DISCUSSION
Our results show that the occurrence of the three Amazonian canid species C. thous,S. venaticus and A. microtis may be negatively affected by climate change in the future, with losses of their entire suitability area in South America from 15% to 91% and in the Brazilian Amazon losses ranging from 18% to 56%. However, the effects can be differentiated between the species according to their ecological demands and the environmental variables that can modulate their climatic niches. For C. thous and S. venaticus , the climate variables and those associated with anthropogenic disturbances that modulate their niches today, may not act in the same way in the future. Thus, the rapid environmental changes might function as ecological traps for these species (Robertson and Hutto, 2006), and in the future they may no longer survive in most of the areas considered suitable for them today.
Among the three species A. microtis has the most restricted distribution confined to the Amazon Forest and the one with the least projected loss of suitable climatic areas in the future. Apparently, this species is a specialist in forest habitats, in particular in continuous and preserved forest (Michalski, 2010; Leite-Pitman and Willians, 2011). Probably, the loss of suitability for this species is directly related to the loss of deforestation and forest degradation (Rocha et al., 2020). In addition to the distribution of the climatic niche of this species mainly coinciding with the most conserved regions of the Amazon Forest, our results showed a strong influence of Forests and Water on the occurrence of A. microtis . The association of this species with watercourses is also known from the literature and is suggested by interdigital membranes which may be an adaption for moving on soft soil (Leite-Pitman and Williams 2004; Castelló, 2018). Although the proximity to urban areas was not one of the main variables determining the occurrence of A. microtis , we found that this variable was negatively related to its occurrence, corroborating its low tolerance to disturbed habitats (Michalski, 2010; Leite-Pitman and Williams, 2011; Rocha et al., 2020). Thus, the loss of suitability areas in the future for A. microtis cannot be related to the ecological trap theory. This species will probably no longer occur in unfavorable habitats due to its fitness for higher-quality forest habitats.
For A. microtis , one of the consequences of deforestation and human pressure is increased hunting (Constantino, 2016), including hunting with domestic dogs. Thus, disease transmission by domestic dogs will pose a threat to A. microtis (Schenck and Stail, 2004; Leite-Pitman and Williams, 2011). Our results corroborate those of Rocha et al. (2020), who expected a loss of 30% of the A. microtisdistribution area by 2027, due to forest loss. Based on the last three years of high deforestation rates in the Amazon (PRODES, 2021), and using a pessimistic projection for the future, the loss of area of occurrence of this species could be much more severe.
Considering habitat selection, an animal cannot always make the best choice about where to live (Kristan, 2003; Stamps and Swaisgood, 2007), but its initial choice may later affect the individual’s survival and its reproductive success. However, the ultimate reason that determined the choice may have not been evident when the choice was made (Hutto 1985; Kristan, 2003). In the case of altered environments in the Amazon, it is possible that the attractiveness of these habitats by species such as C. thous or S. venaticus , are detaching themselves from the species’ aptitude for survival and reproduction. The preference of these species for anthropogenic environments was evident from our GLM analyses, where natural open areas negatively influenced the occurrence of C. thous , while anthropogenic open areas had a positive impact. The same was apparent for S. venaticus where anthropogenic open areas had a positive effect on its occurrence. These lower-quality habitats may at some point be just as attractive as preserved native habitats. This can happen when animals whose behaviors have been shaped by exposure to different environmental conditions in their evolutionary history, are suddenly confronted by altered or novel environments (Iwasa and Levin, 1995). In these situations where impoverished environments become more attractive and ”bait” individuals, these environments are called ecological traps (Gates and Gysel, 1978; Robertson and Hutto, 2006).
Of the three species surveyed, C. thous appears to be the least demanding in terms of habitat, with aptitude for open environments and a high tolerance for anthropogenic habitats (Ferraz et al., 2010). In the Amazon, this species is common in natural non-forest environments, such as in areas of Canga Metalófila in Carajás, State of Pará (Carvalho et al., 2014), or in areas of the Amazon Cerrado in the State of Amapá (Coelho et al., 2014). In our occurrence database from the Amazon, we had no record of C. thous in areas of continuous forest and outside naturally open environments. On the other hand, C. thoushas been commonly recorded in areas with eucalyptus (Coelho et al., 2014) and oil palm plantations (Mendes-Oliveira et al., 2017), and may occur at the edge of adjacent forests, but not inside them. It also occurs in mining areas, with a high level of degradation and locations in the initial phase of forest recovery in the region of Paragominas State of Pará. Interestingly, two individuals of C. thous were recorded in 2020 within the city of Belém, State of Pará, in the Amazon (Mendes-Oliveira, personal comm.). Cerdocyon thous does not seem to be distributed within forests, and deforestation has favored its expansion in the Amazon biome. The same occurred with Chrysocyon brachyurus , (maned wolf), which expanded its extent of ​​occurrence in southwestern Amazonia (see Silva-Diogo et al., 2020), precisely due to the conversion of forests into open areas in southeastern Amazonia (Silva-Diogo et al., 2020).
Our current model showed more outstanding suitability of C. thousto more open environments, such as Cerrado and more anthropized open areas in the Amazon, with restrictions to more forested areas, such as in the central and northwestern portions of the Amazon. The GLM results also corroborate the conclusion that C. thous has been positively influenced by urban areas and anthropogenic open areas and has been negatively by forest. However, this species was also negatively influenced by natural open areas, suggesting a greater fitness for anthropogenic habitats over natural areas. Although canids generally do not show cursor adaptations like those of ungulates (Smith and Savage, 1965), C. thous has a more cursorial form than the other two canid species studied in this research (Penrose, 2019). Longer limbs favor more extended travel, especially in open areas. Faced with a scenario of degradation and deforestation, this may be a morphological adaptation that favors the entry of this species into degraded, previously forested environments. Another characteristic that confers an advantage to the ecological plasticity of C. thous , aboutA. microtis and S. venaticus in the Amazon, is its greater tendency to omnivory. The ability of this species to have a poor or meatless diet for long periods, feeding opportunistically on what the environment offers (Macdonald and Courtenay, 1996; De Oliveira, 2009), allows it a more remarkable ability to thrive in a broader range of backgrounds and conditions. Considering all these characteristics, we would imagine that the deforestation of the Amazon Forest has been favoring C. thous . However, in the long term, this species may be be “baited” by lower-quality environments, reducing its success and leading it to extinction in certain areas in the future.
Considering our future projections of climate suitability for C. thous , we observed a significant loss of suitable areas, especially in regions already open or currently degraded by anthropogenic action, mainly in the southern limits of the Amazon Forest and the northeast of this biome. These climate niche loss projections suggest that impoverished environments currently suitable for C. thous may not perpetuate in the future. Thus, these new anthropogenic environments may be working as an ecological trap for C. thous in the Amazon.
The same may happen with S. venaticus . This species is considered rare (De Moura Bubadué et al., 2016), and occurs in forest habitats and forest edges (Zuercher et al., 2005), open areas such as savanah (Zuercher, 2001), and altered habitats (DeMatteo and Loiselle (2008). Despite not being a habitat specialist, S. venaticus is the most carnivorous canid in South America. And this is due to the presence, in individuals of this species, of a large zygomatic arch, a short snout with a reduced number of molars and a pointed lower first molar (LM1), important adaptations for hypercarnivory, which increases the efficiency of meat consumption (Ewer, 1973; Van Valkenburgh and Koepfli, 1993; Beisiegel and Zuercher, 2005; De Moura Bubadué et al., 2016)
Thresholds for the occurrence of S. venaticus appear to be less restrictive than for A. microtis but more stringent than forC. thous . DeMatteo and Loiselle (2008) mention that 20% of the potential distribution areas of S. venaticus were associated with degraded areas, using land-use data from 1992 to 1993. However, Michalski (2010) recorded S. venaticus only in continuous regions of Forest, even applying an effort using camera traps in fragments adjacent to continuous areas. In our study, the current model showed broad distribution suitability in South America for this species, but with a significant loss in the future (the most extensive loss among the three species), especially in the Amazon region. The future loss of climatic niches in the Brazilian Amazon is concentrated both in forest areas and areas already anthropized, indicating that this species may suffer from the loss of high and low-quality habitats. However, the GLM indicated that S. venaticus is positively influenced by forest, urban areas, and anthropogenic open areas. This suggests that S. venaticus may also be involved in an ecological trap when it chooses lower quality habitats that may not be enough to maintain its survival in the future. However, unlike C. thou , S. venaticus also uses forest habitats, which allows it to survive in areas that can be protected in the future in the Amazon.
Evolutionarily seen, the climate has been a more decisive factor than competition in explaining morphological variation and habitat options in canids (De Moura Bubadué et al., 2016). They showed that the skull morphology of A. microtis and S. venaticus is more related to warm, humid, and less seasonal environments than C. thous . The use of different habitats and different diets reduces competition between them. In this way, the coexistence of these species can be compromised not by competition between them but, on the contrary, by differences in climatic adaptations and habitat options, as our results show.
Ecological traps occur when species erroneously choose niches where their suitability is lower than in others after changes in the environment (Robertson et al., 2013). Considering the evolutionary history of species, environments altered by human actions could provide false clues to their adaptation (Sclaepfer et al., 2002). In the medium and long term, these ecological traps can lead to significant decrease or even extinction of populations in these habitats before adaptation to the new modified environment has occurred (Hale and Swearer, 2016). ForC. thous and S. venaticus , we believe that rapid changes in landscapes (habitat degradation) will lead to “wrong choices” for environments that will soon have large climatic variations, leading these species to be trapped in these ecosystems, without time to adapt to them.
Further, both S. venaticus and A. microtis appear to be forest-dependent and must rely on the goodwill of decision-makers to be maintained in the future. However, even though C. thous is least dependent on the Amazon Forest, this species is probably the one that will be most affected in the future due to the ecological traps that the region has offered. Speothos venaticus , can also suffer from the same process, but if the forest is maintained, the impact on this species can be less intense. Using the canid species as a model, we could deeply investigate these ecological effects that can affect a large part of the Amazonian fauna in the current scenario. The future perspectives for preserving terrestrial vertebrate fauna in this region are pretty pessimistic. When it comes to the conservation of impoverished regions in the Amazon, the theory of ecological traps must be studied and discussed at the same level that habitat loss is considered a decisive criterion for biodiversity threat.