The process of vocal communication involves the transfer of information between individuals through an acoustic signal (Craig 1999; Alvar 2015), which, in the case of birds, is essential, since they are used for the recognition of their partner and conspecifics (Tubaro and Segura 1989; León et al. 2014; Isaksson 2018). Urban environments emerge as new and constantly expanding ecosystems, exposing wildlife to selective pressures that often differ drastically from those in which they have evolved (Lowry et al. 2013). In this context, within cities, there are gradients of degrees of urbanisation, some characterised by houses with semi-natural private spaces and high vegetation cover (e.g. gardens), while others have areas covered by buildings of various heights and surfaces highly impermeable to water (Marzluff 2001; MacGregor-Fors and Schondube 2011; Isaksson 2018; Bernat-Ponce et al. 2022). Vegetation is one of the main variables influencing bird diversity in urban ecosystems; trees, shrub and herbaceous cover positively influence bird abundance (Leveau and Leveau 2006; Benito et al. 2019), as these cover areas are associated with food availability, nesting sites, shelter and perching sites (Benito et al. 2019; da Silva et al. 2021). Other standard land covers in urban environments, such as pavement and buildings, also influence bird abundance in the city. In general, areas with a high construction density have a low abundance of native birds because habitat quality is considered poor for these species (Amaya-Espinel et al. 2019; Benito et al. 2019; Apfelbeck et al. 2025).

Anthropogenic disturbances vary within cities as green areas (i.e. land surface predominantly covered by vegetation), trees and buildings are not distributed homogeneously (Benito et al. 2019), generating environments that may hinder vocal communication in birds to varying degrees, particularly in species that depend on vocalisation in their intraspecific relationships, which can result in a series of effects on behaviour, distribution and population density (Calderón-Quiroga et al. 2020; Suarez-Romero 2021; Bernat-Ponce et al. 2022). Additionally, urban areas generate acoustic environments characterised by high levels of ambient noise (> 56 dB; Juaréz et al. (2020)), which can affect vocal communication and pose a challenge for species that depend on vocalisations in their relationships (Slabbekoorn and Peet 2003; Mendes et al. 2017; Juaréz et al. 2020). These high levels of anthropogenic noise can interfere with the detection of relevant biological sounds, such as songs and calls, which could make them more vulnerable to predation, alter their behaviour, in terms of finding breeding territories, defence mechanisms, foraging, communication between pairs and adult-offspring and compromise their chances of reproductive success (Krause 2008; Francis et al. 2011; Halfwerk et al. 2011; Ortega 2012; León et al. 2014; Templeton et al. 2016; Mendes et al. 2017; Suarez 2021; De Kort et al. 2024). Such impacts can ultimately affect population stability and community composition in urban environments (León et al. 2014; Carral-Murrieta et al. 2020; Suarez 2021).

Morton (1975) proposed that habitat structure has influenced the evolution of the acoustic characteristics of songs, a hypothesis known as the Acoustic Adaptation Hypothesis (AAH). This hypothesis holds that, since sound propagation varies according to habitat type, natural selection favours those sounds that maximise sound dispersion from the source. The AAH has been evaluated in urban environments, where multiple species have been reported to have the flexibility to adjust the spectrotemporal characteristics of their vocalisations to avoid being masked by the structure of the urban environment (e.g. vegetation structure, reflective surfaces) or by human-associated factors such as anthropogenic noise (Halfwerk and Slabbekoorn 2009; Dowling et al. 2012; Redondo et al. 2013; Gough et al. 2014; Juaréz et al. 2020; Hao et al. 2024; Díaz-Lezama et al. 2025).

In this context, to communicate effectively in urban environments, some birds may: (1) switch to higher-frequency song types (e.g. Parus major, Halfwerk and Slabbekoorn (2009)); (2) sing during times of lower anthropogenic disturbance (e.g. Erithacus rubecula, Fuller et al. (2007); Isaksson (2018)) or (3) omit low-frequency elements, which may not be well transmitted in noisy environments (e.g. Haemorhous mexicanus, Bermúdez-Caumatzin et al. (2009, 2011)). Species described as not only surviving in urban environments, but doing so successfully, occurring in large numbers with high reproductive values, are known as “adapters” or “exploiters” (McKinney 2006; Humphrey et al. 2023).

The Austral Thrush (Turdus falcklandii) is a widespread species in Chile (Jaramillo et al. 2005), inhabiting a wide range of habitats and is abundant in urban environments (Medrano et al. 2018). This positions it as a good model species to evaluate the effect of urban disturbances on its vocal communication. In this study, we evaluate the influence of four urban variables relevant to vocal communication on the bioacoustic parameters of the territorial song of T. falcklandii. Considering that the AAH postulates that the spectral characteristics and vocal repertoire of its vocalisations are influenced by the environment, we expect to find variation in these parameters associated with greater urbanisation.

We analysed 222 songs from 75 individual male Austral Thrushes (mean = 2.96 songs per individual, range: 2-3). Ambient noise levels across the 75 sampling sites ranged from 40.4 to 79.9 dBA (mean +/- SE: 56.3 +/- 1.2 dBA), encompassing both relatively quiet residential areas and busier urban zones.

The number of elements in the territorial song varied between individuals with a range of 3 to 21 elements (mean = 7.95, SD = 3.51; Fig. 2). The mean duration of the song was 4.35 s (SD = 2.12), with a maximum frequency of 10,821 Hz (SD = 4,230), minimum frequency of 1,613 Hz (SD = 241) and peak frequency of 3,158 Hz (SD = 299). Element delivery rate averaged 1.95 elements per second (SD = 0.45, range: 1.15–3.45). Song duration and number of elements were highly correlated (r = 0.87), while other bioacoustic variables showed weaker intercorrelations (|r| < 0.45; Suppl. material 1: fig. S3).

Random effects for individual identity accounted for substantial variance in most response variables, with intraclass correlations (ICC) ranging from 0.39 (peak frequency) to 0.76 (maximum frequency), confirming the importance of accounting for repeated measures within individuals (Table 1).

From the GLMMs, we found that noise level had a significant positive effect on the rate of element delivery (elements per second): birds in noisier environments sang with faster element rates (model-averaged beta = 0.134 elements/s per SD increase in noise, 95% CI: 0.034–0.234, P = 0.009; standardised effect size = 0.25; Fig. 3, Table 1). Ground cover type also significantly influenced element rate, with birds in bare soil areas (ground-cover type 3) singing approximately 0.46 elements/s faster than those in grass-dominated areas (P = 0.003).

In contrast, noise, vegetation cover, building height and ground-cover type had no significant effects on song duration, peak frequency, minimum frequency, maximum frequency or the total number of elements per song (all P > 0.05; Table 1). The relationship between noise and minimum song frequency showed a non-significant positive trend (model-averaged beta = 47.6 Hz per SD increase in noise, 95% CI: -17.8 to 113.0, P = 0.15; standardised effect = 0.19). The interaction between noise and vegetation cover was not significant for minimum frequency (chi-square = 0.12, P = 0.73).

Model-averaged effect estimates with 95% confidence intervals for all response variables are presented in Table 1. For spectral parameters, all confidence intervals for urban disturbance variables included zero and standardised effect sizes were uniformly small (|d| < 0.20).

Our results reveal a nuanced response of Austral Thrush territorial songs to urban conditions: while spectral parameters remained unaffected by urbanisation variables, the temporal structure of song delivery was significantly influenced by ambient noise levels. Birds in noisier environments sang with faster element delivery rates, representing a behavioural adjustment to urban acoustic conditions that has not been previously documented in this species.

The significant effect of noise on element delivery suggests that Austral Thrushes modify the temporal pacing of their songs in response to urban noise, even while maintaining consistent spectral characteristics. This temporal compression of song structure may represent an alternative strategy to the frequency shifts commonly reported in other species. Similar adjustments in temporal pacing have been reported in other taxa; for example, some species increase song rate or reduce inter-song intervals in noisy conditions (Brumm and Slabbekoorn 2005). However, findings on element delivery rate are mixed. Slabbekoorn and den Boer-Visser (2006) found that urban great tits (Parus major) sang songs with fewer notes than forest birds, potentially an adaptation to reduce reverberation effects. In contrast, Ríos-Chelén et al. (2015) found no significant relationship between noise levels and trill rate in red-winged blackbirds (Agelaius phoeniceus), suggesting that temporal adjustments may be species-specific. Our finding that Austral Thrushes increase element delivery rate in noisier environments adds to this growing body of evidence that birds may employ diverse temporal strategies in response to urban noise.

The concurrent effect of ground-cover type on element rate suggests that multiple environmental factors may interact to shape temporal aspects of song delivery in urban Austral Thrushes. Birds singing in areas with bare soil delivered elements approximately 0.46 elements/s faster than those in grass-dominated areas. Notably, this effect was independent of ambient noise levels: bare soil areas were quieter on average (51.5 dBA) than grass-dominated areas (58.7 dBA), indicating that the ground-cover effect is not simply a proxy for noise exposure. Concrete and mixed ground cover showed non-significant positive trends in the same direction and the difference between bare soil and concrete effects was not statistically significant. The consistent positive direction across all non-vegetated substrates suggests that open or hard ground surfaces may generally favour faster element delivery, though only the bare soil effect was statistically detectable with our sample sizes (n = 30 songs for bare soil vs. n = 65 for concrete).

The Acoustic Adaptation Hypothesis predicts that animals in open habitats should produce signals with higher repetition rates compared to those in densely vegetated closed habitats, where reverberation and sound degradation are more severe (Morton 1975; Wiley and Richards 1982; Boncoraglio and Saino 2007). Ground surface composition can influence sound propagation through differential reflection and absorption: vegetated surfaces absorb acoustic energy while hard surfaces, such as bare soil or concrete, produce stronger reflections (Aylor 1972; Richards and Wiley 1980). However, flat reflective surfaces like concrete also generate reverberation that can blur rapidly delivered temporal elements (Bradbury and Vehrencamp 1998; Naguib 2003), potentially counteracting any tendency towards faster delivery rates. This may explain why concrete areas did not show a significant effect despite their expected acoustic reflectivity. Job et al. (2016) found that urban structure influenced spectral, but not temporal features of Chipping Sparrow (Spizella passerina) songs, noting that temporal features of trill songs are highly vulnerable to degradation through reverberation. In contrast, bare soil — which is reflective, but with irregular surface texture — may create acoustic conditions that favour faster element delivery without the reverberant blurring associated with flat concrete surfaces.

Alternatively, the ground-cover effect may reflect ecological or behavioural correlates rather than purely acoustic mechanisms. Bare soil areas in urban Concepción likely represent open, exposed microhabitats, such as garden patches, park edges or disturbed areas where birds may experience heightened vigilance or territorial pressure, prompting faster song delivery as a behavioural response. Further experimental work manipulating ground cover, while controlling for other habitat variables, would be needed to distinguish between acoustic and ecological explanations for this pattern.

In contrast to element delivery rate, spectral parameters of Austral Thrush songs showed no significant association with urban disturbance variables. This lack of spectral adjustment contrasts with patterns reported for other thrushes, such as the pale-breasted thrush (Turdus leucomelas) and Eurasian blackbird (Turdus merula), where individuals in urban environments have been found to sing at higher minimum frequencies (Mendes et al. 2011; Mendes et al. 2017). Similarly, in great tit (Parus major), urban populations produce shorter songs compared to rural counterparts (Slabbekoorn and den Boer-Visser 2006). Such vocal adjustments are thought to enhance signal transmission under the specific constraints of urban environments (Brown and Handford 2000; Brumm and Slabbekoorn 2005; Mendes et al. 2017). Shorter songs may reduce overlap with reverberant echoes from buildings and other hard surfaces, thereby improving signal clarity in acoustically complex urban habitats (Slabbekoorn 2013). When sound reflects off buildings and pavement, longer vocalisations are more susceptible to temporal smearing, where echoes from earlier portions of the signal interfere with later portions (Naguib 2003). By producing shorter songs, birds may minimise this self-masking effect and maintain signal integrity.

However, not all species respond to noise in the same way. For example, Peris and Pescador (2004) found that, in more than half of the passerine populations they studied near roads with varying noise levels, breeding densities did not differ significantly between areas with different noise levels. Interestingly, because noise propagates more effectively in open habitats (Reijnen et al. 1996), one might expect species in these environments to be more exposed to anthropogenic noise. However, species naturally adapted to open habitats typically produce vocalisations with acoustic features — such as higher frequencies, simpler structures and faster temporal patterns — that are already optimised for transmission in low-reverberation environments (Morton 1975; Boncoraglio and Saino 2007). These pre-existing adaptations may confer resilience to urban noise masking, as their signals are less dependent on the acoustic refugia provided by vegetation. This could explain the higher abundance of open-habitat passerines near noisier roads (Peris and Pescador 2004) - species with vocal repertoires pre-adapted to open environments may tolerate or even thrive in noisy roadside habitats where forest-adapted species with lower-frequency, more reverberant songs are acoustically disadvantaged.

These findings suggest that habitat preferences and ecological adaptations may mediate the degree to which noise impacts vocal communication – potentially helping to explain why T. falcklandii, a generalist species often found in open and semi-open habitats, does not exhibit spectral modifications in response to urban noise in our study. The Austral Thrush may already possess song characteristics that are robust to urban acoustic challenges or spectral shifts may not be an effective strategy for this species given its vocal structure.

Urban landscapes are often composed of dense exotic vegetation (Borgmann and Rodewald 2005), trees and buildings (Warren et al. 2006), all of which can hinder acoustic transmission (Dowling et al. 2012; Hao et al. 2024; Díaz-Lezama et al. 2025). However, in some species, these structural elements appear to have little impact on vocal behaviour (Dowling et al. 2012). For example, Melospiza melodia and Cardinalis cardinalis showed no significant variation in song attributes across differing urban habitat structures and noise (Dowling et al. 2012; Desirée and Rodewald 2015). Similarly, in our study, the territorial song of T. falcklandii was unaffected by variation in vegetation cover, built-up area, building height or soil structure – suggesting a comparable insensitivity to urban structural features. These urbanisation characteristics could have parallels with natural systems, such as streams, canyons and windy environments – where birds have long adapted their vocal strategies (Warren et al. 2006). However, concrete and glass canyons are more likely to reverberate sounds further (Warren et al. 2006), enhancing the propagation of urban background noise, while also possibly benefitting the transmission of acoustic signals, as the communication distance may be greater in open urban areas with impermeable surfaces that act as sound channels, thereby enhancing signal propagation (Gall et al. 2012). This dual effect raises the possibility that certain urban structures may inadvertently facilitate vocal communication under specific conditions.

The habitat structure and spatial arrangement of urban green spaces is a crucial factor influencing bird-song behaviour, jointly shaping their acoustic niche in urban environments. Vegetation can absorb, reflect and scatter sound, thereby dampening anthropogenic noise and creating acoustic refuges (Watts et al. 1999; Bormpoudakis et al. 2013; Darras et al. 2016; Job et al. 2016; Priyadarshani et al. 2018). Abundant vegetation not only provides suitable habitats and feeding grounds, but can also buffer anthropogenic noise, creating acoustic refugia where birds experience reduced masking of their vocalisations (Watts et al. 1999; Bormpoudakis et al. 2013). By attenuating traffic and industrial noise, vegetated areas may allow birds to communicate effectively at lower amplitudes or maintain larger active spaces for their signals (Hao et al. 2024). However, dense vegetation can also degrade signal transmission through absorption and scattering (Richards and Wiley 1980), suggesting a trade-off where intermediate vegetation density may optimise the balance between noise reduction and signal propagation.

Another potential strategy for coping with urban disturbances is behavioural adjustment in the timing of vocalisations. Some species, such as E. rubecula, shift their singing activity to night-time hours to reduce overlap with periods of high noise (Fuller et al. 2007), while T. merula begins its dawn chorus earlier to the avoid traffic noise and artificial light (Nordt and Klenke 2013). Arroyo-Solís et al. (2013) also found that, in areas with greater noise exposure, Sturnus unicolor and Passer domesticus, advanced the onset of their vocal activity. Our finding that Austral Thrushes increase element delivery rate in noisy areas represents another form of temporal adjustment, operating at the within-song level rather than the daily activity level.

Our findings indicate that urban noise significantly influences the temporal structure of T. falcklandii’s territorial songs, with birds in noisier environments singing at faster element delivery rates. In contrast, spectral parameters, including minimum frequency, maximum frequency and peak frequency, showed no significant association with urban disturbance variables. This pattern suggests that the T. falcklandii adjusts the temporal pacing of its songs rather than their spectral characteristics in response to urban acoustic conditions. The discovery of this temporal adjustment, combined with the null findings for spectral parameters, underscores the importance of examining multiple acoustic dimensions when evaluating avian responses to urbanisation. Future research should explore whether faster element delivery enhances signal detection in noisy environments and whether this behavioural adjustment has fitness consequences for urban-dwelling T. falcklandii.