Correlational ecological niche models have seen intensive use and exploration as a means of estimating the limits of actual and potential geographic distributions of species, yet their application to explaining geographic abundance patterns has been debated. We developed a detailed test of this latter possibility based on the North American Breeding Bird Survey. Correlations between abundances and niche‐centroid distances were mostly negative, as per expectations of niche theory and the abundant niche‐centre relationship. The negative relationships were not distributed randomly among species: terrestrial, non‐migratory, small‐bodied, small‐niche‐breadth and restricted‐range species had the strongest negative associations. Distances to niche centroids as estimated from correlational analyses of presence‐only data thus offer a unique means by which to infer geographic abundance patterns, which otherwise are enormously difficult to characterise.