Human Wildlife Conflict (HWC) is a growing problem that is often exacerbated in areas with an abrupt ‘wilderness’ interface (i.e. from intact native habitat to agriculture; Betts et al., 2017; Di Marco et al., 2018; Woodroffe, 2000). HWC are situations in which the behaviour of a wild animal species poses a direct and recurrent threat to livestock, domestic animals, or game species used by people and, in response, persecution of the “threatening” species ensues (Inskip and Zimmermann, 2009; Zimmermann et al., 2010). Deforestation within intact landscapes opens up human access to wilderness areas, increasing contact between people and wildlife (e.g. improving hunter access, poaching and eliminating prey), and thus triggering situations where humans come into conflict with wildlife (Betts et al., 2017; Di Marco et al., 2018; Graham et al., 2005; Teixeira et al., 2020). These situations worsen with increasing human densities near intact landscapes and usually end with the extirpation of species considered problematic by people (Betts et al., 2017; Di Marco et al., 2018). For instance, historic increases in human density had a strong association with the historic loss of populations of large carnivores in North America (Woodroffe, 2000). The extinction risk of terrestrial mammal species worldwide was also positively associated with human density (Di Marco et al., 2018). Fragmentation of the world's most intact forest landscapes – such as the tropics – is predicted to increase over the coming five decades (Taubert et al., 2018), thus increasing the probability of occurrence and severity of HWC in these wilderness landscapes (Betts et al., 2017; Frank et al., 2019).

Raptors provide an interesting model for analysing how an increase in HWC, as a consequence of the increasing human pressure on natural habitats, is leading to declines of threatened top predator species worldwide (McClure et al., 2018). HWC involving raptors were recorded as early as the sixteenth century with the officially encouraged killing of millions of raptors in many parts of Europe as a control measure to avoid losses in livestock and game species (Broun, 2000; Newton, 1979). Since the early 20th century, HWC has become a threat with significant impact on several raptor populations, leading to local or regional species extinctions in Europe and North America (Bildstein, 2001; Bildstein and Keith, 2008; White et al., 1994). The most remarkable example of the consequences of HWC for raptors was the extinction of the Guadalupe Caracara (Caracara lutosa), an endemic species from Guadalupe Island in Mexico, which was persecuted (in the context of a HWC) to extinction at the beginning of the 20th century (White et al., 1994). Raptor persecution is, nowadays, far from being the widespread activity it used to be in the past in Europe and North America, although conflicts involving raptors are still present and threaten these species in many countries around the world (Donázar et al., 2016; Madden et al., 2019). Effects of raptors on livestock and on populations of game species have been quantitatively well-studied suggesting low incidence of predation, rarely reaching values above 3% in both the proportion of the raptor diet as well as among the livestock mortality causes, suggesting low impacts to people's economies (Aguiar-Silva et al., 2014; Ballejo et al., 2020; Davies, 1999; Kenward, 1999; Madden et al., 2019; Restrepo-Cardona et al., 2019; Sarasola et al., 2010; Valkama et al., 2005). Despite this, as happens in many HWC involving carnivores, conflicts with raptors are widespread and hunters and farmers perceive predatory (and in some cases scavenger) species as harmful (Ballejo et al., 2020; Davies, 1999; Kenward, 1999; Madden et al., 2019; Sarasola et al., 2010; Valkama et al., 2005). This implies the existence of subjacent factors not related to material or monetary losses behind these conflicts (Thondhlana et al., 2020; Zuluaga et al., 2020).

Eagles, one of the most threatened group of raptors in the world, are frequently involved in conflicts with humans due to their large size and their food requirements (McClure et al., 2018; Meyburg, 1986; Newton, 1979). The Black-and-chestnut Eagle (Spizaetus isidori) (hereafter BC Eagle), for instance, is one of the most endangered eagles of the world (BirdLife International, 2020). It is distributed across montane rainforests in the Andes from Venezuela and Colombia to north-western Argentina (Ferguson-Lees and Christie, 2001). This large raptor is globally listed as Endangered, with an estimated population size between 250 and 999 mature individuals (BirdLife International, 2020). The BC Eagle is considered to be very sensitive to habitat loss and fragmentation (Thiollay, 1991), and to human persecution (BirdLife International, 2020; Echeverry-Galvis et al., 2014; Lehmann, 1959; Restrepo-Cardona et al., 2020). BC Eagle is a large forest raptor of 63–74cm in body length, 1500–3500g of weight, and up 180cm of wingspan (Authors unpublished data). The species feeds mostly on arboreal mammals and large-medium sized wild birds, however, domestic fowl (Gallus gallus) are present in almost every nest where the species’ diet has been studied (Aráoz et al., 2017; Lehmann, 1959; Restrepo-Cardona et al., 2019). Retaliation of farmers against the eagles because of livestock losses has been reported as the cause of mortality in most of the 80 BC Eagles found dead in the last 80 years in Colombia (Restrepo-Cardona et al., 2020). Over 30% of these cases occurred in the eastern Andes of Colombia, where a small BC Eagle population resides, seemingly isolated from the northern and southern populations of the species in the country (BirdLife International, 2020; Restrepo-Cardona et al., 2020). This suggests that the conflict between BC Eagle and humans exists and that it may be intense in some areas of Colombia.

In Colombia, around 60% of the original vegetation cover in mid and high montane Andean forests has been lost (Etter et al., 2006). These high rates of deforestation have changed native prey availability and increased the proximity between BC Eagle and human rural populations, thus likely increasing the conflict between local farmers and BC Eagle (Echeverry-Galvis et al., 2014; Restrepo-Cardona et al., 2019; Zuluaga and Echeverry-Galvis, 2016). Historical and current hunting of native prey by farmers, as well the rise of free-range domestic fowl, could also be altering BC Eagle food availability and consequently increasing livestock (i.e. domestic fowl) predation (Lyamuya et al., 2014; Restrepo-Cardona et al., 2019; Woodroffe et al., 2005). Recent studies suggested that the variation in levels of human–BC Eagle conflict in several localities of Colombia may be explained by a negative relationship between the percentage of forest cover and domestic fowl predation by BC Eagle (Restrepo-Cardona et al., 2019, 2020). However, this conclusion is based on a small sample of nests and localities, which could thus have some bias (Restrepo-Cardona et al., 2019). In contrast, other HWC studies in the Neotropical region involving wild top predators with similar habitat requirements as the BC Eagle, indicated that most forested landscapes were positively associated with wildlife attacks on livestock, and thus with higher levels of conflict (Michalski et al., 2006; Soto-Shoender and Giuliano, 2011; Teixeira et al., 2020).

The strategic location of effective mitigation measures to address HWC is as important as the measures themselves (Altringham et al., 2020). In order to make effective decisions on future raptor conservation efforts, and to optimize the use of the limited economic resources, a more comprehensive knowledge of the socio-ecological contexts in which these human–raptor conflicts occur is necessary. As in the case of the BC Eagle populations, other forest raptor species of the Neotropical region are declining primarily due to an increase in HWC (Barbar et al., 2016; Gusmão et al., 2016; Muñiz-López, 2017; Restrepo-Cardona et al., 2020; Sarasola and Maceda, 2006) associated with processes of habitat loss and fragmentation due to agriculture expansion (Grande et al., 2018a; McClure et al., 2018). The mechanisms by which habitat loss and fragmentation, and human persecution in the context of HWC affect a species’ conservation status involves a series of complex social and ecological relationships, which are not completely understood, and thus require socio-ecological studies (Ballejo et al., 2020). The socio-ecological system approach brings together theoretical and analytical techniques from diverse disciplines, including those from social and ecological sciences, to understand complex systems (Binder et al., 2013). Emergent evidence suggests that applying a socio-ecological system approach can inform a better understanding of the socio-ecological contexts in which HWC occur (see Behr et al., 2017; Ceauşu et al., 2019; Dressel et al., 2018; Guerrero and Wilson, 2017; Pooley et al., 2017; Struebig et al., 2018; Teixeira et al., 2020). We anticipate that applying the socio-ecological system approach (Box 1) will improve our current understanding of this specific human–raptor conflict involving the BC Eagle in order to better inform conservation management. The aim of this study is thus to examine the socio-ecological context that exacerbates the human–eagle conflict (i.e. socio-ecological context influencing the people’s tolerance towards BC Eagle domestic fowl predation), in villages of the eastern Andes of Colombia. Our hypothesis is that the people’s tolerance towards this eagle will vary in the different villages influenced by the amount of forest cover (i.e. amount of remaining native forest), the human density, and yearly losses of domestic fowl by BC Eagle in those villages. We predicted that:

• 1.

  Forest cover in villages. Here we postulate two contrary predictions that exclude one another:

  • a.

    People living in more forested villages will have more contact with wildlife and top predators, and thus will suffer higher losses by predation or at least will perceive a higher predation risk for their livestock, therefore they will be less tolerant of the BC Eagle (Betts et al., 2017; Di Marco et al., 2018; Graham et al., 2005; Michalski et al., 2006; Soto-Shoender and Giuliano, 2011; Teixeira et al., 2020).

  • b.

    Conversely, more forested villages will have more diverse and abundant populations of wild prey for the BC Eagle and thus, domestic fowl predation rates will be lower there and farmers will be more tolerant to the eagle. While in more deforested areas wild prey will be scarcer and thus, predation rates of eagles on domestic fowl will be higher there and farmers will be less tolerant of the BC Eagle (Acharya et al., 2017; Artelle et al., 2016; Lyamuya et al., 2014; Restrepo-Cardona et al., 2020, 2019).

• 2.

  Human density in villages. Greater human densities in rural villages imply higher use of natural resources and probably a greater perception of competition between humans and wildlife (Artelle et al., 2016; Betts et al., 2017; Di Marco et al., 2018; Kaswamila et al., 2007; Woodroffe, 2000). Additionally, higher human densities may allow more frequent social interaction, allowing livestock predatory events to become more public, giving locals a higher perception of predatory risk and thus a lower tolerance to the BC Eagle (Bruskotter and Wilson, 2014; Carter et al., 2020; Marchini and Macdonald, 2018).

• 3.

  Livestock losses. BC Eagle is known to prey on domestic fowl (Aráoz et al., 2017; Echeverry-Galvis et al., 2014; Restrepo-Cardona et al., 2019; Zuluaga and Echeverry-Galvis, 2016). Therefore, the people's tolerance towards the BC Eagle should be lower in villages with higher yearly losses of domestic fowl (Restrepo-Cardona et al., 2020).

Box 1.

Socio-ecological system approach applied to human–wildlife conflicts, adapted from Carter et al. (2014).

Several research approaches have been developed and applied to different studies in which the interaction between the social system and the ecological system has been explicitly considered (Binder et al., 2013). We used the integrated socio-ecological systems approach proposed by Carter et al. (2014), for guiding our study on the complex relationships between the Black-and-chestnut Eagle and humans in the eastern Andes of Colombia (Guavio Region). Its conceptualization consists of three main components: the social subsystem, the ecological subsystem and the two-way socio-ecological interactions (or feedbacks). The social subsystem comprises local people and rural villages, and the ecological subsystem comprises wildlife and the land cover characterizing their habitat. The dimensions of each of these subsystems (i.e. local people, rural villages, wildlife and land cover) are interrelated and thus influence the characteristics of each other through socio-ecological interactions.

By transcending a single discipline, this approach can account for the patterns and processes that link people and their activities with wildlife and their habitats. Also, it can identify key relationships and feedbacks between people and wildlife. Finally, the approach facilitates understanding of cross-scale (e.g. spatial, temporal, and organizational) interactions between people and wildlife (Carter et al., 2014).

Our socio-ecological model used to analyze the socio-ecological factors that affect the local people's tolerance towards the BC Eagle, in a region of the Eastern Andes of Colombia, showed the best performance to explain tolerance among several alternative ecological, demographic and socio-economic models. The people's tolerance to the BC Eagle was lower in those villages with higher forest cover, higher human density and higher yearly domestic fowl losses. Forest cover was also positively associated with BC Eagle yearly attacks on domestic fowl. Increasing human density in areas that still hold important forest cover is likely exacerbating the severity of the human–eagle conflict in areas with an abrupt wilderness interface.

Our results suggest that the risk of persecution of BC Eagles in the context of HWC is relatively high in the most suitable habitats of the eastern Andes of Colombia. We found that the people's tolerance was negatively affected by the proportion of forest cover and by the domestic fowl losses by BC Eagle, both higher in the most forested landscapes. Based on knowledge of habitat requirements of the species, it is likely that BC Eagle selects villages with highest proportions of forest cover (i.e. those with minimal deforestation; Thiollay, 1991), where they had a higher impact on domestic fowl. Overall, the proportion of forest cover in the 20 villages was 44%. Around half of the people interviewed showed low tolerance to losing domestic fowl by BC Eagle (52%), but this proportion increased up to 76% in the six villages with the highest proportion of forest cover.. Human persecution was also most frequent in these villages where 8% of people acknowledged to have killed the species. Therefore, our results suggest that it is in those areas with higher forest cover and high domestic fowl losses where conservation action should be prioritized in the eastern Andes of Colombia.

Our outcomes supported the overall evidence that forest cover is positively associated with wildlife attacks on livestock and thus forest cover is indirectly and negatively associated with people's attitudes and tolerance towards predators (Graham et al., 2005; Michalski et al., 2006; Soto-Shoender and Giuliano, 2011; Teixeira et al., 2020). However, a study made in the vicinity of four BC Eagle nests suggested the contrary (Restrepo-Cardona et al., 2019, 2020). This difference could be due to the geographic scale of both studies as well as to its designs. While Restrepo-Cardona et al. (2019, 2020) worked in the vicinity of nests, in both the Central and the Eastern Andes mountains, we worked in several villages only in the Eastern Andes but in a wider area not restricted to the vicinity of nests. This divergence clearly shows the need to consider the approaches and the different scales of analysis. Future studies considering the interplay among forest cover, hunting of native prey by farmers, and prey availability for the eagle (i.e. free-range domestic fowl and wild prey) are necessary to have a better ecological understanding of the system and how it interplays with social factors to increase or reduce local HWC, and thus to improve our socio-ecological evidence to drive decision-making and implementation of conservation measures (Lyamuya et al., 2014; Restrepo-Cardona et al., 2019; Teixeira et al., 2020; Woodroffe et al., 2005).

Low tolerance in villages with high human density where the remaining habitat for the BC Eagle can be scarce could be related to some underlying issues related to a human–human conflict which are also present (Fraser-Celin et al., 2018). HWC are frequently complex and it is well known that mistrust between management agencies and other stakeholders, or the mere communication of predation events among stakeholders can create or aggravate conflicts (Bruskotter and Wilson, 2014; Marchini and Macdonald, 2018). The presence of the BC Eagle in these villages may be less frequent because there is a lower proportion of forest cover (Thiollay, 1991), and thus domestic fowl losses should be lower here. Although we did not find an influence of the proportion of forest cover on the observation frequency of BC Eagle by people, low proportion of forest cover was associated with low BC Eagle attacks on domestic fowl. However, the relationship between high human density and low forest cover may cause a high demand of natural resources which are scarce in these villages, but necessary to small scale subsistence farming production, (i.e. trees for firewood and construction materials for homes and fences, or wooded pastures with forage for livestock). Access to these natural resources may be limited by authorities through regulations and laws which foster discontent and environmental conflicts among stakeholders. These two factors have been shown to influence human–jaguar (Panthera onca) and human–cougar (Puma concolor) conflicts in Sao Paulo, Brazil (Engel et al., 2016), human–black bear (Ursus americanus) conflict in Colorado, USA (Lischka et al., 2018), and human–African wild dog (Lycaon pictus) conflict in Botswana (Fraser-Celin et al., 2018). Further socio-ecological research to deepen the this aspect of the conflict is certainly needed.

Despite being pretty simple, our socio-ecological model performed much better in explaining tolerance than all alternative models. These alternative models included a sample of simple models including only the ecological component, only the social component or the output of the interaction of both subsystems in the number of domestic fowl preyed upon by the eagle, as well as several other models including demographic and socio-economic drivers previously reported in the literature as relevant in human–raptor conflicts (Tables 1–3; see Ballejo et al., 2020, 2019; Cailly-Arnulphi et al., 2017; Restrepo-Cardona et al., 2020). Our results are thus in consonance with the emerging evidence suggesting that applying socio-ecological models to HWC can be informative and beneficial (Behr et al., 2017; Carter et al., 2019, 2014; Ceauşu et al., 2019; Dressel et al., 2018; Guerrero and Wilson, 2017; Pooley et al., 2017; Teixeira et al., 2020). Similarly, a recent study combining people's tolerance for critically endangered Sumatran tigers (Panthera tigris sumatrae) in socio-ecological models, with underlying attitudes, emotions, norms, spiritual beliefs and geographic profiles yielded predictions of tolerance that were 32 times better than models based on social predictors alone (Struebig et al., 2018). Those outcomes are clear examples of how a socio-ecological approach can improve our understanding of HWC with several species, including raptor species of conservation concern. These human–raptor conflicts have been widely studied from the biological sciences perspectives (e.g. Madden et al., 2019; Restrepo-Cardona et al., 2019; Sarasola et al., 2010; Valkama et al., 2005) but have usually ignored the socio-ecological evidence to inform decision-making and implementation of conservation measures (e.g. Ballejo et al., 2020, 2019; Cailly-Arnulphi et al., 2017; Grande et al., 2018b; Restrepo-Cardona et al., 2020), thus hampering the success of the proposed conservation measures.

The Conservation Plan for the BC Eagle in the Guavio Region considers that threats related to farming encroachment into intact native habitat of the species, and the increase of severity of human–eagle conflicts, should be mitigated (see Zuluaga, 2018). Based on the new socio-ecological evidence generated in this work, we suggest that the implementation of conservation measures related to these threats should focus mainly on specific areas where there is a larger risk of human–eagle conflict (i.e. villages with most forest cover, and among them, those with higher human density) to be more cost effective. This strategy would optimize the use of limited economic resources and would help to proportionally decrease BC Eagle mortality by poaching where this threat is higher. In the most forested villages where people have lower tolerance towards BC Eagle, more economic and human resources are needed in order to increase people's tolerance towards BC Eagle. In the same way, environmental education programs for saving the BC Eagle should be targeted to a broad audience (Zuluaga, 2018), and focused on specific areas with historical human–BC Eagle conflict (that could be measured by historical or current records of BC Eagles killed by people; see Nilsson et al., 2020). Approaches based on the science of behavioural change, such as ‘Theory of planned behaviour’ and ‘Theory of change’ have proven to make a more tangible difference for human behavioural changes (Altringham et al., 2020; Center for Theory of Change, 2019; Nilsson et al., 2020) and should be applied. A way to assess the success of the Conservation Plan for the Black-and-chestnut Eagle in the Guavio Region might be to measure whether this program's actions truly reduced the number of BC Eagles killed in those villages with the most intact native forest in the Guavio Region (Nilsson et al., 2020).

Our study represents a good example of how an abrupt ‘wilderness’ interface can exacerbate a human–eagle conflict. The socio-ecological approach allowed us to better understand the complex interplay between people's tolerance, forest cover, human density and livestock losses. In this way, we captured the multiple ecological as well as social dimensions of this human–eagle conflict, identifying the combinations of attributes that should be considered for decision-making and implementing of conservation measures for the BC Eagle. Our findings supported the overall evidence that forest cover is positively associated with wildlife attacks on livestock and were in consonance with the most recent evidence indicating that population declines of top predators as well as of other vertebrate biodiversity can be severely affected by the exacerbation of HWC in areas of wilderness-agriculture interface.

None declared.