Competition across diverse taxa: quantitative integration of theory and empirical research using global indices of competition

Abstract

A main limitation for quantitative assessments of the effect of species richness and competition intensity on ecosystem functions is that the number of model parameters to estimate for a S‐species community grows at least as S 2. We propose and evaluate a novel framework based on a set of global, i.e. for the whole community, indices of competition (GIC) that allows quantitative predictions of community productivity without knowing the entire community matrix. This framework also serves for comparisons among widely different taxa. These GICs were mostly evaluated in communities of primary producers and are expressible in terms of relative yields (mixture biomass/monoculture biomass), such as the relative yield total (RYT) − the total sum of relative yields − and the mean relative yield (MRY). Specifically: 1) we conducted a meta‐analysis of 166 published competition experiments covering a wide variety of communities − bacteria, protists, plankton, plants, fish, insects, mammals. For each experiment we computed the set of GICs. 2) We derived from the Lotka–Volterra competition model (LVCM) analytical expressions for the RYT and MRY in terms of the species richness and the mean value of the interspecific interaction coefficients. 3) We tested the agreement of such expressions with the corresponding empirical GICs computed in 1). The main findings are: First, the RYT (significantly different among taxonomic groups) increases with the species richness S; conversely the MRY (not different among groups) decreases with S. Second, these analytical expressions for the RYT and MRY, besides giving the above qualitative dependence on S, reproduce with accuracy these empirical GIC. Third, we thus show that our general formulas can be used to assess the effects of competition on global community attributes in cases with incomplete information on the LVCM parameters (i.e. when only a subset of these parameters is known).