Extreme natural events such as cold waves, droughts, floods or hurricanes can drastically impact wildlife populations and are expected to become more frequent and more intense in the coming decades. When populations experience abnormally high declines within a short time interval due to such phenomena, the losses can be referred to as natural population die-offs (NPDOs). Although such events have been observed and population declines have been recorded, there is currently little known about which species might be most affected. Using a database of 72 NPDOs from 31 terrestrial herbivorous mammals, we modelled the effects of four biological traits (adult body mass, foraging strategy, home-range area and territoriality) on the degree of population loss (severity) caused by extreme natural events. We found that the susceptibility to large NPDOs decreases with increased home-range size for a given body mass. Foraging strategy was also found to be significantly associated with NPDO severity, with grazers and mixed feeders experiencing larger declines than browsers. Our analyses moreover suggested that wide-ranging browsers might be less susceptible to large NPDOs than browsers with small home ranges. Identifying the traits shaping high biological sensitivity and/or limited adaptive capacity to extreme natural events can help us to identify those populations most likely to become increasingly vulnerable to NPDOs, allowing tailored interventions to be implemented to avoid local extinctions. This will be of the utmost importance for those populations already experiencing high levels of anthropogenic impact and distributed in regions where exposure to extreme natural events is expected to increase in the coming decades.