Effects of Wintering Environment and Parasite-Pathogen Interactions on Honey Bee Colony Loss in North Temperate Regions

Abstract

Extreme winter losses of honey bee colonies are a major threat to beekeeping but the combinations of factors underlying colony loss remain debatable. We monitored colonies in two environments (colonies wintered indoors or outdoors) and characterized the effects of two parasitic mites, seven viruses, and Nosema on honey bee colony mortality and population loss over winter. Samples were collected from two locations within hives in fall, mid-winter and spring of 2009/2010. Although fall parasite and pathogen loads were similar in outdoor and indoor-wintered colonies, the outdoor-wintered colonies had greater relative reductions in bee population score over winter. Seasonal patterns in deformed wing virus (DWV), black queen cell virus (BQCV), and Nosema level also differed with the wintering environment. DWV and Nosema levels decreased over winter for indoor-wintered colonies but BQCV did not. Both BQCV and Nosema concentration increased over winter in outdoor-wintered colonies. The mean abundance of Varroa decreased and concentration of Sacbrood virus (SBV), Kashmir bee virus (KBV), and Chronic bee paralysis virus (CBPV) increased over winter but seasonal patterns were not affected by wintering method. For most viruses, either entrance or brood area samples were reasonable predictors of colony virus load but there were significant season*sample location interactions for Nosema and BQCV, indicating that care must be taken when selecting samples from a single location. For Nosema spp., the fall entrance samples were better predictors of future infestation levels than were fall brood area samples. For indoor-wintered colonies, Israeli acute paralysis virus IAPV concentration was negatively correlated with spring population size. For outdoor-wintered hives, spring Varroa abundance and DWV concentration were positively correlated with bee loss and negatively correlated with spring population size. Multivariate analyses for fall collected samples indicated higher DWV was associated with colony death as did high SBV for spring-collected samples.

Fig 1. Change in colony population score (frames of bees, 1 frame = ~ 2,430 bees) from fall to spring, for indoor (-■-) and outdoor (─●─) wintered colonies.

Means followed by the same letter are not significantly different (P > 0.05; Bonferroni).

Fig 2

Interactions between season (spring and fall) and wintering method (indoor and outdoor) (see Table B in S1 File) for DWV and BQCV concentrations (left axis) and mean abundance of Nosema (right axis) (± standard error). Results of significant slices for each virus are indicated above the bars for slices by wintering method and below the graph for slices by season, (horizontal lines indicate significant slices). Data represent pooled values for brood area and entrance samples.

Fig 3. Interaction between season (spring and fall), wintering method (indoor and outdoor), and sampling location (brood area and entrance) (see Table B in S1 File) for BQCV concentrations (± standard error).

Results of significant slices are indicated above the bars for wintering method and below the graph for season and wintering method, (horizontal lines indicate significant slices).

Fig 4. Interaction between season (fall and spring) and sampling location (brood area and entrance) (see Table B in S1 File) on mean abundance of Nosema (± standard error).

Results of slices are indicated above the bars for sampling location and below the bars for season, (horizontal lines indicate significant slices). Data represent pooled values for indoor and outdoor samples.

Fig 5

Effect of season (fall and spring) on mean abundance of Varroa (left axis) and concentrations of viruses (SBV, KBV and CBPV) (right axis) (± standard error) (see Table B in S1 File). Data represent pooled levels for both wintering methods (indoor and outdoor) and both sampling locations (brood area and entrance). Means followed by the same letter within a parasite or pathogen group are not significantly different (P > 0.05, Slice).