Public perception of ecosystem and social services produced by Sardinia extensive dairy sheep farming systems

Abstract

Dairy sheep farming systems provide a great range of ecosystem services (ESs) and social services (SSs). These are Agro-pastoral Secondary Outputs (ASOs), the promotion of which can help the survival of the systems and the rural regions in which they exist. However, little attention has been paid to understanding which ASOs are recognized by the public, which is the first step to adequately promote them. This study first aims to review previous literature on ASOs relating to livestock in general and dairy sheep farming systems in particular. The literature review, conducted according to the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) framework, revealed significant gaps. Second, the research provides evidence of public perception of ASOs of a given dairy sheep sector—i.e. that developed on the region of Sardinia (Italy)—via a questionnaire distributed to a composite sample of 525 stakeholders. We found that cultural and landscape services are the most appreciated services. Multiple correspondence analysis suggests that appreciation of a specific secondary output would imply the appreciation of all the other outputs. Furthermore, we ran a set of logit regressions where each ASO was related to several socio-economic variables. Findings showed, among others, that the ‘subjective knowledge’ of the Sardinian agro-pastoral reality positively and significantly affects appreciation of all the ASOs. Several implications for practitioners, academics and policymakers are derived from these findings.

Introduction

Ecosystems provide a range of services that significantly impact human well-being, health, livelihoods and survival (MEA 2005; TEEB 2010; Costanza et al. 2014). Several ways of defining ecosystem services (ESs) have been developed, such as ‘the benefits people obtain from ecosystems’ (MEA 2005, p. 26) or as contributions of ecosystem structure and function (in conjunction with other inputs) to human well-being (economic, social and personal well-being) (Burkhard et al. 2012; Burkhard and Maes 2017). In this context, it is important to emphasize that ecosystems provide benefits to people only because of the presence of individuals (human capital), their communities (social capital) and their built environment (built capital). Therefore, the benefits of ESs flow from natural capital to human well-being only through interaction with the other three forms of capital (Costanza et al. 2014).

The Millennium Ecosystem Assessment (MEA) (2005) classified ESs into four categories: provisioning (material or energy outputs), regulating (biophysical processes providing benefits such as climate regulation or water purification), supporting (processes that allow the functioning of other ecosystems that, in turn, provide other services, such as nutrient cycles, soil formation, photosynthesis or pollination) and cultural (recreational, aesthetic and spiritual benefits). Scientists and policymakers use ESs extensively to highlight the importance of the environment in supporting human livelihoods. Therefore, the term is broadly used and ES research has made advances in many areas, from theoretical conceptualization to practical applications (Potschin et al. 2016; La Notte et al. 2017). Numerous studies have investigated how the concept can be used in an agricultural context. In doing so, however, the few ES assessments aimed at decision support appear to be mainly limited to theoretical reflections. Few studies have focused on specific agricultural practices and the gap between ES research and information required to support decisions persists (Holt et al. 2016; Olander et al. 2017; Dendoncker et al. 2018).

However, one cannot ignore the existence of a variety of disservices to and from agriculture (Zhang et al. 2007; Herd-Hoare and Shackleton 2020). Ecosystem disservices (EDs) highlight the negative economic and non-economic effects of nature on human well-being within social–ecological systems (Blanco et al. 2019). The EDs generated by agroecosystems, i.e. “the ecosystem generated functions, processes and attributes that result in perceived or actual negative impacts on human well-being” (Shackleton et al. 2016), in many cases, can lead to impact production and economic losses, compromising food security, income, and thus well-being (Herd-Hoare and Shackleton 2020). The breadth and depth of EDs depend on management practices because, if they are appropriate, can improve and mitigate many agriculture negative impacts (Power 2010). Considering the livestock farming systems, utilization of resources and intensity of the production system can determine disservices; landscaping, hydrology and environmental damage are some examples (FAO 2006; Montrasio et al. 2020).

The present work is focused on the ESs commonly provided by dairy sheep farming systems. Sheep farming represents a traditional activity in many world regions, although the main product (milk, meat, wool, or a combination of the three) may vary depending on the geographical context or the breed raised. The breeding for milk production assumes considerable economic importance in many areas (Pulina et al. 2018). Sheep milk production accounts for about 10.6 Mt, of which 29.5% is from Europe (FAOSTAT 2022), with growth forecasts of around 3 Mt by 2030 (Pulina et al. 2018).

Sheep farming is especially practised in temperate and subtropical areas, concentrated in the Mediterranean and the Black Sea regions and accounting for approximately 211 million sheep (FAOSTAT 2022). In the Mediterranean areas, sheep breeding systems for milk production have been traditionally developed in economically and physically disadvantaged areas (Ripoll-Bosch et al. 2012) and strongly affect the agricultural economy of several regions, where it is often practised using a low-input system. It is important to note that, in general, low-input farming systems are associated with many agricultural practices that provide a greater range of ESs than intensive farming systems (Cooper et al. 2009). Dairy sheep farming is no exception to such rule. Furthermore, the sheep farming systems also produce other externalities of public relevance in terms of social benefits. In other words, production of social services (SSs) in connection with ESs can be recognized from these practices; one example is the role of pastoralism in safeguarding the livelihoods of populations located in disadvantaged and/or rural areas, by ensuring jobs or preventing depopulation. These are public functions that can counteract elements that strongly affect the economic marginalization of rural areas and from which many adverse social, economic and environmental effects derive (O’Rourke et al. 2016; Nori et al. 2017; Quaranta et al. 2020). Pastoralism is commonly recognized as having a multifunctional dimension as it produces a wide range of environmental and social goods and services alongside food (Ripoll-Bosch et al. 2012; Nori et al. 2017; Bernués et al. 2019), thereby providing a valuable opportunity for sustainable development (Meloni et al. 2015). Overall, these externalities can be considered as Agro-pastoral Secondary Outputs (ASOs).

A significant research challenge concerns the evaluation of ASOs produced by dairy sheep farming systems, since ‘the true contribution of agro-pastoralism to societal objectives is not fully accounted for, which provides an explanation for its decreasing role on Mediterranean islands over time’ (Nori et al. 2017, p. 138). The ASOs are not properly valued and priced through market mechanisms, which means that these externalities are ‘lost’ across the value chain and often unrecognized by the final user (Nori et al. 2017). To that and, it is necessary to understand how the public perceives the relationships between such agriculture practices and ASOs and identify those that are the most requested (Bernués et al. 2016, 2019). However, a necessary condition is that the population are aware of such agro-pastoral outputs and appreciate them (Montrasio et al. 2020). The assessment of the values and perceptions of ASOs is of great importance because it is the first step towards developing a shared understanding of the agro-landscape, support change paths in terms of sustainability and understand how agricultural practices affect ASOs that, in its turn, influence agricultural productivity and society (Dendoncker et al. 2018).

In the context described above, this paper evaluates the perception of ASOs generated by the dairy sheep farming system in Sardinia, the second-largest Mediterranean island. Sardinia is one of the main European producers of sheep's milk and contributes to about 67% of Italian dairy sheep production (Autonomous Region of Sardinia 2020) (Italy accounts for 19% of total European milk production—FAOSTAT 2022). Sardinian dairy sheep breeding is usually pasture-based and quite extensive (Vagnoni et al. 2015; Vagnoni and Franca 2018). It uses traditional and low-input practices and maintains the semi-natural habitats they helped create (Ripoll-Bosch et al. 2013; Pulina et al. 2018; Faccioni et al. 2019). Moreover, the Sardinian shepherding model has been at the basis of local cultural identity, provides landscape maintenance and care, contributes to the protection of territory and animal and vegetable biodiversity, and ensures jobs for thousands of people (Furesi et al. 2013; Mattalia et al. 2020).

Despite the growing appreciation of primary and processed sheep dairy products (Nori et al. 2017, p. 136), and its role in the regional economy, the dairy sheep sector operates with low profit margins (Idda et al. 2010). Currently, its profitability often depends on the amount of financial aid made available by the Common Agricultural Policy (CAP) (Idda et al. 2010; Vagnoni et al. 2015; O’Rourke et al. 2016; Quaranta et al. 2020). Sardinian dairy sheep sector is facing a serious threat to its medium- to long-term survival since if the business is scarcely profitable, current and future generations of stock farmers would not be encouraged to continue operating. The resulting impacts could be severe on most social, economic and environmental benefits people obtain from the agro-pastoralism contribution to societal objectives. The scenario presented here could worsen a critical situation already in place because Sardinia is one of the European Union’s (EU) most underdeveloped areas and suffers from a lower Gross Domestic Product (GDP) and industrialization rate than the peninsula (ISMEA 2019). Consequently, it appears to be pivotal to foster sector durability, here understood as the ability to cope with chronic endogenous stress (Dawson et al. 2010), through the sustainability defined by Hodge as ‘the persistence over an apparently indefinite future of certain necessary and desired characteristics of both the ecosystem and the human subsystem within’ (Hodge 1997, p. 9).

The present paper focuses on this area of interest and has two aims. First, it aims to explore the state of knowledge of ESs and SSs related to both zootechnical activities on the whole and to dairy sheep farming systems specifically. A qualitative systematic review of world literature was performed using the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) framework. Second, it aims to evaluate the most recognized ASOs of the extensive Sardinian dairy sheep system. Building on qualitative systematic review findings, the ASOs were identified, and a qualitative survey of residents and non-residents was carried out to ascertain perceptions of the benefits of ASOs. Data were first analysed with a descriptive analysis, followed by a multiple correspondence analysis (MCA) to cluster perceptions of the relevance of ASOs. Finally, the relationship between preferences stated for each secondary output and sociodemographic variables was evaluated using a set multinomial logit regression (MLR).

The paper makes several contributions to the literature, for policymakers and stock farmers. First, as far as we know, this is the first economic study that considers both environmental and social secondary outputs provided by an extensive sheep farming system. Second, this paper helps to bridge the gap both between the growing scientific research on perceptions and values of ESs and the few studies on ES perception concerning specific agricultural practices, and between ES research and information required to support decisions (Bernués et al. 2016; Olander et al. 2017; Dendoncker et al. 2018). It does so by investigating the ESs provided by the Sardinian dairy sheep farming system and overcoming the limitations of previous research by including the assessment of perception of SSs. Third, it addresses recent calls for research on the complex relationship between pastoral activity and rural landscape in the Mediterranean region (Mattalia et al. 2020). As the paper highlights which socio-economic, cultural and environmental outputs of the Sardinian dairy sheep sector are most appreciated, it raises awareness of the agro-pastoral contribution to the pattern of sustainable development in the Sardinian region. Fourth, the paper highlights the appreciation of ASOs and recognizes the existence of supply and demand for a public good that is inadequately remunerated in the market. As a result, the financial aid provided by the CAP as a welfare contribution non-configurable and can make stock farmers aware of their role in maintaining territory and producing public goods.

The paper is structured as follows: Sect. 2 describes the research methodology and sample; Sect. 3 presents the results; Sect. 4 discusses the results and provides a conclusion, outlining the implications for practitioners, academics and policymakers, and recommendations for future research.

Methodology

In order to achieve the aims of this paper, two-step research methodologies were used: a qualitative systematic review on the secondary outputs produced in breeding and an empirical analysis—by submitting a structured questionnaire—aimed to evaluate the perceptions of secondary outputs in extensive dairy system by part of society. Figure 1 provides a visual representation of the aims and methods used in the different stages of the research and the output of the first phase used in the second one.

Fig. 1

Graphical representation of the research model

Systematic review

To explore the first aim, we conducted a qualitative systematic review. Quantitative reviews usually produce meta-analyses and find applications in the case of selected studies that share a highly standardized procedure, centred on similar topics, analyses or research designs, allowing for a direct comparison of results that can be contextualized in terms of effect size (Borenstein et al. 2009). The available scientific literature on the ESs and SSs of zootechnical activities does not fulfil all these conditions. A qualitative literature review based on narrative surveys overcomes problems arising from the wide differences in the literature in terms of methods and results and the relatively limited number of relevant studies available (Baumeister and Leary 1997; Jahan et al. 2016).

We followed one of the approaches to qualitative survey proposed by Baumeister and Leary (1997) and drew up a survey aiming to explore the state of knowledge of the ASO-related aspects of livestock activities. While the literature on ESs in a broad sense has developed in recent years, the same does not apply to zootechnical activities. As our aim was to investigate if sheep farming systems can survive giving value to positive externalities, our review addressed the following research questions:

• 1. Has previous literature analysed the population's perception towards ASOs in the context of agroecosystems?

• 2. Is there evidence for which ASOs provided by the primary sector are most appreciated?

• 3. Has the previous literature provided an economic evaluation of primary sector ASOs, especially concerning breeding systems?

• 4. Are there any references to extensive dairy sheep farming systems?

The review focused on scientific, English-language, peer-reviewed publications after 2010. We used the most common online database, Scopus (Moher et al. 2009), which has previously been used across a wide range of studies because of its uniqueness, its reputable publisher (Elsevier Co.) and because it indexes over 14,000 journals in numerous fields, including social, environmental and economic. The literature search considered all countries and was performed on 8th March 2021. The PRISMA framework procedure (Moher et al. 2009) was adapted for our purposes. The PRISMA statement highlights four steps: identification, screening, eligibility and inclusion criteria.

Concerning the keyword identification process to retrieve as many articles as possible in the target research fields, data were acquired by searching publications in agriculture, business and economy areas, whose title, abstract or keywords contain the following words: (“ecosystem service*” OR “agroecosystem*” OR “multifunctional*” OR “agroecosystem service*” OR “socio-economical system*”) AND (“economic valuation” OR “socio-cultural valuation” OR “perception” OR “socio-economic valuation” OR “social preferences” OR “stakeholder involvement” OR “service value” OR “ecosystem services value” OR “agroecosystem*” OR “agrienvironmental payment*” OR “agri-environmental payment*” OR “choice experiment” OR “survey” OR “agrienvironmental polic*” OR “agri-environmental polic*”) AND ("agriculture” OR “farm*” OR “livestock” OR “landscape” OR “silvopasture” OR “pastoralism”). The truncation symbol “*” covers the variations of the selected terms.

After the screening process to include or exclude based on the criteria previously defined, the eligibility step involved two researchers who examined titles, abstract, methods, results and discussion to verify the articles’ relevance to the study’s research questions and compliance with the inclusion criteria. A third researcher conducted an independent evaluation in case of doubts about inclusion criteria satisfaction. Articles that fulfilled all requirements were analysed.

Clustering of Sardinian extensive dairy sheep system secondary outputs

To investigate the societal awareness and appreciation of social and environmental benefits obtained by Sardinian agro-pastoralism externalities, we first specified a list of ASOs. Based on the findings of a previous systematic qualitative review, a list of ESs relating to other livestock realities suitable to the Sardinian sector was identified and classified as ‘regulating’, ‘supporting’ and ‘cultural’. The ‘provisioning’ services were excluded since they were outside the research scope. The list of externalities of SSs counts the outputs referring to the means of subsistence provided in inland areas (e.g. employment for indigenous people and immigrants), the avoidance of depopulation and the provision of social agriculture activities. Twenty secondary outputs were identified and used to answer our second research question (see Table 4 of Appendix).

A structured questionnaire was created in Google Forms in the Italian language and disseminated across social networks. Online surveys are increasingly used to collect data for research purposes, as they are cheap, fast and efficient in collecting valid data, removing geographic limitations and costs (Wright 2005; Vecchio et al. 2020). Recruitment was based on non-random criteria and influenced by the use of social networks, creating a potential selection/sampling bias, as respondents’ beliefs, interests and strength of feelings about a topic can influence their willingness to participate in a survey (McAleese et al. 2016; Burruss and Johnson 2021). However, since no respondents received incentives for participating in this study, the voluntary response sample allowed us to investigate who has a strong opinion (positive, negative or neutral) about ASOs, therefore understanding genuine appreciation for ASOs.

The form was sent to a sample of Italian people, with a sub-category represented by Sardinian residents and was in the field for 1 month (18 March–18 April 2021). By the closing date for responses, 525 answers were recorded, all valid as the questionnaire could only be sent if all questions were answered. According to Kline (2015), there should be at least 10 valid responses per parameter. Therefore, given the numbers of the final sample (525) and the items (20), the research meets the above prior condition.

The questionnaire comprised three parts. After being welcomed into the study and reassured about the reliable and anonymous use of the data provided, respondents were first given a brief overview of the economic role of sheep farming in Sardinia, as well as how its environmental, social and cultural functions—if preserved and well-valued—can contribute to guaranteeing not only the sustainability of the sheep system but also generating well-being in the populations and the vitality of the territories concerned.

The second part collected the sociodemographic data of respondents. The third aimed to investigate two constructs that influence people's perception: past professional or family experience and subject knowledge. The latter reflects the self-confidence that an individual has in the adequacy of their knowledge (Brucks 1985; Carlson et al. 2009), and within the environmental behaviour context, is of substantial importance for the individual’s formation of perceptions and behaviours (Kaiser and Fuhrer 2003; Frick et al. 2004).

In the last part, respondents were asked to express their appreciation for the ASOs of the Sardinian extensive sheep sector, indicating what level of benefit (none, moderate or high) they recognize from each of them. In other words, respondents valued ASOs based on their individual subjective (group-, time- and location-dependent) assessment (Spangenberg et al. 2014; Spangenberg and Settele 2016).

Data were analysed with three methods. First, descriptive statistics were calculated to summarize the data and provide background information about the variables in the data set. Afterwards, based on the perception of the relevance to the community of the secondary outputs, we clustered preferences using MCA. In so doing will be possible to highlight the most recognized agro-pastoralism ESs and SSs and, especially, to estimate the connections among agro-pastoralism amenities. MCA is an exploratory, multivariate analysis technique considered particularly suitable for categorical data description (Greenacre 1984, 2017; Le Roux and Rouanet 2004; Greenacre and Blasius 2006). It is a technique of interdependence used for dimension reduction in qualitative variables (Benzécri 1992), considered as a multivariate extension of simple correspondence analysis and principal component analysis between quantitative variables, enabling the visualization of perceptual maps.

We used an MCA biplot (Greenacre 2010) to visualize the pattern of relationships of several categorical dependent variables in a two-dimensional plot. An orthogonal rotation of the principal axes was performed to improve the interpretation. MCA is obtained using a standard correspondence analysis on an indicator matrix (X). It is a J × M matrix where J_k is the vector of the levels (three levels: null, moderate, elevate attributed importance) for each K nominal variable (with $\sum {J_{k} = J}$ and where k = 20 nominal variables represented by the considered secondary outputs), and M is the number of observations (525 respondents).

Performing MCA on X will provide two sets of factor scores, one for the rows and one for the columns. These factor scores are, in general, scaled such that their variance is equal to their corresponding eigenvalue.

The analysis should allow us to provide evidence for the relationship between the 20 individuated ASOs by a representation in a low-dimensional space—designed based on few principal components—in order to define some clusters (profiles).

Finally, we evaluated the relationship between preferences stated for each ASO and the sociodemographic variables using an MLR model set. MLR is a highly efficient probability estimation method that can be used to address binary, ordinary or multinomial problems. It is an extension of binary logistic regression that can be used to forecast the probabilities of the different possible outcomes of a multi-way categorical dependent variable (in our case, the variable can assume three levels) given a set of independent variables that can be either dichotomous (i.e. binary) or continuous (i.e. interval or ratio in scale). Using MLR, we can see the influence of sociodemographic variables in determining trends in appreciation of ASOs arising from the Sardinian extensive dairy sheep system. Specifically, each ASO was regressed on the set of variables reported in Table 1.

Table 1 List of the independent socio-economic variables involved in the regressions

Results

Qualitative literature review

Using the PRISMA framework procedure (Moher et al. 2009), 12 papers were considered relevant to our research questions. Figure 2 shows the flow diagram of the selection of relevant records.

Fig. 2

Systematic review of ecosystem services using PRISMA in Scopus database

Table 5 of the Appendix reports a summary of the results for the selected studies according to the research questions. Findings show that only five studies have previously examined the population’s perception towards some ESs in the context of agroecosystems (RQ1), only three provided early indications on preferences among some ESs provided by the primary sector (RQ2), and only four studies provided an economic evaluation of primary sector ecosystem output (RQ3). Four out of 12 articles are attributable to Bernués A. and Ripoll-Bosch R. Four papers have been published in the Land Use Policy journal and three in Ecosystem Services. The main results are in the literature gaps detected. First, none of these studies investigated the social benefit generated from breeding in general. Second, no previous research focused on the secondary output of extensive dairy sheep farming systems on the whole. Third, there is a lack of studies on the economic benefits generated by extensive dairy sheep farming systems. These literature gaps validated the opportunity and importance of the second part of our research.

Sample profile

The 525 Italian respondents, mainly residents, were 48.8% males and 51.2% females; 32.4% were aged between 50 and 59 years old, 24% were under 40 years old, about 23% were aged between 40 and 49 years old, and 7.1% were over 60 years old. Only 5.5% had a lower secondary school education, whereas 23.1% had an intermediate education level, 44.8% had a university degree and 26.7% had a postgraduate degree.

These percentages may result from the voluntary response sample. Regarding occupations, most respondents had a desk job (24%), 15.4% were students, 12.8% were shepherds and 1.7% were farmers. Almost half of the sample stated that they had previously been employed in the agricultural or livestock sector (45.9%) and come from a family with a peasant tradition (54.3%). The subject knowledge was high, since on a scale of one to seven, 65.1% rated themselves at five and above. The purchase frequency of Sardinian products was also very high, as 38.1% and 39.6% declared that they consume Sardinian products at least once a day or once a week, respectively. The data are reported in Table 2.

Table 2 Profile of the sample

Descriptive analysis

The average of the responses for each ASO showed a moderate recognition of the externalities generated by the Sardinian extensive dairy sheep farming system (Table 3). The fact that the standard deviation was slight compared to the mean value demonstrated that the data were concentrated around the mean.

Table 3 Descriptive analysis of ASOS's perception level by residents and non-residents

The tendency to recognize a moderate level of benefit generation from ASOs arose also when analysing the frequency, with two exceptions: the culture and landscape (enhancing cultural identity of Sardinia and safeguard of typical landscape). Most respondents recognized that extensive dairy sheep farming generates high benefits in terms of these two secondary outputs.

Previous studies found differences between residents and non-residents concerning the perception of ESs (López-Santiago et al. 2014; Bidegain et al. 2019; Montrasio et al. 2020). As our sample mainly included residents, we wanted to investigate the possible differences between the respondents. The results showed that non-residents had a higher perception of ASOs than residents. In fact, in addition to the ‘cultural’ and ‘landscape’ ESs, non-residents demonstrated a high perception of the other six ESs (‘biodiversity’, ‘habitats’, ‘fires’, ‘insects’, ‘invasive’, ‘arts’) and also of three SSs (‘territory’, ‘employment of local inhabitants’, ‘employment of immigrants’). It is striking that residents demonstrated a moderate perception of opportunities for social activities (i.e. social farming), which represents a form of resilience and innovation within rural systems and a tool to support the competitiveness of the production system (Nicolosi et al. 2021), and of employment of the foreign workforce, which has a growing relevance, especially in rural and marginal areas (Marongiu 2021).

Multiple correspondence analysis

Two main dimensions were detected to represent the results within the Cartesian plane. Figure 3 shows the plot of component loadings for the survey data. Dimension one showed a Cronbach’s alpha of 0.910 and explained 37% of the total variance, while dimension two showed a Cronbach’s alpha of 0.69 and explained 15.5% of the total variance. Overall, about 50% of the variability was explained by the first two-component model.

Fig. 3

Multiple correspondence analysis biplots

The biplot allows us to investigate how well variables separate groups of objects. Findings indicated that observations were clustered based on the degree of relevance attributed by respondents. By examining the graph, it emerged that the first dimension (the horizontal axis) separated respondents who gave a moderate judgement of ASOs from those who gave more polarized opinions, i.e. null or elevated. Furthermore, ‘biodiversity’ was located much further from its origin than other variables, suggesting that taken as a whole, the others did not share many characteristics of it.

The second dimension (the vertical axis) separates the respondents who did not perceive a contribution from the Sardinian extensive dairy sheep farming system to the ASOs from those who detected a high contribution. The graph shows that the first and second dimensions involved perfect discrimination between the variables. Therefore, the overall picture shows that recognizing the importance of one service involves recognizing the importance of all services provided by the Sardinian system. It was therefore concluded that multifunctionality as a whole was recognized.

Set of logit analysis

Before running a set of MLRs, the multicollinearity test among sociodemographic variables was performed. All 10 sociodemographic variables reported a variance inflation factor (VIF) of less than 2 and a tolerance greater than 0.5. Therefore, there is no risk of multicollinearity or independent variables that are highly correlated and provide the same information in the present study (Hair et al. 2010; Tabachnick et al. 2013).

Each ASO (a trichotomic variable taken as dependent variables) was regressed to the bundle of sociodemographic variables (taken as independent variables) (see Table 6 of the Appendix). Findings showed that all 20 regression models led to significant results, although not all variables were significant in all regressions. In no case did coming from a family with a peasant tradition affect the perception of the Sardinian system over ASOs. Age and gender were rarely estimated among significant findings. In contrast, others variable, such as education level and regional residence, were often significant, although magnitude and sign can vary among the secondary output considered. Only three sociodemographic variables appeared noteworthy, past professional experience, residence and subject knowledge, which were found to be highly significant and positively correlated with all ASOs.

Discussion

The concept of ESs, as well as the research field, is ingrained in strong sustainability thinking (Jacobs et al. 2013) and provides a framework that unravels the complex feedback loops of how pasture affects ESs flows and how, in turn, these ESs are perceived (Lamarque et al. 2014). The multifunctional character of pasture-based sheep farming systems and thus their economic, environmental and social roles are recognized by policymakers and society (Ripoll-Bosch et al. 2012). Nevertheless, the literature analysis through the PRISMA framework confirmed the need to broaden the research on ESs to meet decision-makers’ and practitioners' information needs (Olander et al. 2017). Despite recognizing the pivotal role of sheep farming systems and their basis of the rural economy of Sardinia (Agris Sardegna 2017), there is a lack of studies on the secondary production of extensive dairy farming systems as a whole. Likewise, research on the economic benefits generated by extensive dairy sheep farming systems is missing. The lack of scientific relevance can only have repercussions on the ability of local, national and community policymakers to make informed decisions based on reliable information.

Moreover, the lack of studies focused on social benefits generated from breeding as a whole is surprising. Indeed, extensive dairy sheep farming systems are essential for the social fabric in rural areas. Due to the need to directly breed the sheep, people must live and work in these areas, avoiding depopulation and guaranteeing the defence of the territory. In addition, the rural decline is a global issue; extensive dairy sheep farming systems can tackle such problems by offering job opportunities to foreign workers, who in turn contribute actively to the process preservation of the above traditional practice (Nori et al. 2017).

Previous research (Ripoll-Bosch et al. 2013; Furesi et al. 2013; Leroy et al. 2018; Mattalia et al. 2020) has shown that the dairy sheep breeding system performs various functions which have given rise to several services and, in turn, benefits, for humans, who give a value of those services (Hansen and Pauleit 2014). This work recognizes that the economic value analysis of ESs is a methodology that has been criticized because of a conceptual controversy surrounding the use of economic approaches applied to ecosystems (e.g. Pimm 1997; Pearce 1998; Toman 1998; Viglizzo et al. 2012). The economic approach is anthropogenic and evaluates ecosystems according to human utility. Accordingly, the concept of ESs outlines the various contributions of ecosystems to human well-being, which are the main target of the analysis (MEA 2005). It is acknowledged in this work that, under the ecological approach, every ecosystem has its own intrinsic value, independent of its benefits for humanity and of what people happen to like, want or need (Sagoff 1997). However, the anthropogenic definition of ESs proposed by the MEA (2005) originated from the aim of making natural capital relevant to society and decision-makers; moreover, making an economic evaluation allows for the comparison of the different benefits received by ESs by measuring them and expressing them in a common denominator, also including ESs not traded on the market and which do not directly produce monetary benefits (Ottaviani 2020).

This work recognizes both the need to give value to externalities of the Sardinian extensive dairy sheep farming system and the anthropocentric nature of the framework of ESs (D’Ottavio et al. 2018) and SSs, which requires the assessment of their perception by stakeholders (Leroy et al. 2018). The assessment of ES perceptions is the first step to develop a shared understanding of the agro-landscape and support change paths in terms of sustainability (Dendoncker et al. 2018). Thus, the interviewees' perceptions of the ASOs were estimated, shedding light on possible paths to foster the durability and sustainable development of the target areas. While respondents generally showed a moderate perception of benefits generated by ASOs, significant findings emerged by analysing the differences among residents and non-residents. In effect, because the ‘subjective values are attributed based on the (individual and social) perception of real-world objects’ (Spangenberg and Settele 2016, p. 102), different groups with different worldviews and different relationships with an ecosystem may recognize diverse potential services (Spangenberg et al. 2014; Bidegain et al. 2019).

Our results showed that the residence of the interviewees influenced the perception of SSs and ESs, confirming previous studies (López-Santiago et al. 2014; Bidegain et al. 2019; Montrasio et al. 2020). Non-residents, compared to residents, gave great value to ASOs, especially to those related to supporting and social services. There may be many explanations for this finding. Overall, residents recognized—more than other ASOs—the importance of pasture for the regional economy and society to enhance Sardinia's cultural identity and safeguard its typical landscape. In effect, shepherding is Sardinia's traditional and stereotypical representation (Ferrari 2009). Pasture has been at the focus of Sardinia's cultural identity for millennia and the agro-pastoral landscape, with its product, is one of the major attractions for tourists (Mattalia et al. 2020). Perhaps, it is the reason why residents placed a high value on the landscape and cultural ‘noticeable’ variables. In contrast, they may take for granted all other positive externalities of pasture, unlike the non-residents who make a conscious and targeted territory-use choice (Montrasio et al. 2020). Moreover, residents could also acknowledge several pasture problems associated with economic and environmental implications, such as out-migration and changes in ecologies because of shifts in grazing patterns (Nori and Scoones 2019). Idda et al. (2006) found a sense of social hostility towards the professional category of shepherds and a negatively oriented mentality towards sheep farming entrepreneurs by a portion of residents who carry out other activities that can overshadow all the other benefits generated by the Sardinian dairy sheep system. It could be that some of the residents may not look beyond these problems, resulting in a lack of knowledge that limits their ability to evaluate ASOs (Gundersen et al. 2017). A lack of information on an ASO may involve its ‘unperception’ and, in turn, render its consideration as not important (Bingham et al. 1995).

Another fact worthy of reflection is the cultural ESs ‘that are usually directly experienced and intuitively appreciated’ (Hermes et al. 2018, p. 296). In addition to the ‘culture’ appreciated by all, and ‘art’ appreciated more by non-residents, it was surprising that there was a moderate value attributed to ‘heritage’ (safeguard of cultural heritage) and ‘recreation’. In reference to the first, we need only think of the ‘canto a tenore’, which developed in the pastoral culture of Sardinia and ‘the paths of transhumance’, keepers of a rich social, cultural and environmental heritage, which the shepherds protect with their own activities. Both of these are forms of heritage inscribed in the United Nations Educational, Scientific and Cultural Organization (UNESCO) lists. The findings may mean that most people do not perceive that the Sardinian dairy sheep farming system is a vector for maintaining culture by carrying on traditions that have been going on for centuries.

These findings all highlight the need for a greater narrative effort to challenge the vision of pastoral activity, which too often is seen as old-fashioned and with negative connotations in terms of social environment and living conditions (Pastomed 2007, p. 8). There is a need to favour an increase in knowledge of the role of dairy farming systems and in the handing down of the centuries-old traditions that mark the regional cultural identity.

The moderate perception of the ES ‘recreation’ suggests the need for greater attention both to the management and to the promotion of the areas dedicated to pastoral activity. The durability and sustainable development of the Sardinian extensive dairy sheep systems, and the areas in which they are present, can be achieved by leveraging sustainable tourism, especially that based on natural and cultural assets. In fact, it can favour the exploitation of the territorial capital—meaning as historic and artistic capital, agricultural, pastoral and forest systems, territorial protection, know-how and craftsmanship—which can be a lever to revitalize those territories (Garrod et al. 2006; UVAL 2014) and foster local development (Maretti and Salvatore 2012; Sechi et al. 2020). However, if ‘only human activities and presence can maintain this huge capital’ (Sechi et al. 2020, p. 2), the territorial capital exploitation may be challenging since most Sardinian tourism is concentrated along the coasts and only a small part is in the inner area (ISTAT 2020). The increase in coastal tourism matches a decrease in population of the inner areas of the island (Onni and Cannaos 2017). The role of local policymakers in the implementation of policies to attract people towards these is essential.

Another noteworthy fact is the ‘one for all’ tendency, i.e. that the appreciation of one ASO involves the appreciation of all ASOs. These data can be very useful for policymakers looking to stimulate public recognition of ASOs provided by the Sardinian extensive dairy sheep farming system. The increase in information on ASOs deemed unimportant could lead to their perception and, in turn, appreciation of their importance.

Our results showed that personal provenance from a family with a peasant tradition does not affect perception of ASOs, but previous employment in the agricultural or livestock sector affects three ASOs (‘insect’, ‘territory’ and ‘local employment’). Similarly, being a shepherd or a farmer affects only four ASOs positively (‘biodiversity’, ‘habitats’, ‘carbon’ and ‘invasive’) and affects provision of social agricultural activities negatively. These data align with previous studies that see Sardinian shepherds as ‘unaware gardeners’ who build, maintain and save the landscape unwittingly and play, albeit without knowing it, a social role related to cultural production and re-elaboration (Pitzalis and Zerilli 2013; Mattalia et al. 2020). Local organizations in conjunction with policymakers could adopt actions to raise the shepherds' awareness of the multifunctionality of their activity. On the demand side there is a recognition of the ASOs produced by the Sardinian extensive dairy sheep farming system, which could create value within the company and ultimately increase its involvement in the production of ASOs.

Given the descriptive analysis results regarding other sociodemographic variables, it is not surprising that residence significantly influences perceptions of almost all ASOs. The educational level of the interviewees affects the perception of seven benefits generated by the Sardinian extensive dairy sheep farming system. A strongly significant result concerns the variable ‘subject knowledge’, particularly given previous studies that have found a positive relationship between subject knowledge and objective knowledge, which refers to accurate stored information that an individual possesses (Radecki and Jaccard 1995; Carlson et al. 2009). These data confirm the need to increase public awareness of the role of pastoralism in generating environmental, humans’ health and social well-being.

Conclusion

Shepherding is a primary source of livelihood for several communities in Sardinia and other Mediterranean areas. It has been at the heart of local cultural identity for millennia, provides landscape maintenance and care, and contributes to the protection of the territory and the safeguarding and enhancement of vegetal and animal biodiversity. A performing and sustainable agro-pastoral sector is seen as ‘the best ticket to the future of most Mediterranean islands’ (Nori et al. 2017, p. 145). Along the same lines, this study focused on the demand side, offering an overview of how the public perceives the relationships between the dairy sheep farming system and the territory's environmental and social aspects, demonstrating which ASOs are most appreciated by different beneficiaries and on which strategic policy choices should focus. This is the first step to value these positive externalities that are largely involuntary (UVAL 2014). The assessment of perception and valuation of ASOs could also allow for the setting of an effective communication strategy towards local and general communities about the positive externalities of dairy sheep farming systems.

Sardinian stock farmers provide maintenance and care of the territory almost unconsciously. Recognizing the value and contribution of their secondary products can turn them from unaware to aware territory keepers and voluntary public goods providers (Mattalia et al. 2020), and lay the foundation for ‘bring together territorial protection, development and (living) labour, and ensure that territorial protection evolves from precondition to development process’ (UVAL 2014, p. 42). This study, highlighting the appreciation of ecosystems and social public goods provided by stock farmers that are not adequately valued through the market mechanism, offers an answer to those who see the EU’s agricultural policy support for livestock farming as excessive. The perception of ASOs makes it possible to give them a value and determine the magnitude of the benefits that the Sardinian dairy sheep farming system provides, thus supporting policymakers to make effective, informed decisions and to justify investments and actions. Our findings suggest the possibility to switch from the traditional concept of subsidies to payment for ES provision, which links payments to market demands (i.e. the values attributed to ASOs by society) and can target policy towards specific ASOs (Plieninger et al. 2012; Faccioni et al. 2019).

Finally, to our knowledge, this is the first economic study on secondary outputs provided by extensive sheep farming systems, and it considers a large number of secondary outputs. The voluntary response sample that has tilted the balance of sample in favour of resident limits the generalizability of findings, although it allowed an insight into the views of people who have a strong feeling about the topic. Further research based on stratified sampling could validate our findings. Comparative studies would also allow us to understand the change in perception and how this relates to characteristics of the agricultural landscape and its users. Given the importance of the issue to support the decisions of public and private actors to foster the survival of the dairy sheep sector in Sardinia and other Mediterranean regions, further research is needed to deepen the relationship between pasture and the rural landscape in the Mediterranean region. Such evidence could support rural policies aimed at safeguarding primary sector activities in the European landscape.

This study focused on SSs and ESs provided by dairy sheep farming systems. According to Blanco et al. (2019) and Herd-Hoare and Shackleton (2020), an integrated assessment of ESs and EDs could offer further perspectives for proper management approach and the definition of innovative sustainability policies that consider and integrate the social, economic, and environmental dimensions of livelihoods and land use. Future research may expand our research by identifying EDs arising from dairy sheep farming systems, investigating people's reactions to them, and assessing whether the information obtained offers significant contributions to rethinking sustainability policies towards greater effectiveness and equity.

Finally, further research needs to be conducted in the field of economic evaluations of secondary production of the extensive dairy sheep farming system, such as the monetary valuation of key ASOs and preferences and willingness-to-pay of different groups of people.

Appendix

See Tables 4, 5 and 6.

Table 4 List of secondary outputs

Table 5 Description of selected studies

Table 6 Set of logit regressions