ABSTRACT Multiple drivers contribute to community stability measured as the temporal invariability of total abundance, density, or biomass. However, the direct and indirect effects of these drivers, such as number of species, synchrony between species, species' life strategies, and environmental conditions in communities of terrestrial vertebrates remain unclear. To address this gap, we compiled a worldwide dataset of time series corresponding to 124 natural communities of terrestrial vertebrates including mammals, birds, and reptiles. We found that communities tend to be more synchronous than expected under the theoretical scenario of weak covariance between species because, at the interannual level, shared responses to the environment predominate over compensatory dynamics. Furthermore, piecewise structural equation models reveal that the number of species had a consistent positive effect on stability (either directly or indirectly via synchrony), but the sign and magnitude of the paths involving synchrony changed depending on the metric used and whether detrending was applied or not. Interestingly, we did not find the expected decreasing latitudinal gradient in community stability towards the poles. Our results demonstrate that the synchrony metrics used are not interchangeable and thus, whenever we use multivariate methods to understand the relation between richness, synchrony, and stability, our results may depend on how we define synchrony and how we measure it. Put together, our results also reinforce the idea of the predominance of environmental forcing over compensatory dynamics and the key role of the accumulation of species (statistical averaging) for the temporal stability of ecological communities.
Picó et al. (Mon,) studied this question.