Introduction

Interest in the description of the interaction between anthropogenic and wildlife sounds emerged at the time of Aristotélēs (384 – 322 BC) through underwater acoustics observations [3]. Nearly 2000 years later, Leonardo da Vinci observed the sonic environment of ships and their effect on marine mammals [4]. In 1870, Thomas Edison invented a wax-cylinder recorder, which was used in 1889 for the first animal recordings in Germany [5]. Some years later, in 1892, Richard Lynch Garner recorded primates in a North American zoo [6]. However, anthropogenic noise was not highlighted as a problem until 1960, when the availability of audio cassette recorders made it possible to capture recordings during field studies [7]. In 1978, Flecher & Busnel published one of the first books about the effects of anthropogenic noise on wildlife as a result of the Symposium on the Effects of Noise on Wildlife, organised by ICA-WG4 as part of the 9th International Congress on Acoustics in Madrid, Spain in 1977 [8]. Since then, interest in the topic has increased, especially in recent years, due to the biodiversity crisis and rapid globalisation [9].

Besides wildlife, noise generated, especially on roads, railways, or airports, significantly impacts humans. These effects can transcend nuisance when exposition over a long period becomes a public health issue. The health impacts include annoyance, sleep disturbance, cardiovascular and metabolic problems, and adverse cognitive effects on children during learning [10].

Anthropogenic noise can impact wildlife at individual and population levels on all continents and habitats [11], affecting land and sea wildlife [10]. It can also disrupt ecosystems through changes within populations (e.g., species that use vocalisations for courtship) [12], reducing the ability to reproduce [10] and interactions between species, including prey location and predator detection [12]. In underwater ecosystems, it can affect large mammals' ability to communicate and find food when exposed to the noise of shipping activities [10].

In 2020, the European Environment Agency published Report No. 22/2019 [13], which highlights the potential for anthropogenic noise to impact on both terrestrial and aquatic species. Impacts may be related to physiological and behavioural responses, influencing reproductive success, mortality, emigration and population density. Effects can start at low noise levels (by human definition, e.g. 40 dB(A)) for terrestrial species, such as birds, and the impacts depend on the frequency and noise source [14]. In an area of 1,594,451 km2 across the EEA-33 (excluding Croatia and Turkey), at least 19% of the Natura 2000 protection areas are affected by noise from roads, railways, and aircraft with levels over the expected sound pressure range for quiet areas established by the European Noise Directive [13]. The Natura 2000 network was established based on European legislation (e.g., the Birds Directive [15] and Habitats Directive [16]), which states how special areas of conservation and special protection areas should be managed [15,16,17].

In North America, the US Department of Transportation published a report in 2004 report, no. FHWA-HEP-06–016 [18], which provided a “Synthesis of Noise Effects on Wildlife Populations”. Regarding birds, it considered their hearing range and sensitivity, also how propagation and attenuation of sounds can affect exposure at different forest heights, and species avoidance of traffic-dominated areas mediated by the quantity of traffic and distance from roads [18]. Subsequently, the US Department of Transportation provided reports about noise control measures, such as noise barriers made of glass, these highlighted the issue of bird collisions with this kind of noise barrier, and proposed solutions to avoid bird collisions [18,19,20,21]. The detailed abbreviations and their definitions are listed in Table 1.

Table 1 List of abbreviations used in the paper
Full size table

Motivation for Avian Species Noise Impact Investigation

According to the IUCN’s Red List of Threatened Species [22], the avian class is the only class of animal species that was completely evaluated regarding the number of existing species and species under threat, as observed in Table 2. Therefore, the avian class is the only class allowing a correct anthropogenic impact estimation.

Table 2 Number of terrestrial vertebrate species (adapted from [22])
Full size table

Noise Exposure Characteristics and Effects on Birds

In 2007, a report by the California Department of Transportation explored the effects of highway noise on birds [23]. In this report, Dooling and Popper [23] showed not only the effects caused by highway noise but also the characteristics of the noise source (e.g., their amplitude, frequency range and distance from the receiver, shown as zones (Table 3)). They did not define a distance range for their suggested zones of influence. Instead, they stated that “zone 1” was very near to the sound source and “zone 4” was very far from the source. Other studies regarding traffic noise have also shown the influence of noise amplitudes, the number of vehicles passing per hour and the distance over which effects on birds may be observed [24,25,26]. Based on these studies, it was possible to suggest some distances for the suggested effect zones, where zones 1–2 could be up to 30 m, zone 3 around 100 m, and zone 4 around 1500 m from the noise source.

Table 3 Characteristics of anthropogenic noise for which impacts on birds have been reported (empty cells indicate data not available)
Full size table

As observed in Table 3, the effects on birds depend on noise amplitudes and covered frequency range. Likewise, in humans, the corresponding dose of noise can cause permanent threshold shift (PTS), which is permanent hearing damage, occurring from noise levels over 110 dB(A) in small mammals [27]. Temporary threshold shift (TTS), which is temporary hearing damage, has been observed in birds exposed to noises between 93 and 110 dB(A) in budgerigar (Melopsittacus undulatus) and small mammals [23]. Vocalisation masking has been reported when the noise sources have amplitudes between 85 and 125 dB(A). Physiological responses have been observed with noises between 85 and 140 dB(A), while behavioural responses with noises between 38 and 140 dB(A). The negative impacts on birds start with noise levels at 38 dB(A) related to breeding behaviour [25]. This could indicate that the noise source is disturbing rather than the noise level.

Most noise impact responses are acute [11], especially if the frequency content overlaps with the communication frequencies from birds. For example, it can be mentioned species like the Black-billed magpie, with communication frequencies (CFR) ranging from 0.5 to 3 kHz [33] and Rainbow lorikeet, with communication frequencies between 1 and 1.5 kHz [34], that overlap with traffic noise sources, which generally have a sound peak frequency between 0.5 and 1 kHz [35, 36]. Other species, like Common murre (CFR = 0.1 to 10 kHz) [37], Black-crowned night heron (CFR = 1.5 kHz) [38] and Wood duck (CFR = 1.4 to 1.9 kHz) [39], overlap their communication frequencies with aircraft noise sources that generally have a source peak frequency of 1 kHz [40]. However, species differ in communication behaviour in response to anthropogenic noise (e.g., with traffic spectral components). The differences occur due to the auditory thresholds’ differentiation between species, where some species react, and others do not react to anthropogenic noise [41]. Over time, noise exposure can have other chronic characteristics, which can also degrade mid- and low-frequency hearing abilities, as has been observed in canaries (Serius canarius) [42]. Behavioural responses, such as increased flushing distance during the breeding period [43], have been observed in birds exposed to anthropogenic noise and changes in the frequency and duration of feeding nestlings in playback experiments, which used highway traffic noise around 69.1 dB(A) [44].

Classification of Anthropogenic Noise Impacts on Birds

Over the last two decades, several frameworks, mechanistic schemes and descriptions regarding the impacts of anthropogenic noise on wildlife have been presented [11, 13, 45, 46]. Effects of acoustic stimuli on animal physiology, development, neural function, genetic effects and behavioural changes have been reported in mammals, birds, fish, reptiles and insects [11, 13, 45, 46]. Figure 1 summarises the frameworks and mechanistic schemes regarding the findings about anthropogenic noise impacts on birds. The presented framework is subdivided into four major topics, which are: 1) ‘Acoustic perception/Communication and hearing’; 2) ‘Physiological response’; 3) ‘Behavioural response’; and 4) ‘Impact on fitness’, which will be detailed in the sequence.

Fig. 1
figure 1

Framework of anthropogenic noise impact on birds. Items with an (a) were adapted from EEA [13]; Items with a (b) were adapted from Kight and Swaddle [45]; Items with a (c) are suggested in this study. Items with green frames indicate positive impacts and red frames indicate negative ones. Yellow colour indicates chronic impacts, and grey indicates acute impacts. Blue colours indicate the major impact group

Full size image

Acoustic Perception/Communication and Hearing

Acoustic perception comprises ‘detection of sound stimuli’ and ‘inhibition of sound stimuli’, which are common states of acoustic perception [13]. Bird communication and hearing aspects include: ‘vocal plasticity’, ‘repetition of calls’ and ‘masking of calls’. The repetition of calls is used to help birds that are not visually in contact with each other, to perceive another bird’s approach. Anthropogenic noise interferes with their communication by signal confusion, by overlapping sound signatures from bird’s vocalisations with the anthropogenic noise frequency components [47]. Vocal plasticity is an adaptation in birds’ vocalisation, such as using higher frequencies to avoid and mitigate masking by low-frequency noise [48, 49]. Masking of calls can inhibit birds’ ability to propagate and interpret vocal information since low-frequency noise from the traffic of cars and trains can travel long distances and through vegetation. Depending on the frequency range of the avian vocalisation, the masking of communication frequencies can occur through these noise sources. Another important fact is the observation of tonal components of the noise sources, which, due to the high amplitudes, can also realise a masking process on some bird's vocalisations [50].

Physiological Response

This category has wide ramifications in areas such as: ‘community ecology’, ‘DNA integrity and genes’, ‘cell structure and signalling’, and ‘physiological systems’ [45].

Concerning the aspects related to community ecology [45, 51], it has been shown that anthropogenic noise can impact birth rate, mortality and survival through a ‘decrease of reproductive / breeding success’ ([52,53,54,55,56,57,58,59,60,61,62]).

Evidence of ‘decrease of reproductive / breeding success’ is influenced by heart rate changes, e.g. overexcitation, in Muscovy duck (Cairina moschata) embryos in response to acoustic stimulus [63] and reduction in the numbers of great tit (Parus major) [64] and eastern bluebirds (Sialia sialis) [65] fledglings in response to urban noise. ‘Decrease of nesting success’ due to noise includes negative alterations of nesting communities’ species richness and increasing nest predators' influence on western scrub-jay (Aphelocoma californica) [66]. Regarding the ‘Decrease of survival success’, a study of white-crowned sparrows (Zonotrichia leucophrys), shows that noise does not directly influence the survival success of these birds, but predicted a significant negative influence on body condition, which may have long-term consequences for survival [67].

Sound stimuli, when considered unwanted and harmful for humans and wildlife [68], such as anthropogenic noise, may affect ‘DNA integrity and genes’ by provoking a stress response that leads to DNA damage and gene alteration [45]. Based on observations of infertile eggs [69], it has been suggested that long-term exposure to loud noise damages DNA because females in louder environments produce a higher number of infertile eggs.

Subcategories of ‘Cell structure and signalling’ include the effects of ‘Decrease of immune response’ and ‘Increase of stress’, being a problem when there is a chronic noise exposure. Cell structure damage can occur due to lesions caused in the avian auditory epithelium by noise with pure-tone and broadband components [70]. However, as soon as the damage occurs, new auditory hair cells around the hair cell damage are stimulated and the growth of support and hyaline cells induce the regeneration of avian auditory cells [71]. Biomarkers of stress, such as glucocorticoids, have been shown to elevate in response to noise exposure [72,73,74,75]. Noise as a stressor suppresses the immune function system of young birds, especially during learning periods, due to communication breakdown between adults and their young [45, 76]. House sparrows (Passer domesticus) showed an immune reduction in nestlings exposed to noise [77]. Changes in metabolism were observed with increase of cholesterol and protein levels due to noise stress in domestic hens (G. gallus domesticus) [78]. Additionally, stress-induced changes in glucocorticoid hormones under noisy conditions can interfere with the fitness of populations [79].

The effects on the ‘Physiological systems’ can be observed as possible hearing damage or loss. There is a tendency towards auditory thresholds shifts, when birds are exposed to noise levels above 93 dB(A) over 72 h [23, 80]. This is well documented in the work of Wolfenden et al. [81], which investigated the behaviour of chiffchaffs near airports. They observed a lowering of chiffchaff song frequency for individuals living near an airport compared to those from nearby control sites and hypothesised that noise-induced hearing loss may prevent the airport resident birds hearing higher frequencies. However, laboratory experiments indicate that birds are more resistant to hearing loss and auditory damage than humans and other mammals [80, 82], due to their capacity to regenerate their auditory cells following exposure to loud noises [83].

Behavioural Response

Behavioural ecology deals with the ability of animals to reproduce and survive in their natural environment [84]. Noise can impact the behaviour of birds through ‘Decrease of density/ Abundance of population’ [85,86,87,88,89,90,91,92,93], ‘Behavioural changes/response’ [1, 2] and ‘Reduction of cognitive performance’ [94, 95].

Concerning the ‘Decrease of density/Abundance of population’, the effects of ‘Avoidance response’, ‘Foraging and provisioning efficiency’, can cause ‘Modified space use and movement’, changing the number of species in areas which are affected by anthropogenic noise [13]. The other ‘Behavioral changes/response’ are commonly associated with ‘Flight-fight response’ due to ‘Territory defence’, where during bird fights, they cannot hear their neighbour’s territorial calls, mediated by physiological stress hormone changes (e.g. [81]). Additional behavioural changes are related to ‘Vigilance and anti-predator’ and ‘Temporal activity pattern and sleep’ [13].

Animals move through space due to the use of dead reckoning, landmarks, and cognitive maps, which facilitate their navigation through a mental map-like representation of the environment [96]. Anthropogenic noise can cause a ‘Reduction of cognitive performance’ since the birds affected by noise can have smaller brain regions related to vocal learning; due to altered brain anatomy, they are also more distracted, affecting avoidance or vigilance behaviour [97, 98]. Other cognitive alterations have been observed in Zebra finch (Taeniopygia guttata). Following exposure to noise (110 dBA for > 100 d), the vocal skills learning of juvenile Zebra finches were altered, but these vocal skills subsequently recovered [99].

Impact of Fitness

Anthropogenic noise can have an ‘Impact on fitness’ (e.g., [100, 101]); that is, impacts on ‘Survival’ and ‘Reproductive success rate’ [13]. This happens due to severe hearing frequencies overlapping through anthropogenic noise, called acoustic masking. Birds would not hear significant acoustic signals in the environment, such as predators, competitors or mates, causing them to lose important information about environmental quality, safety, and competition [100]. Another important fact is the overproduction of glucocorticoids, activating the hypothalamic–pituitary–adrenal (HPA) axis, causing the perception of stress in the brain, directly affecting eggs' hatching success [100, 102].

Aim

This systematic review uses the mechanisms of impact framework described above to organise and evaluate the available evidence on noise impacts on birds. The meta-analysis is based on avian species investigated by Peacock et al. [103, 104], which provided a knowledge baseline regarding avian hearing curves, including peak hearing frequency – which is the frequency at maximum amplitude [105] – and dimensions of the hearing system. The information on Peak Hearing frequencies is essential for the verification of how the avian hearing frequencies overlap with and are influenced by anthropogenic noise frequencies, which could cause noise impact on birds, as observed by Engel et al. [106]. A Two Step-Cluster Analysis will provide an optimised noise impact structure considering parameters such as: the main category of impacts, sub-categories of impacts and noise sources. Avian order, species and peak hearing frequencies provide information to characterise impacts across taxa.

Methodology

A PRISMA diagram flow was adopted as a systematic review methodology [107]. Keywords “noise impact” + “noise effect” with species “nominal name” + “common name” were used to structure search of Google Scholar, Pub-Med, SCOPUS, Taylor and Francis, Springer, Wiley online library, Elsevier, ResearchGate, and Academia. The database of Peacock et al. [103, 104] was used to provide names of specific avian species to include within searches. Across the various search engines, a total amount of N = 701,575 candidate papers were identified up to 31st July 2023 using the keywords reported above. Of these, 920 documents which were identified for screening. Duplicate paper exclusion, unmatched topics and keywords or provision of partial information resulted in a pre-selection of 85 papers. After the removal of outputs that had not been subject to a rigorous peer review process prior to publication, 50 papers were carried forward for further analysis (Fig. 2). These papers, which were all published within the last 44 years, provide insights into anthropogenic sound sources that are negatively impacting birds and what the physiological, behavioural, and acoustic (hearing and communication) responses are (Table 4). The classification of noise impacts on birds follows the suggested framework presented in Fig. 1.

Fig. 2
figure 2

PRISMA flow diagram for anthropogenic noise impact on birds

Full size image
Table 4 Anthropogenic noise impact on birds – metanalysis of 50 peer-reviewed works regarding 25 of the 39 avian species cited in Peacock et al. [103, 104] (empty cells in the table indicate data not available)
Full size table

Extracted data from ‘impact’, ‘Impact category’ and ‘noise source’ were analysed using a Two-Step Cluster analysis, which consists of a pre-clustering and a clustering [133]. In the first step, a cluster feature (CF) tree checks if the cluster can be merged with other clusters through a distance criterion. The second step allows the adjustment of clusters according to the desired number of clusters, considering the removal of outliers and considering descriptive, cross-tabulation analysis and correlation results. Afterwards, Chi-squared goodness of fit was calculated to confirm that the generated clusters do not have a random component, validating the result [134].

The Akaike Information Criterion was applied as a Clustering criterion on imputed variables (Table 4), using the Log-likelihood measure of distance. As inputs of the clustering, the following parameters were used: 1) ‘Impact category’ subdivided into ‘physiological’, ‘communication and perception’, and ‘behavioural’; 2) ‘Impact’ which are subcategories from each ‘impact category’, e.g., ‘behavioural changes/responses’, ‘decrease density/abundance of population’, ‘repetition of calls’, ‘masking of calls’, ‘vocal plasticity’, ‘decrease reproductive/breeding success’, ‘reduction cognitive performance’, ‘decrease survival success’, and ‘decrease of nesting success’; 3) noise source, e.g., ‘background’, ‘Drone/UAV/UAS’, ‘white noise’, ‘aircraft/jet’, ‘mining’, ‘human’, ‘fireworks’, ‘anthropogenic’, ‘drilling’, ‘traffic’, ‘wind farm’, ‘construction’, ‘snowmobile’, and ‘military activity’. Parameters such as ‘order’, ‘species’ and ‘peak hearing frequency’ serve as evaluation fields, helping to characterise the receiver (birds) but not influencing the cluster’s power.

Results

The identified noise impacts were categorised using the framework presented in Table 4. The results highlight impacts within the blue coloured section of the framework; that is, ‘Physiological response’ (‘Decrease of survival success’, ‘ Decrease of reproductive/breeding success’, ‘Decrease of nesting success’); ‘Animal behaviour’ (‘Reduction of cognitive performance’, ‘Decrease of density/abundance of population’, ‘Behavioural changes/response’); ‘Acoustic perception – communication and hearing’ (‘Vocal plasticity, ‘Repetition of calls’, ‘Masking of calls’). A descriptive overview of the noise sources and impacts on avian species is presented in Sect. "Descriptive Analysis of the Anthropogenic Noise Impacts on Birds". The Two-Step Clustering (Sect. "Two-step Cluster Analysis of the Anthropogenic Noise Impacts on Birds") groups avian orders, species, noise sources, impacts and peak hearing frequency.

The systematic review yielded data for 20 avian orders and 39 species following the species sampled in Peacock et al. [103, 104] (Table 4). However, of the 39 species, only 25 presented information about anthropogenic noise impact, gathered in 50 peer-reviewed papers and 52 noise impact reports on avian species. The following fourteen species did not present studies related to anthropogenic noise impact during the period of data collection for this systematic review: common merganser (Mergus merganser), common ostrich (Struthio camelus), pied-billed grebe (Podilymbus podiceps), common nighthawk (Chordeiles minor), white-throated swift (Aeronautes saxatalis), yellow-billed cuckoo (Coccyzus americanus), kildeer (Charandrius vociferus), ring-billed gull (Larus delawarensis), great horned owl (Bubo virginianus), blue-naped mousebird (Urocolius macrourus), belted kingfisher (Megaceryle alcyon), laughing kookaburra (Dacelo novaegineae), northern flicker (Colaptes auritus), and red-naped sapsucker (Sphyrapicus nuchalis). Considering the IOC World Bird List [135], the reference work of Peacock et al. [103, 104] does not present all avian orders. The following eight orders were not represented within the analysis of Peacocock et al. [103, 104] and consequently in this analysis: Phaethontiformes, Pterocliformes, Otidiformes, Gaviiformes, Procellariiformes, Ciconiiformes, Trogoniformes, Bucerotiformes.

Descriptive Analysis of the Anthropogenic Noise Impacts on Birds

Results related to descriptive statistics are presented in Figs. 3 and 4. Results concerning sound sources and responses to anthropogenic-generated sounds were not available for two orders: Cuculiformes and Suliformes.

Fig. 3
figure 3

Quantification of reported sound sources that impact avian orders

Full size image
Fig. 4
figure 4

Quantification of anthropogenic noise impacts on birds subdivided according to avian orders

Full size image

Figure 3 shows the sound sources that have been observed to change bird behaviour in different environments. These sources were: ‘background’, ‘drone/UAV/AUS’, ‘white noise’, ‘aircraft/jet’, ‘mining’, ‘human’, ‘fireworks’, ‘anthropogenic’, ‘drilling’, ‘traffic’, ‘windfarm’, ‘construction’, ‘snowmobile’, and ‘military activity’. As a matter of definition, background noise is classified as the opposite of foreground noise or principal sound signal under investigation [106, 136]. Anthropogenic noise is equivalent to human-produced noise, which is considered unwanted and harmful [106, 137], and human noise is related to human-related noises (voice—communication, body—movement and clothing) [106, 137]. The plot from Fig. 3 shows the number of studies related to each sound source, subdivided by avian order. The avian order which has been most studied in relation to different sound sources is Passeriformes (Pa), with 17 studies covering anthropogenic, traffic, snowmobile, and military activity sound sources.

Figure 4 shows the impacts of noise on avian populations that were identified through the Systematic Review. These impacts are ‘decrease of survival success’, ‘reduction of cognitive responses’, ‘vocal plasticity’, ‘decrease of reproductive/breeding success’, ‘density/abundance of population’, ‘repetition of calls’, ‘decrease of nesting success’, ‘behavioural changes/responses’, and ‘masking of calls’. The avian order with the greatest number of studies reporting behavioural changes is Passeriformes (Pa) with 17 studies highlighting ‘decrease of reproductive/breeding success’, ‘reduction of cognitive performance’, ‘density/abundance of population’, ‘behavioural changes/responses’, and ‘masking of calls’.

Two-step Cluster Analysis of the Anthropogenic Noise Impacts on Birds

The Two-Step Clustering results are presented in Fig. 5 and Table 5. As observed in Fig. 5, Cluster 1 had 40 observations (76.9%), and Cluster 2 had 12 observations (23.1%). The ratio between the largest and smallest cluster was 3.33. To confirm the significance of the extraction of these two clusters, a Chi2 goodness of Fit of the clusters was calculated, which confirmed their significance χ2(1) = 15.077, p < 0.001, indicating that the clusters were not formed randomly. The Silhouette measure of cohesion showed good cluster quality (0.6).

Fig. 5
figure 5

Clusters of noise impact on birds and related noise sources

Full size image
Table 5 Clusters subdivision of birds impacted by noise
Full size table

The Two-Step Cluster analysis also indicates the importance of each predictor, where 1.0 indicates a high importance and 0.00 is a low importance. The ‘Impact category’ variable presented high importance (1.0), the ‘Impact’ indicated moderate importance (0.69) and the ‘sources’ presented low importance (0.07).

To verify the effect size of the clusters, a Kruskal–Wallis test was used because the sample was not homogeneous. Additionally, the effect size was estimated through epsilon squared (ɛ2) [138]. The interpretation of the epsilon-squared effect size corresponds to 0.00 < 0.01 (negligible), 0.01 < 0.04 (weak), 0.04 < 0.16 (moderate), 0.16 < 0.36 (relatively strong), 0.36 < 0.64 (strong) and 0.64 < 1.00 (very strong) [139].

The Kruskal-Walis test and the epsilon-squared effect size confirmed the findings of predictor importance of the Two-Step Cluster Analysis, where for ‘Impact category’ H(1) = 39.830, p-value < 0.001, ɛ2 = 0.781 (very strong), ‘Impact’ H(1) = 28.986, p-value < 0.001, ɛ2 = 0.568 (strong) and ‘Sources’ H(1) = 0.000, p-value = 0.982, ɛ2 = 0.000 (negligible).

Regarding the cluster composition, Fig. 5 shows a sunburst graph with the categories of each cluster. Cluster 1 concentrated ‘Impact category’ observations on ‘behavioural’ (n = 35, 87.5%) and ‘communication and perception responses’ (n = 5, 12.5%). The sub-categories from ‘behavioural’ are ‘behavioural changes/response’ (n = 19, 47.5%), ‘decrease density/abundance of population’ (n = 11, 27.5%), and ‘reduction cognitive performance’ (n = 5, 12.5%). ‘Communication and perception’ are subdivided into ‘masking of calls’ (n = 2, 5%), ‘repletion of calls’ (n = 1, 2.5%) and ‘vocal plasticity’ (n = 2, 5%). Related to the noise sources that are causing these impacts, the majority was associated with traffic (n = 11, 27.5%), followed by background (n = 8, 20%), anthropogenic (n = 6, 15%), drone/UAV/UAS (n = 4, 10%), aircraft/jet (n = 3, 7.5%), wind farm (n = 2, 5%), and white noise, mining, human, fireworks, snowmobile, military activity (n = 1, 2.5% each). Cluster 2 observations concentrated on ‘physiological’ responses (n = 12, 23.1%), with 100% of the observations related to ‘Physiological’ aspects subdivided into ‘decrease of nesting success’ (n = 2, 16.7%), ‘decrease of reproductive/breeding success’ (n = 8, 66.6%), and decrease of survival success’ (n = 2, 16.7%). The related noise sources to these impacts are anthropogenic (n = 6, 50%), human (n = 2, 16.8%), and background, aircraft/jet, drilling, and traffic with n = 1 (8.3%) each.

Table 5 shows which ‘noise sources’ are associated with each ‘impact’ and the characteristics of the impacted species through the indication of ‘order’, ‘species’ name and ‘peak hearing frequency’. Cluster 1 is formed by impacts on ‘behavioural responses’ such as ‘behavioural changes/responses’ occurred in association with ‘aircraft/jet’ (2 observations), ‘anthropogenic’ (3 observations), background (3 observations), drone/UAV/UAS (4 observations), ‘human’ (1 observation), ‘mining’ (1 observation), ‘traffic’ (3 observations) and ‘wind farm’ (2 observations) on birds with peak hearing frequency varying from 862 to 2018 Hz. These birds include Anseriformes (An) – wood duck, mallard duck, Passeriformes (Pa) – common grackle, American crow, blue jay, zebra finch, European starling, black-billed magpie, Columbiformes (Cl)—Eurasian collared dove, Phoenicopteriformes (Ph) – Chilean flamingo, Galiformes (Ga) – Indian peafowl, Psittaciformes (Ps) – Budgerigar, and Accipitriformes (Ac) – Turkey Vulture.

There was evidence that ‘Decrease density/population abundance’ was happening due to noises generated by aircraft/jet (1), anthropogenic (1), background (2), fireworks (1), military activity (1) and traffic (5). The birds which presented this form of impact were birds with peak hearing frequencies from the species ranging from 677 to 1764 Hz from the following orders: Pelecaniformes (Pe) – black-crowned night heron; Falconiformes (Fa) – American kestrel; Anseniformes (An) – mallard duck; Passeriformes (Pa) – common raven; American crow (reported with two observations impacts) and Steller's jay; Gruiformes (Gr) – American coot; Accipitiformes (Ac) – turkey vulture; Piciformes (Pi) – hairy woodpecker; Psittaciformes (Ps) – rainbow lorikeet (Table 4).

‘Reduction of cognitive performance’ occurs due to background (1), anthropogenic (1), and traffic noise (3). The affected birds, with frequencies from the species ranging from 1801 to 2018 Hz, were from the orders Charadriformes (Ch)—inca tern and Passeriformes (Pa)—zebra finch (three observed impacts) and European starling (Tab4).

Other impacts on ‘Communication and hearing’, such as ‘Masking of calls’, were associated with anthropogenic (1) and snowmobile (1) noise sources. The species which presented this output were from the orders Charadriformes (Ch) – common murre, and Passerifomes (Pa) – common raven with peak hearing frequencies from the species of 1928 Hz and 793 Hz, respectively. ‘Repetition of calls’ was observed with the association of background noise (1) for Podicipediformes (Po) – western grebe, which has a peak hearing frequency of 1906 Hz. ‘Vocal plasticity’ was observed when there is white noise (1) and background noise (1), for the following orders Anseriformes (An) – mallard duck and Galiformes (Ga) – Indian peafowl, with 1138 Hz and 862 Hz as respective peak hearing frequencies from the species (Tab4).

In Cluster 2, ‘Physiological response’, such as ‘decreased nesting success’, is related to human and anthropogenic noise. The orders for which impacts were reported are Pelecaniformes (Pe)—great blue heron (PHF = 895 Hz) and Falconiformes (Fa)—American kestrel (PHF = 1174 Hz). ‘Decrease reproductive/breeding success’ occurred due to anthropogenic (4), drilling (1), human (1), traffic (1), and background (1). The birds for which this type of impact was reported were birds Gruiformes (Gr) – American Coot, Pelecaniformes (Pe) – great blue heron (two impact observations), white-faced ibis and black-crowed night heron, Falconiformes (Fa) – American kestrel, Piciformes (Pi) – hairy woodpecker and Passeriformes (Pa) – zebra finch, with a variation of peak hearing frequencies from the species from 677 to 1879 Hz. ‘Decrease of survival success’ was reported in relation to aircraft/jet noise (1) and anthropogenic noise (1). The following orders are Gruiformes (Gr) – American coot (PHF = 1301 Hz), and Charadriformes (Ch) – common murre (PHF = 1928 Hz) (Tab5).

Figures 6-8 show the hearing frequency range and peak hearing frequency of the investigated avian species [103], noise source frequency and peak sound frequency [125, 126, 128, 140,141,142] of the observed noise sources that are causing impacts on these avian species. Figure 6 is showing the species, which reported ‘behavioural’ responses and Fig. 7 shows ‘communication and perception’ impacts regarding to noise observed in Cluster 1. Figure 8 shows the species and noise sources which caused ‘physiological’ impacts observed in Cluster 2. As observed in all figures, part of the spectral content of the noise sources overlaps with the hearing frequencies or at least the peak hearing frequencies of the investigated species.

Fig. 6
figure 6

Overlapping of avian frequency range and noise source spectra observed in the behavioural impacts of noise. Legend: ● Peak Hearing Frequency (PHF); ▲ Peak Sound Frequency (PSF); Depends on Predominant Source (DPS); PSF1 = Anthropogenic; PSF2 = Background; PSF3 = Drone/UAV/UAS; PSF4 = Human; PSF5 = Military activity; PSF6 = Traffic; PSF7 = Wind farm; PSF8 = Aircraft/Jet; PSF9 = Fireworks; PSF10 = Mining; Black-crowned Night Heron – PHF = 677 Hz, PSF8 = 1000 Hz; Wood duck – PHF = 1106 Hz, PSF8 = 1000 Hz; American Crow—PHF = 1062 Hz, PSF1 = DPS, PSF5 = 500 Hz; American Kestrel—PHF = 1174 Hz, PSF1 = DPS; Blue Jay—PHF = 1480 Hz, PSF1 = DPS; Common Grackle—PHF = 1604 Hz, PSF1 = DPS; Zebra Finch—PHF = 1879 Hz, PSF1 = DPS, PSF2 = DPS, PSF6 = 500–1000 Hz; Common Raven—PHF = 793 Hz, PSF2 = DPS, PSF6 = 500 – 1000 Hz; Eurasian Collared Dove—PHF = 1269 Hz, PSF2 = DPS; European Starling—PHF = 2018 Hz, PSF2 = DPS, PSF6 = 500 – 1000 Hz; Inca Tern—PHF = 1801 Hz, PSF2 = DPS; Mallard—PHF = 1138 Hz, PSF2 = DPS, PSF3 = 3000 Hz; Chilean Flamingo—PHF = 973 Hz, PSF3 = 3000 Hz, PSF4 = DPS; Turkey Vulture—PHF = 1165 Hz, PSF3 = 3000 Hz, PSF6 = 500 – 1000 Hz, PSF7 = 1–30 Hz / 500-1000 Hz / 10–200 Hz; American Coot—PHF = 1301 Hz, PSF9 = 20 Hz; Indian Peafowl—PHF = 862 Hz, PSF10 = 63 to 125 Hz; Black-billed Magpie—PHF = 1426 Hz, PSF6 = 500 to 1000 Hz; Budgerigar—PHF = 1672 Hz, PSF6 = 500 to 1000 Hz; Rainbow Lorikeet—PHF = 1764 Hz, PSF6 = 500 to 1000 Hz; Steller’s Jay—PHF = 1257 Hz, PSF6 = 500 to 1000 Hz

Full size image
Fig. 7
figure 7

Overlapping of avian frequency range and noise source spectra observed in communication and perception noise impacts. Legend: ● Peak Haring Frequency (PHF); ▲ Peak Sound Frequency (PSF); Depends on Predominant Source (DPS); PSF1 = Anthropogenic; PSF2 = Background; PSF3 = Snowmobile; PSF4 = White noise; Common Murre—PHF = 1928 Hz, PSF1 = DPS; Indian Peafowl – PHF = 862 Hz, PSF2 = DPS; Western Grebe—PHF = 1906 Hz, PSF2 = DPS; Common Raven—PHF = 793 Hz, PSF3 = 200 to 800 Hz; Mallard – PHF4 = 1138 Hz, PSF = same amplitude in all frequencies

Full size image
Fig. 8
figure 8

Overlapping of avian frequency range and noise source spectra observed in the physiological noise impacts. Legend: ● Peak Haring Frequency (PHF); ▲ Peak Sound Frequency (PSF); Depends on Predominant Source (DPS); PSF1 = Aircraft/Jet; PSF 2 = Anthropogenic; PSF3 = Background; PSF4 = Drilling; PSF5 = Human; PSF6 = Traffic; Common Murre—PHF = 1928 Hz, PSF1 = 1000 Hz; American Coot – PHF = 1301 Hz, PSF2 = DPS; American Kestrel—PHF = 1174 Hz, PSF2 = DPS, PSF6 = 500 -1000 Hz; Great Blue Heron—PHF = 895 Hz, PSF2 = DPS, PSF5 = DPS; Hairy Woodpecker—PHF = 1706 Hz, PSF2 = DPS; Zebra Finch—PHF = 1879 Hz, PSF3 = DPS; White-faced Ibis—PHF = 1076 Hz, PSF4 = 3000 Hz; Black-crowned Night Heron—PHF = 677 Hz, PSF5 = DPS

Full size image

Discussion

Anthropogenic Noise Impacts on Birds

Anthropogenic noise has been shown to impact on physiological responses, behavioural responses and acoustic perception – communication and hearing (Table 5). The reported noise sources' associated with the classified impacts on birds are discussed below.

  • a. Anthropogenic noise

This category of noise is human-produced noise, which is unwanted and maybe harmful. This kind of noise can cause: 1) a decrease in the survival success of birds due to antipredator responses, e.g., vigilance or flight on American coot [60]; 2) a decrease in reproductive/breeding success through energetic costs due to disturbance by flight activities related to feeding, therefore there is no energy or motivation for reproductive activities in the American coot [60]. Another reason for the decrease in reproductive/breeding success is the presence of humans near colonies, as observed in great blue heron studies [55]; 3) a reduction of cognitive performance was observed in zebra finch studies through the reduced performance of learning of novel motor skills, as well as fewer copy demonstrations [124]; 4) behavioural changes were observed in zebra finch, common grackle, and American crow due to foraging difficulties [125, 128, 129]; and 5) masking of calls on common murre studies, by the observation that sensitive hearing frequencies overlap with anthropogenic noise [116].

  • b. Human noise

This noise is generated by human communication, clothing and movement, which can affect birds through 1) a decrease in reproductive/breeding success, observed in black-crowed night heron studies, the disturbance and destruction of eggs, chicks and adults [52]; 2) a decrease of nesting success was evident on investigations related to great blue heron, where there was an increase of young chicks exposure predation, causing abandonment of the colony [61]. In the American kestrel, human disturbance caused nesting failure [62].

  • c. Fireworks

The use of fireworks caused alterations in animal behaviour in the American coot [93].

  • d. Military activity noise

Noises generated by military activities (e.g., firing guns, artillery, and explosive ordinances) caused the foraging disruption of American crows, and consequently, this species is observed less frequently near military areas [89].

  • e. Drilling/Construction noise

Drilling adversely affected white-faced ibis' breeding, as reported by Mueller and Glass [53]. There was also a ‘decrease in reproductive/breeding success in hairy woodpeckers, affecting their mating rituals dependent on song, especially for the "song birds" [59].

  • f. Urban/Background noise

Urban and background, the opposite of foreground sound signal, are noises which influenced the: 1) ‘decrease of reproductive/breeding success’ (e.g., the pair bond of zebra finch [57]); 2) ‘reduction of cognitive performance’ where chicks of inca tern altered their begging calls in dangerous environments, which is normal behaviour and a way to adapt to the environment, however in noisy environments, they tend not to hear the alarm calls produced by their parents when there is a predator in the proximities [94]; 3) ‘density/abundance of population’ of mallard ducks [87]; 4) ‘behavioural changes’ were observed in Eurasian collared dove by increase of distances from urban noisy environments [2], increased song level on European starlings [127]; 5) the ‘vocal plasticity’ changed on Indian peafowl by the alteration of male signals [112]; 6) ‘repetition of calls’ in the western grebe [47].

  • g. Traffic noise

Several noise impacts on birds are associated with traffic noise, such as: 1) the ‘decrease of reproductive/breeding success’ in American kestrel due to the difficulty in incubation related to the perception of greater predation risk [58]; 2) ‘reduction of cognitive performance’ of European Starlings, which are emitting more begging calls due to difficulties on communication [126]; 3) alteration on ‘density/abundance of population’ observed in black-crowed night heron [92], turkey vulture [90], hairy woodpecker [85], American kestrel [85], rainbow lorikeet [91], common raven [86], Steller’s jay [85]; 4) ‘behavioural changes’ in budgerigar in which the social behaviour was affected by masking of signals [122], and the black-billed magpie changed the selection of sites for nest choice [130].

  • h. Aircraft/Jet noise

Aircraft and jet noise affected the 1) ‘decrease of reproductive/breeding success’ observed in common murre, which felt disturbed by low-flying aircraft noise during the breeding season [56]; 2) ‘behavioural changes’ in wood duck and observation of sensitiveness [110].

  • i. Drones/UAV/UAS noise

Regarding drone noise, it was observed 1) ‘behavioural changes’ by foraging or head and tail movement of mallard duck [1]; continuous flapping due to stress caused by UAVs in Chilean flamingo [114], in turkey vultures caused disturbance [119] and increased their reaction time to reproductions of UAS noise [120].

  • j. Snowmobile noise

This kind of noise influenced the ‘masking of calls’, which inhibited the propagation and interpretation of animal vocalisations of common raven [50].

  • k. Mining noise

Mining noise affected the behavioural responses of Indian peafowls, as observed in Rathoure [113].

  • l. Wind farm noise

Regarding wind farm noise, ‘behavioural changes’ were observed through the avoidance behaviour [118] and disturbance [117] of turkey vultures.

  • m. White noise

This kind of noise, which shows the same amplitude over all frequencies, caused ‘behavioural changes’ such as foraging in the blue jay [132] and ‘vocal plasticity’ by regulating call amplitude and increasing peak frequency in mallard ducks [109].

From Figs. 5-8 and Table 5, it was observed that the composition of two clusters had a clear subdivision of the impacts of noise on birds. Cluster 1 showed ‘behavioural’ and ‘communication and perception’ impacts. The ‘behavioural’ outcomes affecting birds with peak hearing frequency ranging from 677 to 2018 Hz. The associated sound sources are ‘aircraft/jet’ (\(f\Delta =50\; Hz-5\; kHz, SPF=1000\; Hz)\) [40], ‘anthropogenic’ (\(f\Delta =below\; 2\; kHz)\) [143], ‘background’, ‘drone/UAV/UAS’ (\(f\Delta =2\; kHz-4\; kHz, SPF=3kHz)\) [144, 145], ‘human’ (\(f\Delta =63\; Hz-20\; kHz)\) [146,147,148], ‘mining’ (\(f\Delta =31.5\; Hz-8\; kHz, SPF=63\; Hz;-125\; Hz )\) [149], ‘traffic’ (\(f\Delta =500\;Hz-3\; kHz, SPF=500\; Hz-1kHz)\)[35, 36] and ‘wind farm’ (\(f\Delta =20-200\; Hz)\) [150]. Since most of the observations were related to traffic, background, anthropogenic and done/UAV/UAS, it is possible to realise that the frequency range of the emitted noise sources overlaps with the peak hearing range of the affected birds, which presented behavioural outcomes. Regarding the ‘communication and perception’ outcome, Table 5 shows that affected birds presented peak hearing frequency ranging from 793 to 1928 Hz. The related noise sources that provoke this outcome are ‘anthropogenic’ (\(f\Delta =below\; 2\; kHz)\) [143], ‘background’, ‘snowmobile’ (\(f\Delta =1\; kHz-1.5\; kHz)\) [151] and ‘white noise’. For this outcome, the most reported noise sources were ‘anthropogenic’ and ‘background’ noise, which also have emission frequencies which overlap with the peak hearing frequencies of the affected bird species. Cluster 2 showed physiological-related outcomes on bird species with peak hearing frequency range between 677 and 1928 Hz. These outcomes were associated with ‘human’ (\(f\Delta =63\; Hz -20\; kHz)\) [146,147,148], ‘anthropogenic’ (\(f\Delta =below\; 2\; kHz)\) [143], construction – ‘drilling’ (\(f\Delta =below\; 800\; Hz)\) [152], ‘traffic’ (\(f\Delta =500\; Hz-3\; kHz)\)[35, 36], ‘background’, ‘aircraft/jet’(\(f\Delta =50\; Hz-5\; kHz)\) [40] noise sources. Again, the most reported noise sources in the investigated studies which inform physiological outcomes on birds are ‘anthropogenic’, ‘traffic’, ‘background noise’ and ‘human’ noise, in which the emission frequency comprehends the peak hearing frequencies of the affected bird species.

Limitations

The main limitations observed in this study are finding works with a complete dataset of the investigated species. As observed in the quantity of screened works, the literature review screened 701,575 and we extracted 50 publications, which corresponded to the species investigated by Peacock et al. [103, 104]. Besides, there are around 11,000 bird species, so our data are a very small sample with many missing orders. This means it is difficult to generalise findings.

Additionally, it was observed that several authors presented their frameworks regarding the classification of anthropogenic noise impact on wildlife ([11], [45], [46], [13]), but there is no specific work defining the framework for impacts on avian species. The classification of impacts on this work was more demanding and uncertain due to the possibility of new findings related to the topic since there as orders which were not investigated regarding the impacts of anthropogenic noise.

Conclusion

Based on an extensive literature review on noise's impact on birds, it was possible to present a framework which classifies the major impacts in different levels, e.g., ‘acoustic perception—communication and hearing’, ‘physiological response’, ‘behavioural response’ and ‘impact on fitness’ which is the consequence of the three first levels of impacts. As observed in the limitations, this framework is the first attempt to present a classification method on impacts on avian species and could be adapted in the future, according to the progress of findings related to the topic.

Negative and positive noise impacts were highlighted through a Two-Step Cluster Analysis. Two Clusters subdivided the noise impacts. Cluster 1 (76.9%), showed negative impacts on birds' behaviour, and negative as well as positive impacts were shown on the category ‘communication and sound perception’. These were caused mainly by traffic, background, anthropogenic, and done/UAV/UAS noise sources. In Cluster 2 (23.1%), most impacts were negative, where physiological responses caused by ‘anthropogenic’, ‘traffic’, ‘background noise’, and ‘human’ noise sources were highlighted.

The presented systematic review categorises the main impacts caused by anthropogenic noise on avian species based on datasets of species that already have information about their hearing capacities, such as hearing frequency range. It also shows how the demonstrated species' hearing frequency range overlaps with the noise sources' frequency range. This information can be used for an effective noise control action plan, ensuring a better sonic environment for humans and wildlife.