The spatial structure of host populations can influence patterns of disease incidence, and the effects of a pathogen on host community dynamics in a spatially heterogeneous community may differ from predictions derived from simple models. Here we examine the potential effects of habitat fragmentation on multi-host pathogens due to changes in landscape connectivity and host community composition. In previous work, a non-spatial model of a generalist plant pathogen, barley yellow dwarf virus (BYDV), suggested that BYDV could reverse the competitive outcome between perennial and annual host grasses, leading to the successful invasion of the competitively inferior annuals. However, continued existence of BYDV requires the persistence of the perennial grass in the community due to its role as a reservoir for infection between growing seasons. To investigate whether heterogeneity in the distributions of the host populations could alter the effect of BYDV on host community interactions, we have modified the BYDV model to include patch dynamics. Patches containing subpopulations of the perennial and/or annual grass species are linked by dispersal of the aphid vector responsible for BYDV transmission. Using this spatial model we examine whether BYDV can persist locally or in a metapopulation framework across a range of host community configurations. We then determine how pathogen-mediated interactions between perennial and annual competitors are altered at the local and regional scale when the host populations are spatial structured.

Results/Conclusions

The ability of BYDV to invade a spatially heterogeneous grassland community depends on both the proportion of perennial grasses and the spatial configuration of perennial and annual grasses. Under most conditions, increasing the proportion of patches occupied by perennials leads to an increase in the regional BYDV prevalence. However, increasing the number of patches containing both annuals and perennials can lead to a decrease in BYDV prevalence and raise the minimum aphid dispersal rate needed for pathogen persistence. Continued presence of BYDV in annual-only patches requires both a minimum level of connectivity with other patches and a threshold level of perennial abundance. The competitive outcome between perennial and annual grasses in local patches can be influenced by regional levels of BYDV, with perennial grasses dominating at low levels and annual grasses dominating at high levels. Finally, we explore the potential application of this approach to other multi-host pathogens such as Lyme disease, West Nile virus, and canine distemper virus.