Modeling pollinating bee visitation rates in heterogeneous landscapes from foraging theory

Pollination by bees is important for food production. Recent concerns about the declines of both domestic and wild bees, calls for measures to promote wild pollinator populations in farmland. However, to be able to efficiently promote and prioritize between measures that benefit pollinators, such as modified land use, agri-environment schemes, or specific conservation measures, it is important to have a tool that accurately predicts how bees use landscapes and respond to such measures. In this paper we compare an existing model for predicting pollination (the “Lonsdorf model”), with an extension of a general model for habitat use of central place foragers (the “CPF model”). The Lonsdorf model has been shown to perform relatively well in simple landscapes, but not in complex landscapes. We hypothesized that this was because it lacks a behavioral component, assuming instead that bees in essence diffuse out from the nest into the landscape. By adding a behavioral component, the CPF model in contrast assumes that bees only use those parts of the landscape that enhances their fitness, completely avoiding foraging in other parts of the landscape. Because foraging is directed toward the most rewarding foraging habitat patches as determined by quality and distance, foraging habitat will include a wide range of forage qualities close to the nest, but a much narrower range farther away. We generate predictions for both simple and complex hypothetical landscapes, to illustrate the effect of including the behavioral rule, and for real landscapes. In the real landscapes the models give similar predictions for visitation rates in simple landscapes, but more different predictions in heterogeneous landscapes. We also analyze the consequences of introducing hedgerows near a mass-flowering crop field under each model. The Lonsdorf model predicts that any habitat improvement will enhance pollination of the crop. In contrast, the CPF model predicts that the hedgerow must provide good nesting sites, and not just foraging opportunities, for it to benefit pollination of the crop, because good forage quality alone may drain bees away from the field. Our model can be used to optimize pollinator mitigation measures in real landscapes.