Modeling associations between migratory movements and seasonal honey bee colony survival: a machine-learning approach with big data
Abstract This study investigates the impact of commercial beekeeping practices, specifically colony movement and co-variates, on honeybee colony survival. We used data on 362,884 colonies from 62 beekeepers over 4 years (2022-2026) to compare the numberof times a given hive was moved to the survival or death of those hives. We found that excessive transportation significantlyreduces the probability of surviving beyond one season. We also found a positive effect between the predicted longevity andthe average days between moves, suggesting movement strategy could mitigate the negative impact of movements on longevity.Finally, we’ve compared 3 machine learning survival models and their combination on C-index and Brier score. We’ve thenselected the best model and data format for the features importance and simulation. We generated 15 different movementsscenarios on 3 simulated beekeeping seasons. The results of our model comparison highlight the importance of a proper datastructure to achieve optimal performance on censored data.
Authors
- Maxime Fraser Franco (ORCID: https://orcid.org/0000-0002-8687-8513)
- Ryan William Kuesel (ORCID: https://orcid.org/0000-0001-6450-939X)
- Brandon Kingsley Hopkins (ORCID: https://orcid.org/0000-0002-1740-7187)
- Nico Coallier (ORCID: https://orcid.org/0009-0005-1798-4661)
Institutions
- Washington State University (US)
Publication Details
- Journal
- Scientific Reports
- Published
- 2026-11-14
- DOI
- https://doi.org/10.1038/s41598-026-78956-7
- Primary Topic
- Insect and Pesticide Research
- Type
- article
- Field-Weighted Citation Impact
- 0.00