Recent work has demonstrated a strong positive relationship between avian species richness and net primary productivity (NPP) over continental to global scales. In North America, this relationship has been shown to hold seasonally as well as spatially, and yet the mechanistic underpinnings are still uncertain. NPP may have a positive effect on species richness through one of two non-mutually exclusive mechanisms. First, the relationship has often been considered in the context of species-energy theory, with the assumption that NPP reflects the amount of resources available to consumers, and that more resources can support more bird species. Second, the continental NPP gradient also reflects a gradient in habitat complexity, from structurally simple deserts and grasslands to more vertically layered deciduous forest. The importance of foliage height diversity in facilitating avian diversity has long been known. We set out to test the relative importance of resource availability and habitat structure for explaining avian patterns of abundance and richness. Over ~30 sites spanning 500 km in the southern Appalachians, we characterized the bird community using point counts, and quantified arthropod availability and habitat structure using a variety of methods. Evaluating these relationships across multiple foraging guilds provides multiple tests of these ideas.
We found that both resource availability and habitat structure were important for explaining variation in avian community structure, but each foraging guild responded to different resource or habitat variables. The number of foliage gleaning birds was better predicted by resource availability (density of caterpillars) than any of our measures of habitat structure. In contrast, the density and species richness of ground gleaners exhibited stronger relationships with habitat structure. Specifically, ground gleaners were more abundant and more diverse at sites with lower canopies and more snags. Bark gleaners exhibited a negative relationship with leaf area index, but were only weakly related to our estimates of bark arthropod density. Despite these patterns, a substantial amount of variance in species richness and abundance remains unexplained. We suggest that this may be due to 1) the patchy distribution of arthropods relative to our ability to sample them, and 2) landscape level effects at scales beyond that of our focal site.