Habitat, not biodiversity, drives some apparent dilution effects: Comparing the relative effects of biodiversity and habitat on pollinator pathogen prevalence

Michelle L. Fearon and Elizabeth A. Tibbetts, Ecology and Evolutionary Biology, University of Michigan, Ann Arbor, MI, Chelsea L. Wood, School of Aquatic and Fishery Sciences, University of Washington, Seattle, WA

Background/Question/Methods
Habitat characteristics are well-known to influence host biodiversity, nutrition, and immune function with likely implications for infectious disease, but research has focused on how biodiversity can influence disease risk (e.g. ‘dilution’ and ‘amplification’ effects). Few studies have evaluated the how habitat may be linked with pathogen prevalence. Here, we propose the ‘habitat–disease relationship’ as a non-mutually exclusive pattern to biodiversity–disease relationships. Habitat may directly influence pathogen prevalence by effecting host traits like nutrition and immune response (‘habitat–disease relationship’) and/or indirectly influence pathogen prevalence by modulating host biodiversity (‘biodiversity–disease relationship’). Therefore, the habitat–disease relationship may contribute to patterns that are normally ascribed to the dilution effect alone. We use a path model to test the relative strength of direct and indirect links between multiple habitat characteristics and pathogen prevalence in a pollinator system with multiple bee hosts and multiple viruses. Previous evidence demonstrates the dilution effect for pollinator viruses, but it remains unclear whether this pattern is driven by habitat characteristics, pollinator species richness, or a combination of both pathways. We assess habitat characteristics associated with pollinator nutrition, including abundant and diverse local floral resources, proportion of surrounding natural area, and landcover diversity.
Results/Conclusions
Habitat factors are directly associated with both positive and negative changes in virus prevalence, showing evidence of a habitat–disease relationship that is separate from the biodiversity–disease pathway. Specifically, increased landcover diversity and floral density are linked with reduced virus prevalence, while greater proportion of natural area, landcover diversity, and floral richness are linked with increased virus prevalence. These habitat characteristics also indirectly mediate biodiversity–disease relationships by promoting pollinator species richness and indirectly reducing virus prevalence, consistent with a habitat-driven dilution effect. Surprisingly, the strength of habitat–disease paths is similar to biodiversity–disease paths, but the directionality of the pathways often opposes each other. The summed net effect of all direct and indirect links between habitat, biodiversity, and virus prevalence resulted in reduced prevalence by 0.8%, 6.5%, and 14.3% for black queen cell virus, sacbrood virus, and deformed wing virus, respectively. Disentangling the relative impact of direct habitat–disease relationships and the indirect effect of habitat on biodiversity–disease relationships is critical to understanding the sources of variation in pathogen prevalence. Focus on biodiversity–disease relationships alone overlooks the potential important direct effects of habitat on disease risk that could augment or obscure observed biodiversity–disease relationships.