Testing whether connectivity stabilizes metacommunities and rescues declining diversity in a 25-y grassland study.

Abstract

Theory predicts that metacommunities, or connected ensembles of local communities, can be stabilized by dispersal in the face of environmental changes. Using a 25-y dataset from 80 grassland sites embedded in a 2,800-ha landscape, we tested for stabilizing effects of connectivity, defined here as the amount of source grassland habitat within dispersal-relevant radii of each site. We found no relationship of connectivity to the temporal variability of total cover or estimated biomass, or to the retention of species diversity over time, and no predicted unimodal relationship of connectivity to species richness. Connectivity did show a predicted positive relationship to temporal species turnover, but this effect disappeared after accounting for livestock grazing history. These grasslands changed toward taller exotic annual grass species and lower species richness over the study period, regardless of connectivity or grazing history, and the richness-reducing shift to taller grasses was consistent with previously documented thermophilization, i.e., shift to warmer and drier species composition. Connectivity in this system appears not to stabilize local communities undergoing directional change.