Predicting Pollen Specialist Bee Species Using Visitation Data
Author Information
Author(s): Colleen Smith, Nick Bachelder, Avery L. Russell, Vanessa Morales, Abilene R. Mosher, Katja C. Seltmann
Primary Institution: University of California, Santa Barbara
Hypothesis
Can we predict pollen specialist and generalist bees from flower visitation data?
Conclusion
Visitation data can be used to predict specialist bee species with high accuracy, while predicting generalist species is more challenging.
Supporting Evidence
- On average, 72.9% of visits made by pollen specialist bees were to their host plants.
- The models predicted specialist bee species with high accuracy (mean 94%).
- Generalist bee species were predicted with lower accuracy (mean 70%).
- Visitation data can enhance understanding of plant-pollinator interactions.
- Range size is predictive of diet breadth in bees.
- Male bees were found to visit non-host plants more frequently than females.
Takeaway
Scientists can use information about which flowers bees visit to figure out which bees are picky eaters and which ones are not.
Methodology
Machine learning models were developed to predict pollen diet breadth in bees using visitation, occurrence, and phylogenetic data.
Potential Biases
Visitation datasets may miss interactions and incorrectly classify generalist species as specialists.
Limitations
The study may have biases due to incomplete visitation datasets and the challenge of accurately classifying generalist bees.
Participant Demographics
The study included 682 bee species native to the United States.
Statistical Information
P-Value
p<0.0001
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website