Pollinator thermal preferences in an early-blooming crop

Insect-mediated pollination is critical to the human food supply, responsible for roughly one third of the food we eat. However, global pollinator populations are being threatened by a complex mixture of stressors, including extreme weather events and climate change. The production of some fruit crops, particularly those that bloom in the unstable weather conditions of early spring, is currently limited by a lack of pollinators. In this study, we assessed how thermal and climatic preferences varied among the bee taxa involved in apple pollination in the Midwestern US. We hypothesized that, when compared to honey bees, native bees would be active under less favorable weather conditions. To test this hypothesis, we used AutoPollS, a new technology that employs computer vision and deep learning neural networks to automatically monitor pollinator activity in the field. AutoPollS units were placed in four different apple orchards during bloom along with weather-monitoring equipment. The units took photos of insects as they visited flowers, and offline analysis later identified the insects to genus. We will present the results of this analysis, providing critical information about pollinator thermal preferences that will be useful for ensuring food security in a changing climate. Supported by NSF PRFB DBI-2305941.