Drought resistance increases with species richness in restored populations and communities

It is unknown to what extent or by what mechanisms introducing biodiversity influences stability of high-stress ecosystems undergoing restoration. Opportunity to investigate patterns of biodiversity and resistance to disturbance in a high-stress environment was presented when severe drought struck a restoration experiment underway on abandoned limestone quarry floors in Ontario, Canada. Experimental communities were previously established within small quarry-floor plots by sowing native grass and forb species considered to be characteristic of rare natural limestone pavements called alvars. Despite adding an identical 18-species seed-mixture to all plots, realized communities varied extensively with respect to the numbers of species established (species richness), the total number of individuals established (community abundance), and the number of individuals belonging to each species (population abundances). We investigated the relationship between species richness and resistance of community abundance to drought, while accounting for background richness–abundance correlation, by contrasting slopes and intercepts of the richness–abundance relationship immediately before vs. 6 weeks after the drought. This relationship was significantly positive prior to drought but 72% steeper in slope following drought, while the abundance intercept exhibited a 44% drop. Plots featuring richer, more abundant communities prior to drought thus suffered considerably less damage than species-poor, low-abundance plots. Population abundance was weakly related to richness prior to drought, but strongly and positively related to richness after the drought. At the individual species level, no species experienced greater losses of abundance with increased plot richness, but six species experienced reduced abundance losses where they co-occurred with more neighbour species. Facilitation or other mechanisms capable of increasing population resistance may thus underlie community resistance in high-stress environments. Though controlled experiments are required to establish causes of relationships reported here, the forms of these relationships suggest that managers may be able to promote resistance in high-stress ecosystems by establishing species-rich communities.