From Winchell et al. (2020): Anoles throughout the Caribbean differ in their tolerance to urbanization. Red colors = urban tolerant, blue colors = intermediate tolerance, green colors = urban intolerant.
Seven years ago I asked for the help of Anole Annals readers as I started to think about how different species of anoles throughout the Caribbean tolerate urbanization. This question, it turned out, was a lot more complex than I had originally anticipated! The idea was simple, find out which species are in urban areas and to what extent they use urban habitat elements, then determine if there is an evolutionary signal in urban tolerance and what traits are correlated with urban tolerance. Many hours of troubleshooting and brainstorming with my coauthors Klaus Schliep, Luke Mahler, and Liam Revell (and years later) and this study is finally out in the journal Evolution: Phylogenetic signal and evolutionary correlates of urban tolerance in a widespread neotropical lizard clade.
Anolis lineatopus, one of many urban tolerant anoles (photo K. Winchell)
Inventorying urban species
To figure out which anole species are tolerant of urbanization, my initial plan was to survey researchers and the literature to score each of the 100+ Caribbean species based on their presence in different types of urban habitats and their habitat use. Although I got a lot of great feedback from this original survey, it left a lot of gaps in the dataset. I needed to find a more objective way to assess urban tolerance.
With the help of Klaus Schliep and Luke Mahler, we decided to examine location records in museum collections (via GBIF) to determine which species had been observed (collected) in urban environments. Because we suspected museum records might be biased towards non-urban habitats, we also examined location records from the citizen science database iNaturalist, which we suspected might be biased in the opposite direction (i.e., people photograph things where they live). For each record, we looked at satellite imagery and scored the observation as urban or non-urban, then tallied the total number of observations and the total number of urban observations per species.
Even with these two data sources, we noticed gaps in our data for some species. So we included a third source, Henderson & Powell’s (2009) book on the Natural History of West Indian Amphibians and Reptiles. This fantastic reference (highly recommended!) gives detailed natural history information and summarizes key features of every anole (and other Caribbean herps) in the Caribbean. Of course, this is more subjective than the location-based data, so Luke and I came up with a scoring system that assigned a set number of urban tolerant or avoid “points” based on key descriptors. For example, if a species was described as being common around houses and often observed on buildings, it would get points for being tolerant of urbanization. In contrast, a species described as having a restricted range and intolerance of anthropogenic disturbance, it would get points for being intolerant.
Analyzing urban tolerance in a phylogenetic framework
We combined these disparate data sources into a logistic model with parameters we set based on the number of urban observations we would need to be certain of urban tolerance and how many total observations we would need to be certain of our species assessment. This resulted in a probability of being an urban avoider or urban tolerant for each species, which we used as our prior probabilities for these states in our phylogenetic model. We then reconstructed ancestral states and missing tip states for urban tolerance in 131 species of Caribbean anoles.
Of course, we don’t mean to say that we attempted to reconstruct the evolution of urban habitat use — anoles are far older than urbanization! Instead, we wanted to understand the evolution of the behavioral, physiological, ecological, and morphological traits traits that influence whether a species will exploit or avoid urban habitat when it arises. The threshold model is well-suited for this type of complex trait. The threshold model assumes that a discrete trait is determined by a combination of continuously valued characteristics. These characteristics may be measurable, unmeasurable, or even unknown. As a taxon accumulates specific trait changes, the species is pushed incrementally closer and closer to the discrete state change (in this case urban tolerance), and the more recently this discrete character state has flipped, the more likely a reversal to the previous state could occur. From this model we can extract a single continuously valued trait, the liability, that underlies the complex trait of urban tolerance.
Urban tolerance in Caribbean anoles, from Winchell et al. (2020).
Traits of urban species
So what did we find? To start, urban tolerance appears to be widespread in Caribbean anoles and has a strong phylogenetic signal. Because of that, we suggest that our approach may be used to predict urban tolerance of species that either have yet to encounter urbanization or for which we are lacking information. This application could be particularly useful for determining which species are likely to be intolerant of urbanization and thus should be prioritized in conservation efforts. At the other end of the urban tolerance scale, we caution that our approach should not be used to predict species that are robust to anthropogenic habitat loss, but rather that it might be useful to identify species that are promising for future urban ecology and evolution studies.
Finally, we used the liability score for each species to try to get a better understanding of what those traits underlying urban tolerance are exactly. Using PGLS we looked for correlations between the liability and a suite of ecological and phenotypic traits. We found that species that are more tolerant of urbanization had higher field body temperatures, fewer ventral scales, more rear lamellae, shorter hindlimbs, and experience warmer and drier climates within their native range. These traits may be key “pre-adaptations” enabling species to colonize urban habitats as they arise and to take advantage of anthropogenic niche space (i.e., on and around buildings). For example, urban habitats tend to be hotter and drier than nearby forest sites, so it makes sense that species with larger ventral scales, higher field body temperatures, and which experience hotter and drier temperatures in their non-urban range would be predisposed to tolerate urban habitats. Similarly, lamellae are important for clinging to smooth surfaces, which may be particularly beneficial in urban habitats dominated by smooth anthropogenic surfaces.
Lastly, we found, somewhat to our surprise, that no one ecomorph seems to be best suited for urban environments. Based on our experience, we had thought that trunk-ground anoles would be more likely to tolerate urbanization, but it turns out that there are a lot of trunk-ground anoles that are intolerant of urbanization and a lot of species from other ecomorphs that are tolerant (think A. equestris or A. distichus)!
