Discussion
Our study shows that variation in human activity can profoundly affect multiple aspects of risk-sensitivity and antipredator responses in wildlife even over relatively small spatial scales. Specifically, ground squirrels residing in areas of greater human activity consistently had shorter FIDs than squirrels residing in adjacent, nearby areas of lower human activity. Importantly, repeated exposure to human activity also influenced the post-FID response, as squirrels from high human disturbance areas also fled shorter distances before stopping to observe a potential threat following their flight response, and if they fled into shelter, were consistently faster to emerge from shelter. Squirrels from high activity sites have likely habituated to high human activity, thus responding less to an approaching human compared to individuals that have less experience with humans. While many other studies have observed animal habituation to non-dangerous humans (Petelle et al. 2013; Vincze et al. 2016; Uchida et al. 2019; Uchida & Blumstein 2021), it is striking that variation in the degree of habituation emerged over a small spatial scale (i.e., within a 1-hectare area). Beyond the strong effects of relative human activity on multiple antipredator behaviors, we also documented consistent individual differences in the suite of repeatable behaviors along a general, shy-bold continuum (Wilson et al. 1994) including high among-individual correlations between FID and post-FID responses for the first time.
Having a highly correlated suite of behaviors may be adaptive or maladaptive, depending on the situation and environment in which the animal resides (Trouilloud et al. 2004; Geffroy et al. 2015; Wilson et al. 2020). In this case, having correlated behaviors throughout the antipredator response may be beneficial when it comes to predators (e.g., when real danger is present, animals should both flee readily and hide for a relatively long time before resuming activity). However, the correlation may exacerbate unnecessary over-avoidance of humans that can be a substantial problem in human disturbed environments (Trimmer et al. 2017; Guiden et al. 2019). Unhabituated animals might both flee too readily from humans (that are not actually dangerous), but also wait too long to emerge from shelter which may mean a lost opportunity in foraging time. On the other hand, getting habituated in terms of both reduced tendency to flee and reduced time spent inactive can be beneficial. However, if getting bolder (here, when habituating to humans) carries over to other situations, this can be ecologically relevant. For example, we know that bolder squirrels pick-up more fleas (Smith et al. 2021) and have increased venom resistance (Holding et al. 2020). Perhaps most notably, it has been suggested that boldness associated with habituation to humans can make animals less wary when encountering actual, dangerous predators, which can obviously be very costly (Lowry et al. 2012; Geffroy et al. 2015). Thus, anthropogenic influences on boldness could influence disease transmission and predator-prey dynamics.
Here, we found that animals that were bolder, apparently due to being habituated to human activity, were also more readily trapped. Other studies have found mixed evidence for an association between boldness and trappability (Biro 2012; Michelangeli et al. 2015; Brehm & Mortelliti 2018). The observation that trappability predicts antipredator behavior can have important implications as a systematic bias when trapping for monitoring, ecological research or pest control (Biro & Dingemanse 2009; Garvey et al. 2020).
Besides strong effects of human activity and individual differences, the three main antipredator responses were largely either unaffected or only weakly affected by the social (presence and number of conspecifics) and environmental contexts (distance to shelter, vegetation cover) that we considered. We predicted that animals that are closer to shelter would exhibit shorter FIDs and that vegetation cover would affect FIDs, but these factors did not significantly affect FIDs. FIDs were, however, affected by the social context (group size). In theory, this effect could have gone either way. If larger groups with “many eyes” (Pulliam 1973) should detect predators sooner, animals in larger groups should flee sooner (i.e., have larger FIDs). On the other hand, the dilution or confusion effects (Bertram 1978) can make animals in larger groups safer in which case, FIDs should be smaller. In fact, FIDs increased with group size, consistent with the former mechanism. This finding is particularly interesting given that California ground squirrels are less vigilant when foraging in groups (Ortiz et al. 2019). Taken together, this suggests that, on average, individuals foraging in groups may benefit from both increased energy intake and earlier detection of threats. However, humans are not a direct threat to this species and consistently fleeing early due to human disturbance can lead to decreased energy intake in the long run. Thus, the social information provided from group foraging may only be beneficial when confronted with an actual predator.
We also found evidence for more subtle context-dependence in these antipredator behaviors. In particular, after fleeing (post-FID response), the decision to ‘stop and look’ versus flee all the way to shelter was largely environmentally context dependent. Animals showed no consistent individual differences in this decision. Instead, they were more likely to flee to shelter if available shelter was closer. Interestingly, juveniles were generally more likely to flee to shelter rather than ‘stop and look’; this presumably reflects their greater vulnerability to predators and, thus, greater fear (Putman et al. 2015).
We believe our findings of high individual-level correlation among multiple stages of the antipredator response are consistent with our general conceptual framework, contributing new insights and opportunities for the study of animal personalities. Future research should continue to investigate these correlations throughout the overall antipredator response in other species and the trade-offs animals face across all stages of prey’s response to risk and not just their initial flight response. Doing so can also provide more insights on the impacts of human activity on wildlife behavior, predator-prey dynamics and ecosystem function.