Colonization-extinction dynamics in protist metacommunities

Background/Question/Methods

Colonization-extinction dynamics are central components of patch dynamics and species sorting processes in metacommunities. However, little is known about the interaction of these two processes. Theory suggests that environmental regulation of species composition is reduced by high extinction rates under disturbance, generating strong residual variation. We examined how extinctions in a heterogeneous landscape of patches affected community composition related to environmental, spatial, and stochastic factors using protist microcosms. We used a 5 × 5 lattice of patches with four disturbance levels, and each landscape contained 50% of light and dark patches to generate environmental heterogeneity. Experiments were started with 1/3 of the patch occupied by each of the two species which dominate differently in each environment, and 1/3 of the landscape being empty. Disturbance was manipulated weekly by replacing different numbers of patches with sterilized medium. Dispersal was implemented weekly by transferring a small amount of medium from one to the next patch from four directions. Each treatment was replicated 5× for a total of 500 communities. We weekly monitored the number of individuals in all patches for 8-weeks and performed a partitioning of the observed variation into unique spatial and environmental components.

Results/Conclusions

Our results showed a negative correlation between two protist species in light and dark environments by species sorting processes. We also found that environmental, spatial, and stochastic factors changed through time in absolute and relative importance depending on the disturbance treatments. In no disturbance treatments, environmental components significantly increased while residual variation decreased over time, suggesting that it took some time for the species distributions to come to equilibrium. In contrast, in intermediate disturbance treatments, environmental components were relatively low and most of the community variation was explained by residuals and spatial patterning. In high disturbance treatments, residual variation explained most of community composition. It has been argued that colonization-extinction dynamics within heterogeneous environments should generate spatial patterning even if environmental filtering is decreased. Our experiments however show that spatial effects due to dispersal generally increased little throughout the experimental period, but random disturbance introduced spatial patterning that was indistinguishable from dispersal-generated spatial patterning in a PCNM analysis. Overall, our findings indicate that patch dynamics in metacommunities with patch heterogeneity should show enhanced residual and spatial variation with enhanced extinction rates, but that determining underlying mechanisms may be difficult in natural systems.