Network theory has emerged as a powerful tool in wildlife epidemiology in recent years, as it provides a framework for incorporating host contact patterns into models of infectious disease spread. Host contact patterns of wildlife species are typically derived from observational studies of their social interactions. In solitary species, contacts are less frequent and thus harder to observe. For such species, pathogen transmission may be limited to certain habitats, such as refugia (e.g. burrows or caves), or sources of common resources (e.g. watering holes), that provide increased opportunity of contact between hosts.
Here we focus on a relatively solitary species, the desert tortoise. An infectious and horizontally transmitting disease (upper respiratory tract disease, URTD) is considered to be a major threat as a strong correlation has been found between prevalence of URTD and declines in desert tortoise populations. Preliminary results from one-to-one transmission experiments suggest long duration close contacts to be important for URTD transmission. Close contacts between individuals that occur above ground, however, are infrequent and usually too short for URTD transmission. We therefore hypothesize extended duration and proximity during burrow sharing events facilitates spread of URTD between burrow sharing animals.
Results/Conclusions
In order to tease apart the behavioral determinant of URTD spread, we first construct network models that represent different mechanisms of disease transmission (i.e., direct versus indirect) in desert tortoises. We next use statistical approach to test which contact network best explains disease prevalence in tortoise populations. Specifically, generalized linear mixed effect models are used to compare degree centrality of nodes across all contact network and determine the best predictor of disease risk among individuals. Our results suggest infection risk in desert tortoises is best explained by asynchronous burrow associations among individuals. We will discuss the implications of these results and our future plans to understand how disease may, in turn, change host behavior and contact networks. In conclusion, our approach allows inference on mechanisms of infection spread between hosts where experimental transmission studies are not conclusive or feasible, and can be easily extended to other wildlife species.