Many infectious diseases of importance to human and wildlife health are shared by a community of host species. In such cases, some host populations often contribute significantly more to transmission than others. Recent studies have noted that the identity of these transmission dominant host species may vary across a landscape. Here we examine how patterns of transmission dominance vary across space in an ungulate host community at the National Bison Range (Montana, USA). We asked two key questions: (1) does differential space use by host species produce spatial heterogeneity in infection risk and (2) does differential space use produce spatial variation in the identities of transmission dominant hosts? We used data on infection intensities of 7 parasitic nematodes to estimate partial force of infection from 6 host species. We then fit a novel species distribution model (the Plug-and-Play model) to host presence data and predicted host population intensity across the range. This model estimates the relative suitability of a given environment by comparing its frequency in the species’ known habitat to that on the landscape. We calculate spatial contributions to force of infection by distributing the previously calculated partial force of infection across the range in accordance to population intensity.
This analysis estimated the relative contribution of each host species to the force of infection of each parasite species at a 30m scale. Statistical analysis showed there to be significant positive spatial autocorrelation (Moran’s I, p<.01) in infection risk across the range, suggesting that different hosts might experience wildly different infection risks. There were also high levels of spatial diversity in transmission dominant hosts. Although American Bison are responsible for over 97% of the force of infection of Cooperia oncophora across the range they are only transmission dominant in 81% of the range. More broadly, the proportion of the range over which a host species is transmission dominant is highly positively correlated with their range-wide partial force of infection (p<.01). As a result parasites which are spread more equally by different species, such as Haemonchus contortus, have extremely high spatial diversity in transmission dominance (Simpson’s Diversity Index=0.771). These results suggest that identification of a transmission dominant host at one scale may not be applicable at other scales. These results motivate further study of spatial dynamics of multi-host disease spread by calling into question the generality of our (not explicitly spatial) understanding of many important disease systems.