Is the productivity of vegetation plots higher or lower when there are more species? Variable predictions from interaction of the 'sampling effect' and 'competitive dominance effect' on the habitat templet

Using a habitat templet model, we predict that the productivity (total biomass) of plots within a plant community may be positively, negatively or not at all related to variation in the number of species per plot, depending on successional stage (time since major disturbance) and habitat carrying capacity (reflecting the total resource supplying power of the habitat). For plots of a given size, a positive relationship between productivity and species richness is predicted in recently disturbed habitats because local neighbourhoods here will have been assembled largely stochastically, usually from a pool of available species with a right-skewed size frequency distribution. Hence, in the earliest stages of succession, plots will have relatively high total biomass only if they contain at least some of the relatively uncommon larger species which will, in turn, be more likely in those neighbourhoods that contain more species (the sampling effect). Among these will also be some of the more common smaller species; hence, these high biomass, species-rich plots should have relatively low species evenness, in contrast to what is predicted under effects involving species complementarity. In late succession, the plots with high total biomass will still be those that contain relatively large species but these plots will now contain relatively few species owing to increased competitive exclusion over time (the competitive dominance effect). In intermediate stages of succession, no relationship between plot productivity and species richness is predicted because the opposing sampling and competitive dominance effects cancel each other out. We predict that the intensity of both the sampling and competitive dominance effects on the productivity/species richness relationship will decrease with decreasing habitat carrying capacity (e.g. decreasing substrate fertility) owing to the inherently lower variance in between-plot productivity that is predicted for more resource-impoverished habitats.