Habitat properties are key drivers of Borrelia burgdorferi (s.l.) prevalence in Ixodes ricinus populations of deciduous forest fragments

Background: The tick Ixodes ricinus has considerable impact on the health of humans and other terrestrial animals because it transmits several tick-borne pathogens (TBPs) such as B. burgdorferi (sensu lato), which causes Lyme borreliosis (LB). Small forest patches of agricultural landscapes provide many ecosystem services and also the disservice of LB risk. Biotic interactions and environmental filtering shape tick host communities distinctively between specific regions of Europe, which makes evaluating the dilution effect hypothesis and its influence across various scales challenging. Latitude, macroclimate, landscape and habitat properties drive both hosts and ticks and are comparable metrics across Europe. Therefore, we instead assess these environmental drivers as indicators and determine their respective roles for the prevalence of B. burgdorferi in I. ricinus. Methods: We sampled I. ricinus and measured environmental properties of macroclimate, landscape and habitat quality of forest patches in agricultural landscapes along a European macroclimatic gradient. We used linear mixed models to determine significant drivers and their relative importance for nymphal and adult B. burgdorferi prevalence. We suggest a new prevalence index, which is pool-size independent. Results: During summer months, our prevalence index varied between 0 and 0.4 per forest patch, indicating a low to moderate disservice. Habitat properties exerted a fourfold larger influence on B. burgdorferi prevalence than macroclimate and landscape properties combined. Increasingly available ecotone habitat of focal forest patches diluted and edge density at landscape scale amplified B. burgdorferi prevalence. Indicators of habitat attractiveness for tick hosts (food resources and shelter) were the most important predictors within habitat patches. More diverse and abundant macro and microhabitat had a diluting effect, as it presumably diversifies the niches for tick-hosts and decreases the probability of contact between ticks and their hosts and hence the transmission likelihood. Conclusions: Diluting effects of more diverse habitat patches would pose another reason to maintain or restore high biodiversity in forest patches of rural landscapes. We suggest classifying habitat patches by their regulating services as dilution and amplification habitat, which predominantly either decrease or increase B. burgdorferi prevalence at local and landscape scale and hence LB risk. Particular emphasis on promoting LB-diluting properties should be put on the management of those habitats that are frequently used by humans. In the light of these findings, climate change may be of little concern for LB risk at local scales, but this should be evaluated further.