Undergraduate Research Conference Proceedings

Damage to bee wings commonly occurs during ecologically-essential foraging and flight activities of bees. Interactions and collisions with native environmental conditions such as wind and vegetation leave wings exposed to potential wear and tear. Morphological traits influence bee flight performance, as bees with larger body sizes demonstrate increased flight distance capacity and foraging range. We analyze the degree of wing wear as an indicator of these activities, predicting that larger bees will accumulate more wing damage from extended exposure to environmental wear. We imaged wings across bee species native to Santa Barbara County. By manually creating masks to isolate the wing, we trained a computer vision model with ground-truth data to identify pixels in wing images and separate them from the background. This image segmentation model allows for shape analysis and comparison of degree of wear in the wing’s margin across various species through this large-scale collection of qualitative data. We also created a manual wing wear scoring index to compare accuracy with this model. Genera such as Bombus and Xylocopa that tend to have larger body sizes are predicted to be assigned, on average, higher scores of wing wear compared with smaller species being assigned lower scores to indicate experiencing less wear. By comparing wear patterns, these correlations could offer insight on species resilience in their native habitats, suggesting that species with larger body sizes may be more vulnerable to environmental stresses that require more demanding flight and foraging efforts. This model could then have potential research applications as bioindicators of the local foraging ecology and environmental stress/fragmentation.

This poster was presented at the UCSB Undergraduate Research Symposium 2025, and at the UCSB URCA Poster Colloqium 2025.