The maintenance of human well-being and ecosystem services depends heavily on the effective conservation of biological diversity (Rands et al., 2010). However, the anthropogenic pressures on global biodiversity continue to increase which has resulted in an overall decline of the wildlife populations (e.g., Ceballos and Ehrlich, 2002). The greatest anthropogenic threats to the species are habitat degradation, fragmentation and destruction (IUCN, 2014). If the most pessimistic extinction rates (5% per decade) are right, more than half of all species on Earth could be extinct within 150 years (Costello et al., 2013).

Therefore, the global biodiversity will hardly survive without effective conservation measures. One of the most powerful tools for species and ecosystems conservation is the establishment and maintenance of protected areas (PAs) (e.g., Naughton-Treves et al., 2005). Protected areas are representative sites of ecological importance for biodiversity, ecological research, environmental education and local economy (CBD, 2011). However, the protected area establishment per se does not guarantee an appropriate representation or persistence of the threatened biota (e.g., Venter et al., 2014; Eduardo et al., 2012). A global assessment of more than 4000 threatened vertebrate species showed that only 15% of the species have their global geographic range adequately covered by protected areas (Venter et al., 2014).

Well-managed protected areas can achieve benefits beyond biodiversity conservation, such as delivering ecosystem services essential for human well-being (e.g., Chape et al., 2008). However, the protection status does not imply in the absence of human impacts and often protected areas suffer direct or indirect threats of human populations living within or adjacent to these sites (Chape et al., 2008). For example, protected areas can register a high number of roadkill that threaten local fauna (Garriga et al., 2012). The impacts of roadkill are more pronounced in large and slow-moving animals that use the roads regularly (Forman et al., 2003; Coffin, 2007).

The giant anteaters (Myrmecophaga tridactyla Linnaeus, 1758) are commonly dead by roadkills in the Cerrado (e.g., Cunha et al., 2010; Freitas et al., 2014). Only one and a half year of monitoring in a stretch of 216km found 30 giant anteaters killed by vehicles (Cunha et al., 2010). Another monitoring of 10 years on highways in the vicinity of the Parque Nacional da Serra da Canastra, a Cerrado protected area, found that on average 2.5 giant anteaters were lost annually by roadkills (Freitas et al., 2014). The mortality caused by roadkills can threaten profoundly the population persistence of giant anteater (Diniz and Brito, 2013). Therefore, the present study aims to evaluate the effectiveness of the national protected areas in the Cerrado to maintain viable giant anteater populations and to analyze the effects of roadkills on the persistence of these populations.

The results presented here have as main assumption the total isolation of populations, i.e. there is no migration among populations. Our results show that populations of 150 individuals are necessary to achieve demographic stability and 210 animals for genetic stability in the Cerrado. There was an expressive difference in the amount of protected areas able to maintain viable populations of the giant anteater in the Cerrado according to population density scenarios.

In the pessimistic scenario (population density of 0.085ind./km2) only three (Araguaia National Park, Serra Geral do Tocantins Ecological Station and Nascentes do Rio Parnaíba National Park) out of 18 protected areas are capable of maintaining viable giant anteater populations with demographic and genetic integrity (Table 1). Although the Grande Sertão Veredas and Serra da Canastra National Parks have size to sustain viable populations despite the potential effects of demographic stochasticity (1765–2471km2, 150–210 individuals), both presented high probabilities of falling below the critical size of 150 individuals (91% and 99%, respectively) (Table 1). Three small protected areas (1–1765km2, <150 individuals) (Pirapitinga Ecological Station, Silvânia National Forest and Contagem Biological Reserve) are not able to maintain any populations of giant anteater, with probabilities of extinction of 100% in the next 100 years. The population of the Lago do Cedro Extractive Reserve also presented a high likelihood of extinction (99%), with a mean time to extinction of about 32 years (Table 1).

In the optimistic scenario (population density of 0.4ind./km2), we found that 11 out of the 18 protected areas have sufficient area to maintain viable populations of giant anteaters (Table 2). The Brasília National Park should offer protection against the effects of demographic stochasticity but probably its population suffers from the negative effects of genetic erosion. However, this protected area had a high probability (94%) of falling below the population quasi-extinction threshold size (150 individuals) in the next 100 years (Table 2). According to this scenario Silvânia National Forest and Pirapitinga Ecological Station still presented a 100% probability of extinction for the next 100 years.

The sensitivity analysis showed that the population of 150 giant anteaters presented a high sensitivity to all factors. The changes in the modeled population output values influenced significantly the population mean growth rate, probability of extinction, mean final population size and expected heterozygosity (Table 3). In this population only the probability of catastrophes had no significant influence on the population mean growth rate. For the population threshold size of 210, the probability of catastrophes, inbreeding depression and reductions in sex ratio did not affect the probability of extinction. The changes in mortality rate and a 10% deviation in the sex ratio for males affected all parameters in both populations (Table 3).

The three roadkill rates (2.5%, 5% and 10%) were evaluated in 15 protected areas and 16 protected areas in the optimistic and pessimistic scenarios, respectively. Our results suggest that roadkill mortality is a serious threat to species persistence (Tables 1 and 2). The removal of individuals resulted in a rapid extirpation of the populations in all protected areas. In both pessimistic and optimistic scenarios, the mean time to extinction was not exceeding 60, 30 and 15 years for roadkill scenarios of 2.5%, 5% and 10% of the population, respectively (Tables 1 and 2).

The protected areas of Brazil might not be effective for the conservation of the some large mammals such as the lowland tapir (Tapirus terrestris) (Eduardo et al., 2012), and might be only reasonably effective for the conservation of the some small mammals (e.g., Brito and Grelle, 2004). The results presented in our study might raise concern because based on the pessimistic scenario only three protected areas have sufficient size to maintain giant anteater populations in the long-term. The Pirapitinga Ecological Station, Silvânia National Forest and Contagem Biological Reserve should have special attention from managers because they have no areas able to maintain any giant anteater population. The persistence of the giant anteater and other large and medium-sized species in these protected areas depends primarily on the habitat availability. Some potential solutions to increase the efficiency of these sites are the increase in size and/or in connectivity between protected areas.

The Lago do Cedro Ecological Reserve and Silvânia National Forest are sustainable use protected areas. These reserves allow the exploitation of natural resources in an environmentally responsible manner, at least in theory. However, a current study demonstrated that many natural resources exploited (in an alleged sustainable manner) in some sustainable use protected areas in Brazil do not obey this criterion (Fernandez et al., 2012). Thus, these protected areas do not have the minimum spatial requirements to ensure the persistence of the giant anteater and other species in the long-term, and the quality of habitat within these areas may not be suitable to the objectives of wildlife conservation.

A genetic study conducted in Emas National Park found that the population of giant anteater of this area has low levels of genetic diversity and high levels of inbreeding (Collevatti et al., 2007). We do not consider the inbreeding depression in our model. However, the effects of inbreeding depression in addition to the demographic and environmental instability could make small populations of giant anteaters even more vulnerable to extinction. This certainly might raise concern particularly in the context of the pessimistic scenario. The environmental heterogeneity and some threats such as hunting pressure and interference from domestic animals were not considered in our analysis. However, giant anteaters are hunted (e.g., Koster, 2008) and are attacked by dogs (Lacerda et al., 2009) in certain areas of its distribution. Both the critical population sizes (150 and 210 individuals) were sensitive to changes in mortality rate. The MVPs sizes identified here do not consider the threats mentioned above and therefore in such cases they cannot be sufficient to guarantee the persistence of species in the next 100 years.

Roadkills are one of the main sources of mortality for terrestrial vertebrates, surpassing even the effects of hunting in some places (Forman et al., 2003). The reduction in population size caused by direct mortality of animals is the primary and most obvious consequence of roadkill. However, the secondary effects range from behavioral changes, population fragmentation, genetic changes to increased risk of local extinction (Forman et al., 2003; Coffin, 2007). The giant anteater is a large mammal of slow movements and a limited vision, characteristics that make it more susceptible to be killed by vehicles. A study in the Brasília National Park showed that the roadkill is the greatest threat to the persistence of giant anteaters, having a more serious negative effect than fire and inbreeding (Diniz and Brito, 2013).

The implementation of mitigation measures against roadkills (e.g., installation of speed bumps and traps, revision of the speed limits, more effective traffic warning signals and construction of wildlife overpasses) was suggested as an important conservation action for a giant anteater population in the Cerrado (Diniz and Brito, 2013). In order to reduce the impacts of roadkills, we strongly recommend that these measures should be adopted in all Cerrado protected areas under the influence of roads. These measures should be complemented by the implementation of monitoring programs (to monitor both wildlife populations and roadkills) in order to estimate qualitative and quantitative impacts of roadkills and to evaluate the effectiveness of the measures implemented.

The authors declare no conflicts of interest.