Occupancy at backyard bird feeders: The influence of land-cover, spatial scale, and urban growth

Gretchen Lynn White, Purdue University

Abstract

Previous studies have shown that birds select habitat based on characteristics at multiple spatial scales. Therefore, presence or absence of a species at a site might be influenced by fine scale characteristics, but also features that occur at larger landscape scales. A species' selection criteria may also vary seasonally. An understanding of bird habitat selection may be especially important as increasing urban growth impacts greater areas of land. To learn more about this phenomenon, I used data from a multi-year, citizen scientist study that monitored birds at backyard bird feeders across the U.S. to provide detection histories for 13 species. I used occupancy modeling to determine the influence of land cover types at multiple spatial scales (500m, 1km, 3km, 5km, and 7km) on species' probability of occurrence in human altered landscapes (e.g. residential backyards) while incorporating detection probability. At each spatial scale, I built models for each season with low and medium intensity urban percent cover as covariates as well as other land cover types (forest, shrub, etc.), that are previously known to be preferred habitat types for each species. In many cases, the strength and directionality (positive or negative) of habitat preference changed across seasons and spatial scale. This trend was especially apparent in respect to urban land cover types for many species. To understand how future urban growth may impact bird species' presence at bird feeders, I created occupancy models to predict any changes in each species' occupancy probability as the amount of urban land cover increases at decadal increments from 2010 to 2100. Urban growth data was provided by models created using an Artificial Neural Network and a raster modeling environment by Tayyebi et al. (2012). All species were predicted to experience a change in mean occupancy probability across all survey locations. Those species for which mean occupancy was predicted to decline in both winter and spring were primarily found in association with forest habitat. Species whose predicted mean occupancy probability declined in spring, but increased when winter were primarily associated with grassland and cultivated crop land cover. Species that had predicted mean occupancy probabilities which increased in both winter and spring were shown to be positively associated with urban areas in initial models. In light of growing human populations and increasing influence of human development, a refined understanding of habitat preferences has the potential to be important for urban planning when avian biodiversity is a priority. These results may provide important insights as we contemplate the impact of future human-avian interactions in human altered habitat. While the species included in this study currently have large and fairly stable populations, changes in their occupancy as a result of urban growth may signal concern for less common species that share similar habitats and may decrease the health benefits to the humans who provide food for them.

Degree

M.S.

Advisors

Zollner, Purdue University.

Subject Area

Wildlife Conservation

COinS