Recent evidence has shown that many arthropod communities around the world are in rapid decline. Though the causes are not entirely clear, many studies point to the expansion and intensification of agricultural systems as a primary driving factor. However, we know very little about the mechanisms by which arthropod communities are structured in agroecosystems, particularly in seasonal environments. Here we aimed to determine: (1) the relative importance of environmental filtering, dispersal limitation, and ecological drift and (2) the roles that local habitat, climatic factors, and landscape-level agricultural intensity play in structuring arthropod communities in a seasonal agroecosystem. To answer these questions, we used a highly comprehensive dataset consisting of five months of arthropod metabarcoding data that were collected from 29 malaise traps at 15 sites in Southern Ontario, Canada. These sites spanned a gradient of agricultural intensity and varied in the representation of local habitat types. Using redundancy analysis and variance partitioning, the arthropod data were related to five months of plant survey data collected on the same sites, climatic data, data on landcover and site management, and spatial and temporal variables that were created via Moran’s eigenvector maps and asymmetric eigenvector maps.
Results/Conclusions
We found very high variation in arthropod community composition between traps and sampling periods (average Jaccard dissimilarities of 0.81 and 0.68, respectively). Spatial, environmental, and temporal factors together accounted for 21% of total variation in arthropod community composition (adjusted R2). Environmental variables played the most important role (R2adj=0.18). Most of this effect was neither spatially or temporally structured (R2adj=0.10), corresponding strongly to local habitat type and plant community composition. However, temporally structured environmental factors also played an important role (R2adj=0.05) primarily due to climatic variation. There were smaller effects of spatially structured environmental factors associated with local habitat and agricultural intensity, as well as pure spatial and temporal effects (all R2adj=0.02). Overall, these results are consistent with strong environmental filtering and possible weak effects of dispersal limitation and ecological drift. Our findings demonstrate that seasonality plays an important but often overlooked role in structuring arthropod communities. There are also clear implications for habitat restoration programs. We showed that that even highly local variation in habitat can have strong effects on arthropod biodiversity in agroecosystems and that at least for the taxa considered here, dispersal would not be highly limiting for the colonization of new habitat.