A scale-independent approach to the richness vs biomass relationship in ground-layer plant communities

It has been proposed that the much discussed 'hump-backed' relationship between species richness and habitat productivity might be an artifact, arising from the use of an anthropocentric sampling scale. We studied this relationship in 27 Estonian grassland and forest ground-layer plant communities by using scale-independent sampling. Richness was estimated simultaneously from 1-m(2) quadrats, and from quadrats inhabited by 500 plant ramets. We also studied richness relative to the size of the actual species pool (relative richness). Modal maximum richness at intermediate standing biomass was observed in both fixed-size and flexible-size quadrats, indicating that the unimodal shape of the studied curve is not merely a sampling artifact. Relative richness varied in a very narrow interval (especially when estimated per 500 ramets), showing that small-scale absolute richness is mostly determined by the size of the local species pool. However, the relationship between relative richness and biomass was slightly parabolic, showing that modal relative richness is expected at intermediate biomass. This could be interpreted as evidence supporting the competition-based explanation of the hump-backed curve, but, more likely, this is an artifact due to possible underestimation of the size of the actual species pool (overestimation of relative richness) in species-rich communities with intermediate productivity. Relative richness is positively correlated with absolute richness in 1-m(2) quadrats, but independent of absolute richness when scale-independent sampling is used. In flexible-size quadrats, richness is predictable as a certain percent of the actual species pool, independent of the density of ramets in the community.