An analysis of trends, uncertainty and species selection shows contrasting trends of widespread forest and farmland birds in Europe

Abstract 1. Composite, multispecies biodiversity indices are increasingly used to report against international and national environmental commitments and targets, the Wild Bird Index being a prominent example in Europe, but methods to assess trends, error and species selection for such indices are poorly developed. 2. In this study, we compare methods to compute multispecies supranational indices and explore different approaches to trend and error estimation, the presentation of indices, and species selection. We do so using population trend data on forest and farmland birds from 28 European countries, 1980–2015. 3. We find relative stability in common European forest bird populations over this period, but a severe decline in farmland bird populations. Altering the benchmark year affects index characteristics and ease of interpretation. We show that using annual species’ indices and their SE s to calculate confidence intervals delivers greater precision in index estimates than bootstrapping across species. The inclusion of individual species within indices has limited leverage on index characteristics, but subjective selection of species based on specialisation has the potential to generate bias. 4. Multispecies indices are valuable policy-relevant tools for describing biodiversity health. Their calculation and presentation need to be tailored to meet specific policy objectives, and they must be supported by clear interpretative information. We recommend methods for indicator analysis, forms of presentation, and the adoption of an objective species selection protocol to ensure indicators are representative and sensitive to environmental change.