Abstract: Community ecology is shifting focus from describing patterns to identifying and understanding the underlying processes. This requires development of a holistic statistical framework able to detangle the factors controlling species distributions. One challenge to statistically detangling assembly processes is that assembly processes may act at different spatial scales. There are various ways to include different scale relationships into a model including: the extent of the study region, the grain or resolution of spatial data, and accounting for spatial autocorrelation. However, one scale relationship that tends to be overlooked in community studies is scale of response. The goal of our study is to determine the spatial scale at which the bee community responds to various types of land use and how that relationship depends on species traits and phylogeny. We analyzed wild bees collected in June through August of 2014 and 2015 at 106 sites across the agricultural region of Alberta, Canada. We used land cover data from the annual crop inventory by Agriculture and Agri-Food Canada to optimize buffer radius (60 m – 5010 m) for each species and calculated the percent land-use type within each buffer. Generalized linear mixed models show that variation in the scale of response is associated with habitat type, functional traits, and phylogeny. Species level variation in scale-habitat relationships may complicate interpretation of community level studies.
