Urban greenspace connectivity drive shifts in host assemblages and tick-borne pathogen infection

Habitat fragmentation is often highlighted as a driver of tick-borne disease hazard and spillover risk via reduction in biodiversity. However, habitat fragmentation can have divergent impacts on host, vector, and pathogen dynamics depending on the distribution of fragment sizes and the levels of connectivity to surrounding habitat, particularly when habitat fragments are embedded in an urban matrix. We examine how extreme habitat fragmentation influences host community composition in an urban landscape and determine its cascading impacts on Ixodes scapularis vector abundance and infection prevalence with human pathogenic Borrelia burgdorferi, Babesia microti, and Anaplasma phagocytophilum. We utilize camera-trapping and live mammal-trapping methods to quantify the availability of vertebrate hosts to questing larval ticks and relate relative host activity to the resulting density of nymphs and nymphal infection prevalence; the combination of these metrics determines the tick-borne disease hazard (i.e. the density of infected questing nymphs - DIN). We found that increased habitat connectivity in urban areas shifted the composition of the host community from human-adapted to forest-dependent species, species which inhabit forested habitats for all or a portion of their lifecycles. The resulting increased encounter probability between ticks and forest-dependent species increased the density of nymphs and nymphal infection prevalence with host-limited pathogens, A. microti and A. phagocytophilum, amplifying local tick-borne disease hazard. Host encounter probability of all species examined did not increase B. burgdorferi nymphal infection prevalence, likely due to this pathogens wider host range; whereas increased deer encounter probability decreased the nymphal infection prevalence of B. burgdorferi. These findings emphasize the importance of host identity, rather than host diversity, in shaping the heterogenous distribution of tick-borne pathogen risk in highly fragmented urban forest patches and suggest a non-linear association between disease risk and host biodiversity.