Demographic stochasticity, invasion rates, and higher order interactions: Detecting the signal of 'emergent' community dynamics in an intrinsically noisy world

Background/Question/Methods

Demographic stochasticity is random variation in birth and death rates among otherwise-identical individuals. Demographic stochasticity creates random deviations away from the population dynamic trajectories that would occur in a deterministic world. It also has more subtle effects. First, its effect can accumulate over time, especially in small populations and weakly stable populations. This leads to population dynamic trajectories that eventually deviate greatly from those that would otherwise have been expected. Second, in systems with nonlinear density dependence, demographic stochasticity can alter the expected population trajectory, as well as creating variance around the expectation. Both these effects should be particularly relevant to the fates of initially-rare invaders, but quantifying them in nature is challenging due to the difficulty of conducting lengthy replicated experiments and teasing out other sources of stochasticity (environmental stochasticity, sampling error). I used competing species of bacterivorous protists in laboratory microcosms to conduct a mutual invasibility experiment. Specifically, I grew replicates of all possible combinations of three competing species. In multi-species combinations, each species was an initially-rare invader in some replicates, and an initially-common resident in other replicates. Laboratory microcosms allow replication, control of environmental stochasticity, and independent quantification of sampling error, thereby allowing effects of demographic stochasticity to be isolated.  

Results/Conclusions

Some invaders rapidly and repeatably invaded some resident communities. However, other invader-resident combinations resulted in rapid invasions in some replicates but slow invasions in other replicates. By fitting stochastic competition models to the time series data, I show that this variability in realized invasion rate arose from the combination of demographic stochasticity and low expected (deterministic) invasion rate. These results highlight that small population size on its own is neither necessary nor sufficient for demographic stochasticity to affect population dynamics. I then go on to use stochastic competition models fitted to the single-species and two-species dynamics to predict invasion rates and dynamics of the three-species community. This procedure tests for higher-order interactions: ‘emergent’ features of community dynamics that cannot be predicted from knowledge of single-species population growth and pairwise interspecific interactions. I find evidence for subtle higher order interactions that could not have been detected without first isolating the effects of demographic stochasticity. Finding ‘emergent’ community dynamics is surprising because protist microcosm communities lack many of the features thought to give rise to ‘emergent’ dynamics. Protist microcosm communities are considerably more complex than is commonly recognized, but modern analytical and experimental tools can tame that complexity.