The reliability of a composite biodiversity indicator in predicting bird species richness at different spatial scales