A multistate capture-recapture model to estimate cause-specific injury and mortality of North Atlantic right whales

Understanding the causes of mortality for a declining species is essential for developing effective conservation and management strategies, particularly when anthropogenic activities are the primary threat. Using a competing hazards framework allows for robust estimation of the cause-specific variation that may exist across multiple dimensions, such as time and individual. Here, we estimated cause-specific rates of severe injury and mortality for North Atlantic right whales (Eubalaena glacialis), a critically endangered species that is currently in peril due to human-caused interactions. We developed a multistate capture-recapture model that leveraged 30 years of intensive survey effort yielding sightings of individuals with injury assessments and necropsies of carcass recoveries. We examined variation in the hazard rates of severe injury and mortality due to entanglements in fishing gear and vessel strikes as explained by year and the age and reproductive status of the individual. We found strong evidence for increased rates of severe entanglement injuries after 2013 and for females with calves, with consequently higher marginal mortality. The model results also suggested that despite vessel strikes causing a lower average rate of severe injuries, the higher mortality rate conditional on injury results in significant total mortality risk, particularly for females resting from a recent calving event. Large uncertainty in the estimation of carcass recovery rate for vessel strike deaths permeated into the apportionment of mortality causes. The increased rates of North Atlantic right whale mortality in the last decade, particularly for reproducing females, puts the species at risk of severe decline. By apportioning the human-caused threats using a quantitative approach with estimation of relevant uncertainty, this work can guide development of conservation and management strategies to facilitate species recovery.