Pathogen abundance model summaries.

Multiple linear regression models were utilized to identify associations between colony-level residual pathogen abundance and the factors monitored in this study (i.e., colony health, percent mite infestation, and beekeeping operation) throughout the monitoring period from January to September 2014. Residual pathogen abundance was used as the response variable in the models to remove the confounding effect of seasonality on pathogen abundance in order to investigate associations between residual pathogen abundance and other covariates. Associations between the residual abundance of each pathogen assayed in this sample cohort and other covariates are denoted by the magnitude and direction of the estimate listed in the table. Operation 2 estimates the residual pathogen abundance in comparison to Operation 1, frame count was used as a continuous variable as an estimate of colony health, and mite infestation (%) represents the number of mites counted per 100 bees. The standard error, t-value, and p-value for each estimate of coefficients (est.) within the final model are provided; p-values <0.05 are indicative of significant associations (bold).