Spatiotemporal realism in the modeling of honey bee pesticide exposure

The potential risk that pesticides pose to honey bees is not disputed, but the question of whether a specific pesticide can be blamed for a particular pattern or incident of colony damage is always controversial. Most recently, this controversy has surrounded the neonicotinoid insecticides and their possible role in unusual patterns of honey bee mortality in Europe and North America. While laboratory experiments have clearly documented toxic effects of neonicotinoids on individual bees, field studies have usually failed to detect toxic effects in free-foraging colonies. To some extent, this dissonance may be due to flaws in laboratory methodology or insufficient statistical power in field studies, but a more fundamental problem lies beneath the uncertainty: any assertion of toxic effects assumes some model of toxic exposure, but a mechanistic model of pesticide exposure in honey bees has never been proposed. We present a descriptive model of honey bee pesticide exposure that mechanistically links the environmental presence of a pesticide with the biological processes of the honey bee colony to refine the concept of “exposure” and discern the pathways by which it may occur. One of the key insights of our model is that exposure dynamics emerge from the fundamentally stochastic processes of honey bee foraging and in-hive food transmission, meaning that must be understood in probabilistic terms. We then apply our model to explain the results of an empirical study of honey bee exposure to seed treatment neonicotinoids released during corn planting.