Introduction: Canine rabies, a fatal viral zoonosis, is now less than 300 kilometres from Indonesian archipelago. To respond to this imminent threat, we need to model how rabies will spread through Australian ecosystems, to develop effective rabies management plans. This will improve our chances of minimising reaction times and containing outbreaks. Aims: To quantify contact rates, abundance, population turnover and dog bite frequency, in order to inform models of canine rabies spread. This is essential for identifying critical control points.

Currently, Australia is free from terrestrial rabies but an incursion from nearby Indonesia, where the virus is endemic, is a feasible threat. Here, we aimed to determine whether the response to a simulated rabies incursion would vary between three extant Australian dog populations; free-roaming domestic dogs from a remote indigenous community in northern Australia, and free-roaming domestic and wild dogs in peri-urban areas of north-east New South Wales. We further sought to predict how different management strategies impacted disease dynamics in these populations. We used simple stochastic state-transition models and dog demographic and contact rate data from the three dog populations to simulate rabies spread, and used global and local sensitivity analyses to determine effects of model parameters. To identify the most effective control options, dog removal andvaccination strategies were also simulated. Responses to simulated rabies incursions varied between the dog populations. Free-roaming domestic dogs from north-east New South Wales exhibited the lowest risk for rabies maintenance and spread. Due to low containment and high contact rates, rabies progressed rapidly through free-roaming dogs from the remote indigenous community in northern Australia. In contrast, rabies remained at relatively low levels within the north-east New South Wales wild dog population for over a year prior to an epidemic. Across all scenarios, sensitivity analyses revealed that contact rates and the probability of transmission were the most important drivers of the number of infectious individuals within a population. The number of infectious individuals was less sensitive to birth and death rates. Removal of dogs as a control strategy was not effective for any population modelled, while vaccination rates in excess of 70% of the population resulted in significant reductions in disease progression. The variability in response between these distinct dog groups to a rabies incursion, suggests that a blanket approach to management would not be effective or feasible to control rabies in Australia. Control strategies that take into account the different population and behavioural characteristics of these doggroups will maximise the likelihood of effective and efficient rabies control in Australia.