AVIAN COMMUNICATION IN URBAN NOISE: CAUSES AND CONSEQUENCES OF VOCAL ADJUSTMENT

Successful acoustic communication requires that sounds propagate through the environment between the sender and receiver; vocalizations that transmit effectively in the habitat in which they are used are favored by natural selection. This “acoustic adaptation” hypothesis (Morton 1975, Wiley and Richards 1978, Richards and Wiley 1980, Ryan and Brenowitz 1985) laid the groundwork for the “sensory drive” concept (Endler 1992), which describes how environment affects the evolution of sensory systems and signals in all modalities. By allowing us to examine animal signals as adaptations shaped by selection, this framework has lead to a greater understanding of the bewildering diversity of animal signals. One of the environmental factors that exerts selection pressure on acoustic signals is ambient noise (Ryan and Brenowitz 1985). To elicit a response from a receiver, signals must be detectable in background noise; the detectability of a signal is determined by the signal-to-noise ratio (SNR) and the masked auditory detection threshold of the receiver (Marten and Marler 1977, Brenowitz 1982, Dooling 2004). Within a given frequency band, signals with an SNR below the detection threshold of the receiver are “masked.” Background noise thus plays a fundamental role in determining which “receivers” can hear a vocalization and the fidelity of the signal received. Noise varies among locations, and there is evidence that many species have evolved signals that maximize the habitat-specific SNR (reviewed in Brumm and Slabbekoorn 2005). For birds in and around urban areas and roads, the background noise is largely anthropogenic. Urban development thus provides a unique opportunity for a “natural” experiment studying how signals change in response to rapid change in the acoustic environment. This natural experiment can inform us about how sensory drive can change signals and about the mechanism by which these changes occur. Understanding this process also has important conservation implications, allowing us to predict how birds will adjust to urban development and potentially to mitigate the effects of this development on communicating birds. In this issue of The Auk,Wood and Yezerinac (2006) present evidence that Song Sparrows (Melospiza melodia) adjust their vocalizations to reduce masking by urban noise, and they propose and discuss several mechanisms by which these changes may arise. In a study of urban Song Sparrows in Portland, Oregon, Wood and Yezerinac found a positive relationship between the minimum frequency of male song and the amplitude of anthropogenic noise. They also found that males shifted more energy into the higher frequencies (4–9 kHz) of their songs in noisy areas. Urban noise is loudest between 1–2 kHz, so both of these responses should serve to decrease masking by shifting the spectral energy of the vocalization away from the spectral energy of the noise (Lohr et al. 2003). These results suggest that Song Sparrows, like several other birds (Slabbekoorn and Peet 2003, Brumm 2004b, Fernández-Juricic et al. 2005), respond to changes in the acoustic environment by altering their songs; we refer to this process as “vocal adjustment.” All known examples of avian vocal adjustment in response to urban noise have involved song, but other types of vocalizations—such as begging calls, alarm calls, and food calls—may also be adjusted (Warren et al. 2006). In this overview, we ask three questions: (1) what features of a bird's vocalization can be adjusted to reduce masking, (2) how do these adjustments come about, and (3) what are the consequences of these changes for individual fitness and population persistence? The answers to these questions depend on the morphological, developmental, and behavioral underpinnings of the vocalization, and the context in which the vocalization is used. This is an area where knowledge of physiology, developmental neurobiology, animal behavior, and behavioral ecology all contribute to understanding how animals adjust (or fail to adjust) to anthropogenic change. Four of the most salient features of animal sounds are the frequency structure, amplitude (i.e., loudness), temporal structure (timing of modulations, notes, and syllables within vocalizations), and timing of vocal delivery (repetition rate of vocalizations, diel patterns). Animals use variation in these features to detect and discriminate relevant sounds from background noise. How can a signaling bird increase its chances of being perceived in a noisy environment? Here, we discuss changes in the frequency, amplitude, and timing of vocalizations that might reduce masking and some of the physical and morphological constraints that might limit an individual's ability to make these changes. Wood and Yezerinac's (2006) findings add to a small but growing body of evidence that oscine birds adjust the frequency structure of their vocalizations to reduce masking by anthropogenic noise; similar shifts have been observed in Great Tits (Parus major; Slabbekoorn and Peet 2003) and House Finches (Carpodacus mexicanus; Fernández-Juricic et al. 2005). These changes mirror those observed in the vocalizations of birds living in areas with high levels of natural noise (e.g., from waterfalls or other animals; Dubois and Martens 1984, Slabbekoorn and Smith 2002b). The three described cases of frequency adjustment in response to low-frequency urban noise involve reduction in the frequency range of songs (i.e., increased minimum frequency but no change in maximum frequency; Slabbekoorn and Peet 2003, Fernández-Juricic et al. 2005, Wood and Yezerinac 2006). A second means by which birds could adjust the frequency structure of their vocalizations to reduce masking is by shifting the entire vocalization to a higher frequency. We are not aware of any examples of this kind of adjustment in response to urban noise; use of this adjustment will be limited by several morphological and kinematic factors that constrain the maximum frequencies that a bird can produce, including head angle, beak gape, and beak shape (Westneat et al. 1993, Palacios and Tubaro 2000, Podos et al. 2004, Nelson et al. 2005). A third means by which birds could adjust the frequency structure of their vocalizations is by changing the relative amplitude of different frequency components. Rabin et al. (2003) found that California ground squirrels (Spermophilus beecheyi) in areas with low-frequency noise from wind turbines shifted the peak energy of their calls from lower to higher harmonics. Wood and Yezerinac (2006) found that Song Sparrows shifted more energy into higher frequencies in noisier areas, but this pattern appears to reflect the use of fewer low- frequency notes rather than a change in the relative amplitude of notes. Laboratory psychoacoustic studies have shown that sounds with a greater bandwidth and higher rate of frequency modulation are more difficult to detect from noise (Lohr et al. 2003). Animals in habitats with high levels of natural noise converge on songs with primarily pure tones (e.g., Dubois and Martens 1984). No examples of these changes have yet been found in response to anthropogenic noise; the studies reporting changes in frequency content of songs measured the minimum and maximum frequency of songs, rather than notes within songs (Slabbekoorn and Peet 2003, Fernández-Juricic et al. 2005, Wood and Yezerinac 2006). This remains an interesting possibility for future study. Birds living in noisy habitats may also increase the signal-to-noise ratio during communication by boosting the amplitude of their vocalizations, a response called the “Lombard effect” (Rabin and Greene 2002, Warren et al. 2006). The Lombard effect appears to be taxonomically widespread among birds and mammals (Brumm and Slabbekoorn 2005) and may be the most common mechanism for increasing SNR in urban noise. Whether birds can increase their amplitude sufficiently to rise above background noise is likely constrained by body size (Brackenbury 1979, Brumm 2004b) as well as by the energetic costs of producing louder sounds (Oberweger and Goller 2001). The temporal structure of a vocalization may also affect the ability of receivers to detect the sound in masking noise, and thus may also be adjusted in urban environments. Laboratory psychoacoustic studies have shown that increased repetition and duration increase the detectability of sounds in white noise; vocalizations of some avian species living in areas with high natural noise have these features (see recent reviews by Brumm and Slabbekoorn [2005], Warren et al. [2006]). Increased duration of vocalizations has been observed in other taxa; Foote et al. (2004) found that killer whales (Orcinus orca) increase the duration of their vocalizations in the presence of boat noise. Interestingly, Fernández-Juricic et al. (2005) found a significant reduction in the number of notes in House Finch songs in noisy areas, which is the opposite of psychoacoustic predictions. Similarly, Wood and Yezerinac (2006) found a weak trend toward fewer notes. Fernández-Juricic et al. (2005) suggested that decreased song length may reflect a trade-off between energetic costs associated with increasing the amplitude of songs and the length of songs; this possibility requires further examination. In addition to adjusting the temporal structure of their vocalizations, birds may adjust the timing of vocalizations (Brumm and Slabbekoorn 2005, Warren et al. 2006). Many species of birds vary the time at which they vocalize to avoid interference from neighboring birds and other sources of noise (Ficken et al. 1974), so this capacity may be well developed. Some sources of urban noise are variable over time, and birds may be able to time their songs to take advantage of small gaps in noise (Popp 1989, Lohr et al. 2003) or diel fluctuations in noise levels (Warren et al. 2006). Avoidance of acoustic interference is one hypothesis for the evolution of the dawn chorus, predicting that its timing may shift with the timing of rush-hour traffic (Bergen and Abs 1997, Warren et al. 2006). Complex responses at the community level may result from temporal shifts of song; for example, suppression of call rates in one species in response to urban noise may stimulate increased song rates in another species, as was recently found in frogs (Microhyla bulteri, Rana nigrovittata, R. taipehensis, and Kaloula pulchra) exposed to anthropogenic noise (Sun and Narins 2005). Another means by which animals may increase the efficacy of communication amid noise is by shifting emphasis to another modality (Brumm and Slabbekoorn 2005). The “backup” or “redundant signaling” hypothesis proposes that animals have multiple sexually selected signals so that if one modality fails to transmit to the receiver (e.g., because of masking), other signals will do the job (Møller and Pomiankowski 1993, Hebets and Papaj 2005). It would be interesting to examine whether males in species with multiple sexual signals emphasize visual displays over acoustic displays in noisy urban environments and whether females and competing males correspondingly shift their attention while assessing these displays. There are many features of a signal that can be changed to decrease masking; what determines which of these changes will be made? We have discussed mechanistic constraints that may limit vocal adjustment. In the following two sections, we will discuss how the bird's developmental program may facilitate or limit vocal adjustment, and the consequences of these adjustments on fitness. Because the variation within a species' existing repertoire will also be shaped by mechanisms of production, developmental processes, and fitness consequences, it is reasonable to predict that vocal adjustment is most likely to occur along an axis (amplitude, frequency, or time) on which a species already exhibits variation in its song. The labile nature of song in many bird species offers the possibility of vocal adjustment in a changing acoustic but how do these adjustments Wood and Yezerinac (2006) propose several for the mechanism of vocal adjustment in Song may as responses to increased ambient noise levels adjustments may occur through mechanisms as during of vocal of masked songs and of vocalizations or of vocalizations that transmit well these adjustments may result from changes in urban bird We will discuss of these the for these mechanisms to in birds with different for vocal these it is to for an observed relationship between ambient noise levels and bird vocalizations in an that birds in which their vocalizations propagate or that the observed relationship is an of the noise level of the (i.e., if lower frequencies are masked by noise, or if amplitude background noise the measured amplitude of the Slabbekoorn and Peet 2003, Wood and Yezerinac 2006). adjustment of vocalizations may be through use of songs from a or through the adjustment of signal as frequency, amplitude, temporal structure, and timing of song (Brumm and Slabbekoorn 2005, Warren et al. Wood and Yezerinac 2006). responses do not use of this mechanism of adjustment will be constrained by an individual's ability to detect masking and make changes to their is important for song structure in birds with a of and and and is a likely for birds to detect The of the Lombard as discussed that at some of adjustment may be used to mitigate the effects of that in environments with and urban noise are to adjustments from adjustments adjustments may result from of signal during of vocal (Rabin and Greene 2002, Wood and Yezerinac which we refer to as the developmental A individual's ability to change the structure of its vocalization to reduce masking will be determined both by its ability to masking and the and duration of vocal of masking is most likely to occur through auditory which plays a role in vocal to song during song development of vocal is variable among species may to and songs their et al. which the for in a changing environment. In species of the of vocal during which birds and their songs they are in and 2004). This may limit an individual's ability to make adjustments to the song. Wood and Yezerinac (2006) that variation in the has been to vocal and has been observed in both and species et al. 1997, Brenowitz 2004). These changes in vocal for vocal adjustment into Birds may also avoid masking by songs for their repertoire Wood and Yezerinac 2006). is by of song during vocal development and of a of the which is into and The to which birds can use to reduce masking on the song selection as well as the timing of Song Sparrows, among other species, use which potentially to song selection to the acoustic environment of their et al. 2001). songs that those of their et al. which that repertoire selection from the environment and the role of background noise in the song selection process is This is not to all because some species during song development (e.g., Marler and 1982, Nelson and in species that it has the to be an important mechanism for Birds may also songs birds a for songs by to other birds during a of song development and and these birds may hear and songs or the of songs that are not masked (Rabin and Greene 2002, Wood and Yezerinac 2006). The timing of the not what birds the bird but also the environment in which it may be an means of songs to background noise if birds song in areas with noise levels similar to those in their (i.e., within noisy areas, or song Laboratory and studies have found that the within the of for many and but that many species, like the Song will to songs to their et al. which adjustment to the acoustic environment their Song may increase the ability of a species to adjust its vocalizations in response to an acoustic environment. the cases described to of vocal adjustment in response to urban noise are in oscine (Slabbekoorn and Peet 2003, Brumm 2004b, Fernández-Juricic et al. 2005, Wood and Yezerinac 2006). among species, the variation of vocal is likely to result in a range of responses to urban noise, and about the of vocal adjustment a vocal development in and birds is a largely process and not to a 2004). birds do not to use auditory in the development of their vocalizations and The relative of vocal development in the opportunity for of changes may be The Lombard effect has been found in studies of birds (Brumm and Slabbekoorn 2005), and many adjust the timing of their vocalizations in response to (e.g., et al. and 1985) and may do the in response to urban noise. from other that do not as frogs which adjust the frequency of their calls to maximize from and that responses may not be to or species with the recent of urban noise and the of song in the vocal adjustments observed to are to have through this possibility be (Warren et al. Wood and Yezerinac 2006). anthropogenic noise a more of selection may with vocalizations or the to adjust their may also receivers with that are for signals in urban noise. For some species, the small found in urban areas may the variation in vocal and that is for selection to shape Wood and Yezerinac (2006) that male Song Sparrows have many to masking of their songs by urban noise, because songs are for and are there also costs to vocal adjustment in of individual fitness and population if are the The ability to adjust vocalizations in response to masking noise may be an to levels of noise in the natural (Brumm and Slabbekoorn 2005), but and its noise is a recent and we that observed responses are in urban No studies have measured the fitness consequences of vocal adjustment to or so for we are with We will discuss some costs of vocal adjustment during and over and in other with We will discuss the of these costs and to conservation of bird communication plays a role in sexual selection and in an of birds and animals must change their vocalizations in noisy this may have effects that many of their Some of the costs of vocal adjustment will be of the of vocalization and the context in which it is For example, more in noisy environments may have energetic costs that decrease the of vocal adjustment and the bird's energy (Brumm Warren et al. Wood and Yezerinac 2006). their songs to higher frequencies may also energetic costs The of frequency adjustment may be further by given that higher frequencies do not propagate as effectively as through urban environments (Warren et al. 2006). We that an increase in frequency to a increase in the area over which the vocalization can be (Brumm and Slabbekoorn 2005, et al. 2005). The auditory of a species is to the frequencies used in communication (e.g., and Dooling there may also be a trade-off between the efficacy of in noise and the efficacy of by receivers and Yezerinac 2006). costs of vocal adjustment will vary on the context in which the signal is used. For example, changing the frequency and temporal content of vocalizations may affect individual and species by as well as the of the signal in a response from the The following discuss the effects of vocal adjustment on and other types of birds in many species use vocalizations to whether a male is an (i.e., species so if males in urban areas adjust the frequency or temporal features of their vocalizations to avoid masking by noise, they may no be by females (Slabbekoorn and Peet 2003, Wood and Yezerinac 2006). birds can discriminate changes in frequency in the studies suggest that frequency changes within of the population do not affect response may have some in adjusting their vocalizations the of by This further that in species where there is more natural variation in the frequency of male vocalizations (i.e., higher is likely to be by vocal adjustment that if songs are as they may be as if females from the population as an of song et al. Wood and Yezerinac (2006) that in might lead to and between urban and of Song Sparrows (see also Slabbekoorn and Peet 2003, Warren et al. 2006). Song has been shown to an important role in between two of Song Sparrows et al. and females in the found to discriminate from with increasing with et al. adjusted songs are to may between urban and The role of in is but it is that not lead to Slabbekoorn and Smith Slabbekoorn and Smith suggested that is more likely to habitats are sufficiently different to et al. 2004). is likely to be in to urban environments. may also lead to between urban and than we might predict on the of which would facilitate variation in urban noise remains a but not a In addition to vocalizations to females use vocalizations to their from among these competing males that adjust their vocalizations in response to urban noise may between and One trade-off is between song amplitude and song the energetic costs of increasing the amplitude of vocalizations may at the high rate by females (Oberweger and Goller Brumm 2004b, Fernández-Juricic et al. 2005, Warren et al. 2006). may also be present males adjust the frequency of their For example, in species in which females males with low-frequency vocalizations (e.g., and males that increase their frequency to avoid masking may increase the number of females detect their vocalizations but decrease their to those Wood and Yezerinac (2006) found that male Song Sparrows increased the minimum frequency of the but not the maximum frequency, to a This adjustment may reduce the of the male song in species in which females frequency bandwidth as an of vocal et al. 2004). may also reduce their by songs from their repertoire to avoid masking, if females repertoire size during (e.g., et al. 2005). females their a and 2005), and if all males adjust their vocalizations, male vocal adjustment may not lead to a decrease in relative the of the male signal as an of male may be Many of the by males during communication with females may also be during communication with competing because vocal adjustment may affect both of and of the content of the signal (i.e., the and to In Song Sparrows, males are than females in response to and will respond with similar toward from to away et al. This that vocal adjustment is more likely to be in by females than by the pattern may be in many other species (e.g., but Nelson and 2004). if vocalizations are by vocal adjustment may costs if it the of a response from For example, vocalizations are used in displays (Morton that shift to higher frequencies or tones to avoid noise may be as there is a shift in the of the Song is important in in Song Sparrows and many other species et al. et al. adjust their songs or low-frequency songs from their repertoire may not have the song types for these neighboring males adjust their songs we have evidence that male birds adjust their songs in response to urban noise, but other of vocalizations are to and and may also be adjusted and Warren 2004, Warren et al. 2006). These alarm calls, begging calls, calls, calls, and food calls and can be given between between and among and between 2004). all of these calls, may between increased SNR and decreased or by calls are to by and vocal adjustment may also affect the of There is evidence that and development have effects on population of at some avian species et al. 1980, et al. et al. 2002, 2003, and 2004, and we about the to which noise to these et al. and can about the role that vocal adjustment plays in increasing or the effects of noise and Warren 2004, Warren et al. 2006). vocal adjustment masking of a vocalization, should we it as an of animals and thus in an urban environment? are we of communication that should be of We these questions we have of the of vocal adjustment to and of the effect of vocal adjustment on This is an area for future There is variation in how species respond to some species and vocal adjustment a to species that have this capacity may more to development (Rabin and Greene 2002, Rabin et al. 2003, Slabbekoorn and Peet 2003, Warren et al. 2006). species that adjust their songs thus by than species that do not This a which requires that we more about the of vocal adjustment. adjustment is likely one among many that affect how animals respond to but understanding its effect may ability to predict the effects of on bird on acoustic communication in urban environments will make an important to conservation at the time, understanding of vocal the of vocal behavior, and how the process of sensory drive the diversity of animal signals. We and for on this Some of the in this from by the at the of and from by the of and