Skip to main content
ARS Home » Northeast Area » University Park, Pennsylvania » Pasture Systems & Watershed Management Research » Research » Publications at this Location » Publication #368349

Research Project: Sustainable Intensification of Crop and Integrated Crop-Livestock Systems at Multiple Scales

Location: Pasture Systems & Watershed Management Research

Title: Summer weather conditions influence winter survival of honey bees (Apis mellifera)in the northeastern United States

Author
item CALOVI, MARTINA - Pennsylvania State University
item GROZINGER, CHRISTINA - Pennsylvania State University
item MILLER, DAVID - Pennsylvania State University
item Goslee, Sarah

Submitted to: Scientific Reports
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/5/2021
Publication Date: 1/15/2021
Citation: Calovi, M., Grozinger, C.M., Miller, D.A., Goslee, S.C. 2021. Summer weather conditions influence winter survival of honey bees (Apis mellifera) in the northeastern United States. Scientific Reports. 11:1553. https://doi.org/10.1038/s41598-021-81051-8.
DOI: https://doi.org/10.1038/s41598-021-81051-8

Interpretive Summary: Despite the profound importance of honey bees to global agriculture and to native ecosystems, the causes of overwintering mortality in colonies of European honey bee in Pennsylvania are not clearly understood. While it is believed that temperature, moisture, quality of the surrounding landscape, and management are important, no previous research has attempted to quantify all of those factors. We used three years of survey data of Pennsylvania beekeepers, and a suite of climatic, topographic and management variables that we expected to describe temperature, moisture, floral resources, and pesticide load in ways relevant to honey bees. The machine learning model we developed explained 74% of the variability in overwintering mortality in honey bee colonies that had been treated for Varroa mites. This model was used to map the probability of overwintering mortality across Pennsylvania, for the three years studied individually, and across a thirty-eight year period. These maps highlight the variability in space and time of the factors that determine honey bee mortality, and the importance of collecting data over many years. These maps are being incorporated into an existing decision support tool, Beescape.

Technical Abstract: The European honey bee is both a crucial pollinator, contributing to agricultural productivity and natural ecosystems, and an agricultural commodity in its own right. Over the past two decades, honey bees have been facing heavy mortality rates globally; a complex suite of factors appear to be responsible, with no clear major driver. Climate is important, both for its effect on the bees themselves, and for its effect on the plants that support them. Surrounding land use, particularly proportion of agricultural and urban land uses, determines forage resource abundance and pesticide exposure risk. Finally, management decisions, such as the use of mite treatment, contribute to colony success and failure. We used three years of data from a survey of Pennsylvania beekeepers to assess the importance of climatic, topographic, land use, and management factors on overwintering mortality in the European honey bee. Random Forest model of overwintering survival had a 74% accuracy; a climate-only model was as accurate as a model that also included topography, land use, and management. The most important variables in either model were growing degree days, maximum temperature, and winter precipitation. The climate-based model was used to create maps of survival probability across Pennsylvania for the three years of the study, and a composite map of survival probability for 1981-2019. Probability of mortality varied greatly in both time and space. Over the 38-yr period for which climate data were available, there was strong geographic variation within Pennsylvania (long-term survival probability 8 - 84%). Although three years of data were not enough to adequately capture the range of possible climatic conditions, the model nonetheless performed well. This approach is suited to understanding complex drivers of survival, and to predicting performance both given current conditions and for projected climate change.