Making quantitative predictions on the yield of a species immersed in a multispecies community: The focal species method

Abstract

• Predicting the species yields of a community is central for its management. • Estimating all the required parameters to attain such goal may be unfeasible. • The focal species approximation requires only a fraction of these parameters. • Predictions are compared against experiments covering a wide variety of taxa. • Predictions using only a fraction of the set of model parameters are quite accurate. The abundance or yield of a species is an ecological quantity of paramount importance when making management and conservation decisions. The linear generalized Lotka–Volterra equations (LGLVE) constitute the simplest theoretical framework for predicting such yields. However, in many ecological communities ( e.g. tropical forests), the number of coexisting species S can be very large and estimating the entire interaction matrix A from experiments is practically unfeasible. Here, I introduce a method for predicting the yield of a particular species of interest, or focal species. The method proposed here only requires data on the interaction coefficients of the focal species k , referring to the elements of row and column k of the approximate focal interaction matrix A ( k ) . These interaction coefficients can be obtained from measuring the yields of species in monoculture ( S experimental treatments) and in biculture experiments involving the focal species ( S − 1 treatments). The remaining elements of A ( k ) are considered equal to − 1. I compare the species yields predicted by the focal species method against a data set of experimental studies I compiled from the literature across several taxa −plants, algae, protozoa and crustacean. The accuracy of this method is slightly lower than that attained when using the entire Lotka-Volterra interaction matrix. The advantage of this method is that it requires a much smaller number of experimental treatments to estimate parameters (of the order of S rather than of S 2 ). Therefore, it can be a useful quantitative tool for applied sciences like agriculture or fisheries.