Modeling the Distribution of Purple Martins in New York State

Abstract

The purple martin (Progne subis) is a migratory songbird of the swallow family. The eastern subspecies (Progne subis subis) breeds from Florida to Southeastern Canada and has an unusual reliance on humans, nesting exclusively in man-made nest boxes. As the subspecies has experienced the most severe population declines in the North-east of its range, we aimed to produce a species distribution model (SDM) of purple martins in New York State to better understand the environmental factors predicting their occurrence. To inform our model we used occurrence data from the eBird citizen science project along with environmental covariates thought to affect habitat selection in aerial insectivores like the purple martin (tree cover, human population density, proximity to water, temperature and precipitation). We built our model using MaxEnt and fine-tuned the output using AICc rankings through ENM Tools. The final model indicated that three environmental covariates have a significant impact on purple martin distribution: distance from water, human population density, and tree cover. Purple martin occurrence strongly negatively correlates with tree cover- as tree cover increases the probability of purple martin occurrence rapidly declines. This is most likely due to purple martin's preference for housing that is placed in open areas with clear flyways. Purple martins have been known to prefer housing near bodies of water which is supported by the model - as distance from water increases purple martin density decreases. Human population density is also an important factor- as population density increases so does purple martin occurrence. This may be due to humans providing purple martins with housing- areas with very few humans will have fewer nesting opportunities. By providing a scientific basis for nest box location at the regional level, our SDM could have a direct conservation outcome. We also plan to use our model to examine how changes in climate and land cover over the past 40 years may have contributed to the decline in purple martins over the same time period.