Composite bird indicators robust to variation in species selection and habitat specificity

Ecological indicators are increasingly being used to monitor the state of the environment. Ideally indicators should quantify and summarise ecosystem function and overall health. However, there is a lack of consensus on the correct approach for the production of these indicators which is crucial to their success in achieving this overall aim. The selection of species and data to include in each indicator are key to the development of indicators. We examine the effects of changing the criteria for species inclusion in a high profile composite indicator based on population trends of breeding birds in England, currently used to represent the status of the overall environment. This is achieved by (i) applying an objective measure of categorisation, (ii) changing the minimum threshold for data quality, and (iii) changing from a species-level categorisation approach to one based on habitat-specific population trends for each broad habitat of interest. We show that the species in the current wild bird indicators, selected by expert judgement, is similar to the objective assessment that we propose. More generally, we show that indicators with the correct species selection are relatively robust to the habitat-specificity of the composite species trends. The most important consideration is the inclusion of scarce species which may be more sensitive to environmental change. Omission of these species can produce an over-optimistic assessment of the health of the ecosystem. We offer a simple but robust methodology that can be applied to a wide range of species and habitats with minimal knowledge of the specified habitat required.