Assemblage-structuring force of species interactions varies spatially and temporally: Co-occurrence analysis of canopy arthropod distributions

Background/Question/Methods Understanding how interactions amongst species contribute to their distribution is central to community ecology. However, it is not understood how these processes may change under different environmental conditions. In this study, we present evidence that suggests interactions amongst canopy arthropods contribute to assemblage structure, depending on both the sampling time and location. We sampled canopy arthropod species abundances on individual trees in three adjacent, monospecific stands of Populus fremontii (Fremont cottonwood), which differ in planting year (2000, 2002, 2005). We then conducted null-model based co-occurrence analysis using the freely available software EcoSim (Kesey-Bear and Acquired Intelligence, Inc.) to test for patterns suggestive of interactions influencing assemblage structure. Environmental heterogeneity amongst trees is low within stands; therefore environmental gradients are most likely not contributing to co-occurrence patterns. Also, all stands were planted recently and closely enough such that evolutionary processes (e.g. allopatric speciation) have not occurred. For the analyses of each stand at each sampling date we used a fixed-row, equiprobable-column matrix permutation algorithm in order to simulate presence-absence matrices in which species co-occurrences are effectively random with respect to species interactions. We chose to use the C-score index of co-occurence because of its robustness to Type I Errors when used with this simulation algorithm. P-values and standardized effect sizes (SES) were based on 5000 simulated matrixes.

Results/Conclusions We found significant variation in SES values across both sampling dates and stands; and three main patterns emerged. 1) The SES values in the youngest stand (2005) were all near zero and none were statistically significant. This could be the result of unresolved species interactions, meaning that species interactions have not had sufficient time to produce consistent distributional patterns. 2) The SES values were primarily negative, indicating a dominating force of facilitative interactions. 3) The 2000 and 2002 stands varied in the size and significance of their SES values with the oldest (2000) stand switching from positive to negative as the season progressed, suggesting a switch from competitive to facilitative interactions driving structure. These results suggest that species interactions are indeed important in structuring canopy arthropod communities, but that the efficacy of these forces can vary across both space and time.