Functional diversity of snakes is explained by the landscape composition at multiple areas of influence

Abstract Roadkill and landscape composition affect snakes at different spatial scales, depending on the functional trait value of the species, which is reflected in the functional diversity indices at the assemblage level. This study evaluated the effect of roads and landscape composition on snakes' functional diversity at different areas of influence (250, 500, 1000, and 2000 m buffer areas). We compared roadkill snake species with those assemblages inhabiting the adjacent vegetation in the Orinoco region, Colombia. We surveyed snakes using transects on the road and adjacent areas on 13 landscapes along the road. We evaluated the effect of 16 landscape metrics at six land cover classes on the snake's functional diversity at four different areas of influence (from 250 to 2000 m around the sampled sites). The functional redundancy index was higher for roadkill species, suggesting that roads eliminate species that play similar roles in the assemblage and ecosystem processes. Likewise, the low values of functional redundancy in the adjacent vegetation call attention to the fact that each species surviving in this transformed landscape has a crucial active role in ecosystem processes in snake assemblages. For roadkill snakes, forest metrics explained changes in functional richness and functional evenness at a 250 m area of influence. In comparison, transient crop and pasture metrics explained changes in functional evenness and divergence at 2000 m. For snakes inhabiting the adjacent vegetation, the cohesion of pasture explained changes in functional richness at 250 m, and forest metrics explained changes in functional redundancy and evenness at 2000 m. Anthropogenic landscape transformation may have a greater effect on snake functional diversity at local scales than roadkill. In savanna ecosystems, the presence of native forest at 2000 m radius around roads promotes the conservation of snake assemblages. However, within a 250 m radius, the risk of snake roadkill increases when the road borders native forest. Therefore, it is necessary to implement wildlife crossing in these sections of the road.

13 landscapes along the road. We evaluated the effect of 16 landscape metrics at six land cover classes on the snake's functional diversity at four different areas of influence (from 250 to 2000 m around the sampled sites). The functional redundancy index was higher for roadkill species, suggesting that roads eliminate species that play similar roles in the assemblage and ecosystem processes. Likewise, the low values of functional redundancy in the adjacent vegetation call attention to the fact that each species surviving in this transformed landscape has a crucial active role in ecosystem processes in snake assemblages. For roadkill snakes, forest metrics explained changes in functional richness and functional evenness at a 250 m area of influence. Anthropogenic landscape transformation may have a greater effect on snake functional diversity at local scales than roadkill. In savanna ecosystems, the presence of native forest at 2000 m radius around roads promotes the conservation of snake assemblages. However, within a 250 m radius, the risk of snake roadkill increases when the road borders native forest. Therefore, it is necessary to implement wildlife crossing in these sections of the road.

K E Y W O R D S
functional redundancy, land use/land cover change, reptiles, roadkill, scale of effect, trait probability density

| INTRODUC TI ON
Road development is one of the main drivers of land use and cover change, leading to habitat fragmentation, degradation, and loss of native species (Dixon et al., 2022;Moore et al., 2023). This anthropogenic drive also increases fauna collision probability, which varies depending on the value of the functional traits related to the species' vulnerability to roadkill (González-Suárez et al., 2018;Rytwinski & Fahrig, 2012). Thus, the change in the composition and dominance of native species in the assemblages due to the roadkill of species with greater vulnerability is reflected in changes not only in taxonomic diversity (Coffin, 2007) but also in functional diversity, which is defined as the value and range of functional traits among the organisms, and how these traits are linked to the ecological functions and interactions of these species with their surrounding environment (Carmona et al., 2019;Mason et al., 2005). However, the composition and configuration of the landscape also mediate the effect that roads can have on assemblages, depending on the degree of complementation or supplementation the native species make in the landscape (Jackson & Fahrig, 2015;Moraga et al., 2019).
Variability in mortality rates of roadkill species has been described across life-history traits such as ecological habits, behaviors, movement patterns, and reproductive times, among others
Para las serpientes atropelladas, las métricas de cobertura de bosque explicaron los cambios en la riqueza y la uniformidad funcional en un área de influencia de 250 m.
En comparación, las métricas de cultivos transitorios y pastos explicaron los cambios en la uniformidad y divergencia funcional a 2000 m. Para las serpientes que habitan la vegetación adyacente, la cohesión de los pastizales explicó los cambios en la riqueza funcional a 250 m, y las métricas de cobertura de bosque explicaron los cambios en la redundancia y la uniformidad funcional a 2000 m. La transformación antropogénica del paisaje puede tener un mayor efecto en la diversidad funcional de las serpientes a escalas locales que los atropellos. En ecosistemas de sabana, la presencia de bosque nativo en un radio de 2000 m alrededor de las carreteras favorece la conservación de los ensamblajes de serpientes. Sin embargo, en un radio de 250 m se aumenta el riesgo de atropellamiento de serpientes cuando la carretera limita con bosque nativo, por lo que es necesario implementar pasos de fauna en estos tramos. animal species may be more vulnerable to vehicle collisions and road mortality than others (Andrews et al., 2015), influencing changes in the functional diversity of assemblages (Trimble & Van Aarde, 2014).
Road mortality is related to reptiles' ectothermic condition because the roads are heat sources (D'Amico et al., 2015). Snakes are some of the vertebrates most affected by roads (Mccardle & Fontenot, 2016;Sosa & Schalk, 2016), and their mortality is influenced by the wide range of life-history traits of species within an assemblage, as well as their habitat preferences, daily and seasonal activity patterns, vagility, and foraging strategies (Bonnet et al., 1999).
Foraging strategy is an important trait that defines how a snake searches for and captures its prey, which can be active foraging (which requires a permanent expenditure of large amounts of energy in the exploration of the environment and the pursuit of prey), or sit and wait (in which the snake ambushes the prey after waiting for a long time without moving within a microhabitat) (Lillywhite, 2014).
In that sense, it is expected to find a greater number of roadkill species whose foraging strategy is to actively search for prey because they have greater dispersal throughout the habitat. Also, the temporal dynamics of foraging reflects the circadian rhythm of the species, which is determined by melatonin levels in the body, and in turn affects internal body functions related to reproduction and exploration of light and temperature gradients in the environment (Lillywhite, 2014). It is expected for nocturnal species to be run over more frequently than diurnal while acquiring heat from asphalt (Das et al., 2007), one of the few surfaces that can maintain heat, which makes it an ecological trap (Mccardle & Fontenot, 2016). Finally, habitat preference may determine species' vulnerability to anthropogenic changes in natural cover. For example, fossorial species may be less affected by human constructions and infrastructure (Lillywhite, 2014) and inhabit pastures, while arboreal species tend to habit on the edges and interior of remnant native forests (Urbina-Cardona et al., 2008). In this sense, terrestrial species are more likely to be victims of roadkill compared to species with different habitat preference. Rincón-Aranguri et al. (2019), evaluated how these life-history traits were related to landscape composition to explain road mortality. However, such an approach fails to clarify how the landscape composition explains roadkill and its effect on functional diversity. In a complementary way, considering fixed traits (such as life history) and field-measured traits that vary within and between populations (Shipley et al., 2016) can give a helpful insight into how the roadkill events structure snake assemblages.
Measuring the functional traits of species living in the adjacent vegetation in contrast with road-killed can be vital to understanding the link between the assembly structure and road risk factors (Lawing et al., 2012). Trait selection is a core topic in understanding the spe-cies´ response to environmental conditions (Shipley et al., 2016), and for snakes, it has been reported that the body size is sensitive to anthropogenic landscape transformation (Doherty et al., 2020;Pfeifer et al., 2017;Todd et al., 2017) and the pattern of road-killed snakes (Mccardle & Fontenot, 2016;Sosa & Schalk, 2016) because of their association with vagility and home range (Andrews & Gibbons, 2005;Bonnet et al., 1999;Doherty et al., 2019). Total size and tail size in snakes are related to a kind of locomotion, habit (i.e., fossorial vs. arboreal), mate acquisition, and habitat preferences (Jayne, 1986;Lawing et al., 2012) and represent the risk of roadkill due to their influence on the shape and velocity of movement over the asphalt on the road (Andrews & Gibbons, 2005). In addition, both traits are good proxies for the matter and energy that everyone contributes to the ecosystem as prey, as a carcass, or as a top predator (Cortés-Gomez et al., 2015;De Miranda, 2017;Lillywhite & Henderson, 1993). Then, larger body size increases the risk of mortality in roadkill events (Roe et al., 2006), which affects the population's fitness (Barron & Andraso, 2001) and predator-prey dynamics (King et al., 2002).
The Colombian Orinoco exhibits a highly heterogeneous landscape with natural regions such as savanna, Moriche palm fragments (Mauritia flexuosa), and gallery forests imbibed within anthropogenic land use areas such as production systems and pastures (Rangel-Ch et al., 1995). Since 1970 this landscape has been transformed by introducing African grasses for cattle ranching (Etter et al., 2010), In this study, we used the snake's body and tail size intraspecific variation to explore the range of values and the size of the variation at the assemblage level that is lost due to the runover. We selected the trait probability density as an analytical approach that considers all field-observed variations on species abundance and functional trait value variation to unbiasedly calculate functional diversity indices (Carmona et al., 2019). Based on the distribution probabilities of both size traits values, it is possible to calculate the size of the functional volume occupied by each species (Functional richness per species or FRic_sp), the functional volume of all species that make up the assemblage (Functional richness or FRic), the equitability in the distribution of the abundances of the different values of the traits (Functional evenness or FEve), the distribution of abundances within the functional volume occupied by each assemblage (Functional divergence or FDiv) (Mason et al., 2005) and the volume of trait space overlap between species in a assemblage (Functional redundancy or FRed; Carmona et al., 2019).
We aimed to identify the variation in functional diversity between road-killed species and those assemblages inhabiting the adjacent vegetation, as a function of changes in landscape composition at 13 surveyed landscapes on the road between Villavicencio city and Puerto López town at the Orinoco region of Colombia. Our specific goals were to: (1) describe the variation in five functional traits (three fixed traits: foraging strategy, temporal dynamic foraging, and habitat preferences; and two traits measured in the field: body and tail length) and the functional richness index per species (FRic_sp) of roadkill snakes and those found alive in the adjacent vegetation; (2) evaluate changes in the functional diversity indexes measured at the assemblage level between roadkill snakes and those inhabiting adjacent vegetation; (3) spatially represent changes on functional diversity indexes of snakes from the road and the adjacent vegetation cover for the 13 landscapes surveyed; and (4) determine the effect of landscape composition on changes in the functional diversity of roadkill snakes assemblages and those found alive inhabiting adjacent vegetation, at four buffer areas of influence (scale of effect; Jackson & Fahrig, 2015).
We expected higher values of functional diversity for snakes inhabiting the adjacent vegetation than for roadkill snakes, suggesting that roads may filter a small portion of the regional pool of snake species that would be functionally similar (i.e., have high functional redundancy sensu Adams et al., 2022). This diversity of roadkill snakes would be explained by metrics related to the extent and connectivity of native forests, mainly within 250 and 500 m of the sampled site. This is because species inhabiting areas with high canopy cover tend to seek open areas to thermoregulate, being run over when exiting onto roads (Andrews et al., 2015). In contrast, the functional diversity of snakes inhabiting the adjacent vegetation was expected to be explained by metrics related to the extent of the different natural and anthropogenic classes (at 1000 and 2000 m around the sampled site) due to the ability of some species to perform landscape complementation and supplementation when land cover types offer refuges and hedgerows for their dispersal (Graitson et al., 2020;Lecq et al., 2017).

| Study site
The study was carried out in an urban-rural area in the Llanos Orientales of the Colombian Orinoco, at the Department of Meta, Colombia, on a paved road (width of 10 m) between Villavicencio city (4°06′41″ N and 73°36′16″ W) and Puerto Lopez town (4°04′51″ N and 72°58′11″ W) (see Appendix S1). The region has a unimodal bi-seasonal rainfall regime, with an annual mean precipitation of 2414 mm and a mean temperature of 25.7°C (Instituto de Hidrología y Ministerio de Ambiente, 2005). The road is between 192 and 400 masl and has an extension of 79.7 km that travels from the foothills of the Eastern Cordillera to the Orinoco natural savannas.
A suburban mosaic settlement dominates the current landscape, added to natural gallery forest remnants, natural savannas with linear palm elements (Morichales), exotic pastures for extensive ranching, and agricultural systems (Romero-Ruiz et al., 2011).

| Snakes sampling
We selected 13 landscapes along the road based on the inspection of satellite images. The landscapes were located at a minimum distance of 1 km (considering Neotropical reptile dispersal; Mendenhall et al., 2014)

| Functional traits and diversity
For each snake found, we measured in the field two continuous morphological traits (Total length, Tail length) following a standard protocol (Rivas et al., 2008). When it was not possible to obtain a value for any of these traits from field data (i.e., because corpses were too damaged, tailless, or individuals escaped), we used the mean value reported in the literature for that species (five out of 119 total individuals found). Both continue traits were used to calculate the functional richness (FRic), functional evenness (FEve), functional divergence (FDiv), and functional redundancy (FRed) as described below.
Patterns of species richness were examined considering the three categorical functional traits (foraging strategy, temporal dy-

| Characterization of landscape composition
We used a vectorial layer from a classified image with a 1:25,000 spatial scale from the GEF-BID project (2016) to evaluate landscape structure and composition on each of the 13 landscapes along the road. Because the vectorial layer comprised Landsat eight images from 2014, and our snake field surveys were conducted from 2017 to 2018, we validated the land use and land cover during our field trips. We modified the land classes manually (adding and removing polygons to update the vegetation type in 2014). Finally, we grouped the vegetation cover types into six land classes: Native Forest, pastures, transitory crops, permanent crops, bodies of water, and urban infrastructure.
For each of the 13 landscapes, we converted the polygons to raster layers with a 6 m × 6 m pixel size according to the scale (1:25,000) following the user manual of FRAGSTATS 3.4. software (McGarigal & Marks, 1995). Each landscape was defined as "a spatial area with a diameter exceeding the dispersal distance of the species of interest" (Driscoll et al., 2013). We selected four buffer distances for each of the 13 landscapes, 2000 m (exceeding the dispersal distance of the most vagile species on the assemblage), 1000, 500, and 250 m. We defined the largest buffer for the landscape analysis considering the highest dispersion reported for some species in the region (Jackson & Fahrig, 2015), of which Eunectes murinus has an average distance per week of 1500 m (Rivas et al., 2016). We clipped the previously classified raster layer at all buffers for the 13 surveyed landscapes. We used FRAGSTATS 3.4. software (McGarigal & Marks, 1995) to calculate 16 metrics (Table 1) for each land class at each of analysis buffers.  Then we calculated the TPD at the assemblage level (TPDc), considering their species' relative abundances on each sampled site. being low on the road and higher further into the adjacent vegetation cover. IDW assumes that each measured point has a local influence that diminishes with distance (Shepard, 1968) and is one of the most used interpolation methods in environmental science (Li & Heap, 2011). We performed a deterministic method instead of a geostatistical method because the number of points we had was insufficient for the second one (Hengl, 2007).

| Variation in the functional traits of roadkill snakes and those found alive in the adjacent vegetation
We found 27 species on the road and 14 species on the adjacent vegetation. Ten (10)  All species that inhabited the adjacent vegetation and 82% of the roadkill species presented an active foraging strategy (Figure 3a).
Regarding the temporal dynamics of foraging, 72% of the species that inhabited the adjacent vegetation were nocturnal, but only 39% of the nocturnal species were road-killed ( Figure 3b). Finally, regarding habitat preferences, all species that inhabited the adjacent vegetation and 82% of the roadkill species were terrestrial (Figure 3c).
The bootstrap region for the roadkill snake assemblage displays a functional space that is nested within the snake assemblage living in the adjacent vegetation (Figure 4).

| Changes in the functional diversity indexes measured at the assemblage level between roadkill snakes and those inhabiting adjacent vegetation
We did not find statistical differences in functional richness

| Spatial representation of changes in the functional diversity indexes of snakes between road and adjacent vegetation cover
We found variability in the index values across the studied landscapes ( Figure 5). The spatial representation of these values for each functional diversity index corroborates this spatial heterogeneity from the road to the adjacent vegetation ( Figure 6).  (Table 3).  (Table 3).

| DISCUSS ION
Anthropogenic changes in land use impose environmental filters on reptile assemblages (Todd et al., 2017), and evidence suggests an increase in functional redundancy and functionally homogenizing (Adams et al., 2022). In this study, we found that the effect of landscape composition on functional diversity was higher than roadkill effects for snake assemblages along 13 landscapes on the road located in the Orinoco region of Colombia.

F I G U R E 2
The relative position of species in the Total Length (left) and Tail Length (right) ratio of roadkill snake species (Black circle) and those from the adjacent vegetation (Gray circles) in the Colombian Orinoco Region. The x-axis represents the difference between the species median trait value (Total Length and Tail Length) and the assemblage trait value.

| Roadkill snakes: From functional traits to functional diversity indices
In contrast with previous reports of lower functional richness and divergence (FRic and FDiv, respectively) in disturbed sites (Adams et al., 2022;Berriozabal-Islas et al., 2017), we did not find statistical differences between roadkill and adjacent vegetation snake assemblages. However, the amount of functional space occupied by roadkill snakes was slightly small compared with the snake as- Our results support the hypothesis that terrestrial species are more likely to be roadkill. However, it should be noted that some snakes that are run over may also be arboreal or fossorial. This pattern of snake roadkill, related to species habitat preference, had been previously reported by Sosa and Schalk (2016) across diverse ecosystems in central Bolivia, who draw attention to the importance of seasonality in rainfall for a better understanding of road ecology. We suggest that the overrun of aquatic and semi-aquatic species (22% of the roadkill species) could be explained because, during the rainy season, the water channels designed to ensure the drainage of roads in Colombia (INVIAS, 2009) overflow and flood the road, allowing the temporary dispersion of these species on the road. Likewise, we believe that, during the rainy season, when the subterranean habitats of fossorial and semi-fossorial snakes (35% of the roadkill species) are flooded, these species are forced to seek higher ground, reaching the road. In both cases, these (semi-)aquatic and (semi-)fossorial snake species are trapped on the road when the water level drops, leaving them at a high risk of being run over. In this sense, it is important that future studies evaluate changes in the value of functional traits and functional diversity indices considering variation in seasonality and even in the type of road construction (Machado et al., 2015).
However, the relationship between daily temporal dynamics in foraging and snake strike is still unclear. The present study's results do not support the hypothesis that nocturnal snakes would be more likely to be run over. We expected higher mortality in nocturnal snakes because, at the beginning of the night, individuals tend to seek the heat accumulated on the asphalt (Das et al., 2007;Hallisey et al., 2022), increasing the probability of being run over by drivers who also only manage to detect the snake on the road when they are too close to avoid it. However, 25% of the roadkill species show a temporal dynamic of diurnal foraging and 36% present a diurnal and nocturnal dynamic. The lack of consistent patterns in snake strike-downs based on their temporal foraging dynamics may be due to behavioral and ecophysiological traits not explored in the present study. Still, they also call for the need for robust databases of species' life-history traits, as they have been consolidated for other groups such as plants (Kattge et al., 2011), fish (Brosse et al., 2021;Mull et al., 2022), birds (Tobias et al., 2022) and amphibians (Oliveira et al., 2017).
The trend to lower value of FEve on roadkill snakes shows that the probabilities associated with tail and total length values were less even (i.e., uneven relative abundances) on the road than for the adjacent vegetation. This result contrasts with the one reported by higher FEve values for lizards in mature natural environments without human disturbance. Snake species tend to occupy peripheral positions within the entire assemblage trait distribution. Therefore, we can suggest that most species contribute to community and ecosystem function.
The highest redundancy (FRed) on roadkill snakes suggests that they occupy the same functional space in the assemblage despite the greatest number of snake species registered. This is a functional space that increases vulnerability to being run over. Roadkill species showed a narrow range (or an aggregated pattern) of size values in the functional space, a pattern like that observed with snake assemblages on burned sites and with thinning operations (Adams et al., 2022).

| Effect of landscape composition on functional diversity at multiple scales for roadkill species and adjacent vegetation assemblages
Recent reviews found that snake species with small size ranges tend to respond negatively to landscape transformation, even though no F I G U R E 4 Bootstrap regions for the assemblages of roadkill snakes (blue triangles) and snakes living in the adjacent vegetation (red inverse triangles).

TA B L E 2
Results from the one-way PERMANOVA analysis of functional diversity indexes between roadkill snakes and snakes inhabiting the adjacent vegetation. The p(perm)-values in bold indicate significant differences in the Pseudo-F test for a specific factor. direct effect from body size was found for some species (Doherty et al., 2019;Todd et al., 2017). Some studies suggest that small individuals move more frequently but travel shorter distances, narrowing their range (Blouin-Demers et al., 2007). Other studies propose a relationship between body size, the degree of habitat specialization, and the risk of extinction of Squamates but with no emphasis on snakes (Böhm et al., 2016;Tingley et al., 2013). In this sense, we believe it is essential to evaluate the effect of landscape composition on the functional diversity of snakes in different areas of influence, representing variation in the dispersal capacities of the species within the assemblage (following Driscoll et al., 2013).

Functional diversity index
We found that the most significant estimated component of variation for all functional diversity indices resides in the residuals ( Table 2), indicating that the variance of these indices within each local landscape is greater than the variance between landscapes. This result suggests that the high heterogeneity in the composition of each landscape shapes the snake assemblages but with differential effects of landscape metrics between roadkill species and assemblages inhabiting adjacent vegetation. The habitat heterogeneity hypothesis, one of the foundational concepts in community ecology, suggests that structurally complex habitats provide more ecological niches and resources for a wide variety of species to exploit, thus supporting more diverse species assemblages (Bazzaz, 1975). Many examples of heterogeneous habitats provide varied resources (structural and provisional) for a larger number of species than more homogeneous habitats (Kissling et al., 2008;McElhinny et al., 2005;Tews et al., 2004).   (Parent & Weatherhead, 2000). In contrast, other species prefer natural open environments (such as savannahs) and even anthropogenic environments (such as cattle pastures), suggesting that habitat quality at local scales, in terms of resource availability, may be more critical for reptiles than the configuration of the whole landscape (Cunningham et al., 2007;Fischer et al., 2004).
We hypothesized that the amount of forest would explain the functional diversity of roadkill snakes at smaller spatial scales (250 and 500 m of the sampling site). The interpolation analysis con-

| CON CLUS IONS
In contrast with our expectation, results did not show higher values for functional diversity indices of snake assemblages inhabiting the adjacent vegetation than roadkill snakes' diversity. Regarding the ecology of the snakes, roadkill is reducing the population sizes of some species even though they have high functional redundancy.
The effect of roadkill is eliminating species that play similar functional roles in the assemblage and the ecosystem processes and services (i.e., from nutrient cycling and energy flow through food chains in their role as prey to pest control, secondary seed dispersal, and pharmacological properties of venoms, among others; Cortés-Gomez et al., 2015;Reiserer et al., 2018). We consider it essential to emphasize that the functional redundancy of snakes in the adjacent vegetation was significantly lower than that on the roads, highlighting these assemblages' high vulnerability to landscape transformation. Despite not finding statistically significant differences in other indices of functional diversity among landscapes, the absolute patterns of the functional diversity index between both snake assemblages varied across the landscape ( Figure 5). Although the absence of differences in the values of functional diversity indices between roadkill snakes and those inhabiting the adjacent vegetation may indicate functional homogenization of snake assemblages in the study landscapes, future research should investigate changes in functional diversity by differentiating between landscape elements that constitute snake habitat (e.g., Mendenhall et al., 2014) at a finer sampling grain. This will help to identify the adequate spatial scale to understand the study phenomenon, conduct the sampling, and run the statistical analyses (as recently proposed by Fletcher et al., 2023).
All functional diversity indexes of roadkill snakes and those from adjacent vegetation were predicted by forest metrics, pastures, bodies of water, and transitory crops, suggesting the need to conduct future studies at the interpopulation level for different assemblage species. Our results reinforce the need to continue exploring the scale of the effect at which habitat quantity is measured (Watling et al., 2020) and to include species response to edge effect, as species in an assemblage may extensively use the anthropogenic matrix, be habitat specialists inside the native forest or be generalists to vegetation cover gradients (Schneider-Maunoury et al., 2016). In this sense, studies of edge effects should be considered in future studies on landscape composition and configuration in biotic communities (Pfeifer et al., 2017).

ACK N OWLED G M ENTS
We thank the Posgrado en Biología and the Grupo Herpetológico de Antioquia from the Universidad de Antioquia for providing permissions and facilities use to MTRA to conduct her Master thesis. We thank the Higher Education Fund from the Gobernación del Meta and the Colombia Biodiversa Scholarship from the Alejandro Ángel Escobar Foundation for providing financial support for this research.
We thank Universidad de los Llanos for the Academic cooperation agreement.

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors declare that there are no potential conflicts of interest concerning this article's research, authorship, and publication.

DATA AVA I L A B I L I T Y S TAT E M E N T
Upon publication, the raw data will be available at "Functional diversity of snakes is explained by the landscape composition at multiple areas of influence", Mendeley Data, DOI: 10.17632/ wp8gw23kd2.1.