Relationships among breeding birds, habitat, and residential development in Greater Tucson, Arizona

Bird abundance trends have been correlated with habitat changes in urban developed areas but have seldom been associated with specific patterns of urban-related habitat changes. We examined breeding bird-habitat relationships in 334 random plots ranging from undisturbed natural to highly developed land in Tucson, Arizona. In each plot we quantified 19 variables describing three land cover patterns (habitat physiognomy, floristics, and spatial relationships of native habitat fragments) and correlated them with abundances of 21 bird species. Abundances of 17 bird species were associated with variables describing land cover pattern. In addition, we correlated abundance, species richness, and evenness for three bird guilds (non-natives, natives, and a native indicator guild) with land cover variables. Housing density best explained the variation in species richness for both the non-native (r(2) = 0.79) and the indicator guilds (r(2) = -0.69), whereas area of Upland Sonoran vegetative cover (r(2) = 0.56) and distance from undisturbed washes (r(2) = -0.56) correlated most strongly with the native-bird group. Finally, we developed and tested regression models predicting species richness for each bird guild. The following variables loaded into the predictive models: house density; percentage cover of paved areas; exotic, Upland Sonoran, and undisturbed riparian vegetation; and distance from undisturbed washes. The models explained 71% of the variation in non-native bird species richness, 56% of the variation in native bird species richness, and 60% of the variation in species richness for the indicator guild of birds. The correlations and regression models can be used to predict species richness responses to future residential development in the Tucson area.