How and why communities change is a central problem in community ecology. While long-term ecological studies increasingly document change in community composition, it remains unclear whether this type of change tends to occur gradually or through rapid reorganization events (e.g. threshold dynamics, or disturbance-driven change). Long-term observational studies have the potential to yield valuable information about the shape of past dynamics and provide insight as to the relative strengths of the often many factors shaping community composition and structure. We used Latent Dirichlet Allocation (LDA) and a change-point model to examine the long-term dynamics of a desert rodent community which is known to have experienced compositional change over a 38-year span. Past work at the site suggests that the community experienced at least one rapid reorganization event (disturbance initiated by a catastrophic flood), and was also influenced by gradual environmental drivers (e.g. shrub encroachment). By quantifying community dynamics over the entire 38-year span, we sought to reveal any additional reorganization events, as well as place the previously known events in a broader context.
Results/Conclusions
We found that this desert rodent community experienced multiple reorganization events, separated by periods of relative stability. While species composition changed dramatically across reorganization events (i.e. change in dominant species), in between reorganization events the community composition was relatively consistent. Despite past work suggesting that gradual shrub encroachment is a significant driver shaping species composition in this community, our results show that the prediction of a gradual, linear response in the rodent community is not well supported. In some ongoing work, we compare the dynamics of this community to an experimentally-altered rodent community existing adjacent to this community. We found that some of the observed reorganization events occurred simultaneously in both communities, suggesting strong influence of external drivers, and other reorganization events were unique to one community, suggesting these changes were contingent upon the particular species present. Continued study of the dynamics of community change will provide important information not only on the processes structuring communities, but will also provide guidance for forecasting how communities will undergo change in the future.