Social media reveal that charismatic species are not the main attractor of ecotourists to sub-Saharan protected areas

Abstract

Charismatic megafauna are arguably considered the primary attractor of ecotourists to sub-Saharan African protected areas. However, the lack of visitation data across the whole continent has thus far prevented the investigation of whether charismatic species are indeed a key attractor of ecotourists to protected areas. Social media data can now be used for this purpose. We mined data from Instagram, and used generalized linear models with site- and country-level deviations to explore which socio-economic, geographical and biological factors explain social media use in sub-Saharan African protected areas. We found that charismatic species richness did not explain social media usage. On the other hand, protected areas that were more accessible, had sparser vegetation, where human population density was higher, and that were located in wealthier countries, had higher social media use. Interestingly, protected areas with lower richness in non-charismatic species had more users. Overall, our results suggest that more factors than simply charismatic species might explain attractiveness of protected areas, and call for more in-depth content analysis of the posts. With African countries projected to develop further in the near-future, more social media data will become available, and could be used to inform protected area management and marketing.

Introduction

Protected area management is crucial to enhance species persistence and reverse the biodiversity crisis1. Resources for protected area management are woefully inadequate2. Nature-based tourism can help generate funding needed to cover important management costs in protected areas3,4,5,6,7. Ecotourism, particularly, has been long promoted for its importance in supporting both biodiversity conservation and economic development8, 9. Sub-Saharan Africa is one of the top ecotourism destinations in the world10. Charismatic megafauna, such as the Big Five (lion - Panthera leo; leopard - Panthera pardus; elephant - Loxodonta Africana; buffalo - Syncerus caffer; black and white rhino - Diceros bicornis and Ceratotherium simum)4, 11, 12 or primates (gorilla spp. - Gorilla gorilla and Gorilla beringei beringei; chimpanzee - Pan troglodytes)13 are considered the main attractor of ecotourists to sub-Saharan African protected areas. Besides supporting management activities, funding from ecotourism can also help reduce important costs human communities pay for living alongside charismatic, yet dangerous, species14.

Besides the presence of charismatic megafauna, a wide range of other characteristics underpin nature-based tourism in protected areas15, 16. These include factors such as broader biodiversity (e.g. species richness17; threatened species and habitat types18; less charismatic biodiversity11, 19) and aesthetic of landscape (e.g. vegetation quality20). Geographical factors, such as accessibility (e.g. travel time15; trails and roads21), or degree of human influence (e.g. cultural landscapes20; overcrowding19) are also considered important. Furthermore, the socio-economic conditions of a country (e.g. political stability) also affect ecotourism visitation22, 23.

Thus far, studies assessing factors affecting tourists’ visitation patterns have focused on protected area visitation statistics15, 23. However, information about visitor numbers are generally costly (e.g. through survey-based methods) or difficult to collect (e.g. most parks are open access for recreation)5. Therefore, this information is often available only for few, well-known, protected areas. Alternatively, social media are increasingly being used as a cost-effective and rapid way to explore tourists’ visitation patterns24,25,26 or hotspots of human activities27. While data on visitor numbers can be scarce5, social media data are widespread and can, in some cases, be used as a proxy for tourism visitation rates24, 27. Therefore, data from social media can potentially be used as a new way to investigate which factors affect protected area attractiveness at continental or even global level. This is the challenge we addressed here.

In this study, we explored which socio-economic, geographical and biological factors explain social media use in sub-Saharan protected areas. Particularly, we were interested in understanding whether the number of charismatic species was an important contributor to social media usage in protected area. To do this, we used georeferenced Instagram pictures, posted within sub-Saharan African protected areas during 2015 to explore the effect of potential biological (i.e. richness of charismatic megafauna, richness of other biodiversity, vegetation cover), geographical (i.e. accessibility, elevation, population density) and socio-economic (i.e. Human Development Index [HDI]) factors on the density of active users, posts and likes (see framework in Fig. 1). In particular, we used generalized linear models with site- and country-level deviations to explore 1) which protected area and country level factors affect the use of social media; and 2) whether different explanatory variables explained the three response variables (i.e. the density of users, posts and likes). A total of 969 protected areas located in 41 countries (Table S1 in Appendix S1), for which social media data were available, were included in the analysis (Fig. 1). For almost half of the countries, we mined social media from protected areas covering ≥ 50% of the total area designated as protected (Figure S1 in Appendix S1). A total of 92,832 posts, posted by 55,756 active users, and liked 6,373,836 times were analyzed.

Figure 1
figure1

Logical framework of the study. For each protected area with data available from social media, biological (green arrows), geographical (orange arrows) and country level (blue arrow) attributes were also obtained and used in the generalized linear model as explanatory variables. Maps were created in QGIS 2.8.1 (URL http://www.qgis.org/en/site/). All images were generated by the authors.

Results

The 6 top-ranked models, for each of the three response variables, describing the use of social media in protected areas are summarized in Table 1. The most important variables affecting social media usage across all models were HDI, accessibility, population density and the mean vegetation cover (Fig. 2). The country-level HDI was the strongest predictor, with coefficients up to four times higher than the other variables, across all models (Fig. 2). The positive sign of the coefficient indicates that social media usage was higher in protected areas in more developed countries. Accessibility was the second strongest variable predicting density of active users and posts. Specifically, accessibility had a negative sign, meaning that social media usage was higher in more accessible protected areas (Fig. 2). Moreover, population density was positively affecting the use of social media across all models (Fig. 2), meaning that social media use was higher in protected areas with higher density of people living around them. Vegetation cover had a negative sign, meaning that protected areas with more dense vegetation had lower social media use, and, in particular, pictures received fewer likes (Fig. 3). While less charismatic species (other biodiversity) was found to be less important compared to the other variables (Fig. 3), it was found to be statistically significant for number of active users in protected areas (Fig. 2). Specifically, less charismatic species (other biodiversity) had a negative sign, meaning that protected areas with higher species richness had lower densities of active users. The other variables considered in the models were less important (Fig. 3).

Table 1 Top-ranked predictors of social media usage within Sub-Saharan Africa protected areas.
Figure 2
figure2

Beta coefficients of best predictors, averaged among the 6 top models of each response variable explaining use of social media in protected areas. The red bars show the confidence interval for each coefficient. The number over each bar are p-values and refer to the statistical significance. Figure S2 in Appendix S1 shows the values corresponding to this figure.

Figure 3
figure3

Overall weights of relative importance of 6 top predictors, averaged among top 6 models of each response variable.

For density of active users, the top-ranked model had an Akaike’s information criterion (AIC) weight of 0.75, explaining 54% of the deviance. For density of picture posted, the AIC weight was 0.78 and the deviance explained approximately 51%. For the density of likes the AIC weight was 0.93 but the deviance explained was 38%.

Discussion

Overall, we found that richness of charismatic species had no influence on the use of social media in sub-Saharan Africa’s protected areas. This means that the number of highly iconic species which can be potentially found in a protected area, did not affect protected area visitors’ posting on social media. Interestingly, protected areas with higher richness of other species had fewer users posting on social media. Meanwhile, other factors, including both the socio-economic condition of countries and the geographical characteristics of the site, were more important in explaining the use of social media in sub-Saharan protected areas. In particular, protected areas located in more developed countries, which were more accessible and with more people living nearby, had higher densities of active users and posts on social media. Finally, protected areas with more open vegetation had higher densities of likes on social media.

While large-bodied mammals are considered the most important flagships for conservation in sub-Saharan Africa11, 28, we found that their presence did not affect the amount of active users, posts and likes on social media across sub-Saharan Africa’s protected areas. Besides charismatic wildlife-viewing, many tourists may also prefer visiting protected areas for their cultural, recreational value29, and visit places which allow for activities, such as hiking or biking, which are normally forbidden in parks where charismatic, dangerous animals are present21. Other studies show that, when looking at the content of pictures shared across different types of nature-based destinations, a variety of cultural uses, including recreation and aesthetic appreciation, are the most common subject among pictures30. In accordance, we found that areas with open vegetation (generally attractive to people as they allow views in the distance31), had higher use of social media, and received more likes, across different protected areas in sub-Saharan Africa. Viewsheds are key aspects of the visual landscape affecting visitor’s experiences32 and part of the sense of place sought by tourists in protected areas33. In addition, while in other regions (e.g. Finland18), or contexts (e.g. people expressing willingness to visit17), biodiversity appeared to underpin tourism attractiveness of protected areas, we found that areas with higher richness of species had fewer users active on social media. In our study area, higher species richness is found in moist tropical forests34 of central Africa’s countries, where protected areas receive fewer tourists due to less developed infrastructures (e.g. roads, cellphone coverage) and political or security issues35. Therefore, information about the use of social media in relation to the presence of species may be further explored in future studies. At the same time, content analysis of pictures may help reveal stronger relations between social media use and e.g. charismatic species.

The socio-economic condition of countries affects tourism patterns worldwide, with higher number of tourists visiting wealthier nations15. Similarly, we found that social media use in sub-Saharan Africa’s protected areas followed the same pattern, with more usage in countries with enhanced socio-economic conditions. Lack of provision of services and remoteness may discourage tourists’ visitation in the first place36. Moreover, gaps in mobile network coverage and the lack of smartphone devices may limit the geographical representativeness of social media data37, 38. As tourism expansion and economic growth of nations are interrelated39, social media potential will increase as many of these countries will also improve their development. Meanwhile, information obtained from more frequented sites, and from where data from social media is available, could be used as a first approximation for similar areas where data are scarce.

At a site level, our results show that variation of social media use across different protected areas well reflects tourist’s behavior in relation to geographical attractiveness of protected areas across sub-Saharan Africa. Similarly to previous studies, we found that better accessibility and higher density of people living nearby protected areas positively affect not only visitation rates15, 25, but also the use of social media. Highly populated areas around the borders of protected areas might also imply higher provision of tourists’ services and infrastructure36, including cellphone coverage40. On the other hand, such areas may be subjected to higher human pressure, such as environmental alteration, depletion of resources41, and threat to biodiversity, such as edge effect, especially in smaller areas42. Data from social media may be used to identify and monitor the use of sensitive locations by tourists, in order to inform conservation and management.

Different metrics of social media have been used to explore various aspects of tourists’ behavior, such as active users for assessing visitation24, amount of posts for exploring geographical hotspots of preferences27, 43 and likes to investigate engagement with specific subjects from the broader network44, 45. We found that all these metrics are affected by the same predictors is sub-Saharan Africa. However, vegetation openness was more important to receive more likes, while species richness was less important to explain higher densities of users. Deviance explained by our best models, especially for likes, suggests that other aspects not considered in this study may also influence the use of social media in protected areas. For example, individuals’ personalities and behavior on social media46, and the content of pictures47, may affect appreciation of pictures. Moreover, opportunities for biodiversity-related activities, such as hiking or camping, which were not considered in this study, might also be important aspects affecting social media usage, as they affect tourists’ decision-making19. However, posts on social media may not reveal the socio-economic background of different protected areas users. Future studies will require a more accurate differentiation between e.g. tourists, researchers, managers, and inhabitants. Future studies should also explore the profile of the social media users, e.g. by implementing deep learning algorithms, to overcome this limitation.

In conclusion, our results show that social media data can potentially be used as a first approximation to understand spatial preferences of tourists for nature-based experiences across protected areas in sub-Saharan Africa. Socio-economic development of countries and geographical characteristics of each site, and not the presence of species, were key aspects affecting visitation and attractiveness in sub-Saharan Africa’s protected areas. The potential of using social media data to inform conservation and ecotourism in sub-Saharan Africa will likely increase in the future, as some countries will improve their socio-economic conditions. Meanwhile, protected area managers and other conservation stakeholders in areas where social media are more commonly used, may take advantage of data uploaded by tourists to monitor the spatio-temporal variations of the use of cultural services, and inform conservation and ecotourism marketing. For example, social media data may help understand interests in biodiversity-related activities and be used to promote ecotourism in sites which lack charismatic species19. Content analysis of social media may help understand preferences for species48, and help identify more attractive protected areas in Africa, where ecotourism can be used as a tool to support conservation49. However, further analyses are needed in order to better understand the relationship between biodiversity and social media use in protected areas, including validating social media content with traditional surveys48, and exploring potential biases in the social media user population.

Methods

Study area and social media data

The framework of our study is presented in Fig. 1. We downloaded geo-referenced borders of sub-Saharan Africa’s protected areas from the World Database on Protected Areas (WDPA) (https://www.protectedplanet.net/ Accessed on June 2016). We considered all protected areas were data from social media was available (Fig. 1).

For each protected area, we collected geo-referenced pictures posted on Instagram within the border of the area (Fig. 1). Only sites over 10 square kilometers15 were considered in order to avoid biases related to social media location inaccuracy37. Data were accessed through the application programming interface (API) (https://www.instagram.com/developer/) available from the platform. Geo-referenced posts were sampled each first week of every month of the year 2015. We collected three metrics of social media usage, i.e. total number of active users (users who had posted at least 1 picture per day is counted once per day), posts (pictures), and likes of pictures posted in the area.

Potential predictors of social media use in protected areas

We selected 8 variables that are considered to affect tourists’ preferences for nature-based tourism in protected areas, and which could potentially be related to social media use (Table 2). These variables were site specific, i.e. biological and geographical, or country-specific, i.e. socio-economic (Fig. 1). All mapping was performed in QGIS version 2.8.1.

Table 2 Potential predictors used in the GLM to explain social media use by tourists visiting sub-Saharan Africa’s PAs.

Biological factors were considered in order to assess whether biodiversity or landscape-related variables affected the use of social media in protected areas (Fig. 1). Biodiversity variables were obtained by calculating richness (sum of species occurring in the area) of 9,916 species of vertebrates, invertebrates and plants, occurring in sub-Saharan Africa, for each site. Species occurrence maps were obtained from the IUCN Red list database (Accessed in May 2015), which is the latest updated source of information about species ranges that is also publicly available. However, range maps overestimate the true area occupied by species, as it may include areas where species presence is probable but not confirmed50. Charismatic megafauna, which are particularly attractive to tourists in sub-Saharan Africa11, were considered separately from other less charismatic species in order to explore whether the use of social media among tourists was affected by the richness of these species in protected areas. Charismatic mammals included 40 large-bodied mammal species, with average body weight larger than 100 kg51, 52. As other less charismatic biodiversity may also be attractive for different markets of tourists19, we grouped richness of amphibians (999 species), arthropods (750 species), birds (2246 species), reptiles (723 species), plants (603 species) and freshwater fish (3350 species), and mammal species (1245 species) with average body weight smaller than 100 kg together as “other biodiversity”. Moreover, we focused our analysis only on continental Africa, excluding Madagascar and other islands, as we wanted to assess the importance of large-bodied mammals.

Vegetation cover was considered as another biological factor as a proxy for landscape aesthetic. Open vegetation is a key aesthetic attractor for landscape preferences53, affecting tourists’ decision-making for nature-based experiences in protected areas19. We used MODIS Enhanced Vegetation Index (EVI) as a measure for vegetation cover (Table 2) in order to explore whether more open vegetation would also affect the use of social media in protected areas. EVI is optimized for characterizing vegetation state in areas with dense canopy. Data were downloaded for the period of February 2000 to December 2014 at 1 km resolution at the equator.

Geographical variables included accessibility, elevation and population density as potential predictors of social media use in protected areas. More accessible areas receive more tourists than remote ones15. In order to explore whether higher accessibility is also driving the use of social media in protected areas, we calculated mean accessibility values of a 10 km buffer zone, built around each protected area. Values were extracted from a global map of accessibility (Nelson 2008), developed by the European Commission and the World Bank (Table 2). Accessibility values represent the travel time, by land or water, from the nearest major city to each protected area (cities with 50,000 or more people in the year 2000)54. Units of time represents the “cost” of travelling where higher values are more costly and smaller values less costly, thus indicating better accessibility.

Elevation was considered as another geographical attribute, as tourists’ preferences for nature-based destinations may also be influenced by topography. For example, elevation (e.g. costal or mountain areas55) and slope (e.g. hiking opportunities56), may affect aesthetic of landscapes and preferences for nature-based experiences in protected areas. In order to determine whether altitude of protected areas may also affect the use of social media we used data from Aster Global Digital Elevation Model v002 (ASTG TM) at 30 m resolution to extract mean elevation values of each site (Table 2).

Moreover, density of population living nearby was also considered among the geographical variables, as tourists visitation rates is positively affected by population density15, 57. More populated areas are more likely to provide infrastructures, such as mobile phone coverage40, which can affect spatial patterns of social media use. In order to understand whether population density outside protected areas may also affect social media usage inside the area, we estimated mean population densities around a 10 km buffer zone built around each area. Population density values were extracted from the Global Rural-Urban Mapping Project, Version 1 (GRUMPv1) (http://sedac.ciesin.columbia.edu/data/set/grump-v1-population-density) and estimated at 10 km resolution at the equator (Table 2).

Finally, we considered the socio-economic condition of countries as potential predictor of social media use in protected areas. Gross domestic product of countries affects tourists visitation rates23, with fewer visitors in poorer and politically instable areas. While tourism is increasing worldwide, international tourism in Africa has decreased by 3% in 2015, due to slow economic growth and struggles related to health and security35. However, trends change among sub-Saharan Africa countries. In order to explore whether the different socio-economic conditions of countries would affect social media use in protected areas, we considered the HDI, developed by UNEP (Table 2). The HDI is the result of a geometric mean between three indexes of human development, i.e. life expectancy, education and gross national income per capita, in each country. The index was chosen in this study as it summarizes information about countries’ socio-economic condition, and represents an official indicator based on data sources provided by major statistical agencies of the United Nations.

Statistical analysis

We used an information theoretic approach58 and a generalized linear mixed effect model (GLMM) to explain the use of social media in sub-Saharan Africa’s protected areas. Three response variables, representing different metrics of use of social media, were used, namely density of social media active users, density of posts and density of likes in each protected area. Densities were calculated as number of active users, posts or likes per km2 of the area were they occurred. The GLMM accounted for both fixed and random effects. Biological (richness of charismatic and other biodiversity, and vegetation cover), geographical (accessibility, elevation and population density) and socio-economic (country-related human development index) were used as fixed effect in all the models. Due to high heterogeneity in the spatial distributions, between countries (Table S2 in Appendix S1) and protected areas (Figure S2 in Appendix S1), of our response variables, two levels, i.e. site (protected area) and country of each site, were used as random effects in the models. This is, in order to allow our models to account for spatial variability, by including regression coefficients which are constant across sites and countries. To fit our model, we used a binomial family type with logit-link distribution of errors. As values of the variables had skewness of distributions, all explanatory variables (charismatic megafauna richness, other biodiversity richness, vegetation cover, accessibility of the buffer area, elevation, population density of the buffer area and HDI of country), except vegetation cover (values range between 0 and 1) were log-transformed. We used the Corrgram package in R59, with a cut-off of r = 0.80, to test for correlation among explanatory variables. We only selected variables with the strongest effect on social media usage which were not correlated in order to avoid multicollinearity among variables. Next, we implemented multimodel averaging in the R version 3.0.260 package glmulti 61. Multimodel averaging58 is commonly used in ecology and conservation science to rank, based on the Akaike Information Criteria, all possible fitted models from best to worse and then averaging the coefficients values across models to reduce uncertainty. In addition, we measured the relative importance of the most important predictor variables62 by using Akaike weights over the six top-ranked models and a cut-off of w = 0.80. Percentage of deviance explained by each model was used as a measure of goodness of fit.

References

  1. 1.

    Watson, J. E. M., Dudley, N., Segan, D. B. & Hockings, M. The performance and potential of protected areas. Nature 515, 67–73, doi:10.1038/nature13947 (2014).

  2. 2.

    McCarthy, D. P. et al. Financial costs of meeting global biodiversity conservation targets: current spending and unmet needs. Science (New York, N.Y.) 338, 946–9, doi:10.1126/science.1229803 (2012).

  3. 3.

    Gössling, S. Ecotourism : a means to safeguard biodiversity and ecosystem functions? Ecological Economics 29, 303–320, doi:10.1016/S0921-8009(99)00012-9 (1999).

  4. 4.

    Goodwin, H. J. & Leader-Williams, N. in Priorities for the Conservation of Mammalian Diversity: Has the Panda Had Its Day? (eds Entwistle, A. & Dunstone, N.) 257–276 (Conservation Biology series, 2000).

  5. 5.

    Buckley, R. Parks and Tourism. PLoS Biology 7, e1000143, doi:10.1371/journal.pbio.1000143 (2009).

  6. 6.

    Buckley, R. C., Morrison, C. & Castley, J. G. Net effects of ecotourism on threatened species survival. PLoS ONE 11, 23–25, doi:10.1371/journal.pone.0147988 (2016).

  7. 7.

    Di Minin, E., Leader-Williams, N. & Bradshaw, C. J. A. Banning Trophy Hunting Will Exacerbate Biodiversity Loss. Trends in Ecology & Evolution 31, 99–102, doi:10.1016/j.tree.2015.12.006 (2016).

  8. 8.

    Goodwin, H. In pursuit of ecotourism. Biodiversity and Conservation 5, 277–291, doi:10.1007/BF00051774 (1996).

  9. 9.

    Krüger, O. The role of ecotourism in conservation: panacea or Pandora’ s box? Biodiversity and Conservation 14, 579–600, doi:10.1007/s10531-004-3917-4 (2005).

  10. 10.

    World Tourism Organization. Towards Measuring the Economic Value of Wildlife Watching Tourism in Africa– Briefing Paper. (2015).

  11. 11.

    Di Minin, E., Fraser, I., Slotow, R. & MacMillan, D. C. Understanding heterogeneous preference of tourists for big game species: implications for conservation and management. Animal Conservation 16, 249–258, doi:10.1111/acv.2013.16.issue-3 (2013).

  12. 12.

    Naidoo, R., Weaver, L. C., Stuart-hill, G. & Tagg, J. Effect of biodiversity on economic benefits from communal lands in Namibia. Journal of Applied Ecology 48, 310–316, doi:10.1111/jpe.2011.48.issue-2 (2011).

  13. 13.

    Leader-Williams, N. & Dublin, H. In Priorities for the conservation of mammalian diversity: has the panda had its day (eds Entwistle, A. & Dunstone, N.) 53–81. (Cambridge University Press, 2000).

  14. 14.

    Selier, S.-A. J., Slotow, R. & Di Minin, E. The influence of socioeconomic factors on the densities of high-value cross-border species, the African elephant. PeerJ 4, e2581, doi:10.7717/peerj.2581 (2016).

  15. 15.

    Balmford, A. et al. Walk on the Wild Side: Estimating the Global Magnitude of Visits to Protected Areas. PLoS biology 13, e1002074, doi:10.1371/journal.pbio.1002074 (2015).

  16. 16.

    Deng, J., King, B. & Bauer, T. Evaluating natural attractions for tourism. Annals of Tourism Research 29, 422–438, doi:10.1016/S0160-7383(01)00068-8 (2002).

  17. 17.

    Naidoo, R. & Adamowicz, W. L. Biodiversity and nature-based tourism at forest reserves in Uganda. Environment and Development Economics 10, 159–178, doi:10.1017/S1355770X0400186X (2005).

  18. 18.

    Siikamäki, P., Kangas, K., Paasivaara, A. & Schroderus, S. Biodiversity attracts visitors to national parks. Biodiversity and Conservation 24, 2521–2534, doi:10.1007/s10531-015-0941-5 (2015).

  19. 19.

    Hausmann, A., Slotow, R., Fraser, I. & Di Minin, E. Ecotourism marketing alternative to charismatic megafauna can also support biodiversity conservation. Animal Conservation 1, 208–11, doi:10.1111/acv.12292 (2016). doi:10.1111/acv.12292.

  20. 20.

    Fyhri, A., Jacobsen, J. K. S. & Tømmervik, H. Tourists’ landscape perceptions and preferences in a Scandinavian coastal region. Landscape and Urban Planning 91, 202–211, doi:10.1016/j.landurbplan.2009.01.002 (2009).

  21. 21.

    De Vos, A., Cumming, G. S., Moore, C. A., Maciejewski, K. & Duckworth, G. The relevance of spatial variation in ecotourism attributes for the economic sustainability of protected areas. Ecosphere 7, e01207, doi:10.1002/ecs2.1207 (2016).

  22. 22.

    Naidoo, R., Fisher, B., Manica, A. & Balmford, A. Estimating economic losses to tourism in Africa from the illegal killing of elephants. Nature Communications 7, 13379, doi:10.1038/ncomms13379 (2016).

  23. 23.

    Balmford, A., Beresford, J., Green, J., Naidoo, R. & Walpole, M. A Global Perspective on Trends in Nature-Based Tourism 7, 1–6 (2009).

  24. 24.

    Wood, S. A., Guerry, A. D., Silver, J. M. & Lacayo, M. Using social media to quantify nature-based tourism and recreation. Scientific reports 3, 2976, doi:10.1038/srep02976 (2013).

  25. 25.

    Sonter, L. J., Watson, K. B., Wood, S. A. & Ricketts, T. H. Spatial and Temporal Dynamics and Value of Nature-Based Recreation, Estimated via Social Media. PLOS ONE 11, e0162372, doi:10.1371/journal.pone.0162372 (2016).

  26. 26.

    Chua, A., Servillo, L., Marcheggiani, E. & Moere, A. Vande. Mapping Cilento: Using geotagged social media data to characterize tourist flows in southern Italy. Tourism Management 57, 295–310, doi:10.1016/j.tourman.2016.06.013 (2016).

  27. 27.

    Levin, N., Kark, S. & Crandall, D. Where have all the people gone? Enhancing global conservation using night lights and social media. Ecological Applications 25, 2153–2167, doi:10.1890/15-0113.1 (2015).

  28. 28.

    Lindsey, P. A., Alexander, R., Mills, M. G. L., Romañach, S. & Woodroffe, R. Wildlife Viewing Preferences of Visitors to Protected Areas in South Africa: Implications for the Role of Ecotourism in Conservation. Journal of Ecotourism 6, 19–33, doi:10.2167/joe133.0 (2007).

  29. 29.

    Ament, J. M., Moore, C. A., Herbst, M. & Cumming, G. S. Cultural Ecosystem Services in Protected Areas: Understanding Bundles, Trade-Offs and Synergies. Conservation Letters 0, 1–11 (2016).

  30. 30.

    Richards, D. R. & Friess, Da. A rapid indicator of cultural ecosystem service usage at a fine spatial scale: Content analysis of social media photographs. Ecological Indicators 53, 187–195, doi:10.1016/j.ecolind.2015.01.034 (2015).

  31. 31.

    Kerley, G. I. H., Geach, B. G. S. & Vial, C. Jumbos or bust : do tourists ’ perceptions lead to an under-appreciation of biodiversity? South African Journal of Wildlife Research 33, 13–21 (2003).

  32. 32.

    Barendse, J. et al. Viewshed and sense of place as conservation features: A case study and research agenda for South Africa’s national parks. KOEDOE - African Protected Area Conservation and Science 58, 1–16, doi:10.4102/koedoe.v58i1.1357 (2016).

  33. 33.

    Hausmann, A., Slotow, R., Burns, J. K. & Di Minin, E. The ecosystem service of sense of place: benefits for human well-being and biodiversity conservation. Environmental Conservation 43, 117–127, doi:10.1017/S0376892915000314 (2015).

  34. 34.

    Gaston, K. J. Global patterns in biodiversity. Nature 405, 220–227, doi:10.1038/35012228 (2000).

  35. 35.

    World Tourism Organization. UNWTO Tourism Highlights, 2016 Edition. doi:10.18111/9789284418145 (2016).

  36. 36.

    Puustinen, J., Pouta, E., Marjo Neuvonen, M. & Tuija Sievänen, T. Visits to national parks and the provision of natural and man-made recreation and tourism resources. Journal of Ecotourism 8, 18–31, doi:10.1080/14724040802283210 (2009).

  37. 37.

    Crampton, J. W. et al. Beyond the geotag: situating ‘big data’ and leveraging the potential of the geoweb. Cartography and Geographic Information Science 40, 130–139, doi:10.1080/15230406.2013.777137 (2013).

  38. 38.

    Tufekci, Z. Big questions for social media big data: Representativeness, validity and other methodological pitfalls. ICWSM ’14: Proceedings of the 8th International AAAI Conference on Weblogs and Social Media 505–514 (2014).

  39. 39.

    Kim, H. J., Chen, M.-H. & Jang, S. ‘Shawn’. Tourism expansion and economic development: The case of Taiwan. Tourism Management 27, 925–933, doi:10.1016/j.tourman.2005.05.011 (2006).

  40. 40.

    Aker, J. C. & Mbiti, I. M. Mobile Phones and Economic Development in Africa. Journal of Economic Perspectives 24, 207–232, doi:10.1257/jep.24.3.207 (2010).

  41. 41.

    Gössling, S. Global environmental consequences of tourism. Global Environmental Change 12, 283–302, doi:10.1016/S0959-3780(02)00044-4 (2002).

  42. 42.

    Woodroffe, R. & Ginsberg, J. R. Edge Effects and the Extinction of Populations Inside Protected Areas. Science 280, 2126–2128, doi:10.1126/science.280.5372.2126 (1998).

  43. 43.

    Su, S., Wan, C., Hu, Y. & Cai, Z. Characterizing geographical preferences of international tourists and the local influential factors in China using geo-tagged photos on social media. Applied Geography 73, 26–37, doi:10.1016/j.apgeog.2016.06.001 (2016).

  44. 44.

    Goyal, R., Dhyani, P. & Rishi, O. P. ‘Like’: A Unique Way to Judge User Preference in Social Networking Sites. in Fifth International Conference on Communication Systems and Network Technologies 1056–1059, doi:10.1109/CSNT.2015.182 (IEEE, 2015).

  45. 45.

    Coelho, R. L. F., Oliveira, D. Sde & Almeida, M. I. S. de. Does social media matter for post typology? Impact of post content on Facebook and Instagram metrics. Online Information Review 40, 458–471, doi:10.1108/OIR-06-2015-0176 (2016).

  46. 46.

    Eftekhar, A., Fullwood, C. & Morris, N. Capturing personality from Facebook photos and photo-related activities: How much exposure do you need? Computers in Human Behavior 37, 162–170, doi:10.1016/j.chb.2014.04.048 (2014).

  47. 47.

    Bakhshi, S. et al. Faces engage us: how to get more likes on facebook photo. in Proceedings of the 32nd annual ACM conference on Human factors in computing systems - CHI ’14 965–974, doi:10.1145/2556288.2557403 (ACM Press, 2014).

  48. 48.

    Hausmann, A., Toivonen, T., Slotow, R., Tenkanen, H., Moilanen, A., Heikinheimo, V., Di Minin, E. Social media data can be used to understand tourists’ preferences for nature-based experiences in protected areas. Conservation Letters. 10.1111/conl.12343 (2017).

  49. 49.

    Willemen, L., Cottam, A. J., Drakou, E. G. & Burgess, N. D. Using Social Media to Measure the Contribution of Red List Species to the Nature-Based Tourism Potential of African Protected Areas. Plos One 10, e0129785, doi:10.1371/journal.pone.0129785 (2015).

  50. 50.

    Rondinini, C., Wilson, K. A., Boitani, L., Grantham, H. & Possingham, H. P. Tradeoffs of different types of species occurrence data for use in systematic conservation planning. Ecology Letters 9, 1136–1145, doi:10.1111/j.1461-0248.2006.00970.x (2006).

  51. 51.

    Ripple, W. J. et al. Status and ecological effects of the world’s largest carnivores. Science 343, 1241484–1241484, doi:10.1126/science.1241484 (2014).

  52. 52.

    Ripple, W. J. et al. Collapse of the world’s largest herbivores. Science Advances 1, e1400103–e1400103, doi:10.1126/sciadv.1400103 (2015).

  53. 53.

    Tveit, M. S. Indicators of visual scale as predictors of landscape preference ; a comparison between groups. Journal of Environmental Management 90, 2882–2888, doi:10.1016/j.jenvman.2007.12.021 (2009).

  54. 54.

    Nelson. Travel time to major cities: A global map of Accessibility. Global Environment Monitoring Unit—Joint Research Centre of the European Commission (2008).

  55. 55.

    Scott, D., Gössling, S. & de Freitas, C. Preferred climates for tourism: case studies from Canada, New Zealand and Sweden. Climate Research 45, 61–73, doi:10.3354/cr00774 (2008).

  56. 56.

    Chhetri, P., Arrowsmith, C. & Jackson, M. Determining hiking experiences in nature-based tourist destinations. Tourism Management 25, 31–43, doi:10.1016/S0261-5177(03)00057-8 (2004).

  57. 57.

    Ghermandi, A. & Nunes, P. A. L. D. A global map of coastal recreation values: Results from a spatially explicit meta-analysis. Ecological Economics 86, 1–15, doi:10.1016/j.ecolecon.2012.11.006 (2013).

  58. 58.

    Burnham, K. P. & Anderson, D. R. Model Selection and Multimodel Inference: A Practical Information-Theoretic approach. (New York: Springer, 1998).

  59. 59.

    Wright, K. Plot a correlogram. https://github.com/kwstat/corrgram Available at: https://github.com/kwstat/corrgram (2016).

  60. 60.

    R Development Core Team. R: The R Project for Statistical Computing. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/ (2013). Available at: https://www.r-project.org/ (Accessed: 26th April 2016).

  61. 61.

    Calcagno, V. & Mazancourt, C. De. glmulti : An R Package for Easy Automated Model Selection with (Generalized) Linear Models. Journal of statistical software 34, 1–29, doi:10.18637/jss.v034.i12 (2010).

  62. 62.

    Selier, J., Slotow, R. & Di Minin, E. Large mammal distribution in a transfrontier landscape : trade-offs between resource availability and human disturbance. Biotropica 47, 389–397, doi:10.1111/btp.2015.47.issue-3 (2015).

Download references

Acknowledgements

A.H. was supported by the Amarula Trust funding to the Amarula Elephant Research Programme. E.D.M. thanks the Academy of Finland 2016-2019, Grant 296524, for support. H.T. thanks DENVI doctoral programme at University of Helsinki for support. All authors thank Kone Foundation for support.

Author information

A.H. and E.D.M. designed and wrote the manuscript. A.H., H.T. and V.H. collected and processed data. A.H. and E.D.M. analyzed the data. A.H., T.T., V.H., H.T., R.S. and E.D.M. provided comments and improved the manuscript.

Correspondence to Anna Hausmann.

Ethics declarations

Competing Interests

The authors declare that they have no competing interests.

Additional information

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Appendix S1

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Further reading

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.