Emerging wildlife diseases, such as the recent global outbreak of Highly Pathogenic Avian Influenza, can devastate animal populations, yet identifying transmission pathways and true levels of mortality remains a formidable challenge. We leveraged European carcass counts from the severely impacted Northern Gannet, to which we fit a Bayesian, metapopulation Susceptible-Infected-Recovered-Deceased model, hence uncovering key epidemiological parameters, transmission routes, and overall mortality across the species’ 43-colony metapopulation. Within-colony transmission patterns were consistent with fixed spatial structuring of nests. Between-colony spread was predominantly distance-dependent, in-line with intra-specific transmission, though large-scale outbreak initiation was best explained by inter-specific vectors. We estimate that these multiscale/multispecies dynamics culminated in 33% continental mortality (95% CI: 27–39%) across the Gannet metapopulation. Our approach not only disentangles inter- and intra-specific transmission processes, but offers a scalable tool for real-time assessment of disease spread in wildlife, helping to control threats to biodiversity, agriculture, and public health.