Models reveal the importance of alternative hosts and environmental transmission for emergence of bacterial disease in muskoxen in the Arctic

• Mechanistic model including multiple host species and environmental transmission. • Feces are more important than carcasses for transmission despite shorter persistence. • Alternative hosts could accelerate population decline of vulnerable species. • Flexible framework useful for other environmentally transmitted wildlife diseases. Emerging diseases are an increasing concern for vulnerable wildlife populations. Studying these is particularly challenging in systems with multiple hosts and environmental transmission, but models can provide critical insights into their dynamics and guide further research and management. We built a multi-species model of Erysipelothrix rhusiopathiae , a bacterium that has caused mass mortalities of muskoxen ( Ovibos moschatus ) in the Canadian Arctic, parameterized it with data from the literature, and analyzed it to predict the role of different hosts (muskoxen and Arctic foxes) and the importance of different environmental transmission sources (carcasses and feces). Based on the model’s basic reproduction number, R 0 , we predicted bacteria establish only if they were transmitted directly or via feces and not if bacteria were acquired solely from infected carcases. Accordingly, sensitivity analyses revealed that processes affecting environmental transmission, like bacterial shedding and persistence outside the host, had the greatest influence on R 0 . Numerical solution of the model showed that disease-induced mortality could produce strong population declines, which could be accelerated as alternative hosts decouple transmission to muskoxen from their population density. Our research highlights the need to understand bacterial contamination in environmental reservoirs like soil and ponds, and species overlap; this is particularly relevant given the wide array of potential hosts in this system, which includes foxes, wolves, polar bears, caribou, hares, lemmings, and ravens. Our modelling approach combining multiple data sources provides a flexible and efficient framework to generate actionable insights that support specific research and guide management of similarly complex wildlife diseases.