Extreme changes in the environment can generate high mortalities in wildlife populations. When these mortalities are attributable to extreme natural events, they are referred to as natural population die-offs. Despite growing reports of such die-offs, a consensus on how to define them has not emerged. Furthermore, although anthropogenically caused extreme events are predicted to occur at a higher frequency and intensity compared with natural events, an integrative synthesis assessing their significance for wildlife population viability is lacking. These issues hamper the ability to identify populations most at risk. Here, we propose a functional definition of natural population die-offs, an assessment of extrinsic and intrinsic processes shaping these die-offs, and a framework for assessing the vulnerability of terrestrial mammals to natural and anthropogenically caused extreme events.