Understanding how abiotic, biotic and historical factors shape species distributions remains a central question in ecology, but studies linking biotic factors to continental-scale patterns remain scarce. Here, we present a novel framework for simultaneously testing patterns expected when abiotic, biotic or historical factors drive species range limits. We use ecological niche models to produce empirical estimates of the "Biotic, Abiotic, and Movement" paradigm (BAM diagrams), which previously had only been used theoretically. Based on climatic and pollen data, as well as explicit consideration of dispersal limitations, we implement the framework for a group of North American birds (Oreothlypis warblers) with clear habitat associations. Because the pollen-based predictor variables characterize vegetation, they represent biotic factors needed by each bird species. Although continental-scale patterns of distribution traditionally are attributed to abiotic factors, only one species matched the hypothesis of solely abiotic drivers. In contrast, pollen-based models indicate biotic drivers for two species, correctly predicting their absence in climatically suitable areas. These results highlight the feasibility of considering and quantifying potential effects of biotic interactions on species ranges, especial when interactions can be decoupled from abiotic factors. Furthermore, the availability of pollen data now and in the Holocene highlights the potential of these data to be used to predict range shifts of other organisms tightly dependent on particular vegetation types.