Pax Romana: ‘refuge abandonment’ and spread of fearless behavior in a reconciling world

Abstract

Intense human persecution of wildlife in the past selected for shy individuals that survived in suboptimal refuge habitats. It appears that this long-term conflict is coming to an end in southern Europe, where direct persecution of wildlife has dramatically decreased due to human abandonment of the rural world and an increase in respectful attitudes toward wildlife. We provide a number of empirical examples of ongoing ‘refuge abandonment’ and the expansion of fearless behavior among southern European wildlife, including island and mainland case studies as well as predator and prey examples. We predict that many ecological refugees will increasingly abandon their historical refuges (of which many currently coincide with protected areas), and that the frequency of fearless behaviors will increase via cultural habituation and natural selection in populations of mammals and birds that were formerly persecuted. We also suggest that this has been a nonlinear process, which began considerably earlier than it became evident, with a long latent inertial regime, and a relatively rapid transition to a state of refuge abandonment and occupation of new, more anthropogenic habitats. The process of refuge abandonment and expansion of fearless behaviors will bring a new paradigm regarding human–wildlife relationships. It will also force changes in policies to protect high-quality sites that are currently unoccupied and unprotected. Refuge abandonment may form a completely new research agenda and should be taken into account when studying patterns of wildlife dispersal or dynamics of spatially structured populations, among other relevant topics.

Introduction

We suggest in this review that many 21st century southern European terrestrial wildlife species still live like Bronze and Iron Age human refugees (Bocherens et al., 2015) in their hill forts, due to huge historical human persecution. The magnitude of this persecution can be exampled in that between 1955 and 1961 the Spanish government paid for the carcasses of 784 vultures, 1033 eagles, 20 228 other raptor types, and more than 500 000 other birds in 10 provinces of Spain, which equate to only 27% of the country's surface area (see Martínez-Abraín et al., 2008a). However, we will also show that wildlife is now showing a trend toward refuge abandonment. A similar trend for human populations took place during the so-called Pax Romana (Roman Peace), a historical process that lasted some 200 years. It began with the emperor Augustus and lead to the abandonment of hill forts by many Iron Age people that urbanized the open plains under the protection of the Roman Empire. We will also show that wildlife is showing an increase in the frequency of bold (i.e., fearless, aggressive, and exploratory) behaviors, rather than shy behaviors, in populations and is moving closer to anthropic areas. This trend is now becoming apparent after six or seven decades of rural exodus and nonlinear wildlife recovery. We predict that this process is leading to a new paradigm in human–wildlife relationships. The main cause of this change is that direct persecution of wildlife, which has lasted for centuries or millennia, has been negligible during the last few decades. This has been caused by the application of conservation policies and the development of awareness campaigns but also, due to the abandonment of the rural world (i.e., rural exodus or demographic transition), with the gathering of most people in a few major cities. This process has generated huge areas of free land formerly occupied by people who would make intensive use of natural resources and exclude wildlife as competitors or enemies. As a result, Europe is witnessing the return of formerly threatened large herbivore and carnivore mammals, a process known as ecological rewilding (Navarro & Pereira, 2012; Deinet et al., 2013; Chapron et al., 2014; Pereira & Navarro, 2015; Milanesi et al., 2017). It is important to note that our impact on wildlife nowadays is mostly unintentional, such as in the development of infrastructure (Martínez-Abraín et al., 2008a, 2009, 2013; Karamanlidis et al., 2015; González et al., 2016). This factor works in conjunction with the possibility that living closer to humans may provide animals with large amounts of predictable anthropogenic food subsidies (Oro et al., 2013) as well as lower predation risks due to the scarecrow effect (Leighton, Horrocks & Kramer, 2010). A parallel can be traced in the marine environment where seascapes of fear (Laundré et al., 2014) created by human fishing promote the development of a timidity syndrome among individuals that then vanishes within marine-protected areas (Arlinghaus et al., 2016).

There is some evidence that the harvesting of wild animals often creates strong selection differentials in exploited populations (Allendorf & Hard, 2009). Behavior of individuals in populations is a polymorphism that can be roughly divided into ‘bold’ and ‘shy’ personalities in reference to their relationship with humans as potential predators (Geffroy et al., 2015). Human persecution of wildlife may cause strong selective forces against bold individuals with better dispersal abilities, stronger prospective behavior (Réale et al., 2007; Biro & Post, 2008), and therefore a higher likelihood of human encounters. On the contrary, respectful human presence (such as in nature-based tourism) can lead to cultural habituation and selection that promotes bold individual personalities (boldness syndrome; Geffroy et al., 2015; Arroyo, Mougeot & Bretagnolle, 2017).

The quantitative genetic architecture of the bold–shy continuum has been reported to be the result of correlated evolution via underlying genetic correlations (Oswald, Singer & Robison, 2013). Behavioral personality traits such as bold–shy or exploratory have been found to be highly repeatable and heritable in long-lived organisms (Patrick, Chamantier & Weirmerskirch, 2013) and across taxa (Dochtermann, Schwab & Sih, 2015) as well as having fitness consequences that depend on the interaction between sex and year quality (Patrick & Weimerskirch, 2014). This polymorphism also seems to be involved in the functioning of historical biogeographical processes that could have been driven by bold (fearless and exploratory) samples of the founder populations (Canestrelli, Bisconti & Carere, 2016).

Our suggestion is based on the idea that human activity in our rural past selectively removed bold individuals that were less afraid of human presence and could be dangerous or pernicious for us. Human beings unintentionally created a process of selection that (1) promoted shy genotypes (i.e., individuals that were afraid of human presence and survived only in suboptimal refuges in remote places far from human settlements) (Arlinghaus et al., 2016), (2) forced many animal species into living as refugees, (3) forced refugees to be mainly nocturnal, and (4) reduced population sizes, among other specific changes such as changes in diet, mating behavior, and body size to help in avoiding humans.

It is known that individuals of continental prey species show naïve behavior toward recolonizing predators after several decades of their eradication and that this behavior can be altered within a single generation (Berger, Swenson & Persson, 2001). This suggests that epigenetic mechanisms could also be linked with the shy–bold continuum, and therefore, rapid abandonment of refuges could be possible even in small populations (see Riyahi et al., 2015, 2017). Additionally, individuals are also capable of having cultural transfer of information (Avital & Jablonka, 2000; Teitelbaum et al., 2016), especially in social species. Therefore, having only a few bold individuals in the population would most likely suffice to be able to see the fast spreading of the fearless strategy via imitation, learning, and habituation, if advantageous. Even though data are lacking, we propose that the process has occurred through a rapid, nonlinear transition to a state where species distributions are characterized by their proximity to anthropogenic sites where resource availability can be high. Overall, it is to be expected that less timid genotypes will have a demographic advantage in nonaggressive contexts (Smith & Blumstein, 2008; Geffroy et al., 2015). It is important to emphasize that the process of fear acquisition and loss is asymmetrical. It takes a long time to select for shy individuals (probably centuries of persecution), whereas only a few decades can cause the return of bold behaviors into populations by means of cultural and epigenetic mechanisms.

We will present some examples of this process of refuge abandonment and the expansion of bold behavior from Spain, a southern European country that has experienced a rapid change from animal persecution to preservation during the last three decades (Martínez-Abraín et al., 2008a). We will also complement these examples from Spain with information from other circum-Mediterranean countries. Mediterranean countries have been one of the cradles of civilizations, with high human population densities and records of their impacts on wildlife that date back several millennia (see Jennison, 2005; Coll et al., 2010). It is difficult to find, in any other part of the world, a rapid and heavy shift from wildlife persecution to wildlife conservation that parallels those exhibited by southern European countries (Martínez-Abraín et al., 2008a), but it is to be expected that this will be the case of many African, Asian, and Latin American countries in the near future. Data to support our predictions and statements come mostly from the limited pool of available gray literature and from unpublished monitoring programs. We separate island and mainland examples because we expect that behavioral plasticity of island faunas will be lower than that of those living on mainland sites because their population sizes have been smaller since their foundation (or have been persecuted by humans more intensively over time) and, therefore, a smaller genetic diversity is expected.

Island examples of refuge abandonment and shifts toward bold behavior

Eleonora's falcon Falco eleonorae is a small social and migratory raptor that breeds on cliffs on Mediterranean islands only when there is noticeable human presence on their breeding sites. However, on small islets without the presence of humans (and without any terrestrial predators), the falcons nest directly on the ground (Martínez-Abraín et al., 2002; Urios & Martínez-Abraín, 2006). Therefore, the presence of cliff-nesting pairs on these islets is an indication of the past and/or present human impact. We predict that Eleonora's falcons breeding on islands where current human presence is respectful will abandon cliffs to nest directly on the ground. This will most likely lead to population increases as well as high densities of Eleonora's falcons on small islets breeding on the ground. This can be seen on Mogador island (off the coast of Essaouira, Morocco), where 1217 nests were counted in 2016 in only 450 ha (http://www.magornitho.org/2016/09/record-breeding-falco-eleonorae-mogador/).

Ground-nesting, island seabird species, such as yellow-legged gulls Larus michahellis, show plasticity to withstand human presence and have smaller flight initiation distances in colonies where respectful human presence is frequent due to a process of rapid cultural habituation to harmless people (Martínez-Abraín et al., 2008b). There is an increasing trend in ground-nesting gulls in Spain to abandon islets and islands (former refuges) and to either move back to their original continental habitats or to colonize anthropic environments, such as cities, villages, marinas, harbors or salt pans. Yellow-legged gulls in Galicia (NW Spain) showed a decreasing trend of 3% in the Cies Islands National Park during the period 1991–2015 (Arcea, 2006, Barros, 2015), whereas the annual growth of the urban population in the city of Vigo was 10% for the same time period (Arcea, 2013). Another example of ground-nesting gulls is that of Audouin's gulls Ichthyaetus audouinii that were only present as refugees on small islets of the western Mediterranean in the 1970s and 1980s. However, Audouin's gulls eventually started leaving refuges and breeding on continental sites and their population grew exponentially until reaching a plateau (Figs 1 and 4e,f).

Figure 1

Growth of Audouin's gull Ichthyaetus audouinii Spanish populations in refuge areas (round dots) and outside refuge areas (square dots) from 1980 to 2013.

Black vultures Aegypius monachus typically nest in green oak trees Quercus ilex or in cork oaks Q. suber on mainland Spain where most of the European population currently breeds (De la Puente, Moreno-Opo & Del Moral, 2007). However, a small remnant population of this large scavenger species breeds on cliff pines on the island of Majorca (Sánchez-Artez, 1996). We think that this is a suboptimal location for black vultures because weather conditions are much tougher than in oak forests. In addition, there are plenty of unoccupied green oak forests on the island that could allow the formation of denser colonies for this semicolonial species. It is likely that the extant population is composed of individuals with shy genotypes that breed far from any human presence in oak forests, where all types of human activities (i.e., charcoal extraction, logging, pig farming growth, cast manufacturing, and hunting, among others) were carried out until the 1960s, when tourism became the main economic activity on the island. The population has been reinforced with individuals from mainland Spain, and it is slowly recovering from almost extinction, likely due to enhanced genetic behavioral variability. Four pairs of the 33 recorded in 2016 now breed in tall pines away from (although near to) the cliffs and close to highly frequented hiking pathways (J. Muntaner, pers. comm.). If this behavior spreads throughout the population, much larger population growth rates are to be expected because, until now, they may have been habitat-limited in their cliff refuges, as large cliff pines were scarce. Food, however, is not a limiting factor as the environmental authorities provide this in special vulture restaurants, and, furthermore, mortality is very low. We predict that most vultures will move to the oak forests within the next few decades. This relocation would be further encouraged by the non shy genotypes of restocked mainland individuals via cultural or social imitation.

Mainland examples of refuge abandonment and increased prevalence of bold behaviors

We showed in the past, by means of a meta-analysis, that large raptors tended to place their nests on cliffs rather than on trees when roads (as a proxy of human disturbance) were in proximity to nests (Martínez-Abraín et al., 2010). Therefore, we would predict a shift from large raptors breeding mostly on cliffs to tree nesting in the future. This prediction is further supported by a marked expansion in the surface area covered by forests since the abandonment of the countryside in Europe (Ameztegui, Brotons & Coll, 2010; Fuchs et al., 2015). For example, Spain's forested areas have increased from 142 100 km² in 1987–1996 to 185 786 Km² in 1998–2008 (a 30.7% increase). This increase in surface area has also affected tree density, with an increase in forest biomass (42–50 m³ ha⁻¹), and tree growth (i.e., trees with a diameter >75 mm measured at 1.3 m of height from 329 to 375 individuals per ha), during the same period (http://www.magrama.gob.es/es/biodiversidad/servicios/banco-datos-naturaleza/informacion-disponible/tablas_resumen_IFN3.aspx).

Both Bonelli's eagles Aquila fasciata and golden eagles Aquila chrysaetos have increased the proportion of nests placed in trees vs. cliffs in Spain over time. In 1990, only 1.8% of Bonelli's eagles nests (n = 679) were placed in trees (Arroyo, Ferreiro & Garza, 1995), whereas, in 2005, this percentage increased to 4.6% (n = 733) plus four nests that were built on top of electric power poles (Del Moral, 2006). During this period, the overall Spanish population of Bonelli's eagles was approximately constant (1990: 677–754 pairs vs. 2005: 733–768 pairs). Conversely, in Portugal, most pairs are now breeding in trees (70%, n = 120), even on exotic tree species such as Eucalyptus and on trees located close to houses; the species is expanding toward the north with new pairs breeding on trees (Palma, 2011). In regard to golden eagles in Spain, the percentage of nests located in trees between 1988 and 1989 was 7.2% (n = 983) (Arroyo, Ferreiro & Garza, 1990), this percentage increased to 17% (n = 1553) in the 2008 census (Del Moral, 2009). This trend appears to be continuing in the present day as shown, for example, by the breeding information available for the province of Alicante (SE Spain) where golden eagles have moved from having just one tree-nesting pair to five pairs between 2000 and 2016, without showing an increase in population size (A. Izquierdo, pers. comm.; Figs 2 and 4c,d). In Catalonia (NE Spain), golden eagles were already rare and restricted to the Pyrenees in 1881, although a few years earlier they were reported to breed on top of trees in the Montseny massif, along the Catalan precoastal range (Ferrer, 2016).

Figure 2

Historical growth of the ratio of golden eagles Aquila chrysaetos breeding on trees vs. golden eagles breeding on cliffs in the province of Alicante (Eastern Spain).

The golden eagle population in Spain has increased from c. 1300 pairs in 1988–1989 (Arroyo et al., 1990) to an estimated number of 1553–1769 pairs in 2008 (Del Moral, 2009). This increase has affected all altitudinal ranges. However, it has not been proportionally equal for all altitudinal ranges (χ² = 22.46, d.f. = 5, P < 0.05) (see Fig. 3). Specifically, a substantial increase is seen in golden eagles within the 0–400 m range (studentized residual in contingency table: +2.9) and a substantial decrease within the 801–1200 m range (studentized residual in contingency table: −3.8). This means that population growth in golden eagles has occurred in parallel with a movement of breeding pairs toward lower altitude zones, where the majority of people live in Spain. Average density in peninsular Spain in 2015 was 87.5 inhabitants per km², whereas average density in the 0–400 m altitude range was 213.7 inhabitants per km² (http://centrodedescargas.cnig.es/CentroDescargas/equipamiento.do?method=descargarEquipamiento&codEquip=8). That particular altitudinal range showed the highest human population increase (17.9%) between 1990 and 2005. A similar pattern was reported by Haller (1996) for the Italian Alps of golden eagles lowering altitude of their nesting sites. The author attributed this to a reduction in human persecution of the species rather than to an increase in golden eagle densities.

Figure 3

Altitudinal zonation of golden eagles Aquila chrysaetos breeding in Spain in 1989 (solid black bars) vs. 2008 (solid gray bars).

Spanish imperial eagles Aquila adalberti also appear to be experiencing similar changes. They increased from a minimum of 38 breeding pairs in Spain in 1974 (restricted to mountainous protected areas) to 198 pairs in 2004, becoming increasingly more common in the plains outside protected areas (González et al., 2008) and closer to areas of human activity (Morandini & de Benito, 2015). Recent expansion (>500 pairs in 2016, M. Ferrer, pers. comm.) is taking place in intensively managed pine forests surrounded by cereal crops in the lowlands of Castilla-León (Rojo et al., 2013) where their preferred prey (rabbit) is more abundant. Similar displacement from mountain refuges to agricultural areas in the lowlands has also been reported for its sister species, the eastern imperial eagle Aquila heliaca in Hungary, where the breeding population has increased from 15 to 25 pairs in the 1980s to more than 100 pairs in 2009 (Horváth et al., 2014). Fledging success was higher in the lowlands due to larger prey densities and shorter distances to hunting grounds (op. cit.). It is most likely that this trend of increase in the number of pairs breeding in the lowlands and closer to people corresponds to the advance of ecological succession in protected highland forests that decreases prey abundance for eagles. It may also be due to the higher prey productivity in lowland managed forest and croplands, together with a change in human attitudes (decreased persecution outside protected areas) in relation to large predators during the last few decades. Refuged large eagles are showing an unexpected behavioral plasticity to habituate to nonharmful human presence and are gaining fitness benefits thanks to their plasticity in the form of increased reproductive success and an increase in the number of pairs. A corollary to this case is that local negative trends are not always a cause of concern. They can simply signify that conditions to host wildlife are improving elsewhere and that the overall situation for the species is actually improving; a fact to be taken into account by wildlife managers. Eastern imperial eagles appear to be aware of patch quality on large spatial scales and choose to breed and forage in the highest quality patches following an ideal-free distribution that most likely maximizes their fitness (see Oro, 2008).

There has been a marked population recovery of Eurasian otters Lutra lutra in Spain during the last few decades (Jiménez et al., 2008). Otters are not only expanding out of their former refuges in the upper stretches of rivers (where they were not hunted so heavily and water was free of pollution) but also rapidly getting used to human presence in artificial environments (Martínez-Abraín & Jiménez, 2016; Fig. 4a,b). In the Spanish Mediterranean river basins, percentage occurrence of otters in artificial reservoirs has moved from 32% in the mid 90s to 77% between 2014 and 2016, being now higher than that in rivers (53–59%) (Delibes, 1990; Ruiz-Olmo & Delibes, 1998; Jiménez et al., 2008; López-Martín, 2008; López-Martín & Jiménez, 2008; SECEM, unpubl. data). Newspapers reporting cases of otters spotted during daytime in harbors, marinas, dams, rivers, or streams crossing villages and cities, university campuses, or golf courses are becoming increasingly more common in recent years in Spain. Another mammal thought to be a strict specialist in the upper stretches of rivers, the small Pyrenean desman Galemys pyrenaicus, has recently been found in the middle and lower stretches of rivers in NW Spain, in parallel with the improvement of water quality after decades of investment into water purification systems (Xunta de Galicia, 2014).

Figure 4

Some examples of animal species that used to live as forced refugees in the past, but that are now leaving their refuges. Eurasian otter Lutra lutra, a historical refugee in the upper stretches of rivers (a), now making extensive use of artificial reservoirs (b); Bonelli's eagles Aquila fasciata with historical nests mostly on cliffs (c) now increasingly breeding on trees (d); Audouin's gulls Larus audouiii formerly a refugee in small Mediterranean islets (e) now breeds on mainland marinas, salinas, and harbors as well as beaches and dune fields (f); Monk seal (Monachus monachus) forced to use small cliff caves for reproduction (g) is now increasingly making use of protected open beaches close to caves (h). Photo credits: Alejandro Martínez-Abraín (a), Charles Key (b), Alejandro Izquierdo (c,d), Columbretes Nature Reserve/Generalitat Valenciana and Pepe Greño (e,f), Miguel Angel Cedenilla/CBD-Hábitat and Zeidán Ranger/CBD-Hábitat (g,h). [Colour figure can be viewed at zslpublications.onlinelibrary.wiley.com]

Human persecution of brown bears Ursus arctos has led to reduced aggressiveness, reduced group size, reduced body size, vegetarian diet, and nocturnal activity according to Fernández-Gil (2014). This is most likely the result of a long process of selection of shy genotypes throughout history, together with cultural learning (Wheat & Wilmers, 2016; Benazzo et al., 2017). However, bears seem to be losing this historical fearfulness and are continuously getting closer to anthropic sites in NW Spain where females increasingly seek the proximity of roads and villages when they have cubs. This is most likely to help prevent males from killing the youngsters, as sexual infanticide is very common in these small populations (Fernández-Gil, 2014; G. Palomero/Fundación Oso Pardo, pers. comm.). Furthermore, some individuals are known to make use of feeders deployed to feed calves and large dogs (G. Palomero/Fundación Oso Pardo, pers. comm.). These changes in behavior are leading to an increase in bear-based tourism (Penteriani et al., 2017). This type of tourism has a long tradition in North American National Parks where hunting has been forbidden for many years (Herrero et al., 2005). Behavioral changes toward boldness have also been reported for Scandinavian bears following changes in the attitudes of humans (Svenson, 1999). Hence, in the future (if human attitudes do not change for the opposite direction), we would expect to have larger Iberian bears, forming larger groups, showing more diurnal activity, with a more carnivorous diet and ranging closer to people.

The critically endangered Mediterranean monk seals Monachus monachus are also known to have become refugees in caves located along continental coasts, in archipelagos and even in caves on the smallest islets of large archipelagos, due to a history of intense human persecution on their original beaches (González, 2015). Selection in favor of shy behaviors was probably intense. It is likely that monk seals were already refugees by the Greek and Roman times, although they were still found on beaches, by the thousands, during the XV century in some remote places such as the Canary Islands, Madeira, and the Western Sahara (Johnson & Lavigne, 1999). Thanks to local protection during the last few decades, monk seals from Cape Blanco Peninsula (Mauritania–Morocco) are now using beaches outside caves (mostly after birth for pup breastfeeding) which provides fitness benefits through the avoidance of the detrimental effects that rough seas have on offspring survival in caves (Johnson & Lavigne, 1999; González & M'Barek, 2004) (Fig. 4g,h).

Refuge abandonment in prey species

The abandonment of historical refuges and the successful expansion of bold genotypes are also taking place among vertebrate prey species formerly forced to be refugees in areas free from any human presence (see Geffroy et al., 2015). This could be the case for wild boars Sus scrofa that became extinct in some Mediterranean regions in the past (Jiménez, 2012) but that are now common in large Mediterranean cities such as the metropolitan area of Barcelona, Toulouse, and Berlin. Requests for the regulation of this situation already exist in France (see Ribier et al., 2012). Roe deer Capreolus capreolus were formerly shy inhabitants of forested areas containing low human density but are now a common cause of road accidents in Spain (Lagos, Picos, & Valero, 2012). Female red deer Cervus elaphus can now be observed a short distance from the roadside, breastfeeding their fawns in protected areas such as Monfragüe National Park (a traditional hunting area protected in 2007) in western Spain. In such areas, long lines of tourists can be seen taking pictures of red deer with their cellular phones (own unpublished observations), a case that resembles that of white-tailed deer Odocoileus virginianus in highly visited North American state nature parks where flight initiation distances of the deer are short (Sutton & Heske, 2017). This process also affects capercaillies Tetrao urogallus. Some capercaillie individuals have been reported to appear in villages, where they are not normally found, and to exhibit extreme fearless behaviors among humans. Such bold individuals, rather than being cases of some pathological condition, were most likely present in the past as well, but due to their boldness were quickly removed by human hunters. However, today, people treat bold individuals with respect and their genes for boldness could spread throughout the population. A final example of refuge abandonment by prey species involves great bustards Otis tarda, the heaviest flying bird species in Europe. Formerly persecuted by hunters and poachers in subsistence economies (see Alonso et al., 2005 for evidence of persecution in Morocco), the great bustard can now be found close to paved roads, freeways, villages, or factory parks in Spain (own unpublished observations) after its population recovered due to a hunting ban applied in 1980 (Alonso et al., 2003; Alonso & Palacín, 2010).

Final remarks

We think that the present change in human attitudes (relaxed persecution) toward wildlife, combined with the large scale abandonment of rural livelihoods in southern European countries during the last 3–4 decades, is leading to an increase in the abandonment of historical refuges by formerly shy-selected wildlife as well as encouraging a closeness of wildlife to areas used by humans. We also expect to see changes in the circadian rhythms of wildlife that could become progressively more diurnal. Brown bears in the Italian Alps show activity schedules that are the reverse of the human presence in their territories (Groff et al., 2017) and dingoes change their activity schedules when persecuted by humans (Brook, Johnson & Ritchie, 2012). However, raccoons Procyon lotor have become diurnal to take advantage of the food provided by tourists, who offer no threat to the raccoons, at the most highly visited National Park in Costa Rica (Manuel Antonio NP) (Farrera, 2017). Therefore, these changes in activity schedules are also to be expected in southern European prey and predator species, following the changed human attitudes in this part of the world. For example, Eurasian otters in NW Spain are not only leaving their former refuges in the middle and upper river stretches and getting closer to anthropic sites but they also show diurnal behavior along coastlines, rivers, streams, and reservoirs (own unpublished data). Wolves are already seen during the day, at close distance from a car, in some places in central and NW Spain where human presence is scarce and harmless (Fernández-Gil et al., 2010; R. Valle, pers. comm.) despite they are generally elusive because they are still persecuted due to interaction with ranchers (Llaneza, López-Bao & Sazatornil, 2012). The first stages of this process are already taking place in southern Europe, but we predict that they will also take place in many other areas of the developing world following the ongoing concentration of people in large cities and continued economic development. We propose that this is a positive development for our future, peaceful coexistence with wildlife within the frame of reconciliation ecology (Rosenzweig, 2003).

However, this coexistence will also bring about new challenges to our relationships with wildlife (Carter & Linnell, 2016). For example, some cases of attacks on humans by bold brown bears have already been reported in the Italian Alps (Groff et al., 2017) and the Pyrenees involving reintroduced individuals from Slovenia (http://eng.fundacionosopardo.org/). In France, some 35 000 car accidents due to wildlife were recorded in 2008, 36% involving wild boars, 17% roe deer, and 8% red deer (Ribier et al., 2012). Barbary macaques Macaca sylvana adjust their daily activity to coincide with tourists in Gibraltar from whom they get given food (O'Leary & Fa, 1993), but such interactions are not free of risk as these primates have an 88% seroprevalence for simian foamy virus that can infect humans (Engel et al., 2008). This paradigm shift will also bring about new pressures to protect or manage current high-quality unprotected areas (‘empty sites’), to where former refugees are relocating. This may all constitute a new research agenda (a new thinking model) and future research, which interpreted from this viewpoint, could reveal the triggers of this singular process in more detail, separating genetic and cultural components and timing its speed of development. The process of refuge abandonment has implications for conservation, behavioral ecology, population dynamics, spatial ecology, and the evolution of many vertebrate species. It should be taken into account in research agendas of ecological patterns explaining nonlinear anticipative responses of animal populations, habitat selection, models of spatial individual distribution with competition, dynamics of spatially structured populations or density-dependent dispersal, in, at least, the more industrialized countries. Although Europeans, once again, returned to their hill forts in medieval times, after the fall of the Roman Empire, let us hope that European wildlife will not need to do the same in the future.

Acknowledgments

We are thankful to Alejandro Izquierdo, Michel Cedenilla, Carlota Viada, Jordi Muntaner, Miguel Ferrer, Juan Carlos del Moral, Xunta de Galicia, Xurxo Mouriño (Arcea), Javier Ezquerra, Simone Tenan, Juan Carlos Alonso, Enrico Bassi, Guillermo Palomero/Fundación Oso Pardo, and Sébastien Renou for providing information on some of the examples used here. Josep Alós and Marta Vila provided constructive comments on an early draft. Jose M. López and Santiago Palazón (SECEM) provided the unpublished records of the otter surveys in Spain. Catherine Andrés helped us getting geographical information. AMA was supported by an Isidro Parga Pondal contract from Xunta de Galicia. Funding was also provided by Universidade da Coruña and Xunta de Galicia (reference GRC2014/050). Charlie Key reviewed the English language.