Evidence that multiple anthropogenic stressors cumulatively affect foraging and vigilance in an urban-living bird

Climate change and anthropogenic noise are two of the most widespread human-induced stressors affecting wildlife populations globally. However, the effects of these stressors are rarely investigated together, despite the fact that they often co-occur, particularly in urbanized areas and can have a multitude of adverse effects on species. Here, we compared the effects of heat stress and anthropogenic noise, both when presented alone and together, on the behaviour of wild Western Australian magpies, Gymnorhina tibicen dorsalis . Birds were presented with a playback of background noise and a playback of anthropogenic (plane) noise under both heat stress and nonheat stress conditions. Consistent with previous studies, we found both heat stress and anthropogenic noise reduced foraging and increased vigilance behaviours of magpies. Importantly, exposure to these two stressors simultaneously led to a greater change in behaviour, revealing that the simultaneous occurrence of two anthropogenic stressors had a more marked effect on behaviour compared to either stressor alone. This study provides some of the ﬁ rst evidence of the additive impact of heat stress and anthropogenic noise on the behaviour of a wild urban-dwelling bird species. Such ﬁ ndings highlight the importance of considering multiple stressors when looking at the effects of human-induced rapid environmental change on animals, since anthropogenic stressors rarely occur in isolation.

Climate change and anthropogenic noise are considered to be two of the most widespread and detrimental human-induced stressors affecting wildlife populations globally (Foden et al., 2013;Kunc & Schmidt, 2019;Parmesan & Yohe, 2003;Shannon et al., 2016;Urban, 2015).The rate at which global temperatures have increased has risen from an average of 0.08 C per decade from 1880 to 1980 to an average of 0.18 C per decade since 1981 (NOAA National Centers for Environmental Information, 2023).Similarly, anthropogenic noise is increasing faster than ever before due to rapidly expanding human populations and resultant urbanization (Barber et al., 2010;Shannon et al., 2016;World Health Organization. Regional Office for Europe, 2011, p. 106).Despite these two human-induced stressors often co-occurring, and the fact that both are increasing at an unprecedented rate, the combined impact of these stressors on wild animals has not yet been investigated simultaneously (Sumasgutner et al., 2023;Verrelli et al., 2022).Recent reviews have recognized this research gap and posit that studies investigating multiple simultaneous stressors are vital for a comprehensive understanding of the effects of human-induced rapid environmental change on wildlife populations, and thus for the development of appropriate management plans for wildlife in urban areas (Lacy et al., 2017;Lopez et al., 2023;Orr et al., 2020;Sumasgutner et al., 2023;Verrelli et al., 2022).
Heat stress induced by rising temperatures often leads to behavioural changes, especially in avian species that may be more susceptible to the adverse effects of increased temperatures due to their small body size and diurnal lifestyle (du Plessis et al., 2012;Edwards et al., 2015;McKechnie & Wolf, 2010).As temperatures rise, heat dissipation behaviours (such as panting and wing spreading) often increase to offload heat and prevent hyperthermia (du Plessis et al., 2012;McKechnie & Wolf, 2010).Such increases in heat dissipation often occur concurrently with decreases in other important behaviours, such as foraging (Conradie et al., 2020;du Plessis et al., 2012;Edwards et al., 2015).Foraging e heat dissipation trade-offs are common in birds, and in some species have been linked to reduced body condition and offspring provisioning (Conradie et al., 2020;du Plessis et al., 2012;Klett-Mingo et al., 2016;van de Ven et al., 2019).In southern pied babblers, Turdoides bicolor, for example, temperatures above 35.5 C cause an increase in time spent dissipating heat, and a decrease in foraging efficiency, which reduces the ability of birds to maintain body condition and provision offspring (du Plessis et al., 2012;Wiley & Ridley, 2016).In male southern yellow-billed hornbills, Tockus leucomelas, higher air temperatures are associated with increased panting, increased occupancy of cooler microsites and decreased foraging efficiency, leading to reduced body mass gain in hornbills as daily maximum temperature increases (van de Ven et al., 2019).
Similar behavioural trade-offs have been attributed to the presence of anthropogenic noise.Many species reduce their foraging effort and simultaneously increase the time they spend vigilant when exposed to anthropogenic noise (Eastcott et al., 2020;Evans et al., 2018;Kern & Radford, 2016;Quinn et al., 2006).Such a foraging e vigilance trade-off has been found in dwarf mongooses, Helogale parvula (Eastcott et al., 2020;Kern & Radford, 2016), zebra finches, Taeniopygia guttata (Evans et al., 2018) and chaffinches, Fringilla coelebs (Quinn et al., 2006).Increased vigilance during noise is hypothesized to occur either because species view noise as a threatening stimulus or as a mechanism to compensate for the reduced perception of acoustic signals in noise by increasing detection of visual stimuli (Cresswell, 2008;Kern & Radford, 2016;Ratnayake et al., 2021).Regardless of the mechanism, this alteration to behaviour caused by anthropogenic noise may affect the predation e starvation trade-off and could in turn reduce body condition and offspring provisioning, leading to potential negative effects on population viability (Cresswell, 2008;Kern & Radford, 2016;Lacy et al., 2017;Quinn et al., 2006).
While previous research has found high temperatures and anthropogenic noise significantly alter the behaviour of avian species, the simultaneous effect of these combined stressors on behaviour has not yet been investigated (Sumasgutner et al., 2023;Verrelli et al., 2022).The aim of this study was therefore to investigate how the simultaneous occurrence of high temperatures and anthropogenic noise affect foraging and vigilance in wild animals.Western Australian magpies, Gymnorhina tibicen dorsalis, a passerine well known for its ability to persist in urban environments, were used as the study species for this experiment.Previous work on this subspecies of Australian magpie found that temperatures above 27 C resulted in significant reductions in foraging behaviour and increases in heat dissipation (Edwards et al., 2015).Similarly, a trade-off between foraging and vigilance has also been identified in this species as a result of anthropogenic noise, with birds increasing their vigilance and decreasing foraging when anthropogenic noise exceeded 50 dB (Blackburn et al., 2023).Since magpies are readily exposed to high temperatures and anthropogenic noise in the urban areas they inhabit, and have already been shown to be negatively affected by both of these stressors when they occur in isolation (Blackburn et al., 2022(Blackburn et al., , 2023;;Edwards et al., 2015), they are an ideal study species for investigating the effects of the simultaneous occurrence of these stressors.In this study, we used playback experiments to investigate how a combination of heat stress and anthropogenic noise conditions affect the foraging and vigilance behaviour of wild Western Australian magpies.We predicted that both heat stress and anthropogenic noise would increase the vigilance and reduce the foraging behaviour of magpies, and that the simultaneous occurrence of these stressors would have a greater effect on behaviour than when either stressor occurred alone.

Study Species and Site
The Western Australian magpie is a medium-sized passerine (250e400 g) found in urban parklands in the southwest of Western Australia (Johnstone, 2004).These highly vocal birds live in stable cooperative groups of 2e12 individuals and defend year-round territories (Pike et al., 2019).This subspecies is sexually dichromatic, allowing sex to be discerned visually after individuals achieve full adult plumage at approximately 3 years of age (Dutour & Ridley, 2020;Edwards et al., 2015).
The study population consisted of nine magpie groups located in the Perth suburbs of Guildford and Crawley, WA.All groups occur in residential areas and live in parks and/or bushland within this urban matrix (Edwards et al., 2015).This population was founded in 1995, and has been studied consistently since 2013, and focal individuals are thus habituated to the presence of humans (Ashton et al., 2018).Most individuals in the study population are ringed (magpies are captured and ringed for individual identification using a walk-in trap) and therefore individually identifiable.Unringed individuals are identifiable due to distinctive plumage or scarring on their beaks.As this population has been studied for many years, we have detailed life history and ringing data for individuals, including information regarding what year birds hatched.This allows us to calculate the age for most birds tested.For individuals whose age was unknown (i.e.birds that emigrated into the population as adults), we assigned a minimum age of 3 years to that bird when it was first noted in our long-term life history database as an adult in the study population because juveniles only attain adult plumage at 3 years posthatching.All study groups live in urban areas, and experience daily anthropogenic noise from sources such as transport networks, construction and machinery (Blackburn et al., 2023) (see Table A1 in Appendix for details on noise levels in each territory).During the austral summer of 2022/2023, temperatures in this area were 1.4 degrees higher than average, with a mean maximum temperature of 32.4 C (Australian Government Bureau of Meteorology, 2023b).Mean annual temperatures in this area are predicted to rise by 0.5e1.2C by 2030 (Department of Primary Industries & Regional Development, 2021), an increase that could be detrimental to magpies; a species that is already threatened by rising temperatures (Blackburn et al., 2022;Edwards et al., 2015).

Playback Track Preparation
Plane recordings used for this experiment were collected in Guildford, WA (À31.9016810 S, 115.971540 0 E), from a location directly under the flight path of planes leaving Perth Airport, and within the territory of one of our study groups.All magpie groups located in Guildford (seven of the nine groups used in this study) are located within 4 km of Perth airport and are therefore exposed to acute levels (approximately 60e90 dB) of airplane noise daily.Background noise was collected from the same location as plane noise, but when there were no anthropogenic noise events (e.g.plane noise, traffic noise) occurring.Audio was recorded using a RODE NTG-2 directional microphone set within a Blimp suspension windshield system and attached to a Roland R-07 wave/MP3 recorder at a sampling rate of 44.1 kHz.A DIGITECH Micro Sound Pressure Level Meter was used to measure the amplitude of both plane and background noise from the location it was recorded at.The amplitude of plane noise ranged between 80 and 90 dB, and background noise varied from 45 to 50 dB.
Playback tracks were prepared in Audacity version 3.0.2(https://www.audacityteam.org/).Playback tracks consisted of 5 s of silence, followed by 35 s of either background noise or plane noise.Each playback presentation consisted of one 40 s silence and noise cycle.A duration of 35 s was chosen as this is the typical duration of noise from a plane passing directly overhead (G.Blackburn, personal observation).Playbacks were normalized in Audacity to ensure they were broadcast at the same amplitude they would normally be heard at (80e90 dB for plane noise and 45e50 dB for background noise).

Playback Presentations
Playback presentations were conducted using a paired design, whereby each of the two playback treatments (plane noise and background noise) were presented once while focal birds were exhibiting signs of heat stress (heat stress condition; characterized by birds displaying heat stress behaviours such as panting or wing spreading for 25% of the time during playback presentation; Blackburn et al., 2022), and once while birds were not exhibiting signs of heat stress (nonheat stress condition), giving four treatments: background noise e heat stress (BH), background noise e non heat stress (BN), plane noise e heat stress (PH) and plane noise e non heat stress (PN).The order of these treatments was randomized, and all individuals (N ¼ 29 birds, 116 playbacks) received all four playback treatments.These 29 individuals consisted of 18 female and 11 male birds, six unringed and 23 ringed, and their group sizes ranged from three to 10 adults.All treatments for an individual were completed within 3 weeks of each other to control for any differences due to seasonality.Treatments were presented to birds a minimum of 20 min apart, to minimize habituation to playbacks (Silvestri et al., 2019).
Playbacks were presented to individuals only when they were foraging, nonvigilant and in social isolation (to minimize the potentially confounding effects of other group members reacting to the playback, all other magpies were >10 m away; Ashton et al., 2018).Playbacks were presented only when birds were on the ground in a shaded area, to control for any additional heat effects of foraging in the sun, and near the centre of their territory so that birds were not vigilant to extragroup conspecifics in neighbouring territories.A UE BOOM2 speaker was connected via an aux cord to a Fiio M6 portable music player and was placed 10 m from the focal bird.Prior to the beginning of playback, a RS Pro RS42 digital thermometer was used to measure air temperature in the shade near the focal bird.In addition, total adult group size and the number of group members within 40 m of the focal bird were recorded.Humidity data were taken from the Bureau of Meteorology records from the nearest weather station, 4 km away (Australian Government Bureau of Meteorology, 2023a).Playbacks were only initiated after 1 min without magpie vocalizations or anthropogenic noise exceeding 50 dB in amplitude, as measured with a DIGITECH Micro Sound Pressure Level Meter.The behaviour of focal birds was recorded using a Panasonic HC-V180 video recorder from 60 s prior to the start of playback, throughout playback presentation and for 60 s postplayback.
The video recordings of the playback experiment were assessed blind by a video analyst for the following behaviours: the time birds spent vigilant in the 60 s preplayback (birds were recorded as vigilant when the head was oriented up or they were scanning their environment; Beauchamp, 2015), whether birds fled in the 60 s after playback had started (postplayback), the time birds spent vigilant in the 60 s postplayback and the time birds spent foraging in the 60 s postplayback.

Bioacoustic Recorders
Song Meter Mini recording units (Wildlife Acoustics Inc., https:// www.wildlifeacoustics.com/)were deployed in each territory to determine the soundscape to which magpies in each territory were exposed.This allowed us to determine whether the birds' normal soundscape affects their response to anthropogenic noise playback.Recording units were deployed in each territory for 2 weekdays and 2 weekend-days and were set to collect sound data continuously with a sampling rate of 44.1 kHz and a gain of 12 dB.Recording units were mounted in trees 3e5 m high, with microphones positioned parallel to the ground.Recordings from these units were then analysed using Kaleidoscope Pro software 5.4.9 (Wildlife Acoustics Inc.).Within each 60 min sampling period, the minimum, maximum and mean sound pressure levels (SPL) and the cumulative sound energy level (SEL) were extracted.These measures were then averaged across all 60 min samples in each territory (N ¼ 96 per territory), to get an average hourly minimum SPL, maximum SPL, mean SPL and cumulative SEL for each territory.

Statistical Analysis
Analysis of factors affecting magpie behaviour following playback was conducted via model selection using generalized linear mixed models (GLMMs) in R studio version 4.2.0 (RStudio Team, 2022).GLMMs were run with the glmmTMB package (Brooks et al., 2017).Group and bird identity were included as random terms in all models.
For analysis of time spent vigilant in the 60 s preplayback, and time spent vigilant and foraging in the 60 s postplayback, the response variable was proportional, with time (s) spent vigilant as the response variable and total time of observation (60 s) as the binomial dominator.An investigation of the factors affecting time spent vigilant and time spent foraging postplayback was conducted on a subset of individuals (N ¼ 100 playbacks across 25 birds) that remained in sight for the full 60 s postplayback (i.e. did not flee out of sight following playback).These 25 birds consisted of 10 male and 15 female birds, six unringed and 19 ringed.Group sizes for these birds ranged from three to 10 individuals.For the analysis of factors affecting whether individuals fled or retreated from the speaker in response to playback, a binomial GLMM with a 0, 1 response variable was used (where 0 ¼ did not flee or retreat from the speaker and 1 ¼ did flee to cover or did retreat from the speaker).All models included treatment condition (BH, BN, PH, PN), adult group size, track order, adults present within 40 m, air temperature, humidity, time of day (hour), age, sex and average hourly minimum SPL, maximum SPL, mean SPL and cumulative SEL of the focal individual's territory as predictors.
Model selection using Akaike information criterion values corrected for small sample size (AICc) was conducted for each response variable to determine which candidate models best explained variation in the data.Terms with 95% confidence intervals that intersected zero when tested alone were not considered to be good predictors of data variation and were excluded from further additive models.Models containing interactions between terms were tested based on their suitability as plausible biological hypotheses (Burnham & Anderson, 2002; i.e. group size may alter how birds respond to anthropogenic noise and heat stress, so group size and treatment interactions were included in all analyses).Where multiple terms were highly correlated (e.g.air temperature and time of day), only the term with the lowest AICc when tested alone was used in further additive models; however, all terms were retained for use in interaction models (Harrison et al., 2018).Models were compared to a null model containing only the intercept and random terms.A top model set was created including all models within 2 AICc of the best model.In cases where multiple models with the same terms were in the top model set, the simplest model was chosen as per Harrison et al. (2018).
Post hoc comparisons using the function emtrends within the R package emmeans (Lenth, 2019) were used to obtain contrasts between treatment condition and continuous predictors in cases where significant interactions were found between treatment condition and other continuous explanatory factors.P values adjusted for multiple comparisons via the Tukey method were also calculated using the emtrends package.

Ethical Note
This study was approved by the University of Western Australia Animal Ethics Committee (approval number 2021/ET000272).Subjects (N ¼ 29 individuals) for this study were wild habituated magpies in urban areas and are used to human presence as well as the presence of anthropogenic noise such as that used in this experiment.Subjects were free to fly away at any point during experimentation.All birds observed in experiments returned to normal behaviours (foraging, perched or preening and nonvigilant) within 5 min after playbacks.Fledgling and adult magpies in this population were captured for ringing using a walk-in trap baited with mozzarella cheese under ABBBS Banding Authority Number 3558.Individuals were only captured for ringing if it was not possible to distinguish them from other group members by distinct plumage anomalies.

Time Spent Vigilant in the 60 s Preplayback
Individuals spent significantly more time vigilant in the 60 s preplayback under heat stress (BH and PH) conditions compared to under nonheat stress (BN and PN) conditions (Tables A2 and A3, Fig. A1).Those from smaller groups also spent more time vigilant preplayback (Table A2).The interaction between adult group size and treatment condition was also significant, with post hoc comparisons revealing that individuals from larger groups spent significantly less time vigilant under nonheat stress (BN and PN) conditions, but not under heat stress (BH or PH) conditions (Table A4, Fig. A2, see Appendix Table A5 for full model set).

Time Spent Vigilant in the 60 s Postplayback
Individuals spent significantly less time vigilant in the 60 s postplayback under background noise, nonheat stress conditions compared to any other treatment conditions (Table 1, Fig. 1, Table A6).Post hoc comparisons reveal that individuals spent significantly more time vigilant when exposed to both stressors (plane noise þ heat stress condition) compared to any other treatment condition (Table A6, Fig. 1).
There was a group size effect on vigilance, with individuals from smaller groups spending more time vigilant in the 60 s postplayback (Table 1) than in the preplayback condition.The interaction between adult group size and treatment condition was also significant, with individuals from smaller groups spending more time vigilant under BN, PN and PH conditions, but less time vigilant under the BH condition (Table A7, Fig. 2).

Time Spent Foraging in the 60 s Postplayback
Magpies spent more time foraging in the absence of anthropogenic stressors (treatment ¼ background noise, nonheat stress conditions) compared to under any other conditions (Table 2, Fig. 3).Post hoc comparisons reveal that birds spent less time foraging while experiencing both plane noise and heat stress compared to under any other treatment conditions (Table A9, Fig. 3).A6).
G. Blackburn et al. / Animal Behaviour 211 (2024) 1e12 Adult group size also significantly affected the time spent foraging, with birds from larger groups spending more time foraging (Table 2).The interaction between playback treatment and adult group size was significant (Table 2), whereby birds from larger groups spent more time foraging under BN, PN and PH conditions, but less time foraging under the BH condition (Table A10, Fig. 4).

Flee Response
The likelihood of birds fleeing to cover in the 60 s postplayback was significantly affected by treatment condition (Table 3).Birds were significantly more likely to flee under plane noise þ heat stress conditions compared to under background noise, nonheat stress conditions (Table 3, Fig. 5).

DISCUSSION
In this study, we found that both heat stress and anthropogenic noise affected the short-term vigilance and foraging behaviour of Western Australian magpies.Magpies spent more time vigilant and less time foraging when they were experiencing heat stress and/or anthropogenic noise compared to when experiencing neither stressor.These stressors had an additive effect on magpie behaviour, whereby birds exhibited a greater change in behaviour when both stressors co-occurred compared to when either of them occurred in isolation.Our study is the first, to our knowledge, to investigate the simultaneous effects of these two pervasive and rapidly increasing stressors on a wild population.A9).

Behavioural Response to Heat Stress
Magpies significantly increased their vigilance and decreased their foraging behaviour under heat stress conditions compared to background noise, nonheat stress conditions.This decrease in time spent foraging supports previous work on this species, where a decrease in foraging effort as temperature increased was observed (Edwards et al., 2015).Edwards et al. (2015) hypothesized that reduced foraging behaviour resulted from a trade-off with increased heat dissipation behaviours as temperatures rose; however, findings from our study suggest that reduced foraging may also be in part due to a trade-off with vigilance.Individuals spent significantly more time vigilant both before and after playback under heat stress conditions, compared to nonheat stress conditions.This increase in vigilance during heat stress may occur for two main reasons.Firstly, increased vigilance during heat stress conditions may be related to the fact that magpies are ground foragers (Edwards et al., 2015;O'Leary & Jones, 2002), and therefore their ability to concurrently forage and dissipate heat (particularly via panting) is limited.As vigilance behaviours may be less energetically costly than foraging at these high temperatures when birds are dissipating heat, individuals may instead invest in vigilance over foraging.Alternatively, corticosteroids, stress hormones related to the antipredator response across taxa (Cockrem & Silverin, 2002;Narayan et al., 2013;Thaker et al., 2009), can be significantly heightened in individuals experiencing heat stress (Marchini et al., 2016;Moagi et al., 2021).The heightened vigilance seen in our study may arise from increased corticosteroid levels in heat-stressed birds causing an increase in vigilance behaviour via a similar mechanism to that involved in the escape response (Cockrem & Silverin, 2002).

Behavioural Response to Anthropogenic Noise
Anthropogenic noise playback also significantly affected magpie behaviour.Individuals spent less time foraging and more time vigilant under anthropogenic noise playback conditions compared to background noise nonheat stress conditions.Similar effects of anthropogenic noise on behaviour have been documented across taxa (Evans et al., 2018;Kern & Radford, 2016;Merrall & Evans, 2020;Quinn et al., 2006;Rabin et al., 2006).Heightened vigilance in response to anthropogenic noise may arise because magpies view such noise as a threatening stimulus (Ratnayake et al., 2021).Alternatively, since anthropogenic noise has been shown to mask important auditory cues and impede acoustic communication in several species, (Grade & Sieving, 2016;Kleist et al., 2016;Templeton et al., 2016;Zhou et al., 2019), increased vigilance may arise as a way of maximizing the visual information birds can take in from their environment when the perception of auditory cues may be impeded.Future studies investigating the response of magpies to important auditory cues (e.g.alarm calls) during both anthropogenic noise that overlaps and anthropogenic noise that does not overlap in frequency with these cues (sensu Zhou et al., 2019) are necessary to disentangle the mechanisms behind increased vigilance during anthropogenic noise in this species.

Behavioural Response to Simultaneous Heat Stress and Anthropogenic Noise
Under the multiple stressors of plane noise and heat stress, individuals spent significantly more time vigilant and less time foraging compared to under any other condition, including conditions in which only one of these stressors was occurring (background noise, heat stress conditions or plane noise, non-heat stress conditions).Magpies were also significantly more likely to flee to cover under the plane noise and heat stress condition.Heat stress and anthropogenic noise therefore have an additive effect on magpie behaviour, leading to a greater reduction in foraging and a greater increase in vigilance behaviour compared to the effect of either stressor in isolation.Such behavioural changes under these stressors have the potential to lead to declines in both body condition and offspring provisioning in magpies, an effect already observed in a number of bird species due to heat stress conditions alone (Clark, 1987;du Plessis et al., 2012;van de Ven et al., 2019;Wiley & Ridley, 2016).Declines in body condition and offspring provisioning could in turn affect reproductive success and therefore viability of populations exposed to multiple anthropogenic stressors (Ridley et al., 2021).Investigating the simultaneous effects of heat stress and anthropogenic noise on body condition, offspring provisioning and reproductive success of magpies is therefore an important next step, particularly as temperatures increase and urbanization continues to expand into the future.

Group Size Effect
Finally, we also identified an effect of adult group size on vigilance and foraging behaviour.Birds from larger groups spent less time vigilant and more time foraging compared to those from smaller groups; however, this effect was only present under certain  3).
treatment conditions.While the decreased vigilance and increased foraging seen in larger groups is consistent with the many-eyes hypothesis, which states that one of the benefits of group living is having more individuals vigilant and scanning the environment (Beauchamp, 2008;Caraco et al., 1980;Wang et al., 2021), the fact that this group size effect was the opposite under background noise and heat stress conditions suggests that this effect may be altered by heat stress.However, as this group size relationship was present under plane noise and heat stress conditions, it may depend on not only heat stress, but also the occurrence or absence of other anthropogenic stressors.This is supported by the fact that adult group size was only significantly positively related to foraging in the absence of heat stress and loud anthropogenic noise, and not when either or both anthropogenic stressors were present.A possible reason for this pattern is that individuals experiencing environmental stressors revert to reliance on their own vigilance and information from the environment, rather than relying on the vigilance of group members.In social species such as magpies, one proposed benefit of group living is a potential buffering effect against harsh environments or rapid environmental change (Blumstein et al., 2023;Komdeur & Ma, 2021).However, our results suggest that living in a larger group may not be beneficial for magpies facing multiple anthropogenic stressors, at least in terms of their behavioural responses to these stressors.

Conclusion
We have documented an additive effect of anthropogenic noise and heat stress on the short-term behaviour of an urban-living wild bird, the Western Australian magpie.Individuals foraged significantly less and were vigilant significantly more when experiencing both anthropogenic noise and heat stress simultaneously, compared to when they were only experiencing a single anthropogenic stressor or neither stressor.Our study is the first to look at the combined effects of two commonly co-occurring anthropogenic stressors on a wild animal population and highlights the importance of considering the simultaneous effects of multiple stressors on wildlife, particularly as human-induced rapid environmental change accelerates into the future.A3).A2.N ¼ 116 playbacks across 29 birds.

Figure 1 .
Figure 1.Proportion of time birds spent vigilant in the 60 s postplayback in relation to treatment condition.BN ¼ background noise playback, nonheat stress conditions; BH ¼ background noise playback, heat stress conditions; PN ¼ plane noise playback, nonheat stress conditions; PH ¼ plane noise playback, heat stress conditions.The box plots show the median and 25th and 75th percentiles; the whiskers indicate the values within 1.5 times the interquartile range and the circles are outliers.N ¼ 100 playbacks across 25 birds.*P < 0.05 (post hoc analysis, TableA6).

Figure 2 .
Figure 2. Proportion of time birds spent vigilant in the 60 s postplayback in relation to the interaction between adult group size and treatment condition.BN ¼ background noise playback, nonheat stress conditions; BH ¼ background noise playback, heat stress conditions; PN ¼ plane noise playback, nonheat stress conditions; PH ¼ plane noise playback, heat stress conditions.Points represent raw data; fitted lines and 95% confidence intervals are generated from the output of the top model presented in Table 1.N ¼ 100 playbacks across 25 birds.

Figure 3 .
Figure 3. Proportion of time birds spent foraging in the 60 s postplayback in relation to treatment condition.BN ¼ background noise playback, nonheat stress conditions; BH ¼ background noise playback, heat stress conditions; PN ¼ plane noise playback, nonheat stress conditions; PH ¼ plane noise playback, heat stress conditions.The box plots show the median and 25th and 75th percentiles; the whiskers indicate the values within 1.5 times the interquartile range and the circles are outliers.N ¼ 100 playbacks across 25 birds.*P < 0.05 (post hoc analysis, TableA9).

Figure 4 .
Figure 4. Proportion of time birds spent foraging in the 60 s postplayback in relation to the interaction between adult group size and treatment condition.BN ¼ background noise playback, nonheat stress conditions; BH ¼ background noise playback, heat stress conditions; PN ¼ plane noise playback, nonheat stress conditions; PH ¼ plane noise playback, heat stress conditions.Points represent raw data; fitted lines and 95% confidence intervals are generated from the output of the top model presented in Table 2. N ¼ 100 playbacks across 25 birds.

Figure A1 .
Figure A1.Proportion of time birds spent vigilant in the 60 s preplayback in relation to treatment condition.BN ¼ background noise playback, nonheat stress conditions; BH ¼ background noise playback, heat stress conditions; PN ¼ plane noise playback, nonheat stress conditions; PH ¼ plane noise playback, heat stress conditions.The box plots show the median and 25th and 75th percentiles; the whiskers indicate the values within 1.5 times the interquartile range and the circles are outliers.N ¼ 116 playbacks across 29 birds.*P < 0.05 (post hoc analysis, TableA3).

Figure A2 .
Figure A2.Proportion of time birds spent vigilant in the 60 s preplayback in relation to the interaction between adult group size and treatment condition.BN ¼ background noise playback, nonheat stress conditions; BH ¼ background noise playback, heat stress conditions; PN ¼ plane noise playback, nonheat stress conditions; PH ¼ plane noise playback, heat stress conditions.Points represent raw data; fitted lines and 95% confidence intervals are generated from the output of the top model presented in TableA2.N ¼ 116 playbacks across 29 birds.

Table 1
Top model set of candidate terms affecting time spent vigilant in the 60 s postplayback All models included group and bird ID as random terms.The top model set includes terms within 2 AICc of the best model.Coefficient estimates ± SE and 95% confidence intervals (CI) are given below the top model set.N ¼ 100 playbacks to 25 birds.BN ¼ background noise playback, nonheat stress conditions; BH ¼ background noise playback, heat stress conditions; PN ¼ plane noise playback, nonheat stress conditions; PH ¼ plane noise playback, heat stress conditions.For the full model set of candidate terms tested, refer to TableA8.

Table 2
Top model set of candidate terms affecting time spent foraging in the 60 s postplayback À0.41, À0.09 PH * Adult group size À0.29 0.09 À0.46, À0.11All models included group and bird ID as random terms.The top model set includes terms within 2 AICc of the best model.Coefficient estimates ± SE and 95% confidence intervals (CI) are given below the top model set.N ¼ 100 playbacks across 25 birds.BN ¼ background noise playback, nonheat stress conditions; BH ¼ background noise playback, heat stress conditions; PN ¼ plane noise playback, nonheat stress conditions; PH ¼ plane noise playback, heat stress conditions.For the full model set of candidate terms tested, refer to Table A11.

Table 3
Top model set of candidate terms affecting whether individuals fled to cover in the 60 s postplayback AICc of the best model.Coefficient estimates ± SE and 95% confidence intervals (CI) are given below the top model set.N ¼ 116 playbacks across 29 birds.BN ¼ background noise playback, nonheat stress conditions; BH ¼ background noise playback, heat stress conditions; PN ¼ plane noise playback, nonheat stress conditions; PH ¼ plane noise playback, heat stress conditions.For the full model set of candidate terms tested, refer to Table A12.
Figure 5. Probability of individuals fleeing to cover in the 60 s postplayback in relation to treatment condition.Error bars represent SEs.N ¼ 116 playbacks on 29 magpies.BN ¼ background noise playback, nonheat stress conditions; BH ¼ background noise playback, heat stress conditions; PN ¼ plane noise playback, nonheat stress conditions; PH ¼ plane noise playback, heat stress conditions.*P < 0.05 (Table

Table A1
Minimum, maximum, mean and cumulative sound pressure level measurements for each territorial group SPL ¼ sound pressure level; SEL ¼ sound energy level.Measurements were obtained from Wildlife Acoustic Sound Meter mini recording units.AICc of the best model.Coefficient estimates ± SE and 95% confidence intervals (CI) are given below the top model set.N ¼ 116 playbacks across 29 birds.BN ¼ background noise playback, nonheat stress conditions; BH ¼ background noise playback, heat stress conditions; PN ¼ plane noise playback, nonheat stress conditions; PH ¼ plane noise playback, heat stress conditions.Estimate represents slope of the levels of interaction between treatment condition and adult group size.N ¼ 116 playbacks across 29 birds.BN ¼ background noise playback, nonheat stress conditions; BH ¼ background noise playback, heat stress conditions; PN ¼ plane noise playback, nonheat stress conditions; PH ¼ plane noise playback, heat stress conditions.DAICc are provided for each candidate model.Only the model within 2 AICc of the top model and with predictors whose 95% confidence intervals did not intersect zero was included in the top model set and is highlighted in bold.N ¼ 116 playbacks on 29 birds.

Table A6
Post hoc analyses of time spent vigilant in the 60 s postplayback Estimate represents slope of the interaction between treatment condition and adult group size.N ¼ 100 playbacks across 25 birds.BN ¼ background noise playback, nonheat stress conditions; BH ¼ background noise playback, heat stress conditions; PN ¼ plane noise playback, nonheat stress conditions; PH ¼ plane noise playback, heat stress conditions.DAICc are provided for each candidate model.Only the model within 2 AICc of the top model and with predictors whose 95% confidence intervals did not intersect zero were included in the top model set and is highlighted in bold.N ¼ 100 playbacks on 25 birds.

Table A9
Post hoc analyses of time spent foraging in the 60 s postplayback Estimate represents slope of the levels of interaction between treatment condition and adult group size.N ¼ 100 playbacks across 25 birds.BN ¼ background noise playback, nonheat stress conditions; BH ¼ background noise playback, heat stress conditions; PN ¼ plane noise playback, nonheat stress conditions; PH ¼ plane noise playback, heat stress conditions.soundpressure level; SEL ¼ sound energy level.All models included group and bird ID as random terms.Corrected Akaike information criterion (AICc) andDAICc are provided for each candidate model.Only the model within 2 AICc of the top model and with predictors whose 95% confidence intervals did not intersect zero was included in the top model set and is highlighted in bold.N ¼ 100 playbacks on 25 birds.Full model selection output for candidate terms affecting whether individuals fled to cover in the 60 s postplayback SPL ¼ sound pressure level; SEL ¼ sound energy level.All models included group and bird ID as random terms.Corrected Akaike information criterion (AICc) and