How important is nectar in shaping spatial variation in the abundance of temperate breeding hummingbirds?

AIM: Our aim was to test the degree to which nectar production predicts hummingbird abundances at large spatial scales compared with other large‐scale environmental variables. LOCATION: Arizona, Colorado, New Mexico and Utah, USA. METHODS: We surveyed nectar producing flowers at 67 sites in the summer of 2008 and converted flower densities to nectar production using data obtained from the literature. We derived a model of nectar production and used this to create a nectar production map for the study region. We then tested the degree to which nectar production predicted the abundance and occupancy of black‐chinned (Archilochus alexandri) and broad‐tailed (Selasphorus platycercus) hummingbirds with zero‐inflated Poisson regression. Abundance data were taken from the North American Breeding Bird Survey. We compared the predictions made from nectar production to those made from temperature, precipitation, growing degree‐days, elevation and primary productivity. RESULTS: We found that black‐chinned hummingbird abundance was best predicted by the abundance of conspecifics in a surrounding 20‐km neighbourhood as opposed to any of the environmental variables. Their occupancy varied independently of any underlying spatial or environmental variation. Broad‐tailed hummingbird abundance was best predicted by average temperature. Nectar was a weak predictor of both species' abundance and occupancy. MAIN CONCLUSIONS: Of the variables we measured, no single one is a key predictor of spatial variation of hummingbird abundance. It is possible that breeding abundances respond to local habitat characteristics that do not correlate strongly with large‐scale environmental variability. Within the breeding season, hummingbird abundance and occupancy may depend on factors unrelated to nectar production and hummingbirds may not disperse to track spatial variation in nectar production.