Our study on 81 bird species highlights the need to investigate how populations across a species' geographic range respond to habitat transformations. The study identifies four sensitivity patterns, including higher sensitivity at range edges, equal sensitivity, higher sensitivity at core ranges, and different sensitivity in populations at intermediate range positions or suitability levels. These patterns cannot be explained by dispersal ability, lower habitat specialization, or extension distribution range, suggesting that other factors are driving the observed sensitivity patterns. Our findings underscore the importance of considering the interactions between biogeographic and landscape factors on population abundances, which can result in either antagonistic or synergistic effects. These interactions can change the strength and direction of populations' responses to habitat transformations depending on their biogeographic position or environmental suitability. It is important to note that these results are specific to the dataset and analysis methods used, and further research may be needed to fully understand the relationship between species traits and sensitivity to environmental stressors. Nevertheless, our study provides valuable insights into how different populations of a species may respond to habitat transformations, highlighting the need for conservation strategies that take into account the varying sensitivities of different populations.
Henle et al. (2017) and Banks-Leite et al. (2022) propose biological mechanisms to explain the contrasting patterns of genetic diversity and differentiation observed in species across their ranges. These patterns can be influenced by multiple factors, including environmental suitability, dispersal capacity, and biotic interactions. The interaction of these drivers determines the extent to which a species can retain or lose its genetic diversity under habitat transformations or range shifts. When a species expands into naturally fragmented habitats or experiences slow natural habitat loss and fragmentation, the resulting environmental pressures may select for different beneficial alleles in different populations, leading to increased genetic diversity. However, when habitat transformation occurs rapidly due to human activities, such as deforestation or urbanization, genetic adaptation may not occur fast enough, resulting in decreased genetic diversity.
Moreover, the biotic interactions among species vary across their ranges, indirectly impacting genetic variability, genetic differentiation, and population sensitivity across a species' geographic range. The drivers influencing genetic diversity and differentiation may act synergistically or independently. Thus, a comprehensive understanding of the complex interaction among multiple factors is crucial to predict the consequences of habitat transformations and range shifts on genetic diversity and differentiation of species.
However, our study did not support the commonly assumed link between dispersal capacity and sensitivity pattern. Most species distribution models and predictions of extinction risk assume equal sensitivity to habitat transformations across geographic range or environmental suitability (Henle et al. 2004; Valladares et al. 2014; Boakes et al. 2018), but few studies have provided evidence to support this assumption. Valladares et al. (2014) simulations suggest that equal sensitivity among populations could occur when all populations of a species have the same magnitude of phenotypic plasticity and unlimited dispersal ability, resulting in higher gene flux across the distribution range (Vucetich and Waite 2003). Some evidence of genetic homogeneity across the distribution range supports this explanation (Johannesson and André 2006; Eckstein et al. 2006).
The highest sensitivity of range edge populations (or at the lowest environmental suitability) has distinct biological reasons for this pattern. For instance, it may arise from less suitable and lower-quality habitat available toward the range edge or may increase due to higher environmental instability along the geographic gradient (Holt and Keitt 2000, 2005; Hardie and Hutchings 2010). Under these stress conditions, populations should adapt less to new environmental changes (low resilience) due to lower genetic variability and stronger effects of genetic drift (Sagarin and Gaines 2006; Henle et al. 2017; Macdonald et al. 2017; Prieto-Ramirez et al. 2020). Therefore, populations at the range edge should be more sensitive compared to the range core. If habitat transformations interact synergistically with environmental suitability or biogeographic position, the decline in population size intensifies (e.g., Drymophila ochropyga and Xiphocolaptes albicollis, Fig. 3).
The highest sensitivity of core populations, as shown in Fig. 3c, may be due to their pre-adaptation to survive in harsh and spatially-temporally stochastic environments. This pre-adaptation is thought to result from selective pressures that vary across geographic ranges or environmental conditions. These pressures arise from trade-offs in resource allocation, conservation, and constraints imposed by resistance mechanisms, as predicted by theoretical models such as those developed by Safriel et al. (1994b) and Hoffmann and Blows (1994). Populations at the range edges or in habitats with lower environmental suitability are subject to extreme and variable environmental conditions, which create selection pressures for resistance to these stresses. This results in increased genotypic and phenotypic variation among the populations. These adaptations for resistance may improve the response of species to habitat transformations, as seen in the case of Tangara sayaca in Fig. 3c. There is growing evidence of adaptive evolution in sink habitats and genetic differentiation towards the range edge, supporting this explanation. For example, studies by Holt and Keitt (2005), Eckert et al. (2008) and Gaston (2009) have shown that range edge populations can differ in phenotypic characters and genetic structure. In some cases, they exhibit higher levels of individual fitness and phenotypic plasticity than most core populations, as demonstrated by Yakimowski and Eckert (2007) and Valladares et al. (2014)
Habitat transformations can have varying effects on environmental stress conditions, depending on the type of interactions (synergistic or antagonistic) and habitat preferences of species. For example, forest specialist species with lower dispersal ability, such as Drymophila ochropyga and Antilophia galeata, have shown a synergistic decline in abundance due to habitat transformations (Warren et al. 2001; Holt and Keitt 2005). However, for other species, such as Myiothlypis flaveola, Myarchus ferox, and Phaethornis squalidus (Fig. 5cb, Table S6), the landscape effect has opposite directions depending on their biogeographic position or environmental suitability. This opposite effect may seem contradictory to the expectation that individuals always select optimal habitat conditions to achieve the highest fitness performance (i.e., highest environmental suitability or lowest habitat transformations). However, when considering biotic interactions in habitat selection, suboptimal habitats may sometimes be the best choice to avoid intra- or interspecific competitors (Jacob et al. 2018; Banks-Leite et al. 2022) Thus, the effects of habitat transformations on species abundance and distribution are complex and depend on a variety of factors, including species' habitat preferences, biotic interactions, and the type of interactions between environmental stressors. Understanding these complex interactions is critical for effective conservation and management of biodiversity in the face of ongoing habitat transformations.
Previous studies have attempted to predict which species are most vulnerable to habitat transformations by analyzing morphological and biogeographical traits (Henle et al. 2004; Hatfield et al. 2018; Boakes et al. 2018). However, contrary to expectations, none of these traits tested in the present study, proved to be a reliable predictor of sensitivity to habitat transformations. It is possible that our approach, which used species-averaged traits to link with sensitivity groups, is not the most effective method for evaluating this relationship. Banks-Leite et al. (2022) suggest that a more promising approach is to examine population-level traits and how they vary across environmental gradients, as well as how they interact with habitat transformations. Another approach, proposed by Henle et al. (2004), is to evaluate trait interactions to better understand species' sensitivity response. In our study, we only tested the additive effect of these traits, but it is important to recognize that traits can also act synergistically or antagonistically on species sensitivity (Davies et al. 2004). Therefore, we recommend that future studies fill this knowledge gap by testing trait interactions to better understand the sensitivity response of species to habitat transformations.
Implications of sensitivity variation to habitat transformations
The varying responses of species to habitat transformations have significant implications for conservation strategies. Current conservation programs often implement uniform measures over a wide geographic range or avoid range-edge areas. However, our findings indicate that such strategies may not be effective for 36% of species due to the interactions between biogeographical and landscape factors that can alter the strength and direction of habitat transformation effects. This means that the minimal amount of habitat or landscape connectivity required to maintain viable populations will vary depending on the spatial position of the populations within a species' geographic range. Therefore, we recommend that species sensitivity to habitat transformations should not be extrapolated from one region to another without testing the effect of biogeographic range position on species abundance. We also urge wildlife managers to focus on protecting and restoring land along the range edge where several species have populations that are more sensitive to habitat transformations. In the case of Atlantic Forest endemics, these vulnerable areas are mainly in the transition region between the Atlantic Forest and the Cerrado.
Overall, our findings suggest that conservation strategies should be tailored to the specific needs of each species by considering their spatial position in the geographic range and the interactions between biogeographical and landscape factors. By doing so, we can ensure that conservation efforts are targeted and effective in protecting the most vulnerable populations of species.