu excellence accreditation 2018–2022 (Ref. MDM-2017-0714), funded by MCIN/AEI/10.13039/ ecological benefits, however, subsidies should ensure 501100011033/, and by the Basque Government through the BERC farmers engage in sustainable practices (Navarro and 2018-2021 program. N.A. and A.P were also financed by the program López-Bao 2019). This requires active management plan- Ramon & Cajal from the Spanish Ministry of Economy, Industry and ning, the spatial distribution of grazing livestock, or re- Competitiveness (RYC-2011-08593; RYC-2017-22143). storing abandoned grazing lands to avoid under-grazed Data availability All data generated or analyzed during this study are areas. included in this published article and its supplementary information files. 47 Page 12 of 13 Agron. Sustain. Dev. (2022) 42:47 Bernués A, Ruiz R, Olaizola A, Villalba D, Casasús I (2011) Code availability Not applicable. Sustainability of pasture-based livestock farming systems in the European Mediterranean context: synergies and trade-offs. Livest. Sci. 139:44–57. https://doi.org/10.1016/j.livsci.2011.03.018 Declarations Bond WJ, Stevens N, Midgley GF, Lehmann CER (2019) The trouble with trees: afforestation plans for Africa. Trends in Ecology & Ethics approval Not applicable. Evolution 34:963–965. https://doi.org/10.1016/j.tree.2019.08.003 Botreau R, Farruggia A, Martin B, Pomiès D, Dumont B (2014) Towards Consent to participate Not applicable. an agroecological assessment of dairy systems: a proposal for a set of criteria suited to mountain farming. Animal 8:1349–1360 Consent for publication Not applicable. Bravo-Lamas L, Aldai N, Kramer JK, Barron LJR (2018) Case study using commercial dairy sheep flocks: comparison of the fat nutri- Conflict of interests The authors declare no competing interests. tional quality of milk produced in mountain and valley farms. LWT- Food Science and Technology 89:374–380. https://doi.org/10.1016/ j.lwt.2017.11.004 Open Access This article is licensed under a Creative Commons Cornelissen JHC, Lavorel S, Garnier E, Díaz S, Buchmann N, Gurvich Attribution 4.0 International License, which permits use, sharing, adap- DE, Reich PB, ter Steege H, Morgan HD, van der MGA H, Pausas tation, distribution and reproduction in any medium or format, as long as JG, Poorter H (2003) A handbook of protocols for standardised and you give appropriate credit to the original author(s) and the source, pro- easy measurement of plant functional traits worldwide. Australian vide a link to the Creative Commons licence, and indicate if changes were Journal of Botany 51(4):335–380. https://doi.org/10.1071/BT02124 made. The images or other third party material in this article are included Correa J, Postma JA, Watt M, Wojciechowski T (2019) Soil compaction in the article's Creative Commons licence, unless indicated otherwise in a and the architectural plasticity of root systems (review paper). J Exp credit line to the material. If material is not included in the article's Bot 70(21):6019–6034. https://doi.org/10.1093/jxb/erz383 Creative Commons licence and your intended use is not permitted by Deléglise C, Loucougaray G, Alard D (2011) Spatial patterns of species statutory regulation or exceeds the permitted use, you will need to obtain and plant traits in response to 20 years of grazing exclusion in sub- permission directly from the copyright holder. To view a copy of this alpine grassland communities. J Veg Sci 22:402–413. https://doi. licence, visit http://creativecommons.org/licenses/by/4.0/ . org/10.1111/j.1654-1103.2011.01277.x Dumont B, Ryschawy J, Duru M, Benoit M, Chatellier V, Delaby L, Donnars C, Dupraz P, Lemauviel-Lavenant S, Méda B, Vollet D, Sabatier R (2019) Review: associations among goods, impacts and References ecosystem services provided by livestock farming. Animal 13(8): 1773–1784. https://doi.org/10.1017/s1751731118002586 Aldai N, de Renobales M, Barron LJR, Kramer JKG (2013) What are the Ekanem E, Mafuyai M, Clardy A (2016) Economic importance of local trans fatty acids issues in foods after discontinuation of industrially food markets: evidence from the literature. Journal of Food produced trans fats? Ruminant products, vegetable oils, and synthet- Distribution Research 47(856-2016-58209):57–64. https://doi.org/ ic supplements. Eur J Lipid Sci Technol 115:1378–1401. https://doi. 10.22004/ag.econ.232302 org/10.1002/ejlt.201300072 Epelde L, Lanzén A, Mijangos I, Sarrionandia E, Anza M, Garbisu C Aldezabal A, Moragues L, Odriozola I, Mijangos I (2015) Impact of (2017) Short-term effects of nongrazing on plants, soil biota and grazing abandonment on plant and soil microbial communities in aboveground-belowground links in Atlantic mountain grasslands. an Atlantic mountain grassland. Appl Soil Ecol 96:251–260. https:// Sci Rep 7. https://doi.org/10.1038/s41598-017-15345-1 doi.org/10.1016/j.apsoil.2015.08.013 Falchero L, Lombardi G, Gorlier A, Lonati M, Odoardi M, Cavallero A Aldezabal A, Pérez-López U, Laskurain NA, Odriozola I (2019) Grazing (2010) Variation in fatty acid composition of milk and cheese from abandonment negatively affects forage quality in Iberian Atlantic cows grazed on two alpine pastures. Dairy Sci Technol 90:657–675. grasslands. Pirineos 174:e042. https://doi.org/10.3989/pirineos. https://doi.org/10.1051/dst/2010035 2019.174002 FAO (2018) Shaping the future of livestock. In: 10th Global Forum for Amores G, Pérez-Elortondo FJ, Albisu M, Barron LJR (2021) To what Food and Agriculture (GFFA), Berlin, pp 18–20 extent do environmental or technological conditions affect the sen- Gésan-Guiziou G, Alaphilippe A, Aubin J, Bockstaller C, Boutrou R, sory differentiation of raw ewe milk cheeses produced in valley or Buche P, Collet C, Girard A, Martinet V, Membré JM, Sabbadin mountain farms? J. Dairy Sci 104:301–307. https://doi.org/10.3168/ R, Thiollet-Scholtus M, van der Werf HMG (2020) Diversity and jds.2020-18358 potentiality of multi-criteria decision analysis methods for agri-food Andonegi A, Garmendia E, Aldezabal A (2021) Social multi-criteria research. Agron. Sustain. Dev. 40:44. https://doi.org/10.1007/ evaluation for managing biodiversity conservation conflicts. Land s13593-020-00650-3 Use Policy 109:105658. https://doi.org/10.1016/j.landusepol.2021. Giampietro M (2004) Multi-scale integrated analysis of agroecosystems. 105658 CRC Press, Boca Raton, FL, USA AOAC (1990) Official methods of analysis, 15th edn. Association of Harwatt H, Ripple WJ, Chaudhary A, Betts MG, Hayek MN (2020) Official Analytical Chemist, Washington D.C. USA Scientists call for renewed Paris pledges to transform agriculture. Batalla I, Knudsen MT, Mogensen L, Hierro Ó, Pinto M, Hermansen JE The Lancet Planetary Health 4:e9–e10 (2015) Carbon footprint of milk from sheep farming systems in Hayek MN, Harwatt H, Ripple WJ, Mueller ND (2021) The carbon Northern Spain including soil carbon sequestration in grasslands. opportunity cost of animal-sourced food production on land. Nat Journal of Cleaner Production 104:121–129. https://doi.org/10. Sustain 4:21–24. https://doi.org/10.1038/s41893-020-00603-4 1016/j.jclepro.2015.05.043 Herrero M, Havlik P, Valin H, Notenbaert A, Rufino MC, Thornton PK, Benoit M, Sabatier R, Lasseur J, Creighton P, Dumont B (2019) Blummel M, Weiss F, Grace D, Obersteiner M (2013) Biomass use, Optimising economic and environmental performances of sheep- production, feed efficiencies, and greenhouse gas emissions from meat farms does not fully fit with the meat industry demands. Agron global livestock systems. Proc Natl Acad Sci 110:20888–20893. Sustain Dev 39:40. https://doi.org/10.1007/s13593-019-0588-9 https://doi.org/10.1073/pnas.1308149110 Agron. Sustain. Dev. (2022) 42:47 Page 13 of 13 47 Hostiou N, Vollet D, Benoit M, Delfosse C (2020) Employment and Renting H, Marsden TK, Banks J (2003) Understanding alternative food farmers’ work in European ruminant livestock farms: a review. networks: exploring the role of short food supply chains in rural devel- Journal of Rural Studies 74:223–234. https://doi.org/10.1016/j. opment. Environ. Plan A. 35(3):393–411. https://doi.org/10.1068/a3510 jrurstud.2020.01.008 Revello Chion A, Tabacco E, Giaccone D, Peiretti PG, Battelli G, Lee JM, Woodward SL, Waghorn GC, Clark DA (2004) Methane emis- Borreani G (2010) Variation of fatty acid and terpene profiles in sions by dairy cows fed increasing proportions of white clover mountain milk and “Toma piemontese” cheese as affected by diet (Trifolium repens) in pasture. In: Proceedings of the New Zealand composition in different seasons. Food Chem 121:393–399. https:// Grassland Association, vol 66, pp 151–155. https://doi.org/10. doi.org/10.1016/j.foodchem.2009.12.048 33584/jnzg.2004.66.2552 Ripoll-Bosch R, Díez-Unquera B, Ruiz R, Villalba D, Molina E, Joy M, Lomba A, Strohbach M, Jerrentrup JS, Dauber J, Klimek S, McCraken DI Olaizola A, Bernués A (2012) An integrated sustainability assessment (2017) Making the best of both worlds: can high resolution agricul- of Mediterranean sheep farms with different degrees of intensification. tural administrative data support the assessment of high nature value Agric Syst 105:46–56. https://doi.org/10.1016/j.agsy.2011.10.003 farmlands across Europe? Ecol Indic. 72:118–130. https://doi.org/ Sayre NF, McAllister RR, Bestelmeyer BT, Moritz M, Turner MD (2013) 10.1016/j.ecolind.2016.08.008 Earth Stewardship of rangelands: coping with ecological, economic, MacDonald D, Crabtree JR, Wiesinger G, Dax T, Stamou N, Fleury P, and political marginality. Front Ecol Environ 11:348–354. https:// Lazpita JG, Gibon A (2000) Agricultural abandonment in mountain doi.org/10.1890/120333 areas of Europe: environmental consequences and policy response. J Singh RK, Murty HR, Gupta SK, Dikshit AK (2012) An overview of Environ Manage. 59:47–69. https://doi.org/10.1006/jema.1999.0335 sustainability assessment methodologies. Ecol Indic 15:281–299. Madelrieux S, Bergeret A, Fillion L (2018) Forms of territorial https://doi.org/10.1016/j.ecolind.2011.01.007 embeddedness in dairy value chains Case of the Chartreuse massif Ulbrich TLV, Southgate DAT (1991) Coronary heart disease: seven die- (French Alps): geographical and historical perspectives. Open Agric tary factors. Lancet 8773:985–992. https://doi.org/10.1016/0140- 3:618–631. https://doi.org/10.1515/opag-2018-0065 6736(91)91846-M Malagon-Zaldua E, Begiristain-Zubillaga M, Onederra-Aramendi A (2018) Measuring the economic impact of farmers’ markets on local Valdivielso I, Bustamante MA, Buccioni A, Franci O, Ruiz de Gordoa economies in the Basque Country. Agriculture 8(1):10. https://doi. JC, de Renobales M, Barron LJR (2015) Commercial sheep flocks – org/10.3390/agriculture8010010 fatty acid and fat-soluble antioxidant composition of milk and Manzano P, White SR (2019) Intensifying pastoralism may not reduce cheese related to changes in feeding management throughout lacta- greenhouse gas emissions: wildlife-dominated landscape scenarios tion. J Dairy Res 82:334–343. https://doi.org/10.1017/ as a baseline in life-cycle analysis. Climate Research 77:91–97. S0022029915000369 https://doi.org/10.3354/cr01555 Valdivielso I, Albisu M, de Renobales M, Barron LJR (2016a) Changes Meuret M, Provenza F (2015) How French shepherds create meal se- in the volatile composition and sensory properties of cheeses made quences to stimulate intake and optimise use of forage diversity on with milk from commercial sheep flocks managed indoors, part-time rangeland. Animal Production Science 55:309–318. https://doi.org/ grazing in valley, and extensive mountain grazing. Int Dairy J 53: 10.1071/AN14415 29–36. https://doi.org/10.1016/j.idairyj.2015.09.007 Navarro A, López-Bao JV (2018) Towards a greener Common Agricultural Valdivielso I, Bustamante MA, Aldezabal A, Amores G, Virto M, Ruiz Policy. Nat Ecol Evol 2:1830–1833. https://doi.org/10.1038/s41559- de Gordoa JC, de Renobales M, Barron LJR (2016b) Case study of a 018-0724-y commercial sheep flock under extensive mountain grazing: pasture Navarro A, López-Bao JV (2019) EU agricultural policy still not green. derived lipid compounds in milk and cheese. Food Chem 197:622– Nat Sustain 2. https://doi.org/10.1038/s41893-019-0424-x 633. https://doi.org/10.1016/j.foodchem.2015.10.133 Nuñez MA, Davis KT, Dimarco RD, Peltzer DA, Paritsis J, Maxwell BD, Valdivielso I, de Renobales M, Aldai N, Barron LJR (2017) Changes in Pauchard A (2021) Should tree invasions be used in treeless ecosys- terpenoid composition of milk and cheese from commercial sheep tems to mitigate climate change? Frontiers in Ecology and the flocks associated with seasonal feeding regimens throughout lactation. Environment 19:334–341. https://doi.org/10.1002/fee.2346 J Dairy Sci 100:96–105. https://doi.org/10.3168/jds.2016-11761 Odriozola I, García-Baquero G, Laskurain NA, Aldezabal A (2014) Vance ED, Brookes PC, Jenkinson DS (1987) An extraction method for Livestock grazing modifies the effect of environmental factors on soil measuring soil microbial biomass C. Soil Biol Biochem 19:703– temperature and water content in a temperate grassland. Geoderma 235– 707. https://doi.org/10.1016/0038-0717(87)90052-6 236:347–354. https://doi.org/10.1016/j.geoderma.2014.08.002 Virto M, Bustamante MA, Ruiz de Gordoa JC, Amores G, Fernández- Odriozola I, García-Baquero G, Etxeberria A, Aldezabal A (2017a) Caballero PN, Mandaluniz N, Arranz J, Nájera AI, Albisu M, Pérez- Patterns of species relatedness created by competitive exclusion de- Elortondo FJ, Barron LJR, de Renobales M (2012) Interannual and pend on species niche differences: evidence from Iberian Atlantic geographical reproducibility of the nutritional quality of milk fat grasslands. Perspect Plant Ecol Evol Syst 28:36–46. https://doi.org/ from commercial grazing flocks. J Dairy Res 79:485–494. https:// 10.1016/j.ppees.2017.07.002 doi.org/10.1017/S0022029912000490 Odriozola I, García-Baquero G, Fortin MJ, Laskurain NA, Aldezabal A Waghorn GC, Tavendale MH, Woodfield DR (2002) Methanogenesis (2017b) Data from: Grazing exclusion unleashes competitive plant from forages fed to sheep. In: Proceedings of the New Zealand responses in Iberian Atlantic mountain grasslands. Dryad, Dataset Grassland Association, vol 64, pp 167–171. https://doi.org/10. 20:50–61. https://doi.org/10.5061/dryad.232ss 33584/jnzg.2002.64.2462 Philippidis G, Sanjuán López AI, Ferrari E, M'barek R (2014) Employing Wang L, Delgado-Baquerizo M, Wang D, Isbell F, Liu J, Feng C, Liu J, social accounting matrix multipliers to profile the bioeconomy in the Zhong Z, Zhu H, Yuan X, Chang Q, Liu C (2019) Diversifying EU member states: is there a structural pattern? Spanish Journal of livestock promotes multidiversity and multifunctionality in man- Agricultural Research 2014 12(4):913–926. https://doi.org/10.5424/ aged grasslands. PNAS 116(13):6187–6192. https://doi.org/10. sjar/2014124-6192 1073/pnas.1807354116 RECAN (Red Contable Agraria Nacional) (2015) Resultados definitivos. Ministerio de Agricultura Pesca y Alimentación. https://www.mapa. gob.es/es/estadistica/temas/estadisticas-agrarias/ Publisher’s note Springer Nature remains neutral with regard to jurisdic- recan2015resultadosdefinitivos_nacionalyccaa_tcm30-380079.pdf tional claims in published maps and institutional affiliations.