This is a post-peer-review, pre-copyedit version of an article published in Sustainability Science. The final authenticated version is available online at: https://doi.org/10.1007/s11625-019- 00672-1 Digital co-construction of relational values: understanding the role of social media for sustainability Article type: Review article Authors: Fulvia Calcagni1, Ana Terra Amorim Maia1, James John Timothy Connolly1,2, Johannes Langemeyer1,2 1 Authors affiliations: Institute of Environmental Science and Technology (ICTA), Universitat Autònoma de Barcelona (UAB), Edifici Z (ICTA-ICP), Carrer de les Columnes s/n, Campus de la UAB, 08193 Cerdanyola del Vallès, Spain 2 Hospital del Mar Medical Research Institute (IMIM), Carrer Doctor Aiguader 88, 08003 Barcelona, Spain Corresponding author fulvia.calcagni@uab.cat Abstract There is a deeply relational aspect to the systems people employ for sorting through and prioritizing plural values assigned to social-ecological interactions. Spurred by interpersonal relationships and adhesion to societal core values, such as justice and reciprocity, relational values go beyond instrumental and intrinsic approaches to understanding human behaviour vis-à-vis the environment. Currently, this relational dimension of values is entering the spotlight of the Cultural Ecosystem Services (CES) literature focusing on non-material benefits and values people derive from ecosystems, such as aesthetics and sense of place. Relational values foster reflections on appropriateness and morality of preferences and respective behaviours in contributing to collective flourishment across space and time, holding implications for social-ecological justice and sustainability. Recently, several studies explored the potential of using social media data for assessing values ascribed to CES, but did not look at how this emerging approach could contribute to an enhanced understanding of relational values. In order to take up this goal, we conducted a systematic review, screening 140 publications and selecting 29 as relevant for exploring the extent to which relational CES values are inferable through social media. Our results show that social media data can reveal CES values’ plural and relational dimension. Social media platforms, thus, can be understood as new arenas for the co-construction of values, where relational values stemming from social-ecological interactions are negotiated and defined. Yet, work on their implications for social-ecological justice and sustainability needs to be extended. Keywords Cultural Ecosystem Services, Social Media analysis, Relational values, Sustainability Introduction 1 Social values, as a system of preferences, principles, and virtues co-constructed and held 2 in common by the members of a social group, are critical to the endeavour of 3 sustainability in that they are closely linked to people’s behaviour (Chan et al. 2012b; 4 Manfredo et al. 2016). A subgroup of social values influences which non-material benefit 5 humans prioritize as a result of interactions with and within their natural environment. 6 Since the Millennium Ecosystem Assessment (2005), these non-material benefits are 7 often referred to as cultural ecosystem services (CES) (Chan et al. 2012a; Milcu et al. 8 2013). Values related to CES are increasingly assessed to highlight the importance of 9 natural assets for sustaining human wellbeing through land-use planning, environmental 10 decision making, and ecosystem-based management (MEA 2005; Chan et al. 2012a; 11 Dickinson and Hobbs 2017). This focus on CES represents an important cornerstone of 12 the wider ecosystem services (ES) framework, allowing ES research to move beyond the 13 stalemate between the “new conservationists” advocacy for the instrumental value of 14 nature and the traditional conservationist claim for protecting nature based on its intrinsic 15 value (Klain et al. 2017). 16 Mediated through human senses and perceptions (MEA, 2005), the intangible benefits of 17 CES are shaped by social values and direct human behaviour in ways that defy the 18 intrinsic-instrumental dichotomy. The systems people employ for sorting through social 19 values make certain provisioning and regulating ES more cognitively accessible (Chan et 20 al. 2012a; Milcu et al. 2013; Dickinson and Hobbs 2017) and, thus, motivate nature 21 conservation and stewardship (Andersson et al. 2014). Yet, the intangibility and 22 incommensurability of values ascribed to CES make them ill-suited to be measured in 23 monetary terms and difficult to be appropriately assessed and incorporated into processes 24 of structured decision-making (Chan et al. 2012a; Milcu et al. 2013; Dickinson and Hobbs 25 2017). Monetary approaches for CES valuation - e.g. the travel cost method, hedonic 26 pricing, and willingness to pay - primarily aim to protect nature by internalizing 27 environmental values into markets. However, such approaches compartmentalize ES into 28 discrete units for marginal valuation (Chan et al. 2012b; Milcu et al. 2013) and assume 29 objectivity in measurement (Raymond et al. 2014). This, critics argue, potentially results 30 in a destructive commodification of nature (Kallis et al. 2013; Milcu et al. 2013) and risks 31 obscuring context-dependent, inter-connected, reciprocal, and plural values that actually 32 shape how humans relate to nature (Chan et al. 2012b; Hernández-Morcillo et al. 2013; 33 Raymond et al. 2014). 34 In this context, Chan et al. (2016) describe values assigned to CES as non-consumptive, 35 non-replaceable, socially-constructed and inherently relational. 36 The relational dimension of CES values manifests when people relate with and within 37 nature, and emerges from a set of preferences, principles and virtues that are 38 fundamentally social and ethical (Kenter et al. 2015). Relational CES values may generate 39 a desire to preserve a landscape due to its close connection with a certain cultural identity, 40 or to engage in green stewardship due to a communally-expressed appreciation for such 41 activities. The relational dimension of values ascribed to CES, thus, associates societal 42 choices, principles and corresponding behaviours to people’s shared sense of justice, care, 43 reciprocity and responsibility towards one another, including humans, non-humans, and 44 ecosystems (Díaz et al. 2015; Chan et al. 2016; Klain et al. 2017; Pascual et al. 2017). 45 Relational values ascribed to CES are, thus, the central conceptual pathway through 46 which ES connects with efforts to build just and sustainable social-ecological systems. 47 Yet, the implications of relational CES values for social-ecological justice – an emerging 48 notion wherein distributional-, procedural-, and recognition-based claims for justice 49 emanate from neither strictly social nor strictly environmental implications, but always 50 from the interplay between the two - and sustainability are still to be conceptualized in 51 order to provide a well-grounded research framework for empirical studies in this area. 52 Often, non-monetary, participatory and deliberative valuation approaches are proposed to 53 account for the multidimensionality of relational values ascribed to CES, building on the 54 longstanding theories of communicative rationality (Raymond et al. 2014). Such methods 55 consist of individuals taking part in collective and elicited processes of CES valuation 56 and co-construction of meanings (Fischer and Eastwood 2016). From these collective co- 57 constructions the researcher can either assemble stated values (e.g. through survey or 58 focus groups) or deduce revealed values (e.g. by analysing behaviours) (Raymond et al. 59 2014; Kenter et al. 2015). These collective processes of valuation, mostly aiming to assess 60 CES distribution and distributional justice across society and space, can also respond to 61 claims of recognition and procedural justice in that they embrace diverse stakeholders 62 and their preferences in a participatory way (Schlosberg 2007). Nevertheless, 63 participatory and deliberative approaches to CES valuation are highly sensitive to the 64 proper representation and empowerment of different social groups and have limited 65 spatial and temporal scope, which may result in imperfect simulations of the process 66 through which values are negotiated within communities (Schafer and Gallemore 2015; 67 Maraja et al. 2016). 68 In an attempt to overcome these limitations, revealed values are increasingly derived from 69 more widely representative social media (Hamstead et al. 2018; Ilieva and McPhearson 70 2018; Langemeyer et al. 2018; Lenormand et al. 2018). Already leveraged to empirically 71 test social science theories (De Nadai et al. 2016), social media represents a digital arena 72 where members of virtual communities share and exchange multimedia content. Social 73 media content can include any information shared on a digital platform, including 74 pictures, tags (text descriptions and geolocalization), or running or biking tracks, for 75 instance. At its root, this content reflects individual user values regarding tangible and 76 intangible aspects of the environment, such as landscape aesthetics, outdoor recreation, 77 cultural identity, and sense of place (Guerrero et al. 2016). 78 When aggregated, social media content related to CES becomes doubly relational. First, 79 the content reflects the inherently relational aspect of values ascribed to CES within wider 80 society. People are reflecting their prior communally shaped notions of what should be 81 valued as they post to social media. Also, because digital communication platforms are 82 co-constructed and shared among the members of a community, the process of producing 83 the content is embedded in a given set of norms that adds a second layer of relationality 84 onto social media data. As individuals share their personal experiences, they expose their 85 social-ecological perceptions and activities to public appraisal and comment within the 86 digital community. In doing so, according to findings from social psychology research, 87 they are motivated and affected by the perceived presence of others (Ames and Naaman 88 2007) and, in seeking alignment with social values of the group, implicitly express 89 relationality based on sentiments of care and reciprocity for the preferences of others. For 90 some, deepening this type of communal relationality relative to the environment is 91 essential for efforts to build just and sustainable social-ecological systems (Chan et al. 92 2016). 93 Apart from revealing the relational dimension of values ascribed to CES, this process of 94 co-construction of meaning (Fischer and Eastwood 2016) holds further implications for 95 assessing social-ecological justice and sustainability. Drawing on Kenter et al. (2015), 96 exchanging multimedia data related to CES on social media platforms can be viewed as 97 a “digital”, non-deliberative and collective valuation approach. This approach ideally 98 overcomes the above-mentioned limitations (e.g. representativeness, power relations, 99 etc.) of other participatory and deliberative valuation approaches and allows for a more 100 inclusive elicitation of values (procedural and recognition justice). Also, since each user 101 produces social media data individually, it might provide a finer picture of values held by 102 diverse people or, when aggregated, by different social groups, offer a useful basis for 103 assessing distributional justice. Moreover, the high spatial and temporal frequency and 104 scale at which social media data can reveal relational values attributed to CES allow 105 researchers to account for users’ behavioural response to ecosystem changes and to the 106 cultural stimuli expressed through social media. Thus, the increased resolution of social 107 media data will likely provide a wide set of complementary information to plan for 108 sustainable social-ecological systems (Ilieva and McPhearson 2018). 109 In essence, social media data assessments are based on capturing components of social 110 processes through various functions, such as sharing, liking, and commenting, and 111 through multiple types of content including, e.g. photos, tags, and posts that represent the 112 ongoing co-construction of relational values ascribed to CES. Thus, based on a review of 113 existing work on CES and social media, we propose this developing approach as an 114 important new empirical basis on which to conceptualize relational values. In particular, 115 drawing on Kenter et al. (2015), we highlight how multimedia content co-constructed by 116 the users of a social media platform through a non-deliberated process exposes the 117 relational nature of social values assigned to CES, whether those values are classified as 118 transcendental (e.g. symbolic, spiritual), contextual (e.g. aesthetics, recreation), non- 119 monetary, other-regarding and/or communal. In sum, because of being co-constructed 120 through a collective valuation process that aligns individual with communal values 121 ascribed to CES (e.g. landscape aesthetic values negotiated among members of a digital 122 community), social media data is especially suited to exposing relational values (see Fig. 123 1). Figure 1. Relational values co-construction through interactions on social media. 124 We view relational values expressed through social media data through a bounded 125 relativist ontological lens. As content is shared within social groups whose definition is 126 bounded in space and time, values arise from people’s interaction with nature and among 127 themselves (Moon and Blackman 2014). These values are constructed from a 128 combination of the subjects’ experiences and the wider societal constructs that shape 129 these experiences – in essence, reflecting a process at the interface between constructivist 130 and subjectivist epistemologies. Given this approach, we are motivated by the hypothesis 131 that social media is a fertile ground for observing relational values ascribed to CES and 132 arising from their collective negotiation. To demonstrate this, we first examine the 133 relevance of social media data in assessing plural/multiple values related to CES and in 134 unveiling their relational dimension. In addition, we specifically focus on how social 135 media can sharpen our understanding of social-ecological justice and sustainability 136 related to relational CES values. Finally, we highlight opportunities and limitations in 137 using social media data for assessing relational values. Materials and methods 138 We performed a literature review of studies using social media data for CES assessment 139 with a focus on the potential for examining relational values. Particularly, we analysed 140 each study’s respective achieved goals, the challenges encountered, and further research 141 suggested in order to assess the potential for addressing relational issues. We performed 142 a systematic, structured quantitative literature review of peer-reviewed articles1, 143 following a replicable procedure. The following criteria guided the search and selection 144 of relevant papers: 145 (i) clear mention of CES; 1 Including one master thesis (Catana 2016) and one peer-reviewed conference proceeding (Goldberg 2015). 146 (ii) use of data retrieved from social media platforms, such as Flickr, Wikipedia or 147 OpenStreetMap, representing the user’s revealed values; 148 (iii) assessment (e.g. quantification, valuation, mapping) of at least one CES, as well as 149 development or discussion of a framework or application. 150 Accordingly, we searched for studies that included in their title, abstract or keywords 151 terms pertaining to two main categories (see Appendix A for details on search terms). 152 The first category restricted the focus of our study to CES. As there are several CES 153 classifications (see Haines-Young & Potschin, 2018; MEA, 2005; TEEB, 2010), we 154 deliberately chose the most general and less detailed search terms. The second category 155 included all the terms that were found to be synonymously used with social media data 156 in scientific publications, e.g. crowdsourced data. 157 We excluded studies not directly relating to the CES framework because we wanted to 158 address the framework’s inherent dimension of relationality stemming from nature- 159 society interactions and its specific aim to shape environmental policy-making for 160 sustainability. Other papers referring to, for instance, scenic route or landscape 161 perceptions rather than CES, either focus strictly on the methodological innovation of the 162 assessment (Levin et al. 2015; Hao et al. 2016), on the potential computing advances 163 (Stefanidis et al. 2013; Chen et al. 2017) or do not address nature-society interactions 164 (Girardin et al. 2008; García-Palomares et al. 2015). We also deliberately excluded 165 studies employing active research approaches to collect primary data, including (active) 166 citizen science approaches, participatory GIS, interviews, focus group discussions and 167 questionnaires. In addition, we discarded publications in languages other than English, 168 those whose full-text could not be found and conference abstracts. 169 We ran an advanced search on Web of Science and retrieved 58 publications meeting our 170 criteria in November 2017. We supplemented these articles with an additional 23 articles 171 that were either still in press or were found to have performed CES assessment through 172 social media but not as their main analysis, so the chosen search terms were not in their 173 abstract or keywords and were not extracted by Web of Science. Among those 81 articles, 174 we selected 22 as directly relevant for our study according to criteria i, ii, and iii (above). 175 We then performed a second round of searches based on those 22 articles. We screened 176 the title of the articles cited by and citing the 22 selected articles and identified 48 new 177 potentially relevant manuscripts. Seven of those 48 met our three criteria. Repeating the 178 same procedure of screening the citations of those 7, we found 11 new potentially relevant 179 articles but none met criteria i, ii, and iii and thus were outside of our study focus (see 180 Appendix B for a detailed diagram of the search). Through this iterative three-stage 181 procedure we reviewed a total of 140 potentially relevant articles and reached saturation 182 of those that met our specific criteria. 183 Only the 29 articles that met all three criteria were included in detailed analyses. Among 184 those excluded, 56 were addressing relevant topics (such as tourism and recreation, scenic 185 and cultural value), but not clearly referring to the CES framework (e.g. see Barry, 2014; 186 Dunkel, 2015; Levin et al., 2015; Seresinhe et al., 2017). The remaining 55 articles were 187 related to other topics, mostly because the acronym “CES” used for the search is valid 188 also in disciplines such as medicine or statistics (e.g. Syahid et al. 2016). 189 For conducting the analysis of the relevant articles we used a standardized assessment 190 protocol (see Appendix C) that allowed us to use the most significant and frequent codes 191 to synthesize and explain large segments of data. The assessment protocol was based on 192 predefined questions regarding general information (publication data, case study location, 193 spatial and temporal scales, data sample) and detailed questions concerning the number 194 and type of CES assessed, the method of assessment, the aim and the further gaps 195 identified by the study. In particular, with these questions, we wanted to understand what 196 motivated the study and whether our hypothesis about the suitability of social media in 197 inferring plural and relational CES values was in some way acknowledged or proven by 198 the selected studies. In addition, we coded the publications based on keywords (e.g. 199 “cultural footprint”, “plurality”, “context-specificity”, etc.) and concepts retrievable from 200 the text (e.g. shared conceptualization, co-construction of values, justice, strategy for 201 conservation, etc.) that aligned with our research objectives. The set of keywords and 202 concepts was continuously updated during the course of the analysis by identifying the 203 synonymous terms used by the different authors. 204 We acknowledge that the screened publications do not include every paper that mentioned 205 CES in relation to social media, and therefore may miss some insights. However, the final 206 sample does allow us to gain a broad and, we believe, representative overview of the most 207 significant literature for drawing reliable conclusions on recent social media-based 208 approaches to CES research. In the sections below we highlight the strongest of these 209 conclusions. Results and discussion Overview and general patterns 210 Overall, the number of publications addressing CES through social media has grown 211 conspicuously since 2012 (see Fig. 2). By nationality, the plurality of the 32 case studies 212 addressed in the 29 papers selected for careful study were located in the USA (n=5), 213 followed by the UK and Singapore (n=3), Africa, Argentina, Denmark, Switzerland, 214 France and Finland (n=2) and then Australia, Japan, Estonia, Germany, Greece, Hungary, 215 Ireland, Spain, and Sweden (n=1) (see Fig. 3). The spatial scale of the studies ranged from 216 global (n=2) to regional (n=15) to urban (n=6) and to local (n=6) (see Appendix C), and 217 data were generally analysed across several years (see Appendix C), with the starting year 218 depending on when the different platforms for social media were launched (see Appendix 219 D). Most of the studies used the pictures shared either on Flickr or on Panoramio, with a 220 small number of them combining the two (n=6). Some studies compared Flickr with 221 original (e.g. survey, interviews, participatory GIS) or official data from statistical or 222 cartographic entities (e.g. land cover map) (n=4). A few studies also used Instagram as a 223 data source, some in combination with other platforms (n=3), some others with original 224 or official data (n=2) (see Fig. 4). 225 Regarding the methods of assessment, many studies performed a visual content analysis 226 of geolocated pictures (n=8), several processed the data using statistical (n=11) and geo- 227 statistical tools (n=7), and some used the available data to model the distribution of where 228 data was missing, both across space and time (n=5). The goal of most studies was to 229 perform correlation analyses between CES and either landscape features (n=13), social 230 groups (n=3) or ecosystem stress (n=1). Other studies aimed to compare the differences 231 between social media and traditional data sources in performing CES assessment (n=5) 232 and to evaluate the different advantages in using each social media platform (n=2). In a 233 few other studies, an analysis of trade-offs and co-benefits was performed (n=2) and some 234 evaluated the spatial distribution of CES (n=6) for either informing decision-makers or 235 prioritizing areas for scenic conservation (see Appendix C). Table 1 Summary of the 29 reviewed articles on CES and social media CES Study Study description assessed reference Landscape Oteros- CES assessment, trade-offs and synergies among them and plurality rozas et al. identification of the landscape features underpinning their of value (2017) provision (recreation Tenerelli et Assessment of CES distribution and correlation with , al. (2016) landscape composition aesthetics, Martínez CES hot-spots assessment and trade-offs, synergies and sense of Pastur et al. correlation with social and biophysical variables place, (2015) social and Catana CES assessment in protected landscapes spiritual (2016) values) Levin et al. Examining the potential of crowdsourced data for (2017) assessing protected area importance. Compares and evaluates multiple crowdsourced data with protected area visitor counts Richards Assessment of CES distribution and correlation with and Friess landscape composition at fine spatial scale (2015) Thiagarajah Assessment of CES change over time et al. (2015) Guerrero et Assessment of CES spatial distribution for informing al. (2016) urban green space governance Landscape Richards Development of a novel method for spatial CES aesthetics and Tunçer assessment (2017) and Kothencz et Comparison with surveyed data for assessing CES recreation al. (2017) predictor value for visitors' level of satisfaction and the self-reported quality of life Derungs and Development of bottom-up approaches to describing Purves landscapes, land cover and land use by building spatial (2016) folksonomies van Zanten Continental comparative assessment of the different social et al. (2016) media platforms ability in predicting CES appreciation Landscape Tammi et al. Supplementary, non-monetary mapping of aesthetic ES aesthetics (2017) index Tenerelli et Assessment of the distribution of CES among users with al. (2017) different provenance and correlation with landscape composition Yoshimura Comparison between CES supply and demand and Hiura (2017) Figueroa- Assessment of CES spatial distribution Alfaro and Tang (2017) Casalegno Assessment of spatial covariation between supporting, et al. (2013) provisioning, regulating and cultural ES Goldberg, Point of Interest identification and prioritization for scenic (2015) conservation Cultural Gliozzo et Assessment of CES spatial distribution, hot/cold spots Identity al. (2016) identification and comparison between the spatial behaviour of different platform users Sherren et Description of the interdisciplinary state-of-the-art that is al. (2017) converging to enable new tools for Social Impact Assessment (SIA), using hydroelectricity as a case study Recreation Cord et al. Assessment of CES distribution and correlation with (2015) landscape composition Allan et al. Quantification of spatial distribution of recreational (2015) facilities and correlation with ecosystem stress assessment Upton et al. Assessment of accessibility to recreational forests and (2015) recreation demand modelling Recreation Heikinheim Comparison with surveyed data to assess the added value and o et al. of social media analysis in reveal correlation with social ecotouris (2017) group characteristics m Hausmann Exploration of which socio-economic, geographical and et al. (2017) biological factors explain social media use Sonter et al. Assessment of nature-based recreation within conserved (2016) lands using social media, analysing its predictor value for visitation rate and its correlation with landscape composition Wood et al. Use of data from social media to predict visitation rates at (2013) sites around the world comparing it to empirical data, such as type of attraction, income-level, temporal changes, etc. Willemen et combining photo counts with species range data to al. (2015) determine the protected areas with the highest potential to attract wildlife tourists Recreation Ghermandi Investigate patterns of public use in natural treatment and (2016) systems education Figure 2. Number of publications per year Figure 3. Geographical distribution of the case studies Figure 4. Number of studies adopting the different social media platforms Social media assessment of plural CES values 236 Quantitatively assessing the plurality of CES values has always challenged researchers, 237 both timewise and in terms of costs. Here we unveil the advantages in using social media 238 data for addressing this challenge. Among the selected papers, the majority (more than 239 60%) include an assessment of multiple CES values and, in some cases, of their spatial 240 co-presence, ranging from a minimum of two to a maximum of eight values (see Fig. 5). 241 These studies confirm social media as a suitable data source for understanding the 242 context-dependency and holistic nature of CES values. Despite the fact that studies 243 adopted different systems of classification (e.g. Haines-Young and Potschin, 2018; MEA, 244 2005; TEEB, 2010), it is possible to highlight connections between the CES values 245 assessed and the methods implemented. Among studies that restricted their analysis to 246 one or two main CES, there was a general agreement on the need to recognize that there 247 are many ways in which the environment is perceived (Tenerelli et al. 2017). These 248 articles express an interest in exploring the multiple means of perception for better 249 integrating non-expert conceptualizations of landscape into policy (e.g. Derungs and 250 Purves, 2016). Figure 5. Number of studies assessing the different CES 251 Among the studies assessing more than two CES, visual content analysis of geolocated 252 crowdsourced pictures is the most applied methodology. Visual content analysis allows 253 the researcher to assess a wide spectrum of CES, including landscape aesthetics (Martínez 254 Pastur et al. 2015; Thiagarajah et al. 2015; Catana 2016; Guerrero et al. 2016; Tenerelli 255 et al. 2016); recreation and ecotourism (Catana 2016; Tenerelli et al. 2016; Oteros-Rozas 256 et al. 2017); cultural heritage and social and spiritual values (Oteros-Rozas et al. 2017); 257 social relation and species existence value (Richards and Friess 2015; Catana 2016), local 258 identity (Martínez Pastur et al. 2015); and sense of place (Guerrero et al. 2016). In 259 addition, studies complementing social media data with primary crowdsourced data, such 260 as those obtained from a public participation GIS (PPGIS) workshop (Levin et al. 2017) 261 or surveyed and mined from archives (Thiagarajah et al. 2015), reveal the potential to 262 provide a wider spectrum of CES values, adding conservation, therapeutic, wilderness 263 and inspirational values to the list. 264 Yet, confirming previous findings (Hernández-Morcillo et al. 2013; Milcu et al. 2013), 265 among CES studies aesthetic and recreational values are, in general, the most frequently 266 quantified, either in single or multiple value assessments. Many of the studies assessed 267 either only aesthetics (more than 30%) (Casalegno et al. 2013; Goldberg 2015; Figueroa- 268 Alfaro and Tang 2017; Tammi et al. 2017; Tenerelli et al. 2017; Yoshimura and Hiura 269 2017), only recreational values (more than 30%) (Allan et al. 2015; Cord et al. 2015; 270 Upton et al. 2015), or both (more than 20%) (van Zanten et al. 2016; Kothencz et al. 2017; 271 Richards and Tunçer 2018). 272 Apart from the multiple CES values that social media have thus far proven suitable to 273 assess, many studies showed also the potential of these methods to account for the 274 different predictors or explanatory variables that help to understand how CES values are 275 constructed. Some studies explore the environmental variables that enable specific CES 276 values, such as how complexity and “naturalness” determine attractiveness in landscape 277 aesthetics (Tenerelli et al. 2017), while others focus on the extent to which species 278 richness (Willemen et al. 2015; Hausmann et al. 2017) or landscape features (Tenerelli et 279 al. 2016) influence recreation, tourism, or aesthetic potential. In addition, social media 280 data allow researchers to correlate plural CES presence to predictor variables, such as 281 accessibility (Thiagarajah et al. 2015; Upton et al. 2015; Willemen et al. 2015; Ghermandi 282 2016; Guerrero et al. 2016), population density (Hausmann et al. 2017), type of habitat 283 and the presence of human artefacts (Gliozzo et al. 2016; Guerrero et al. 2016), scenic 284 spots (Tenerelli et al. 2016), and natural protection areas (Levin et al. 2015; Catana 2016). 285 Similarly, Oteros-Rozas et al. (2017) recognize the context-specificity of CES and their 286 results show a positive relationship between landscape diversity and CES diversity, 287 thereby, verifying previous findings (Casalegno et al. 2013; Gliozzo et al. 2016; Tammi 288 et al. 2017). Moreover, Yoshimura and Hiura (2017) and Tenerelli et al. (2017) found 289 further evidence of the different preferences expressed by foreign and local users, 290 although this finding is not confirmed by other works (Richards and Friess 2015). Social media assessments revealing relational values 291 A person’s perspective on place or landscape has recently been argued to result from 292 interactions that bridge transcendental (or held) and contextual (or assigned) values 293 (Levin et al. 2017). The argument builds on the concept of relational value, which, 294 although already present in the environmental psychology and sociology literature 295 (Stephenson 2008; Graham et al. 2013), was first explicitly attributed to CES in Chan et 296 al. (2016) and is, thus, relatively new and clearly not present in the relevant articles 297 published before that date, nor explicitly assessed by those analysed in this study. Only 298 Catana (2016) refers to the dimension of relationality in order to show how human values 299 are connected to perceptions, preferences, and ultimately to well-being. However, in 300 order to verify our hypothesis on the specific potential of social media data to infer this 301 so far neglected value dimension, we looked for similar concepts or for promising 302 assessment methods in the reviewed articles, despite the fact that they were primarily 303 addressing other objectives. 304 Many studies expressed the need to capture the meanings that people collectively assign 305 to landscapes and that regulate inter-societal relationships involving nature. In this regard, 306 social media data is seen as a valuable source of information about shared 307 conceptualizations and about the process of culture creation relative to the natural 308 environment, as in the case of frequently used tags and hashtags (Derungs and Purves 309 2016; Guerrero et al. 2016). More specifically, some studies suggest that sharing a 310 landscape picture on a social media platform is a form of “digital interaction” that adds 311 another collective dimension to social values, contributing to peoples’ shared image of 312 landscapes and, consequently, attachment to nature and to one another (Oteros-Rozas et 313 al. 2017). These attachments may also be based on their historical and cultural 314 background (Guerrero et al. 2016). People sharing content on social media are, indeed, 315 responsible for influencing their “digital receptors” with what is referred to as their 316 cultural ecosystem footprint (Gliozzo et al. 2016). In addition, social media not only 317 serves as a platform to dispute and share relational CES values, but also stores the process 318 of value creation that generates heritage, allowing the persistence of CES values and 319 counteracting the “extinction of experiences” of nature in modern societies(Miller 2005). 320 In addition, the different forms of interaction allowed on social media, motivate people 321 to co-construct values in diverse manners (Cord et al. 2015). In this regard, we also found 322 some evidence of a correlation between the relational and plural attributes of values. In 323 platforms allowing voluntary and non-restricted participation, such as Flickr, Instagram 324 and Panoramio, people are motivated to share data because of the global visibility they 325 obtain and, thus, tend to express plural and context-specific values (Gliozzo et al. 2016; 326 Guerrero et al. 2016). On the contrary, when a platform has a compiling purpose intended 327 to provide a specific output to decision-makers, such as Geograph, data are more 328 homogeneously distributed and less informative of people’s multiple held values (Gliozzo 329 et al. 2016). Therefore, across the reviewed literature, the majority of the authors 330 recognize the correlation between the collective process of co-construction of meaning 331 associated with social-ecological interactions and the expression of relational principles 332 of care, reciprocity and responsibility towards nature and others. This strain of findings 333 in the literature affirms that social media platforms are suitable arenas for negotiating and 334 capturing relational values assigned to CES. Furthering social-ecological justice and sustainability by inferring relational values from social media platforms 335 The production of values is part of a socialization process that occurs through repeatedly 336 engaging in countless experiences and phases of learning, either formal or informal. This 337 process “embrains” the spontaneous responses and cultural practices that allow 338 individuals and groups to adapt to their social-ecological surroundings without much 339 effort or deliberation (Gliozzo et al. 2016; Manfredo et al. 2016). Therefore, since we 340 assume that value attachment to places motivates people’s actions and the consequent 341 effects on their surroundings (Yoshimura and Hiura 2017), we look at this process with 342 particular attention. 343 In social media, apart from their values and emotions, people share digital and geolocated 344 traces of actions driven by underlying values and, further eased by Internet functionalities, 345 influence each other (Gliozzo et al. 2016). The process of mutual influence in ascribing 346 values to CES provided by places or activities can lead others to personally experience 347 them (Cord et al. 2015) or not (Goldberg 2015; Gliozzo et al. 2016), such as for CES that 348 do not require a physical interaction to be experienced (e.g. cultural heritage, existence 349 value, and spiritual values) (Richards and Friess, 2015). Hence, some studies show 350 empirical evidence of these processes of co-construction of values, mutually influenced 351 behaviour in interacting with nature, and consequent co-production of ES (Fischer and 352 Eastwood 2016), demonstrating that CES values expressed on social media cluster around 353 popular scenic (Goldberg 2015) or recreational spots (Cord et al. 2015) and widely known 354 species (Willemen et al. 2015). 355 In this regard, some studies suggest using “likes” and ratings associated with social media 356 data (Gliozzo et al. 2016; Hausmann et al. 2017) or simply the number of times each 357 picture posted on a social media platform has been visualized (Goldberg 2015), or its 358 location has been reached (Cord et al. 2015), by another user as a proxy for CES hot-spot 359 identification. Mapping CES hot-spots helps identify areas where the services are most 360 highly valued (Goldberg 2015; Guerrero et al. 2016) and whether this results in ecosystem 361 stress (Allan et al. 2015), providing useful information to prioritize areas for conservation 362 (Hausmann et al. 2017) and cultural services management (Guerrero et al. 2016). 363 In addition, several of the reviewed studies have highlighted the lack of methodological 364 approaches for addressing social-ecological justice and sustainability. Some argue for 365 performing a demographic profile of social media users in order to account for procedural 366 and recognition justice in the assessment, as well as assessing distributional justice by 367 accounting for variables such as gender, social class, age or area of residence (Gliozzo et 368 al. 2016) and specifically seeking to reach less represented user groups (Guerrero et al. 369 2016). Others propose to perform an assessment of cross-cultural differences (Cord et al. 370 2015). Some studies recommend monitoring the trends of social preferences towards CES 371 exploring their evolution across geographic and temporal scales (Wood et al. 2013; 372 Martínez Pastur et al. 2015; Derungs and Purves 2016; Guerrero et al. 2016), or to 373 perform scenario-based simulations (Wood et al. 2013) and develop means for 374 quantifying ecosystem resilience over time (Allan et al. 2015). This would help correlate 375 changes in visitation rates with changes in ecosystem health, site access, infrastructure 376 development and alternative management regimes. Finally, some studies recognize the 377 potential of social media data for revealing city dwellers’ preferences and values in order 378 to respond to important challenges for place-based culture and well-being (Guerrero et al. 379 2016; Tenerelli et al. 2017), and plan for healthy green spaces (Kothencz et al. 2017). All 380 of these are promising directions for translating CES knowledge derived from social 381 media data into a more coherent and systematic understanding of relational CES values 382 and of their role for social-ecological justice and sustainability. Opportunities and limitations in the use of social media data 383 Because social media data can be collected passively, its greatest benefit for questions of 384 internal validity is that it provides a research route that compensates for shortcomings 385 affecting data obtained through more active collection approaches, such as surveys, 386 interviews or photo elicitation (Guerrero et al. 2016). Geolocated social media data, 387 indeed, reveals perspectives that arise from directly experiencing the environment at the 388 same time as it is being evaluated (Tenerelli et al. 2017) and is less costly and time- 389 consuming (Yoshimura and Hiura 2017). In addition, because it is collected across a wide 390 variety of the population and can have high spatial resolution thanks to the embedded 391 geotag (van Zanten et al. 2016; Oteros-Rozas et al. 2017) and be reported in real-time 392 (Gliozzo et al. 2016; Oteros-Rozas et al. 2017), it has higher representativeness over space 393 and time than many alternatives. 394 395 Moreover, geolocated social media data are especially useful because the digital 396 interactions that they allow are believed to spur participants to mutually influence gradual 397 changes in their values, associated behaviour and, eventually, produced environment. 398 Such changes, to the extent that they are transferred to the policy level, have been often 399 indicated as necessary to achieve global environmental sustainability (Stern et al. 1999; 400 Manfredo et al. 2016). In addition, given that most of the people using social media are 401 urban dwellers (Guerrero et al. 2016; International Telecommunication Union 2016), this 402 data source is valuable in that it allows researchers to assess CES values held by people 403 with gradually declining opportunities of interaction with nature (Dickinson and Hobbs 404 2017). 405 Of course, there are limitations to the use of geolocated social media data as well. For 406 example, some studies recognize the limitations that arise when the researcher interprets 407 data in a one-directional way (Derungs and Purves 2016; Oteros-Rozas et al. 2017) and 408 because of the temptation to see patterns in the available data where none actually exist 409 (what is known as apophenia) (Wood et al. 2013). This limitation could be partly 410 addressed through the establishment of unified coding protocols for social media data. 411 Others highlight the spatial bias as a result of data gaps in places characterized by poor 412 data (Catana 2016; Levin et al. 2017) or poor reliability of the geotag (Oteros-Rozas et 413 al. 2017). Questions emerge also due to data representativeness. Several existing digital 414 divides and fashions in the use of certain social media platforms, such as those related to 415 age, gender and income level, might mislead the analyst because the data accounts only 416 for behaviours and perceptions of certain profiles and social groups (Wood et al. 2013; 417 Allan et al. 2015; Martínez Pastur et al. 2015; Willemen et al. 2015; Oteros-Rozas et al. 418 2017; Tenerelli et al. 2017). For example, recent studies found a strong gender imbalance 419 (64% male; 36% female) in the users of the photo-sharing platform Flickr who responded 420 to a user questionnaire (Lenormand et al. 2018). Such perceptions might further be 421 influenced by specific individuals, groups or private corporations that, by implementing 422 communication or market strategies on social networks, reach their target audience and 423 influence values creation and data availability (Oteros-Rozas et al. 2017). 424 All of these limitations direct us to consider how social media deals with issues of uneven 425 power relations, which is perhaps the central issue impacting recognition, procedural, 426 distributional justice outcomes (Schlosberg 2007). Especially with regard to procedural 427 justice, a poor accounting of the effect of uneven social power relations has long been 428 central to the urban scholarship critique of communicative rationality (Huxley and 429 Yiftachel 2000). These limitations are believed to decrease with the continuing increase 430 in number and awareness of social media users and in data accuracy allowed by newer 431 technologies (Guerrero et al. 2016; Tenerelli et al. 2017) and the combination of different 432 data sources. However, private monopolies on social media data bare the risk of 433 restricting the scientific analysis of data, for instance by platforms like Facebook and 434 Instagram, and, thus, limit a less biased societal representation of relational values 435 through social media-based research. This relates to questions of ethics, privacy and 436 copyright, which have been raised in relation to accessing social media data (Guerrero et 437 al. 2016), concerns also addressed by the latest European General Data Protection 438 Regulation (COM/2018/043), and which so far lack a more specific discussion from the 439 angle of sustainability science. Conclusions and further recommendations 440 With this study, we aimed to contribute to the emerging questions of relationality within 441 sustainability research. We found evidence of social media platforms serving as valuable 442 data sources for revealing the multiple values that people assign to the environment. In 443 addition, we showed the dimension of relationality within plural CES values. We propose 444 a novel conceptualization that relates relationality to the collective processes of co- 445 construction of values ascribed to CES and, which is visible within social media data. We 446 found that the reviewed literature frequently points at the ability of social media data to 447 reveal people’s willingness to share their experiences online. This is believed to be 448 significant in influencing the co-construction process of plural CES values, people’s 449 interactions with and within the environment and, consequently, the co-production of ES, 450 proving useful information on value and behaviour for landscape and urban planning. 451 However, social disparities are reproduced in this process and, once translated into action, 452 can have implications for social-ecological systems justice and sustainability (e.g. 453 increased visitation rate in touristic spots and consequent gentrification and ecosystem 454 stress). Hence, this study calls for further exploring the different social and environmental 455 factors at play, and specific biases characterizing data sources, in order to enable the 456 potential of social media data to inform just and sustainable landscape planning and 457 management. Future research should also focus on the potential of social media-based 458 approaches to explore the path from value to action, assessing the influence of values 459 created on social media in enhancing people’s agency toward the collective improvement 460 of their well-being (see Hicks et al. (2016)) and, eventually, to activate citizens in a 461 process of co-production of nature (Linders 2012; Guerrero et al. 2016). Acknowledgements We acknowledge financial support from the 2015-2016 BiodivERsA COFUND call for research proposals through the Spanish Ministry of Science, Innovation and Universities (PCIN-2016-002) and from the European Research Council (Greenlulus 678034). F.C. thanks the AGAUR Catalan governmental agency (Grant number 2018FI_B00635) and the Institute for the right to university studies in Lazio, Laziodisu (Grant “Torno Subito 2017” number 7425-18092017) for the funding received to support this study. A.T.A.M. acknowledges support by the European Commission through an Erasmus Mundus scholarship (JEMES CiSu UAB2016/No. 1). J.J.T.C. thanks the Spanish Ministry of Sciences, Innovation, and University’s Subprogram of Juan de la Cierva Incoporacion (IJCI-2016-31100). We also thank the reviewers for their valuable remarks. Bibliography Allan JD, Smith SDP, McIntyre PB, et al (2015) Using cultural ecosystem services to inform restoration priorities in the Laurentian Great Lakes. Front Ecol Environ 13:418–424. doi: 10.1890/140328 Ames M, Naaman M (2007) Why we tag: motivations for annotation in mobile and online media. Proc SIGCHI Conf Hum factors Comput Syst 971–980. doi: 10.1145/1240624.1240772 Andersson E, Tengö M, McPhearson T, Kremer P (2014) Cultural ecosystem services as a gateway for improving urban sustainability. Ecosyst Serv 12:165–168. doi: 10.1016/j.ecoser.2014.08.002 Barry SJ (2014) Using social media to discover public values, interests, and perceptions about cattle grazing on park lands. Environ Manage 53:454–464. doi: 10.1007/s00267-013-0216-4 Casalegno S, Inger R, DeSilvey C, Gaston KJ (2013) Spatial Covariance between Aesthetic Value & Other Ecosystem Services. PLoS One 8:6–10. doi: 10.1371/journal.pone.0068437 Catana AV (2016) Using social media to assess cultural ecosystem services generated in protected areas in Patagonia Chan KMA, Balvanera P, Benessaiah K, et al (2016) Why protect nature? Rethinking values and the environment. Proc Natl Acad Sci 113:1462–1465. doi: 10.1073/pnas.1525002113 Chan KMA, Guerry AD, Balvanera P, et al (2012a) Where are Cultural and Social in Ecosystem Services? A Framework for Constructive Engagement. Bioscience 62:744–756. doi: 10.1525/bio.2012.62.8.7 Chan KMA, Satterfield T, Goldstein J (2012b) Rethinking ecosystem services to better address and navigate cultural values. Ecol Econ 74:8–18. doi: 10.1016/j.ecolecon.2011.11.011 Chen C, Chen X, Wang Z, et al (2017) ScenicPlanner: planning scenic travel routes leveraging heterogeneous user-generated digital footprints. Front Comput Sci 11:61–74. doi: 10.1007/s11704-016-5550-2 Cord AF, Roeßiger F, Schwarz N (2015) Geocaching data as an indicator for recreational ecosystem services in urban areas: Exploring spatial gradients, preferences and motivations. Landsc Urban Plan 144:151–162. doi: 10.1016/j.landurbplan.2015.08.015 De Nadai M, Staiano J, Larcher R, et al (2016) The Death and Life of Great Italian Cities: A Mobile Phone Data Perspective. In: 26th International ACM Conference on World Wide Web (WWW) Derungs C, Purves RS (2016) Characterising landscape variation through spatial folksonomies. Appl Geogr 75:60–70. doi: 10.1016/j.apgeog.2016.08.005 Díaz S, Demissew S, Carabias J, et al (2015) The IPBES Conceptual Framework - connecting nature and people. Curr Opin Environ Sustain 14:1–16. doi: 10.1016/j.cosust.2014.11.002 Dickinson DC, Hobbs RJ (2017) Cultural ecosystem services: Characteristics, challenges and lessons for urban green space research. Ecosyst Serv 25:179–194. doi: 10.1016/j.ecoser.2017.04.014 Dunkel A (2015) Visualizing the perceived environment using crowdsourced photo geodata. Landsc Urban Plan 142:173–186. doi: 10.1016/j.landurbplan.2015.02.022 Figueroa-Alfaro RW, Tang Z (2017) Evaluating the aesthetic value of cultural ecosystem services by mapping geo-tagged photographs from social media data on Panoramio and Flickr. J Environ Plan Manag 60:266–281. doi: 10.1080/09640568.2016.1151772 Fischer A, Eastwood A (2016) Coproduction of ecosystem services as human-nature interactions-An analytical framework. Land use policy 52:41–50. doi: 10.1016/j.landusepol.2015.12.004 García-Palomares JC, Gutiérrez J, Mínguez C (2015) Identification of tourist hot spots based on social networks: A comparative analysis of European metropolises using photo-sharing services and GIS. Appl Geogr 63:408–417. doi: 10.1016/j.apgeog.2015.08.002 Ghermandi A (2016) Analysis of intensity and spatial patterns of public use in natural treatment systems using geotagged photos from social media. Water Res 105:297– 304. doi: 10.1016/j.watres.2016.09.009 Girardin F, Blat J, Calabrese F, et al (2008) Digital footprinting: Uncovering tourists with user-generated content. IEEE Pervasive Comput 7:36–44. doi: 10.1109/MPRV.2008.71 Gliozzo G, Pettorelli N, Haklay M (2016) Using crowdsourced imagery to detect cultural ecosystem services: a case study in South Wales, UK. Ecol Soc 21:. doi: 10.5751/es- 08436-210306 Goldberg L (2015) Utilizing Crowdsourced Georeferenced Photography for Identification and Prioritization of Areas for Scenic Conservation. In: Buhmann E, Ervin SM, Pietsch M (eds) Digital Landscape Architecture. pp 268–275 Graham S, Barnett J, Fincher R, et al (2013) The social values at risk from sea-level rise. Environ Impact Assess Rev 41:45–52. doi: 10.1016/j.eiar.2013.02.002 Guerrero P, Møller MS, Olafsson AS, Snizek B (2016) Revealing Cultural Ecosystem Services through Instagram Images: The Potential of Social Media Volunteered Geographic Information for Urban Green Infrastructure Planning and Governance. Urban Plan 1:1. doi: 10.17645/up.v1i2.609 Haines-Young R, Potschin M (2018) Common International Classification of Ecosystem Services (CICES) V5.1 Guidance on the Application of the Revised Structure Hamstead ZA, Fisher D, Ilieva RT, et al (2018) Geolocated social media as a rapid indicator of park visitation and equitable park access. Comput Environ Urban Syst 72:38–50. doi: 10.1016/j.compenvurbsys.2018.01.007 Hao X, Wu B, Morrison AM, Wang F (2016) Worth thousands of words? Visual content analysis and photo interpretation of an outdoor tourism spectacular performance in Yangshuo-Guilin, China. Anatolia 27:201–213. doi: 10.1080/13032917.2015.1082921 Hausmann A, Toivonen T, Heikinheimo V, et al (2017) Social media reveal that charismatic species are not the main attractor of ecotourists to sub-Saharan protected areas. Sci Rep 7:1–9. doi: 10.1038/s41598-017-00858-6 Heikinheimo V, Minin E Di, Tenkanen H, et al (2017) User-Generated Geographic Information for Visitor Monitoring in a National Park: A Comparison of Social Media Data and Visitor Survey. ISPRS Int J Geo-Information 6:85. doi: 10.3390/ijgi6030085 Hernández-Morcillo M, Plieninger T, Bieling C (2013) An empirical review of cultural ecosystem service indicators. Ecol Indic 29:434–444. doi: 10.1016/j.ecolind.2013.01.013 Hicks CC, Levine A, Agrawal A, et al (2016) Engage key social concepts for sustainability. Science (80- ) 352:38–40. doi: 10.1126/science.aad4977 Huxley M, Yiftachel O (2000) New paradigm or old Myopia? Unsettling the communicative turn in planning theory. J Plan Educ Res 19:333–342. doi: 10.1177/0739456X0001900402 Ilieva RT, McPhearson T (2018) Social-media data for urban sustainability. Nat Sustain 1:553–565. doi: 10.1038/s41893-018-0153-6 International Telecommunication Union (2016) Measuring the Information Society Report 2016 Kallis G, Gómez-Baggethun E, Zografos C (2013) To value or not to value? That is not the question. Ecol Econ 94:97–105. doi: 10.1016/j.ecolecon.2013.07.002 Kenter JO, O’Brien L, Hockley N, et al (2015) What are shared and social values of ecosystems? Ecol Econ 111:86–99. doi: 10.1016/j.ecolecon.2015.01.006 Klain SC, Olmsted P, Chan KMA, Satterfield T (2017) Relational values resonate broadly and differently than intrinsic or instrumental values, or the New Ecological Paradigm. PLoS One 12:1–21. doi: 10.1371/journal.pone.0183962 Kothencz G, Kolcsár R, Cabrera-Barona P, Szilassi P (2017) Urban green space perception and its contribution to well-being. Int J Environ Res Public Health 14:. doi: 10.3390/ijerph14070766 Langemeyer J, Calcagni F, Baró F (2018) Mapping the intangible: Using geolocated social media data to examine landscape aesthetics. Land use policy 77:542–552. doi: https://doi.org/10.1016/j.landusepol.2018.05.049 Lenormand M, Luque S, Langemeyer J, et al (2018) Multiscale socio-ecological networks in the age of information. PLoS One 13:1–16. doi: 10.1371/journal.pone.0206672 Levin N, Kark S, Crandall D (2015) Where have all the people gone? Enhancing global conservation using night lights and social media. Ecol Appl 25:2153–2167. doi: 10.1890/15-0113.1 Levin N, Lechner AM, Brown G (2017) An evaluation of crowdsourced information for assessing the visitation and perceived importance of protected areas. Appl Geogr 79:115–126. doi: 10.1016/j.apgeog.2016.12.009 Linders D (2012) From e-government to we-government: Defining a typology for citizen coproduction in the age of social media. Gov Inf Q 29:446–454. doi: 10.1016/j.giq.2012.06.003 Manfredo MJ, Teel TL, Dietsch AM (2016) Implications of human value shift and persistence for biodiversity conservation. Conserv Biol 30:287–296. doi: 10.1111/cobi.12619 Maraja R, Barkmann J, Tscharntke T (2016) Perceptions of cultural ecosystem services from urban green. Ecosyst Serv 17:33–39. doi: 10.1016/j.ecoser.2015.11.007 Martínez Pastur G, Peri PL, Lencinas M V., et al (2015) Spatial patterns of cultural ecosystem services provision in Southern Patagonia. Landsc Ecol 31:383–399. doi: 10.1007/s10980-015-0254-9 MEA (2005) Ecosystems And Human Well-Being - Synthesis Milcu AI, Hanspach J, Abson D, Fischer J (2013) Cultural Ecosystem Services : A Literature Review and Prospects for Future Research. Ecol Soc 18:44. doi: http://dx.doi.org/10.5751/ ES-05790-180344 Miller JR (2005) Biodiversity conservation and the extinction of experience. Trends Ecol Evol 20:430–434. doi: 10.1016/j.tree.2005.05.013 Moon K, Blackman D (2014) A Guide to Understanding Social Science Research for Natural Scientists. Conserv Biol 28:1167–1177. doi: 10.1111/cobi.12326 Oteros-Rozas E, Martín-López B, Fagerholm N, et al (2017) Using social media photos to explore the relation between cultural ecosystem services and landscape features across five European sites. Ecol Indic 94:74–86. doi: https://doi.org/10.1016/j.ecolind.2017.02.009 Pascual U, Balvanera P, Díaz S, et al (2017) Valuing nature’s contributions to people: the IPBES approach. Curr Opin Environ Sustain 26–27:7–16. doi: 10.1016/j.cosust.2016.12.006 Raymond CM, Kenter JO, Plieninger T, et al (2014) Comparing instrumental and deliberative paradigms underpinning the assessment of social values for cultural ecosystem services. Ecol Econ 107:145–156. doi: 10.1016/j.ecolecon.2014.07.033 Richards D, Tunçer B (2018) Using image recognition to automate assessment of cultural ecosystem services from social media photographs. Ecosyst Serv 31:318–325. doi: https://doi.org/10.1016/j.ecoser.2017.09.004 Richards DR, Friess DA (2015) A rapid indicator of cultural ecosystem service usage at a fine spatial scale: Content analysis of social media photographs. Ecol Indic 53:187–195. doi: 10.1016/j.ecolind.2015.01.034 Schafer JG, Gallemore CT (2015) Biases in multicriteria decision analysis: The case of environmental planning in Southern Nevada. Environ Plan C Gov Policy 34:1652– 1675. doi: 10.1177/0263774X16629675 Schlosberg D (2007) Defining Environmental Justice. Oxford University Press, New York Seresinhe CI, Moat HS, Preis T (2017) Quantifying scenic areas using crowdsourced data. Environ Plan B Urban Anal City Sci. doi: 10.1177/0265813516687302 Sherren K, Parkins JR, Smit M, et al (2017) Digital archives, big data and image-based culturomics for social impact assessment: Opportunities and challenges. Environ Impact Assess Rev 67:23–30. doi: 10.1016/j.eiar.2017.08.002 Sonter LJ, Watson KB, Wood SA, Ricketts TH (2016) Spatial and temporal dynamics and value of nature-based recreation, estimated via social media. PLoS One 11:1– 16. doi: 10.1371/journal.pone.0162372 Stefanidis A, Crooks A, Radzikowski J (2013) Harvesting ambient geospatial information from social media feeds. GeoJournal 78:319–338. doi: 10.1007/s10708-011-9438-2 Stephenson J (2008) The Cultural Values Model: An integrated approach to values in landscapes. Landsc Urban Plan 84:127–139. doi: 10.1016/j.landurbplan.2007.07.003 Stern PC, Dietz T, Abel T, et al (1999) A value-belief-norm theory of support for social movements: The case of environmentalism. Hum Ecol Rev 6:81–97. doi: https://www.jstor.org/stable/24707060 Syahid A, Tareq MA (2015) A penny for your thoughts : A preference modelling case study in R. In: 12th International Conference on Innovation and Management Tammi I, Mustajärvi K, Rasinmäki J (2017) Integrating spatial valuation of ecosystem services into regional planning and development. Ecosyst Serv 26:329–344. doi: 10.1016/j.ecoser.2016.11.008 TEEB (2010) The Economics of Ecosystems and Biodiversity: Mainstreaming the economics of Nature: A synthesis of the approach, conclusions and recommendations of TEEB. Tenerelli P, Demšar U, Luque S (2016) Crowdsourcing indicators for cultural ecosystem services: A geographically weighted approach for mountain landscapes. Ecol Indic 64:237–248. doi: 10.1016/j.ecolind.2015.12.042 Tenerelli P, Püffel C, Luque S (2017) Spatial assessment of aesthetic services in a complex mountain region: combining visual landscape properties with crowdsourced geographic information. Landsc Ecol 32:1097–1115. doi: 10.1007/s10980-017-0498-7 Thiagarajah J, Wong SKM, Richards DR, Friess DA (2015) Historical and contemporary cultural ecosystem service values in the rapidly urbanizing city state of Singapore. Ambio 44:666–677. doi: 10.1007/s13280-015-0647-7 Upton V, Ryan M, O’Donoghue C, Dhubhain AN (2015) Combining conventional and volunteered geographic information to identify and model forest recreational resources. Appl Geogr 60:69–76. doi: 10.1016/j.apgeog.2015.03.007 van Zanten BT, Van Berkel DB, Meentemeyer RK, et al (2016) Continental-scale quantification of landscape values using social media data. Proc Natl Acad Sci 113:12974–12979. doi: 10.1073/pnas.1614158113 Willemen L, Cottam AJ, Drakou EG, Burgess ND (2015) Using social media to measure the contribution of red list species to the nature-based tourism potential of african protected areas. PLoS One 10:1–14. doi: 10.1371/journal.pone.0129785 Wood SA, Guerry AD, Silver JM, Lacayo M (2013) Using social media to quantify nature-based tourism and recreation. Sci Rep 3:. doi: 10.1038/srep02976 Yoshimura N, Hiura T (2017) Demand and supply of cultural ecosystem services: Use of geotagged photos to map the aesthetic value of landscapes in Hokkaido. Ecosyst Serv 24:68–78. doi: 10.1016/j.ecoser.2017.02.009 Table legend Table 1 Summary of the reviewed articles on CES and social media ........................... 14 Figure legend Figure 1. Relational values co-construction triggered by each individual sharing content on social media platforms (self-elaboration). ................................................................. 8 Figure 2. Number of publications per year .................................................................. 17 Figure 3. Geographical distribution of the case studies ................................................ 17 Figure 4. Number of studies adopting the different social media platforms .................. 18 Figure 5. Number of studies assessing the different CES ............................................. 19