A toolbox to explore the composition of species dynamics behind multi‐species indices

Abstract

Abstract In the light of declining biodiversity, monitoring its fate is essential for conservation strategies. Aggregation of temporal change of different species into multi‐species indices such as geometric means makes it possible to identify species groups that are at risk as well as those that are doing well. However, aggregated indices mask the between‐species variability in the temporal trajectories, which could be of high relevance for conservation actions. We propose a toolbox, available as an r package, to investigate compositions of species dynamics in geometric mean multi‐species indices. The toolbox is based on a dynamic factor analysis which uses species dynamics and their uncertainty to (1) identify common latent trends in those species dynamics, (2) display the variability of species dynamics and (3) extract clusters of species with similar dynamics within the species groups used for the indices. We apply the toolbox to common breeding birds in Sweden and explore the variability in dynamics among species included in EU‐official indices for farmland and woodland species, highlighting clusters of species with related dynamics previously hidden by averaging. The toolbox is designed to be applicable to a wide range of ecological monitoring data. By enabling a deeper exploration of the structure behind existing indices, we may refine our understanding of biodiversity change to better inform subsequent conservation policies.

Keywords

• biodiversity index
• DFAclust r package
• dynamic factor analysis
• species clustering
• species dynamics variability