SYMP 22-1 - Using spatially- and temporally-explicit species distribution models to explore the role of tropical cyclones in shaping Puerto Rican Anolis lizard distributions

Driven by climate change, tropical cyclones – which include hurricanes, tropical storms, and tropical depressions – are predicted to change in intensity and frequency through time. Given these forecasted changes, developing an understanding of how tropical cyclones impact insular wildlife is of heightened importance. Previous work has shown that extreme weather events may shape species distributions more strongly than climatic averages, however given the course spatial scale at which tropical cyclone data is often reported, the influence of tropical cyclones on species distributions has yet to be explored.

Using tropical cyclone track data from the National Hurricane Center (NHC), weather data (rainfall, windspeed, wind gust) from 16 local Puerto Rico weather stations, and species occurrence records from the Global Biodiversity Information Facility, we developed spatially- and temporally-explicit species distribution models (SDMs) to analyze the role of tropical cyclones in shaping present-day distributions of Puerto Rico’s ten Anolis lizard species (anoles). To accomplish this, we extracted timestamps from the NHC that corresponded to each tropical cyclone that came within a 500 km radius circle from the center of Puerto Rico. These timestamps were then used to extract weather data from the three closest weather stations to each occurrence record for all storms that occurred within the one-year window prior to when each occurrence record was recorded. Using these data, we created seven variables to represent the intensity and frequency of tropical cyclones, including, for example, variables related to the number of days windspeed was greater than 50 km/hr. We then collated all tropical cyclone variables with variables related to landcover, biomass, climate, canopy cover, and geology. Two machine learning SDM algorithms – MaxEnt and Random Forests – were used to assess model performance and variable importance in models with and without the tropical cyclone variables.

Our results show that variables related to the number of days an anole species experienced extreme weather conditions are periodically more important in shaping its distribution than variables related to climatic averages or landcover. The results for one of our study species, Anolis cristatellus, show that the Area Under the Curve (AUC) metric for the best-fit MaxEnt models increased from 0.63 to 0.75 when the seven tropical cyclone variables were incorporated into the model. Our findings suggest that incorporating data on tropical cyclones into SDMs may be important for modeling insular species that are prone to experiencing these types of extreme weather events.