How powerful are tests of neutral theory? Detecting density dependence in ecological communities

Background/Question/Methods

 It is currently accepted that different processes controlling species abundances can generate the same community-level patterns. Whether or not such patterns can be used to discriminate between underlying mechanisms is still controversial since an extensive study comparing community-level patterns would be very difficult and has never been done. Here we present a simple procedure for performing a power calculation that can be applied to a wide variety of non-neutral models. We illustrate the applicability of this technique to the particular case of density dependence. A stochastic community model introduced by Haegeman and Loreau [1] is modified so that for some parameters it generates data that are indistinguishable to a well studied neutral model for which a likelihood is known [2].

Results/Conclusions

 As the density-dependence is increased we show how it becomes possible to discriminate between community-level patterns and to detect whether an underlying non-neutral mechanism exists. This enables us to determine the sample size needed to distinguish between alternative processes of a given intensity or to give an upper-bound for such processes if they were present in real systems that appear consistent with neutral theory. This procedure can be applied to any non-neutral processes other than density dependence. We hope that such an approach will help in designing experiments with the right sampling effort and to contribute to improving our knowledge of the mechanisms that shape communities.

[1] Haegeman, B. and Loreau, M., (2011). A mathematical synthesis of niche and neutral theories in community ecology. J. Theoret. Biol., 269, 150-165.

[2] Etienne, R.S., (2005) . A new sampling formulat for neutral biodiversity. Ecol. Lett., 8, 253-260.