Biological control agent attack timing and population variability, but not density, best explain target weed density across an environmental gradient

Spatial variation in plant–herbivore interactions can be important in pest systems, particularly when insect herbivores are used as biological control agents to manage invasive plants. The geographic ranges of the invasive plant alligatorweed (Alternanthera philoxeroides) and its biological control agent the alligatorweed flea beetle (Agasicles hygrophila) do not completely overlap in the southeastern USA, producing spatial heterogeneity in interaction strength that may be related to latitude-correlated environmental gradients. We studied this system near the range margin of the alligatorweed flea beetle to test whether spatial variation in alligatorweed density was best explained by agent mean or maximum density, variability in agent density, agent attack timing, or a combination of biological control and environmental (i.e., weather) variables. The pattern that emerged was that mean agent and host densities were negatively and positively associated with latitude, respectively. Variability in agent density increased with latitude and was positively correlated with host density. We further discovered that agent first attack timing was negatively correlated with winter and spring temperatures and spring and summer precipitation, and positively correlated with seasonal temperature extremes, which was then directly influential on agent density and variability in density, and indirectly on host density. This study demonstrates that, contrary to common wisdom, weather-related timing of agent activity and population variability, but not agent mean density, contribute to the spatial heterogeneity observed in alligatorweed populations.