Bird communities, roads and development: prospects and constraints of applying empirical models

Our objectives were to explain the prospects and constraints of applying empirical models that relate bird community metrics to broad-scale characteristics of roads and development. We explored the practical value of regression models that were derived for a large protected area in the Chihuahuan Desert. These models related bird species richness, relative abundance, or probability of occurrence to total length of roads within each of two spatial extents (1- and 2-km radii), distance to the nearest road, distance to the nearest development, or the two-way interactions of these variables. Empirical models can be used to inform conservation decisions, to parameterise simulation models for conservation planning, to identify threshold levels of road and development variables, and to determine the focus of management experiments for confirmatory hypothesis testing and improvement of model realism.