Vector-borne viral infections of ruminant livestock, including bluetongue virus, occur across much of northern and eastern Australia. Historical surveillance data suggests that these and other arthropod-borne viruses (‘arboviruses’) are introduced to Australia, perhaps on a regular basis, in insects borne on moist tropical monsoons originating in Asia. Although surveillance programs in northern Australia monitor for evidence of entry of new viruses and vectors, little is known of their source, their pathways of entry or the likelihood of establishment of the viruses or vectors once they reach Australia.
Understanding from where novel bluetongue viruses and their vectors originate, the potential frequency and temporal nature of these introductions and their likely site of first establishment are crucial for managing the risks associated with their entry. This thesis aimed to answer these questions by looking at long-distance windborne Culicoides dispersal within the region.
Pathways of long-distance Culicoides movement were first assessed using a migration model. This preliminary study showed that long-distance dispersal was a plausible means of entry of Culicoides into northern Australia and that the likely high risk period for such events is December – March. Subsequently, an atmospheric dispersion model was used to expand upon these initial findings. Potential dispersal was evaluated for the high risk period of fifteen consecutive years (1995-2010), from nine putative source sites across the eastern Indonesian archipelago, Timor-Leste and southern Papua New Guinea. The results highlight important differences in the potential frequency and spatio-temporal nature of dispersal from each of the putative source sites.
Case studies of exotic Culicoides and novel bluetongue virus incursions into northern Australia were evaluated using the dispersal model, based on collections made from ongoing, active arbovirus surveillance. In five of the six cases evaluated a likely dispersal event was identified and, in general, the findings were consistent with the hypotheses drawn from the atmospheric dispersion modelling results. Alternative methods of model evaluation are discussed in the thesis, and contribute to recommendations for ongoing research.
To determine the likelihood of bluetongue virus incursion via this pathway, the dispersal results were then incorporated into a risk analysis framework, along with other relevant factors, including environmental variables, and ruminant and Culicoides population densities. Due to the variable nature of data availability for the risk analysis, Bayesian networks were chosen as the method of risk assessment.
The analysis showed that overall the risk of bluetongue virus incursion via long-distance Culicoides dispersal is greatest to the north of the Northern Territory (also known as the “Top End”), and the Kimberley region of Western Australia. The source region posing the greatest risk includes the Indonesian archipelago from Sumba (in the west) to Tanimbar (in the east), including the island of Timor. Whilst there is some variation in months of greatest risk, dependent on the source or arrival area, the overall highest risk period comprises December-March.
The risk analysis highlighted the need for more precise information on bluetongue status and Culicoides populations at source sites. Whilst the latter is probably dependent on a number of environmental and climatic conditions, an understanding of Culicoides population dynamics in response to these factors in the tropics and sub-tropics remains elusive. This is therefore an important area for further research. Additionally, a better understanding of the duration of survival of Culicoides during dispersal, and factors contributing to the timing of their descent are important for continuing improvement of the dispersal modelling so as to better target surveillance efforts.
Whilst the research for this thesis focussed on Culicoides dispersal and transmission of bluetongue virus, the spatial and temporal patterns of dispersal are broadly relevant to a range of insect vectors within the region and the methodology is applicable to surveillance for these vectors and the arboviruses they transmit.