Can correlated population trends among forest bird species be predicted by similarity in traits?

Context. Many conservation decisions rely on the assumption that multiple populations will respond similarly to management. However, few attempts have been made to evaluate correlated population responses to management or to identify traits that could be used to predict correlations. These assumptions are central to the use of the 'population indicator-species concept' (the idea that population trends of one species can be used as an index of trends in other species) for measuring the effects of key ecological drivers. Aims. We investigated correlations among bird population trends in a mixed podocarp-hardwood forest in New Zealand in which introduced mammalian pests are controlled. We analysed trends in the abundance of 18 bird species (primarily passerines) over a 10-year period, using data from 5-min bird counts. Methods. We used a Bayesian modelling approach to identify short-term correlations in population trends among species and to investigate whether ecological traits can be used to predict these correlated trends. Key results. Population increases were detected in 9 of the 18 bird species over the 10-year period of the study. Population trends were correlated for 10% of species pairs (of which 81% were positive correlations). Correlations among seven of the nine species that increased in abundance were always positive; these species form a potential indicator pool. However, traits were not useful for predicting correlated population trends. Conclusions. Bird species affected by a shared ecological driver (predation) can exhibit correlated population trends when introduced predators are controlled, but correlations cannot be predicted by similarity in ecological traits. Implications. We advocate for testing consistency of correlations at multiple sites so as to validate the evidence-based use of the population indicator-species concept as a cost-effective alternative to monitoring whole communities.