Quantifying spatial structure in a food web module: Opuntia cacti and their specialist herbivores

Background/Question/Methods

Spatial structure can be used to make inferences about biological and unmeasured processes. We characterize the relative importance of environmental factors, biological processes (i.e., species interactions), and spatial autocorrelation to observed species co-occurrence patterns in a model system of two species of native cacti (Opuntia pusilla and the threatened O. stricta) and two specialist cactus moths in Florida. The native moth Melitara prodenialis and the invasive South American Cactoblastis cactorum share similar life history traits and are the principal cactus-feeding insects found at our coastal study site. Using stratified adaptive cluster sampling, we set up 1358 1-m²plots at seven habitat patches within Guana Tolomato Matanzas National Estuarine Research Reserve (near St. Augustine, Florida). We recorded species presence/absence at these plots four times over the past 2 years and quantified cactus abundance in a random subset of plots in 2013. Our environmental variables include elevation, primary productivity (using Normalized Difference Vegetation Index), and distance to water. We quantified spatial structure using Moran Eigenvector Mapping. We used generalized linear mixed models to quantify the relative contributions of the environment, species co-occurrence, and spatial structure to variation in cactus abundance for each of the two cactus species.

Results/Conclusions

Our preliminary studies (data collection is ongoing) show that cactus occurrence varies among patches: O. stricta is relatively common on all while O. pusilla is rare to absent in the dune habitat patches but common on islands.  Both cactus species co-occur at 55 plots. More O. stricta populations were infested by the invasive moth compared to O. pusilla, but no such difference was observed for the native moth. For both cactus species, the environmental factors (elevation, NDVI, and distance to water) did not significantly explain variation in cactus abundance. Instead, variation in O. stricta abundance was best explained by spatial autocorrelation.  Opuntia pusilla abundance was best explained by spatial autocorrelation as well as by O. stricta abundance, suggesting the impact of negative species interactions on the smaller-statured of these two Opuntia species. Possible mechanisms include the spillover of insects from O. stricta or direct competition with O. stricta may be important drivers of O. pusilla abundance. Our data suggest that spatial processes play an important role in structuring our community module.  We will continue to explore the causes and consequences of spatial autocorrelation.