Resource information
CONTEXT: Weather surveillance radars (WSR) have been used to locate roost sites used by Purple Martins (Progne subis) for decades. Improvements in radar data processing and accessibility now make it possible to monitor roosts over a broad spatial scale. OBJECTIVES: We sought to locate all of the Purple Martin roosts in North America and to use the data to evaluate (1) the land cover types associated with roosts (2) relationships among roost persistence, land cover type, and regional population trends. METHODS: We used mosaicked images of radar reflectivity based on the NEXRAD WSR network to locate 234 Purple Martin roosts that were active between 2009 and 2014. Of these roosts, we ground-truthed a subset of 57 with site visits and reports from citizen scientists. We assigned roosts to different classes based on local land cover, and used a variety of statistical and spatial analyses to address the objectives listed above. RESULTS: Roosts were mainly associated with forest, cropland, urban, and water land cover types, with cropland being the most common. There was an apparent preference for urban sites, and urban roosts were associated with the high year-to-year persistence. We found no correlation between roost persistence and regional population trends in data from the North American Breeding Bird Survey (BBS). CONCLUSIONS: Although they use a diverse array of roosting habitats, urban roosting areas appear to be increasingly important for Purple Martins. Persistence of urban roosts was high, which aligns with the species’ unique natural history and its association with human societies.
