Social media reveal visitors’ interest in flora and fauna species of a forest region

ABSTRACT Visitor interests can be crucial to understanding humans’ connectedness in nature. We analysed the relationship between people and flora and fauna species (native and exotic) through YouTube videos of a forest region (southern Patagonia) posted by visitors from different parts of the world. We characterised the species of flora and fauna observed by the visitors and calculated the time that appears in the videos as a proxy for their connectedness to biodiversity. The biodiversity observations were contrasted against visitors’ sociodemographic characteristics (age and gender) by the Van der Waerden test and multivariate analyses. We created a sociogram that showed connections among species through visitor links to these data. Our results reveal different degrees of relationship between species, where some exotic ones were more preferred than natives (Van der Waerden test p = <0.100). Differences in the linkage to the flora and fauna species were related to the age and gender of the visitors. Visitor interests are modulated by access to ecosystem types (e.g. forests) and species’ commonness/rarity and docility. Gender and age had less influence on the interests than expected, but it determined differential values on native and exotic diversity. Three groups of species emerged from the sociogram based on the visitor connectedness to flora and fauna species, evidencing high connections among native trees (Nothofagus spp.), exotic beavers (Castor canadensis), and native geese and ducks (e.g. Chloephaga picta). The novel approach utilised provides valuable data that can be used to test the influence of gender and age on the different biodiversity interests. This information has potential applications for nature conservation by detecting unsponsored biodiversity and ecosystem types that could be promoted, capturing the visitor interests, and improving the offer of visitor activities according to gender/age observations.


Introduction
Biodiversity contributes to ecosystem services essential for human well-being (Acharya et al. 2019).It is also attractive for recreation and tourism, constituting an important cultural ecosystem service (Tolvanen et al. 2020).Visitors' interests on biodiversity are often recorded in photos, paintings, and drawings, and engraved in memory (Nobre 2001;Grzyb et al. 2021).With the access to technological innovations and the development of social media, people share their interests vertiginously (Krause et al. 2011;Monkman et al. 2018).For example, a video posted on YouTube social media may contain a user's record of hundreds of species simultaneously (Huertas Herrera et al. 2021).These interests can be linked with the sociodemographic characteristics of visitors (e.g.gender and age) (Ryan and Harvey 2000;Ma et al. 2018), as a proxy of the underlying forces of motivation and satisfaction obtained from biodiversity (Lee et al. 2010;Deng and Liu 2021;Grilli et al. 2021).
One of the most reliable predictors of attitudes and empathy for animals is gender (Paul 2000;Herzog 2007).Non-human species are ubiquitous to human life and permeate a diversity of social contexts by offering, among other things, entertainment, leisure, and companionship (Amiot and Bastian 2015).According to Graça et al. (2018), women are typically more concerned than men about the well-being of non-human animals and the environment.Also, women are more selective than men (Trivers 1972), and often have marked preferences and requirements for certain physical traits (e.g.height).These preferences may also be related to different hormone levels (e.g.oxytocin) in women compared to men, which influences how much affection and empathy they have towards non-human species (such as some mammals) (Miller et al. 2009;Barchi-Ferreira and Osório 2021).For instance, gender differences in earlier studies (e.g.Cooper et al. 2015;Soga et al. 2016;Hosaka et al. 2017) suggest that women can be interested in particular animals (e.g.birds or mammals), whereas men can be more interested in other sorts (e.g.insects or fish).Besides, young people are usually more interested in birds, insects, and reptiles, while adults are generally more interested in birds, mammals, and fish (Nord et al. 1998;Soga et al. 2016;Rosa et al. 2018); Ward et al. (1998) also found that both adults and children preferred large-bodied animals.Direct experiences with nature are considered efficient in promoting positive attitudes toward biodiversity conservation (Zhang et al. 2014).By analysing the visitor interests recorded by themselves, we can also trace connections (e.g.species that most attracted the attention of visitors) between the observed species with socio-ecological system characteristics (e.g.cultural and biodiversity attributes) (Huertas Herrera et al. 2021).Such links can be implemented into conservation strategies, ecosystem management and tourism (Ma et al. 2018).Regarding conservation management, this information can be a tool to increase positive attitudes towards wildlife influenced by gender and age (Bencin et al. 2016;Akasaka et al. 2022), identifying preferences for visited places regarding its characteristics (Ho et al. 2005;Dhungana et al. 2022).For example, for park managers, this could be useful for educational initiatives by including more women and or training young people with proactive attitudes on management plans.Women's participation in natural resource management groups increased collaboration, solidarity and conflict resolution in groups and increased groups' ability for self-sustaining collective action (Westermann et al. 2005).
In this context, social media are a way of understanding cultural phenomena related to nature conservation (Roberge 2014;Toivonen et al. 2019).But, why so much emphasis on social media?On the one hand, social networks can orient human behaviour in favour of conservation (e.g.environmental protection campaigns) and incite changes in public policies (e.g.www.change.org)within the sustainability framework (Tenkanen et al. 2017;Bergman et al. 2022).On the other hand, social media are a huge data distribution that may contribute to understanding cultural trends towards biodiversity (Di Minin et al. 2015;Hausmann et al. 2017) by observing public perceptions for achieving better conservation strategies (Soriano-Redondo et al. 2017); that is, to study a prism of human-nature interactions (Wood et al. 2013;Di Minin et al. 2015;Jarić et al. 2020).Such is the case with 'culturomics' (e.g.https://www.conservationculturomics.com/)(Ladle et al. 2016;Jarić et al. 2020), which seeks to understand human culture through the analysis of social networks using, among other things, frequencies used of words or photographs (Wood et al. 2013;Ladle et al. 2016).In the case of YouTube, the content of the videos can be examined to understand which images and messages resonate with a wide audience, thus informing managers about future conservation efforts (Vins et al. 2022).This reveals cultural patterns derived from data mining, where social media content is in fact a way of studying people's attitudes about biodiversity (e.g.iEcology, https://www.i-ecology.org/)(Hausmann et al. 2017).This is a trend that includes cultural aspects such as the valuation of ecosystem services (Richards and Friess 2015;Gould et al. 2019), where the adoption of different research approaches to cultural ecosystem services can contribute to solving real-world problems in managing human-nature interactions (Milcu et al. 2013).
Thus, understanding people's interests is fundamental to the management of nature for conservation (Daniel et al. 2012;Mccallum and Bury 2013;Wood et al. 2013).Data is so far available to study patterns of behavior and tourist visits (Tenkanen et al. 2017).Sociodemographic factors also can help in understanding visitors' interests.For example, people's empathy and mobility patterns (the way people move and spend their money) have been linked to several sociodemographic factors, such as age and gender (Lenormand et al. 2015;Sommerlad et al. 2021).The experience with nature indeed has been associated with socioeconomic status, gender, and age (Bratman et al. 2019).In this context, social media has the potential to widely disseminate conservation messages and be a powerful tool for mobilising social change to conserve biodiversity (Bergman et al. 2022).And, given the large volume of data that exists in social networks (e.g.videos, photos, comments) can largely contribute to guiding what strategies for conservation, for whom and where (Roberge 2014;Huertas Herrera et al. 2021;Akasaka et al. 2022).Our study, which builds on the findings of these prior studies, provides evidence in favour of the need to measure screen time in front of species of flora and fauna to capture the visitor's interest in the region's biodiversity.Moreover, it can contribute to a novel and practical tool for evaluating the relationships between people and biodiversity.
This study aims to analyse people's interests in flora and fauna species by tracing the video records of visitors in a forest region in southern Patagonia, Argentina.Central questions for this study were: (i) Which species origin (e.g.native or exotic) was most preferred by visitors?, (ii) Do these visitor's interests differ according to their age (youth or adult) and gender (men or.women)?, and (iii) Based on visitor interests, how the video-captured species are related, and how these link among species could be related to some species features (e.g.life forms of plants, animal types)?We expect to understand present or emerging human-nature interaction in one of the most sensitive biomes on the planet.We also expect to identify gender and age influence on biodiversity interests, since gender/age distinction could help to improve conservation management as well as specific oriented tourism offers, e.g.gender or age preferences for diverse species may be useful for evaluating visitor satisfaction with biodiversity in order to develop long-lasting empathy with non-human species among locals for nature conservation.In this study, we considered biodiversity to all species of native and exotic (species introduced intentionally or accidentally by humans) flora and fauna.Not all exotic species are invasive, but some are naturalised (e.g.Taraxacum officinale) and become attractive species for visitors; e.g.horses and cattle are exotic domestic animals in the study area, but they are not invasive species.

Study area
This study was conducted in the city of Ushuaia (Argentina) and its surroundings (e.g.Tierra del Fuego National Park) at southern Patagonia (54° 40' to 54° 53' S, 67° 54' to 68° 36' W; Figure 1(a)).It comprises one of the last well-conserved wilderness regions on the planet (Watson et al. 2018), where Nothofagus trees conform to the southernmost forests in the world.The area covered by this study was about 1,500 km 2 with elevation ranging from 0 to ~1,500 m a.s.l.The vegetation zones host woody habitats such as temperate deciduous and evergreen Nothofagus forests and open ecosystems such as grasslands, shrublands, and peatlands (Toro Manríquez et al. 2020).This area is visited by thousands of visitors each year, that access to several natural places near the city and carry out many touristic activities (e.g.trekking, camping, bird watching) in contact with the natural landscapes, as well as flora and fauna species without further economic restrictions (e.g.sighting of marine biodiversity which is carried out by boat).For more details, see Huertas Herrera et al. (2021).

Sampling design and data collection
We reviewed 100 videos on the YouTube platform (www.youtube.com)posted between 2010 and 2020 by visitors from different countries worldwide (Appendix A).To do this, we searched videos based on the keywords 'Tierra del Fuego National Park' and/ or 'Ushuaia Tierra del Fuego'.We translated these keywords into the 109 languages found in Google Translate (www.translate.google.com) to capture videos from any visitor of any country of the world.Then, we did a visual analysis of the content of the videos, estimating the screen time (seconds) in front of species of flora and fauna (Figure 1(b)).For videos that simultaneously show multiple biodiversity categories at the same time, the time span was assigned to each species equally.Flora included species of trees, shrubs, cushions, graminoids, forbs, bryophytes and lichens, hemiparasites, and fungi, while fauna included wader, geese and ducks, grebes, woodpeckers, falcons, passerines, parrots, herons, vultures, rodents, rabbits, cats, canids, odd-toed ungulates, and bumblebees.The species composition of each biodiversity category and the occurrence frequency in the analysed videos are listed in Appendix B. Then, both flora and fauna categories were split according to the species origin into native or exotic flora and fauna.We focused only on land and/ or coastal flora and fauna species, but not on marine biodiversity because not all visitors had access to catamarans or cruise ships.We determined sampling size (quantity of videos) calculating data video representativeness.We used 'species accumulation curves' (Chao1 method) obtained from EstimateS 9.0 (Colwell et al. 2012) software to estimate the representativeness.As the species accumulation curve stabilises between 85 and 95, we considered 100 videos a good sampling size for our study (tourists will no longer record new many species).Indeed, we slightly boosted our sampling efforts to identify a higher number of rare species, as some are difficult to observe (e.g.mammals such as foxes).Therefore, in the case of the study area, it is a representative sample of local biodiversity that tourists recorded on video (i.e.virtual biodiversity); certainly, it is not analogous to local biodiversity representation from a conventional ecological research perspective, where to exist traditional methods to study the landscape species assemblage (e.g.alpha, beta, gamma diversity analyses).There will undoubtedly be rare species; however, this research focused on measuring species that tourists were interested in seeing.Field verifications were carried out between November 2019 and February 2020 to confirm the presence of such flora and fauna (screenshot of the species) in the studied zone, and to distinguish species with similar characteristics (e.g. two or more species of herbaceous plants with similar flower characteristics, such as colour).Vascular plant identification follows Moore (1983) and Correa (1969Correa ( -1998, bryophyte identification follows Müller (2009) and Hässel de Menéndez and Rubies (2009).Lichens, fungi, and fauna identification follows SIB Parque Nacional Tierra del Fuego (APN 2018), and bird identification follows Narosky and Yzurieta (2010).For sociodemographic characterisation of people, we categorised the visitor ages into young (18-35 years) and adults (>36 years).For the gender characterisation, we considered two categories: men and women.To obtain this information, an extensive investigation was conducted using the YouTube platform (e.g.video comments), complemented with Facebook (Facebook Inc., United States), Instagram (Facebook Inc., United States), or Google (Alphabet Inc., United States) (see more details on methodology in Huertas Herrera et al. 2021).We only evaluated videos where it was evident who was shooting the biodiversity shots.The sociodemographic variables could be defined because the same person who created the video also created the YouTube channel.Besides, the video per se, the YouTube channels provide additional information, such as profile pictures, links to other social networks, and comments from friends, which allows knowing the age and gender of the person.There were videos of couples or groups of people.In those cases, if there was not a person who clearly recorded the video and who could be assigned sociodemographic attributes, the video was not considered.We only considered videos in which it was possible to identify the age and gender of the person who made them, and we discarded others in which there was doubt about the identity of the person.Other sociodemographic information (e.g. annual income, occupation, education) could not be accurately determined from the videos, therefore it was not possible to be considered in this study.

Data analysis
Descriptive statistical analyses were done in terms of frequency of occurrence (%) to identify which species were most observed by visitors.The estimated screen time (seconds) for different species was calculated in percentage over the total filmed biodiversity time of each video.Then, the mean percentage and standard error (SE) of the estimated screen time were used to elucidate the visitors' interests for native or exotic flora and fauna, according to their age (youth or adult) and gender (men or women).We used the Van der Waerden test (VW) to evaluate statistical differences according to visitor's groups (e.g.men vs. women, and youth vs. adults) and the origin of the species.We used InfoStat software (Di Rienzo et al. 2018) to run these analyses.Then, we used a multivariate statistical ordination technique to graphically represent and highlight the interest of tourists, by age or gender, considering flora and fauna origin.We used the Detrended Correspondence Analysis (DCA) with axis rescaling, through a matrix of 100 rows (videos) and three columns (species origin categories) that were (i) native flora, (ii) exotic flora, and (iii) native fauna and fauna.We selected DCA because it considers simultaneously the variability of the rows and columns (Hill and Gauch 1980;Ludwig and Reynolds 1988), and due to the compositional nature of data.It enables analyses of species and sampling units simultaneously, allowing the evaluation of the species interactions in a single analysis.In this context, DCA is particularly useful for dealing with complex ecological problems, such as the variation of local biodiversity with visitors' interests.The DCA analyses were performed using PC-ORD software (McCune and Mefford 1999).
To trace the connections between the species according to the visitor interests, we create a sociogram using the Gephi 0.9.2 (Bastian et al. 2009) (https://gephi.org).
To do this, we constructed an adjacency matrix using the analysed videos' estimated screen time data for different species (listed in Appendix B).Among the characteristics of the sociogram, we considered: (i) the size of the network (n nodes and n links); (ii) the average degree of links to identify species with many or few connections; (iii) the average degree with weights to measure the average number of species connections; (iv) modularity to detect groups within the network; and (v) the centrality of the vector itself (eigenvector centrality) to measure the importance of the species considering others with which they are linked.The distribution algorithm was Force Atlas 2. Such algorithms function as minimal energy algorithms, where nodes (each species) were repelling loads and links try to attract them with those joined nodes.These allow the grouping creation of species in 'groups', which can be primary (the most important 'groups' in the network) or secondary (the least important 'groups' in the network).

Results
A total of 24 young visitors and 76 adult visitor videos were analysed, where 20% were filmed by women and 80% by men.The entire flora and fauna species recorded included a number (n) of 57 species, 81% native and 19% exotic species (Appendix B).The species accumulation curve was useful to calculate the representative number of species based on the interests of tourists (e.g. which species are more identified in the videos, or which species captured more attention and more time in the videos).The species accumulation curves showed that videos were enough to capture most of the species of the studied area (Appendix C), which was stabilises between 85 and 95 videos.Considering flora, all trees, shrubs, cushions, hemiparasites, fungi, bryophytes, and lichens were natives, while exotic species were belonging to forbs (n = 4) and graminoids (n = 1).All bird species were natives, while recorded mammals (n = 6) and insects (n = 1) were exotic.The species with the highest frequency of occurrence (%) were Nothofagus native trees (83%), the exotic beavers or their impacts (e.g.dams of Castor canadensis) (44%), and many species of geese and ducks (mainly upland goose, Chloephaga picta) (41%) (Figure 2).
Visitor interests varied for native and exotic flora and fauna, with the general greatest values for native flora, and the lowest for exotic flora (Table 1 and Figure 3).VDWT showed that demographic characteristics did not greatly influence visitor interests (p > 0.217).However, there was a difference between young and adult interests for native flora (p = 0.092), with more videos from youths than adults, as well as young/adult and men/women interests (p = 0.045 and 0.006, respectively) concerning exotic fauna, showing adults and men greater interests than youth and women.
The DCA analyses (total variance = 0.9377) highlighted which combinations of biodiversity attributes were more registered by visitors of different ages or gender (Figure 4).According to age, young visitors were more interested in native flora and fauna than exotic ones: no video of young visitors captured only exotic flora and fauna, and almost all of them registered some native plants (except 4 videos that only filmed native fauna).Meanwhile, adults showed a higher interest in native flora, more often combined with exotic flora and fauna.Both genders were most interested in all aspects of biodiversity, e.g.mostly linked with native plants, however, men usually combined those observations with exotic fauna and flora, and women were more interested in combine native flora and fauna.
The sociogram constructed with the entire database had 57 nodes (species) and 513 vertices (connections) (Figure 5).Each species was connected with 18 other species within the sociogram graph.The species with the most links were Nothofagus trees (307), beavers (193), and upland geese (169).Three different groups were detected, the first group (located in the middle of the network) with 52.6% modularity, the second with 24.6% modularity, and the third with 22.8% modularity.The first group was integrated by 30 species, where the most important were Nothofagus trees, beavers, upland geese, Cyttaria harioti (fungi), Chiliotrichum diffusum (shrub), Protousnea spp.(bryophytes and lichens), and Equus ferus caballus (odd-toed ungulates) (Figure 5).The second group was conformed of 23 species, where the most important species were Lycalopex Table 1.Van der Waerden test results to evaluate differences among the visitor's interests for the species origins (native vs. exotic) according to their age (young vs. adult) and gender (men vs. women).

Visitor interest and biodiversity valuation
Visitors can provide essential records of the places that they visited (Hausmann et al. 2016;Deng and Liu 2021).In our study area, the Nothofagus trees were the most linked by the visitors (particularly young and women, Figures 3 and 4).This could be due to Nothofagus trees are the dominant and iconic vegetation in southern Patagonia.These trees generate forests with unique characteristics that can be attractive to connect visitors with nature (Martínez Pastur et al. 2017).For example, Nothofagus trees can be deciduous or evergreen (Toro Manríquez et al. 2021), so the colour of their foliage change across the seasons (from green to red, orange, yellow, and brown), producing beautiful and eye-catching landscapes.Complementary, several outdoor activities such as hiking or camping can be performed in these forests (Martínez Pastur et al. 2017).Forests also provide the opportunity to experience other valuable perceptions regarding nature.For example, the Yatana urban forest, one of the most important green areas in the study area, connects cultural and natural heritage.The visitors experience the art and ancestral knowledge of aboriginal culture (Yamana people) in the natural environment of the Nothofagus forest.Another example is the therapeutic use offered by forests, such as shinrin-yoku (forest bathing), which suggests that vigour is more enhanced by walking in forest environments, reducing anxiety and depression problems, among others (Kobayashi et al. 2021).Our results showed how much time biodiversity items are displayed on social media.The three most frequently filmed species (Nothofagus trees, upland goose and beavers), all included in the most important group defined by the sociogram, occur together in the same ecosystem type.This is due to beavers constructing dams in rivers near or within Nothofagus forests.They obtain woody material, forming meadows and lagoons that upland goose currently uses.According to this, accessing of visitors to other less visited ecosystem types (e.g.grasslands, peatbogs, alpine environments) could change visitor interests in flora and fauna species.Birds (e.g.wader, ducks, grebes) and also the fauna provide identity to the visited places, as in this study they were all native and/or migratory.By contrast, all registered mammals and insect fauna were exotic in the video analyses.This could be explained because some native animals are hard to follow or to observe (Krause et al. 2011), but it can also be due to the charisma of some exotic mammals (Hausmann et al. 2016) or to the impressive transformation of the natural landscape that they generated (e.g.ecosystem engineers such as Castor canadensis).This situation may curb people's (e.g.young people) concerns about local biodiversity (Ballouard et al. 2011).In this context, although one of the significant threats to forest biodiversity conservation is the introduction of exotic species, some often have attractive qualities for visitors.This suggests that because of their marked charisma or docility with the humanbeing (Albert et al. 2018), some exotic species are more attractive than native, e.g.horses that are especially likely to be cultural icons that provoke strong charismatic reactions in the visitors who see them (Notzke 2014).
The visitors' age and gender reveal heterogeneous patterns of interests in flora and fauna species, showing a relationship between the native or exotic species, and the sociodemographic characteristics that can be useful to enhance the visiting experience or encourage future visitation (Ma et al. 2018;Zhu et al. 2021).This information can serve to ecosystem managers interested in generating new proposals that motivate and satisfy visitors by age and gender.For instance, Hausmann et al. (2016) found that young women belong to 'other biodiversity experience seekers' visitor groups with a higher perception of the sense of place.This is consistent with our study and can explains the women interests in native vegetation.Besides, men and women differ in their perceptions of nature conservation (e.g.environmental risk, vulnerability to climate change) (Westermann et al. 2005;Nat Clim Change Editorial 2019).In this context, the information on wildlife influenced by gender and age can help park managers better conserve local biodiversity through, for example, preparing a guidance strategy for integrating visitors' wellness motivations with managing protected areas (Ednie et al. 2022;Thi Khanh Chi 2022).
Previous studies showed that the age and gender of visitors could explain the willingness to pay for nature conservation (Martín-López et al. 2007;Cerda et al. 2013).Women tend to express fewer interests in invertebrates and predatory mammals than men (Castillo-Huitrón et al. 2020).Other studies showed that iconic species can be used to attract visitors and promote conservation (Walpole and Leader-Williams 2002).In fact, there may be a greater interest in mammals, birds and plants than for inconspicuous species (e.g.insects) (Albert et al. 2018;Espinosa-Molina et al. 2021).Therefore, visitors can look for something certain, which is what they capture in the videos.For example, the colours of the Nothofagus trees, geese, or even the so-called invasive and problematic beavers, as well as the commonness/rarity, colour and docility of the observed species.

The network analysis of visitors' interests
Particularly in our study area, the network analysis revealed that Nothofagus trees have many connections with other species, but C. canadensis and C. picta would also link them to many other species, and because the species with which they relate are also important.Therefore, this kind of sociogram may differ worldwide, as visitor interests depend on local flora and fauna.It is crucial to consider the scale of the network under analysis (e.g.landscape, region, country) because more ecosystems could contain more species at a larger scale.For instance, we do not include marine biodiversity in our study, which has charismatic species (e.g.whales, dolphins, sea lions) and would probably enlarge our sociogram by generating more groups of related species.Besides, this methodology also implies several weaknesses to consider (Krause et al. 2011;Monkman et al. 2018); e.g.places with not enough visitor videos cannot be adequately evaluated (Huertas Herrera et al. 2021).
Visitor observations can be crucial to understanding better the visitors' interests in some particular species (Grzyb et al. 2021).In this context, the network of connections between interests related to flora and fauna species could help identify people and nature's liking at a regional scale (e.g.associations between species from a place valued by visitors).For example, visitors with greater knowledge about some unknown, uncharismatic, or inconspicuous species could change their interests, drawing more attention and interest to these taxa.A practical example is the 'miniature forests' initiative promoted in Omora Park (Navarino Island, Chile) (Rozzi et al. 2006;Tauro et al. 2021), where the visitors are guided to discover the microcosmos of bryophytes, liverworts and lichens using specific trails, facilitating particular tools to the visitors (e.g.lens).
Tourism managers who do not know well the needs of visitors will hardly be able to arouse people's interest during the trials (Zunino et al. 2020).
A common problem with nature-based tourism is that it only offers what is naturally available, not what is most interesting to visitors.It is necessary to consider whether visitors are more interested in inconspicuous species (Vins et al. 2022).The conscious design of a tourist identity and a place is that it can improve the tourist experience with visitors more aligned with the values of the destination.This implies changing the idea of offering ecosystem services based on the potential supply and not on the demand, which arises from making an unplanned tourist offer to visitors, e.g.what experiences they are looking for.Therefore, local identity based on the local species (biodiversity) and cultural-natural heritage is a key element that demands the knowledge and understanding of the potential ecosystem services that the place offers (Satz et al. 2013;Hølleland et al. 2017).

Potential applications of the methodology employed in this study
We use web data to explore the connectedness of visitors and flora and fauna species.In the context of nature conservation, this study can be relevant for: (i) environmental education and policy.For example, visitors' attitudes towards biodiversity can be used for sustainable ecosystem management and/or policy purposes.In particular, visitors' interests can be helpful in monitoring environments invaded by problematic species.(ii) A social media could show some of likely the connections that regional wildlife has to identify socio-cultural patterns that may be of political interest in nature conservation.(iii) Through the sociogram constructions it is possible to define degrees of travel motivation for native environments that can be described, reflecting visitors' satisfaction from recorded biodiversity of the places.This could reveal conservation actions to consider, such as visitor routes and promotion of the cultural customs.
Preliminary studies showed the importance of contrasting the data obtained in social networks with an on-site interview (Hausmann et al. 2017).This is one of the limitations of the present study, because we did not contact the analysed visitors, or because we did not randomly interview tourists that could validate the obtained information.Thus, our results are based on an indirect method of inquiry about visitors' preferences.However, the survey method is often expensive and has limitations in terms of spatial and temporal scope (Wood et al. 2013;Richards and Friess 2015), which is highly relevant in case of situations such as the COVID-19 pandemic.On the other hand, this work only focused on data from a single social network (e.g.YouTube) and did not consider others such as Instagram, Flickr or Twitter.Consulting multiple social networks could broaden the understanding of visitor preferences and more diverse social groups (e.g.professional photographers using Flickr), but also produce biases according to more use of some social media by some group of people (e.g. more young people using Instagram than adult people).This study shows preliminary results about tourist preferences for a forest region (i.e.southern Patagonia) that could be better confirmed in much larger studies with a greater number of platforms, or with surveys sent out to users to determine their age/gender.However, these are still important to begin to understand the tourist motivations and how to explore them.Despite these limitations, on YouTube finding films using keywords (e.g.Tierra del Fuego National Park, Patagonia) is simple, allowing the experiment to be replicated without the need of Twitter or other social media.Furthermore, most of the videos that have been shared have included helpful information about the user (e.g.visitor's country of origin).

Final remarks
Social networks are data sources that help to improve human understanding of natural systems.Visitor's observations posted on YouTube can be used to trace links between observed species and people, and these connections could lead to ecological management systems based on social media user observations.We empirically show that the most important species of the visitor interests of this study were plants (e.g.native Nothofagus trees), birds (all native species), and mammals (all exotic species), but also fungi, bryophytes and lichens which were important for the visitors.The visitors' interests in flora and fauna species were related to age and gender.Besides, from the visitor observations of the biodiversity, it is possible to create a network that shows the relationship between people and nature.This information is an example of easily accessible data mining that has a practical application to studying human behaviour, animals, and plants, with their respective connections, and a proposal for the analysis of regional biodiversity assessment.This information also provides opportunities to engage visitors with flora and fauna and help local park managers better conserve biodiversity (e.g.increasing positive attitudes towards wildlife influenced by gender and age).

Figure 1 .
Figure 1.The study area in southern Patagonia, Argentina (a) Illustration of the video collection (b) where a visitor records a flora species biodiversity (Nothofagus trees).The red line represents the screen time (seconds).

Figure 3 .
Figure 3. Mean percentage and standard error (vertical bars) of the screen time (seconds) of the flora and fauna species recorded by visitors, considering the visitor age (youth vs. adult) and gender (men vs. women).Different letters show differences between groups based on the Van der Waerden test at p < 0.100.

Figure 4 .
Figure 4. Detrended correspondence analysis (DCA) comparing the visitor interests for the species origins (native flora, native fauna, and grouped exotic flora and fauna) based on the score of visitor's age (a) and gender (b) Black arrows indicate the number of videos capturing only one biodiversity aspect, by age or gender category.

Figure 5 .
Figure5.The sociogram shows the nodes (species) and the connections between species obtained from visitor interests.Group 1 (the most important in the network) is shown in purple, 2 in green, and 3 in pale blue (most peripheral).The importance of some species and interactions in the network is reflected as a function of size.For more details see Appendix B.