Quantifying the temporal ecology of plant-antagonist interactions using hierarchical regression models

Daniel B. Turner and William C. Wetzel, Department of Entomology, Michigan State University, East Lansing, MI, Daniel B. Turner and William C. Wetzel, Ecology, Evolution, and Behavior Program, Michigan State University, East Lansing, MI, Daniel B. Turner and William C. Wetzel, Kellogg Biological Station, Michigan State University, Hickory Corners, MI, William C. Wetzel, AgBioResearch, Michigan State University, MI, William C. Wetzel, Department of Integrative Biology, Michigan State University, East Lansing, MI

Background/Question/Methods
Time is a fundamental axis in our understanding of ecological processes like fluctuations in resource availability, how and environmental conditions promote or constrain population growth, and why species may outcompete others. Many studies to date address the effects of temporal characteristics such as age, phenology, and intrinsic growth rates on ecology, but the timing of species interactions rarely is manipulated in experiments to explain deterministic trajectories in trait values and community dynamics. We test how the timing and order of antagonism, a type of priority effect, alters subsequent community interactions using the tall goldenrod (Solidago altissima)-herbivore system in two field experiments. In the first experiment, we apply herbivory damage from a specialist mesophyll-feeding insect (Slaterocoris sp.) to plants in a common garden. Throughout our second season, we spray a jasmonic acid, a plant defense hormone, solution on plants up to three times to test the effects of multiple antagonistic events. In each experiment, we observed plant tissue lost to herbivory and covered with fungal pathogens and measured plant growth traits. We apply hierarchical Bayesian regression models that incorporate time as group- and population-level effects to estimate immediate and lagging responses to antagonism throughout the tall goldenrod growing season.
Results/Conclusions
Our experimental approach and models indicated one or more antagonistic events can shift the trajectory of species interactions. In our first season, mesophyll-feeding significantly decreased fungal pathogen presence in the weeks immediately following induction, but those effects are diminished by the end of the growing season. Early season mesophyll-feeding did not stunt plant growth or result in higher leaf tissue chewing damage. In our second season, the timing and frequency of jasmonic acid spray events shifts S. altissima growth traits and community interactions as the growing season progressed. For example, plants sprayed with jasmonic acid early in the season grew smaller than those sprayed later in the season. By factoring time explicitly into our experiment and statistical models, we gained a deeper understanding about how the timing and frequency of ecologically relevant events affects community dynamics. These events can shift the direction and magnitude of the strength of species interactions, but these effects may only be detectable while accounting for time. These findings indicate that temporally explicit experimental and statistical approaches can be applied not only to understand historical contingency in ecological communities, but also to predict species’ resilience in response to disturbance and other forms of rapid environmental change.