Discussion
Understanding the dynamics of small populations is critical for the effective conservation of at-risk species. Previous work has demonstrated an increase in extinction proneness in declining populations (Fagan & Holmes 2006). Here, we corroborate preexisting theoretical and empirical studies on the extinction dynamics of populations and, additionally, show for the first time that the rate of the extinction vortex can be altered by body size.
Reinforcing previous findings, our results show that the proximity of a population to extinction is dependent on the logarithm of population size (Lande 1993; Fagan & Holmes 2006). This suggests that the proximity to extinction decreases at an increasing rate as a population declines, indicative of an extinction vortex. Accordingly, care should be taken to maintain populations at high densities to avoid self-reinforcing spirals to extinction and to maximize the probability of long-term persistence (Fagan & Holmes 2006). We cannot rule out the influence of body size; the best fitting models did not provide a significantly improved fit compared to those that included an interaction term between body size and population size (Table 2). The positive coefficients for the interaction between populations size and body size (Table 2), suggests population size becomes increasingly important in determining the distance from extinction as body size increases. As such, smaller-bodied species appear to be more vulnerable to imminent extinction across a greater range of population sizes, in agreement with previous studies reporting greater population persistence among species with slower life history traits (Newmark 1995; Saether et al. 2005). Our relatively small dataset may account for the fact that this is not the clear best performing model in this analysis.
According to the extinction vortex, genetic deterioration and Allee effects are expected to result in proportionally larger declines as population size diminishes (Brook et al. 2008). Indeed, we found an increase in the year-to-year per capita rate of decline as population shrinks (Table 2; Fig. 1). The implication of this is that even with conservation intervention, species that fall into the extinction vortex may struggle to be saved and require a non-linear increase in the magnitude of the change required to save a population as it moves towards extinction. Well-studied populations on the verge of extirpation support this; the decline of the Florida panther population (Puma concolor coryi ) was only reversed after the introduction of several individuals translocated from healthy populations leading to the restoration of genetic diversity (Johnson et al. 2010). In practical terms, this emphasizes the need for early conservation intervention, with a strong focus on ensuring species do not fall into the extinction vortex.
Our results suggest that the key question of when a species is at risk of rapidly collapsing to extinction is not only a function of population size, but is also affected by the body size of the species; we found evidence that small body size exacerbates the rate of decline in geometric growth rate as population size declines (Fig. 1). This was true for all groups (Table 2), suggesting a similar trend across the vertebrate phylum. Though it is acknowledged that extinction risk is an emergent property of the interaction between biological traits and the type of threatening process (Owens & Bennett 2000; Isaac & Cowlishaw 2004; Price & Gittleman 2007; Brook et al. 2008; Davidson et al. 2009; Ripple et al. 2017), our findings may seem at odds with the frequently reported positive association between body size and extinction threat level (e.g. IUCN threat status) (Gaston & Blackburn 1995; Bennett & Owens 1997; Cardillo et al. 2005; Liow et al. 2008; Dirzo et al. 2014). However, the extinction risk of highly fecund species is tempered by naturally larger populations (Tracy & George 1992; Newmark 1995); species at the fast end of the life history speed continuum seem to be more vulnerable after controlling for the confounding effect of population size (Cook & Hanski 1995; Johst & Brandl 1997; Saether et al. 2005; Hilbers et al. 2016). A possible explanation is that smaller-bodied species have faster life histories and are more susceptible to stochastic elements (Peltonen & Hanski 2001; Sinclair 2003; Saether & Engen 2002; Saether et al. 2004; Wilson & Martin 2012), therefore they are predisposed to respond faster to the deleterious demographic impacts of genetic decay, Allee effects and other stressors.
The results of our third analysis, investigating population variability through time, supports the idea that stochastic processes are involved in causing the extirpation of these populations (Fagan & Holmes 2006; Brook et al . 2008). That the magnitude of annual population variability is higher in smaller-bodied species is indicative of lower population stability in smaller-bodied species, as has been noted elsewhere (Sinclair 2003). This may also help to explain the significant negative interaction between body size and years to extinction in the non-avian subset (Table 2); given that the population dynamics of smaller-bodied species is inherently more stochastic, any increase in year-to-year variability due to stochastic elements could be less detectable. We suspect that a similar pattern is not observed in our avian subset because of the small variation in body size; the range of body masses in our avian subset (~7kg) is two orders of magnitude smaller than that of our non-avian subset (~350kg).
In conclusion, despite the large disparity in ecological and environmental contexts among populations constituting this study, we find evidence that small body size exacerbates the rate of the extinction vortex, providing one of the first studies to investigate differential vulnerability to the extinction vortex in relation to intrinsic biological traits and, to our knowledge, the first to specifically investigate this in real-life populations. The practical relevance of our findings is highlighted by the fact that species-specific data on body size is arguably the most widely available across all taxa, and our results demonstrate the need for a generally conservative approach to population targets especially in small-bodied taxa with fast life history speeds and a high susceptibility to stochastic processes.