Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

Post-industrial revolution humans produce more and louder noise than any other species on the planet. Given that we are also ubiquitous and numerous, anthropogenic noise has become a dominating feature of most animals’ environments. Animals have evolved to communicate in the presence of natural biological and physical sources of noise, including other animals of the same and different species (see other chapters in this volume). However, most anthropogenic noise differs from natural noise in features including intensity, distribution, persistence and timescale that are likely to make an adaptive response by most species problematic. Therefore anthropogenic noise has the potential to impact conservation.

Another reason to consider that anthropogenic noise will have conservation impacts in addition to being common and new (on an evolutionary time scale) is its documented adverse effects on human health. For example, a recent study in Western Europe indicated that at least one million healthy life years are lost every year from traffic-related noise. These losses are principally through stress-related effects linked to sleep disturbance and annoyance but also include ischaemic heart disease, cognitive impairment of children, tinnitus (WHO 2011) and incident diabetes (Sørensen et al. 2013). These studies and similar demonstrations in humans of the role of noise as a stressor, suggest that comparable effects could occur in other vertebrates. The WHO study (2011) also pointed out that exposure to noise in Europe is increasing whereas other stressors such as exposure to dioxins and benzene are declining. It is not clear whether this relative difference in noise v. other stressors also applies to animal populations, although it is clear that anthropogenic noise is increasing (see Sect. 14.3).

Our chapter is different in content and scope from the others in this book. We aim to appraise the significance of anthropogenic noise for issues related to conservation. Mitigation of, and adaptation to, noise are fundamental processes in communication and signal detection (well demonstrated by the other chapters in this volume). However, anthropogenic noise can increase errors by signal receivers (see Chap. 2) and such errors can reduce individual fitness. Reductions in individual fitness can translate into effects at a population level and therefore become relevant to conservation. However, demonstrating that the potential impact of noise on conservation is realised, particularly through effects on communication, is not straightforward.

We will discuss anthropogenic noise in terrestrial and marine environments with a taxonomic coverage largely limited to birds and marine mammals because these are the groups with which we are most familiar. This chapter is not intended as an exhaustive review of the importance of noise to animal communication and conservation, rather we highlight what we see as the main themes emerging from the growth in this field. Several recent reviews provide more details and other emphases (e.g. Pepper et al. 2003; Warren et al. 2006; Nowacek et al. 2007; Slabbekoorn and Ripmeester 2007; Southall et al. 2007; Barber et al. 2009a, b; Popper and Hastings 2009b; OSPAR 2009b; Goodwin and Shriver 2010; Tasker et al. 2010; Kociolek 2011; Ortega 2012; Slabbekoorn 2013).

As this book amply demonstrates, noise is a common problem for all modalities of animal communication—acoustic, visual, chemical, tactile or electrical (for acoustic see Chaps. 3–10, visual see Chap. 11, electrical see Chap. 12 and chemical see Chap. 13) and includes intrinsic noise in the reception system of receivers (e.g. noise of receptor cells, see Chaps. 3 and 12). We will focus on extrinsic acoustic noise because this is likely to be the only context in which noise will be familiar to those with conservation interests. However, we believe that it would be productive to apply the approach we develop in this chapter to other communication modalities. Consider, for example, noise in the chemical modality. Anthropogenic chemicals have long been included in legal definitions of pollutants and many have, or mimic, biological signalling functions. Therefore, pollution by such chemicals can also be considered as noise in chemical communication systems. A specific example in freshwater habitats is the widespread presence of anthropogenic sex hormone mimicking chemicals and other endocrine disrupters, which have demonstrable behavioural and physiological effects with potential conservation implications (e.g. Tyler and Jobling 2008).

We begin this chapter by characterising terrestrial and marine environments with respect to the potential for anthropogenic noise to have consequences for conservation through communication effects, including differences in sources of noise. We then consider how noise impacts in general have been assessed. The Sect. 14.2 discusses potential and demonstrated conservation impacts of noise through effects on communication. It is subdivided into evidence for proximate costs (with effects at population level inferred) and evidence for population level effects where proximate causes are inferred. In Sect. 14.2 we look at management of anthropogenic noise and mitigation measures; dealing with terrestrial and marine environments separately because we believe that, unlike previous sections, an integrated approach yields fewer additional insights. We conclude by identifying where further work is necessary and interim approaches that can be applied now.

2 Characteristics of Terrestrial and Marine Environments

There are several characteristics of marine and terrestrial environments that affect both the potential for anthropogenic noise to impact communication and the implications for conservation. These range from the physics of sound transmission to the ease of observing impacts and will be considered individually before looking at their combined effects.

2.1 Sound Transmission

The speed of sound in salt water is approximately 1,500 ms−1 whereas in air it is about four and a half times slower at approximately 330 ms−1. Sound is also attenuated less in seawater, especially at lower frequencies and can thus travel considerable distances (Urick 1983; Richardson et al. 1995; Ainslie 2010). In consequence, the active space of acoustic signals is much larger underwater than it is in air. For example, bottlenose dolphin Tursiops truncatus whistles have an estimated active space of up to 25 km (Janik 2000); and fin whales Balaenoptera physalus could communicate over ranges of up to 100 km, depending on conditions (Stafford et al. 2007). By contrast, the active space of a bird’s song would be measured in tens of metres (Lohr et al. 2003; Nemeth and Brumm 2010). As a result, the area over which a given noise might be of concern is much larger underwater than on land; noise from a shipping lane may interfere with communication across a wide area of sea, whereas noise from a highway is likely to interfere with signalling only in the bird territories within a few road-widths of the road.

2.2 Frequencies Used in Communication

The wide range of frequencies emitted by animals is illustrated in this section by marine taxa. At the lower end of the frequency scale are calls in the region of 20 Hz by baleen whales such as fin whales which are presumed to be reproductive displays (Watkins et al. 1987). The higher end of the scale are clicks of more than 300 kHz produced by odontocetes such as whitebeaked dolphins Lagenorhynchus albirostris (Mitson and Morris 1988, see also Rasmussen and Miller 2002) which are used for navigation (echolocation). Consequently, the hearing of most marine mammals investigated to date spans a very wide bandwidth (see Chap. 10). Southall et al. (2007) divided marine mammals into four functional hearing groups. The three families of pinnipeds were placed in one category with a designated hearing range of 75 Hz–75 kHz. Cetaceans were placed in three functional groups (1) low-frequency cetaceans, e.g. fin whale (7 Hz–22 kHz); (2) mid-frequency cetaceans, e.g. bottlenose dolphin (150 Hz–160 kHz); (3) high-frequency cetaceans, e.g. harbour porpoise Phocoena phocoena 200 Hz–180 kHz. This designation of species into functional groups is preliminary as hearing studies with published audiograms are available for ~20 of the 128 species and subspecies of marine mammals. For the species in which hearing has yet to be measured (and this includes all species of baleen whales), hearing range has been derived from the acoustic properties of the emitted signals and anatomical features (see Ketten 1997).

Fish show a more restricted bandwidth of emitted sounds than marine mammals. Most fish signals are well below 1 kHz, albeit with exceptions (Zelick et al. 1999; Popper et al. 2003; Ladich 2008, see Chap. 4). Hearing ability is diverse and dependent on anatomical features. Taxa with no swim bladder, for example sharks and flatfish, are only sensitive to particle motion. Species such as cod Gadus morrhua have swim bladders but no apparent connection between swim bladder and ear. Such species are sensitive to particle motion and pressure. Species such as herring Clupea harengus have tight connections between pressure receptors and inner ear and exhibit high sensitivity and a wide bandwidth extending to frequencies well above 1 kHz (see Popper and Fay 2011). Hearing has been investigated in fewer than 200 of the 30,000 species of fish, so our knowledge of the fish hearing spectrum (see review by Popper and Hastings 2009a) is more limited than of cetaceans (see previous paragraph).

In addition to fish and marine mammals, invertebrates such as decapod crustaceans have been described as being sensitive to sound, i.e. the particle motion component (Popper et al. 2001) and the shore crab Carcinus maenas responds physiologically to playback of ship noise (Wale et al. 2013). Cephalopods are sensitive to frequencies below 20 Hz (Packard et al. 1990; Mooney et al. 2012). Sea turtles have shown hearing capabilities in the lower frequency band (Bartol et al. 1999; Dow Piniak et al. 2012; Lavender et al. 2012). If and to what extent underwater sound is used by marine birds and how sensitive they are to sound is unknown (Dooling and Therrien 2012) although attempts are underway to document underwater hearing in some species (Johansen et al. 2013).

2.3 Use of Sound

Animals use sound for a range of activities including detecting predators and prey, communication, navigation and foraging. Echolocation is well characterised in marine mammals (e.g. Au 1993) and bats (e.g. Jones and Teeling 2006). The use of sound for navigation and orientation is less well characterised in other groups, although it is possible that fish use the surrounding acoustic environment (acoustic scene information) for orientation (Fay and Popper 2000; Montgomery et al. 2006) and infrasound may provide navigation cues for some birds (e.g. Bingman and Cheng 2005). As a general rule, however, on land sound is primarily a tool for communication, while in marine environments it serves a broader range of functions.

2.4 Habitat Biases

Terrestrial habitats differ from underwater habitats in the visibility of effects. One consequence of this difference is that more is known about the immediate effects of noise on land animals than those living underwater, because it is easier to observe (and conduct) experiments with most terrestrial animals, including assessing their hearing ability. A second consequence is that the habitat destruction associated with noise production is more visible on land than underwater and this has contributed to a difference in the perceived relative importance of noise and habitat destruction in the two habitats. On land it is considered that the conservation consequences of habitat destruction around noise sources (e.g. the cleared area around a gas well) are more important than the effects of noise per se (e.g. interference with communication). Underwater, the reverse is often the case as the difficulty of observing habitat destruction associated with noise production may result in such potentially significant conservation effects being overlooked and attention being concentrated on noise alone. Anthropocentric biases are different; terrestrial noise is readily appreciated to affect humans, whereas marine noise is viewed mainly in terms of its effects on animals. This effect is enhanced if the marine species have iconographic status (e.g. humpback Megaptera novaeangliae, blue Balaenoptera musculus and killer whales Orcinus orca), with the result that the well-being of such species is widely considered.

2.5 Intentional v. Incidental Noise production

Most anthropogenic noise is a by-product of activities such as travel (road, shipping and aircraft noise), construction (e.g. pile driving), extraction (e.g. blasting), industrial activity and wind farms (Blickley and Patricelli 2010). Noise resulting from sound that is intentionally introduced is much more common in the marine environment through sonar and geophysical surveys (e.g. airguns), with terrestrial examples limited to alarm and warning sounds. Clearly, the scope for mitigation is greater when anthropogenic noise is an incidental by-product than when it is vital for the outcome of the activity.

2.6 Summary

It will be clear from the rest of this chapter that anthropogenic noise has received far less attention in relation to its impacts on, and conservation implications for, terrestrial animals than marine animals. This is likely a combination of a failure to consider the impact of noise on terrestrial animals due to anthropocentric bias and the presumed greater effects of visible habitat destruction. This bias is also despite the relative ease of observation and measurement of impacts on land. However, as much terrestrial noise is an incidental by-product of our activities (cf. for example the essential role of sound in marine seismic surveys) there may be more scope for mitigation on land.

3 Sources of Noise

At first consideration, anthropogenic noise would seem to differ in several characteristics from natural sources of noise, such as wind, other species, waterfalls, waves and thermal energy (marine environment reviewed by Hildebrand 2009; Ainslie 2010). The first difference is that anthropogenic sounds are often more intense than natural noises (exceptions include large waterfalls, storms, undersea earthquakes, sea floor volcanic eruptions and sperm whale Physeter macrocephalus echolocation clicks, all of which are relatively localised in space and time.) A second difference is that most anthropogenic sounds contain more low frequencies than natural noises (exceptions are high-frequency sounds produced by some machinery and the hiss of tyres on road surfaces). A third difference is the relative commonness of high-intensity impulse sounds produced by anthropogenic sources (naturally occurring exceptions are lightning strikes and echolocation clicks of most odontocetes). Intense impulse sounds such as airgun firing, blasting charges, pile strikes and sonar pings are more likely to have acute impacts including temporary or permanent injury to auditory systems. By contrast, continuous noise including road, ship and aircraft traffic noise, drilling, construction, industrial activities, low- and mid-frequency sonar systems (see Southall et al. 2007; Tasker et al. 2010) and acoustic harassment/deterrent devices are more likely to produce chronic effects such as masking and stress.

Some of the key acoustic characteristics of marine anthropogenic noise are summarised in Table 14.1. The source levels of sounds can provide a first impression of their potential impacts; however, inferring impact from source levels is complicated by two things. First, source levels are usually determined by measuring sound levels in the acoustic far-field and extrapolating back to determine the level at 1 m from the source (see Ainslie 2010). In many cases a simple Xlog (R/1 m) scaling is used and not an actual propagation loss correction. The resulting source level is therefore not independent of the environment in which the measurements were taken and it is difficult to compare results obtained in different studies. Second, effects on living animals are dependent on many other acoustic characteristics in addition to the sound level at the receiver (for a discussion of these in the marine environment, see Southall et al. 2007). Finally, as there is at least one biological source of naturally occurring high-intensity impulse sounds (odontocete echolocation clicks, see previous paragraph), it is possible that marine animals may be adapted to deal with high-intensity impulse sounds.

Table 14.1 Acoustic characteristics of some marine anthropogenic sounds (Adapted from Hildebrand 2009; OSPAR 2009a; CEDA 2011; Thomsen et al. 2011)
Full size table

Anthropogenic sources of noise are increasing in their distribution and abundance. In the US, for example, road traffic nearly tripled between 1970 and 2007 and aircraft traffic, by some measures, more than tripled between 1980 and 2007 (Barber et al. 2009a). Unfortunately, this increase significantly offsets the reduction in intensity of many sound sources (e.g. sound levels from US aircraft engines dropped 20 dB(A) in the past three decades, Bronzaft and Hagler 2010) that resulted from a growing awareness of noise pollution and consequent regulations (discussed below). In the seas, ambient noise levels have increased in several regions over the past decades due to increased ship traffic (e.g. Ross 1993; Andrew et al. 2011).

4 Assessing Noise Impacts

Anthropogenic noise can have many different impacts on individual fitness that can translate into conservation consequences, such as permanent or temporary threshold shifts, flight reactions and disruption of activities such as foraging and migration. We detail three approaches that have been formalised to assess effects of noise on animal populations. These have mainly been applied in the marine environment, but we discuss their actual and potential application to terrestrial environments.

4.1 Zone of Influence Model

This approach to assessing noise impacts is based, at least partly, on the distance between the source and the receiver; the rationale is that sound intensity falls with increasing distance from the source and therefore impacts are likely to lessen, or at least to change, with distance. Richardson et al. (1995) defined a nested series of zones of influence centred on the source (Fig. 14.1):

Fig. 14.1
figure 1

An illustration of the zones of influence model after Richardson et al. (1995). Bold text shows names of zones. The source is at the centre of the concentric circles. Indicative threshold values in dB (*e5 indicates re 1µPa2·s in water) for the boundaries between zones are taken from Dooling and Popper (2007) (terrestrial) and Southall et al. (2007) (aquatic, for pinnipeds and cetaceans). Values in brackets indicate thresholds for impulsive sounds

Full size image

  • The zone of audibility is the most extensive and is defined by the receiver’s ability to detect noise.

  • The zone of responsiveness is the area within which the receiver reacts behaviourally or physiologically to the sound. (For examples of behavioural disruption in a terrestrial environment see Kaseloo and Tyson 2004).

  • The zone of masking is the area where noise interferes with the detection of biologically relevant signals such as echolocation clicks or social signals. It is highly variable.

  • The zone closest to the source is where the received sound level is high enough to cause hearing loss, discomfort or injury. In air, continuous noise >110 dB(A) causes permanent threshold shifts in birds, noise >93 dB(A) causes temporary threshold shifts (Dooling and Popper 2007). The physiological effects of noise exposure on marine mammals and fish are reviewed by Southall et al. (2007) and Popper and Hastings (2009a, b) respectively.

The zones of influence model has been applied in various marine impact studies (e.g. Erbe and Farmer 2000; Madsen et al. 2006a; Thomsen et al. 2006) and formalised for terrestrial habitats (with birds in particular) by Dooling and Popper (2007). However, we have to bear in mind that the relationship between the type of effect elicited and distance to the sound source is not straightforward. One reason is that the complexity of sound transmission (particularly underwater, but also in complex built environments such as cities) inevitably leads to sound fields that are more complicated than the concentric circles of the Richardson et al. (1995) model. A second reason is that while distance between source and receiver might adequately relate to some of the properties of a sound wave (e.g. received sound pressure level and duration), other sound characteristics do not. For example, kurtosis (‘peakedness’; see Southall et al. 2007), rise time and overall pattern of occurrence can also define sound effects and these features do not relate simply to distance to source. Furthermore, studies have shown that physiological effects are related to the dose of exposure, which involves the duration of the exposure (see Southall et al. 2007; Kastelein et al. 2012). This means that physiological effects can potentially occur at sound pressure levels that do not cause a behavioural response when the animals are exposed for a long period. Thus, the influence zone for physiological effects can be larger than the zone of responsiveness (see also WODA 2013). Finally, although zones of noise influence are a very useful starting point in classifying impacts, they can mislead. For example, behavioural reactions might lead to severe consequences such as stranding (see Cox et al. 2006) so that a zone where initial responsiveness occurs might well become the zone of injury or even death.

4.2 Population Consequence of Acoustic Disturbance Model

A second approach addressing how acoustic disturbance could lead to population level consequences of relevance to conservation is the Population Consequence of Acoustic Disturbance model (PCAD model, Fig. 14.2) developed for marine mammals (NRC 2005). The model involves several steps, from a characterisation of the sound source to population effects, but most of the transfer functions are not well understood. For example, acoustic disturbance can lead to disruption in feeding behaviour in cetaceans such as killer whales, but the effects on variables such as survival, maturation and reproduction are largely unknown (see Williams et al. 2002, 2006 for estimated costs of behavioural reactions).

Fig. 14.2
figure 2

Overview of the PCAD Model (NRC 2005). The number of * within the boxes indicate how well the features of the model can be measured. The number of * under the transfer arrows indicate how well the transfer functions are known

Full size image

Similarly, in terrestrial environments the causal link between the immediate effects of noise on signalling and population level effects is indirect, compared to the more extreme effects of noise. Deafness or repeated interruption of foraging is clearly detrimental, and likely to reduce survival or reproductive success. However, it is less clear that slight changes in song structure or difficulties in signal reception associated with effects of noise on communication will cause appreciable harm, once all the other factors that may affect an animal’s fitness are factored in. Demonstrating effects of noise can be hampered by poorly documented study design, both with regards to a proper characterisation of the source signal and the adequate sampling of behaviour (reviewed by Nowacek et al. 2007 and OSPAR 2009a). Even if the causal links between an effect of noise on communication and a decrement in survival and reproductive success can be demonstrated, this stress is only one of a number that can affect population viability, all of which must be weighed to evaluate whether the effect of noise specifically on animal communication should be a conservation concern. Unless these effects can be shown to be as detrimental as the more obviously extreme effects of noise (e.g. flight responses) the effects of noise on communication are less likely to have priority in conservation efforts, particularly given that the most immediate threats to population viability are habitat destruction and fragmentation. In the marine environment there are similar difficulties in weighing effects of noise relative to more immediate pressures such as fishing (including bycatch effects) and physiological reactions to contaminant loads (see Thomsen et al. 2011).

In both terrestrial and underwater environments it is likely that a combination of impacts will produce population level responses, yet methods for assessing cumulative impacts are still in their infancy (e.g. Wright 2009). A further factor to bear in mind is that population census estimates can be highly variable, making it very difficult to detect effects even in areas of high anthropogenic impacts of several sorts (see Thomsen et al. 2011). Variability in difficult to census species (e.g. many cetacean stock assessments) is understandable and compounded by the resolution with which such census estimates can be made. The usual result is an inability to detect change even when a considerable percentage of the population has been lost.

4.3 Risk Assessment Framework

The third approach is a risk assessment framework, which Boyd et al. (2008) have suggested would result in a more systematic approach to noise impact studies. The risk assessment framework involves a stepwise procedure including:

  1. 1.

    hazard identification (characterisation of the potential threats of a source);

  2. 2.

    dose-response assessment (assessment of the quantitative relation between received sound and the effect);

  3. 3.

    exposure assessment (specifying the number of individuals that might be exposed to the hazard);

  4. 4.

    overall characterisation of the risk, leading to risk management with appropriate mitigation measures (details in Boyd et al. 2008).

It looks as though step 1 might be relatively straightforward, although measurements of sound sources need to be standardised much more and some of the more complex issues related to source characteristics (e.g. vertical differences in emitted sound levels) and transmission of sound (e.g. water column vs. sediment transport) need to be explored more thoroughly (see recommendations in IACMST 2006; OSPAR 2009b; Southall et al. 2009; TNO 2011).

The dose–response assessment and exposure assessment (steps 2 and 3) will be more difficult to apply in the marine environment as the distribution of receivers is highly variable and areas of high importance are therefore quite difficult to identify (e.g. Coull et al. 1998; Hammond 2006). The most challenging step might be to assess the relationship between dose (e.g. properties of the received sound) and response, as results from studies investigating the effects of sound on marine mammals, fish and other marine life are, to date, highly equivocal.

4.4 Summary

Assessing conservation impacts of noise is complicated by the need to translate noise impacts on individual (or small group) communication behaviour into effects on individual fitness that will have population level consequences of interest to conservation. We are some way from a robust impact assessment of noise that can be applied to contexts of potential conservation concern; however, the approaches discussed above are a step forward.

5 Noise Impacts: Potential and Documented

Most reviews of the impact of noise list a variety of effects. For example, noise that causes death or injury (e.g. death in herring due to pile driving noise, Caltrans 2001) is clearly of potential conservation concern. However, overviews of impacts of noise (e.g. Table 14.2, marine impacts and effects for fish and marine mammals) can be difficult to interpret. Two important caveats apply to many studies of noise impacts (illustrated here with marine examples, but which apply equally strongly to terrestrial examples). The first caveat is that results can be equivocal, with both documented presence and absence of effects. For example, Nowacek et al. (2007) and Popper and Hastings (2009b) provide examples of well controlled studies which elicited no apparent behavioural or physiological response, even though some studies involved very high received sound levels. The second caveat is that some studies reporting effects have methodological problems which make it difficult to assess their validity (see Popper and Hastings 2009b for examples of fish injured by pulsed sounds of pile driving). This caveat also applies to some behavioural studies where responses were not documented properly and/or received sound levels were unknown (for a discussion see OSPAR 2009a). This emphasises that research on noise-related impacts is still in its infancy even though attempts to standardise methodologies have been undertaken (e.g. Tyack et al. 2004; ANSI/ASA 2009; TNO 2011).

Table 14.2 Overview of documented effects of underwater noise on aquatic life and example studies (adapted from OSPAR 2009a; see also Richardson et al. 1995; Würsig and Richardson 2002; Popper and Hastings 2009b)
Full size table

In this section, we review the empirical evidence for the effects of noise on communication and, ultimately, population viability (rather than direct effects with severe conservation implications such as death and injury). As noted above, establishing the link between anthropogenic noise, communication and conservation is difficult, because the effects are indirect and because many other factors are involved (e.g. Kight and Swaddle 2011). Currently studies have shown two types of effect. First, studies have shown an apparent effect of noise on communication, but the link between the demonstrated proximate cost and an ultimate cost in survival or reproductive success is inferred rather than demonstrated. Second, studies have shown a decrease in population density or diversity in relation to noise, but the relationship is usually a correlation, so factors other than noise or its effect on communication might account for the relationship.

5.1 Evidence of Proximate Costs, with Population Level Effects Inferred

5.1.1 Costs of Threshold Shift

Threshold shifts (i.e. reduced sensitivity to sounds) resulting from exposure to noise are a likely cost of communicating in noise. Permanent damage to hearing (permanent threshold shifts, PTS, see Chaps. 8 and 10) often through damage to hair cells (e.g. pink snapper Pagrus auratus from seismic airgun sounds, McCauley et al. 2003) is considered a form of injury; however, as part of its effect will be through communication, we include it here. Whereas PTS is considered an auditory injury, temporary threshold shift (TTS) represents auditory fatigue (with effects on hair cells, variation in middle ear muscular activity and blood flow that are recoverable; Southall et al. 2007, see Chaps. 4, 8 and 10). Nevertheless, part of the effect of TTS will be through communication. TTS has been documented in a variety of fish and marine mammals (overview in Southall et al. 2007 and Popper and Hastings 2009a, b, Ladich this volume; examples in Table 14.2; for detailed discussion see Chaps. 4 and 10). Both threshold effects are likely to have similar costs to masking discussed in the Sect. 14.5.1.2. It is worth pointing out that in marine mammals TTS can occur at frequencies that are very different from the main frequency of the received sound (see for example Lucke et al. 2009) so conclusions on effects based solely on the frequency spectrum of the emitted sound are problematic. We have presented aquatic examples in this section because most examples on threshold effects come from the marine environment. Also, they are arguably of more concern because of better transmission of sound in water compared to air, and because in terrestrial environments human health concerns might be expected to keep levels below TTS effects in mammals.

5.1.2 Costs of Masking

The most obvious purported cost of communicating in noise is a decrease in survival or reproductive success because of signal masking. As this volume shows (Chaps. 3–10), both signallers and receivers have a range of adaptations to reduce these costs, but these strategies are presumably adaptations to the conditions that prevail in nature, and thus might well fail in the face of anthropogenic noise (but see Cunnington and Fahrig 2013).

In terrestrial animals, signals that are imperfectly detected or discriminated might result in poor predator detection (e.g. Francis et al. 2009), lower mating success (e.g. Bee and Swanson 2007; Samarra et al. 2009; Gross et al. 2010; Gordon and Uetz 2012), smaller territories (e.g. Parris et al. 2009), poorer flock cohesion (e.g. Lohr et al. 2003) or reduced parental care (e.g. Leonard and Horn 2005, 2012; Schroeder et al. 2012; but see Leonard and Horn 2008; Naguib et al. 2013).

In the marine environment, cetaceans are known to use vocalisations to coordinate movements (Ford 1989; Janik and Slater 1998; Miller 2006), in reproductive behaviour (e.g. Payne and Webb 1971; Oleson et al. 2007), and to maintain contact between group members (e.g. Ford et al. 1989) and mothers and their calves (e.g. Smolker et al. 1993). Cetaceans also use passive sonar when hunting, detecting acoustical cues from potential prey (e.g. Barrett-Lennard et al. 1996). Noise could adversely affect all of these uses of sound by masking. However, the role of masking will remain speculative until there is evidence that marine mammals communicate or orientate over the large distances indicated by the enormous active space of their signals. In fish, close range signals such as reproductive calls of some species (e.g. cod Brawn 1961; Hawkins and Rasmussen 1978; Lusitanian toadfish Halobatrachus didactylus Vasconcelos et al. 2007; damselfish Chromis chromis and drums Sciaena umbra Codarin et al. 2009) can be masked by continuous sound and at least potentially disrupt mating and spawning.

In both environments animals have been shown to gather information by eavesdropping on signals of more distant individuals (McGregor 2005), therefore noise has the potential to adversely affect such interactions by reducing the extent of the communication network (see Janik 2005 for marine mammals).

5.1.3 Production Costs of Attempts to Overcome Masking

For signallers, another possible cost of communicating in noise is the cost of changing the signal so that it is less likely to be masked. Such changes include increases in intensity, rate, duration or frequency, all of which might increase the usual costs of signal production, such as energy expenditure or predator attraction (e.g. Gil and Gahr 2002; Parris et al. 2009) as well as increased social aggression (Brumm and Ritschard 2011). It should be remembered, however, that evidence for such signalling costs in terrestrial environments is still scant and controversial for most signalling systems (Searcy and Nowicki 2005; Zollinger et al. 2011). In the marine environment there is evidence of changes in vocalisations in the presence of noise (change in intensity: beluga Delphinapteras leucus Scheifele et al. 2005; killer whale: Holt et al. 2009; and/or change in frequency: right whale Eubalaena glacialis Parks et al. 2011). In the presence of sonar other species change signal duration (humpback whale Miller et al. 2000) and/or frequency (humpback whale Miller et al. 2000). However, the costs for the individuals (Bejder et al. 2009) and their fitness consequences are currently the subjects of discussion.

5.1.4 Increased Stress and Impaired Decision Making

Receivers, too, might bear costs in trying to detect and discriminate signals in noise. Given the many options they have to better perceive masked signals (e.g. Dooling and Popper 2007), the costs they might incur are varied. One cost that is likely to be universal, however, is that the extra effort required to perceive masked signals might cause receivers to miss critical stimuli, such as alarm calls or acoustic cues from predators (Quinn et al. 2006; Rabin et al. 2006). Similarly, the extra cognitive effort needed to process masked signals might impair decision-making more generally, resulting in poor behavioural choices and perhaps physiological stress (Bateson 2007; Kight and Swaddle 2011; Owens et al. 2012; but see Zheng 2012; Crino et al. 2013). Similarly stranding or beaching events that have occurred in species such as Cuvier’s beaked whales Ziphius cavirostris and other species in response to mid-frequency active sonar (Cox et al. 2006) are most likely related to stress and impaired decision making. Interruption of normal behaviour patterns due to playback of pile driving noise to sole Sola solea and cod (Mueller-Blenkle et al. 2010) and in humpback whales due to seismic surveys (McCauley et al. 2000) could induce stress.

5.1.5 Changes in the Location or Timing of Signalling

Both senders and receivers may incur costs when they change the location or timing of communication to avoid interference from noise. If a bird has to sing from higher perches to overcome traffic noise, for example, it may increase its exposure to predators (e.g. Díaz et al. 2011; Halfwerk et al. 2012; see also McLaughlin and Kunc 2013). Similarly, if birds have to shift the timing of vocal behaviour to avoid noisy periods (Bergen and Abs 1997; Fuller et al. 2007; Kaiser et al. 2011; Arroyo-Solís et al. 2013; see Chap. 7) or have to sing more to compensate for masking (Díaz et al. 2011), the change in their overall time budget will almost certainly entail trade-offs with other behaviours (e.g. Conomy et al. 1998; Díaz et al. 2011). Displacement from locations for short and long durations has been found for fish in response to seismic surveys (cod and haddock Melanogrammus aeglefinus Engås et al. 1996) and for cetaceans in response to acoustic harassment devices (e.g. killer whales, Morton and Symonds 2002), marine construction activities (gray whales Eschrichtius robustus Bryant et al. 1984) and pile driving (harbour porpoises, Brandt et al. 2011).

5.1.6 Evolutionary Changes

Noise-induced changes in signal structure, especially long-term learned or evolved changes, as seen in nestling tree swallows Tachycineta bicolor (Leonard and Horn 2008) and, possibly, adult European blackbirds Turdus merula (Slabbekoorn and Ripmeester 2007, but see Mendes et al. 2011), may have evolutionary consequences, for example shifting the preference of females for particular songs or the ability of males to sing preferred songs, thus potentially reducing gene flow between urban and rural populations, for example (Slabbekoorn and Peet 2003; Montague et al. 2012; see also Luther and Derryberry 2012). Growing evidence suggests learned changes in songs may lead to genetic differentiation, presumably via mate choice (e.g. Leader et al. 2005, see more extensive discussion therein and in Slabbekoorn and Ripmeester 2007; Slabbekoorn 2013). Nonetheless, the possibility of anthropogenic noise having evolutionary consequences still remains speculative (for a detailed discussion of urban song divergence in birds, see Chap. 7).

In summary, several effects of noise on animal signalling appear to be costly, but none have been directly linked to survival, reproductive success or any other more direct measure of population viability, despite the theoretical possibilities.

5.2 Evidence of Population Level Effects, with Proximate Cause Inferred

An alternative approach to establishing an immediate effect of noise on communication, and then inferring its possible effects on population viability, is to correlate potential masking noise with measures of population viability, while attempting to rule out other habitat effects. For instance, some studies have shown that reductions in abundance (van der Zande et al. 1980; Eigenbrod et al. 2009; Kaiser et al. 2011) or diversity (Stone 2000; Francis et al. 2009) extend farther from noise sources than one would expect if the reductions were caused by habitat degradation or disturbance. In some particularly convincing studies of the effects of road noise, declines in density or diversity correlated with traffic load, but at distances beyond where direct mortality and disturbance could be an issue, suggesting that signal interference is the most likely cause (Reijnen et al. 1995, 1997; Forman and Deblinger 2002; Peris and Pescador 2004; Jaeger et al. 2005; Eigenbrod et al. 2009; but see Fahrig et al. 1995; Benítez-López et al. 2010). Most convincing of all are studies in which the breeding density (Bayne et al. 2008; Francis et al. 2011; Blickley et al. 2012), diversity (Francis et al. 2009) or reproductive success (Habib et al. 2007; Schroeder et al. 2012) of birds is lower near noisy compared to silent machinery, or near playback of traffic or machinery noise compared to silent controls (see also Barrass 1985, cited in Eigenbrod et al. 2009). One study further shows how such effects of noise on birds can have far-reaching effects on how ecosystems function; Francis et al. (2012) found that key pollinating and seed dispersing bird species change where they forage in response to noise.

Whether the above patterns are caused by the effect of noise on signalling per se, as opposed to some other behavioural effect, is unclear. More direct evidence implicating signal masking comes when the declines are stronger for species that have lower frequency vocalisations and thus are more likely to be masked by anthropogenic noise. Several studies have, indeed, shown this pattern for declines in abundance or diversity with traffic noise (Rheindt 2003; Francis et al. 2009, 2010; Parris and Schneider 2009; Goodwin and Shriver 2010; Herrera-Montes and Aide 2011; Proppe et al. 2013; see also Hu and Cardoso 2009; Hoskin and Goosem 2010), or have shown that declines in reproductive success are best explained by the noise levels in the frequency band that would mask songs, rather than in other frequencies (Halfwerk et al. 2011).

6 Anthropogenic Noise and Environmental Management

There is a stark contrast between marine and terrestrial environments in relation to anthropogenic noise and conservation impacts—noise in the marine environment is now a major issue for both the public and regulators, whereas in terrestrial environments noise receives much less attention in general, and particularly so with regard to conservation. For example, when a new road is proposed, habitat destruction, particularly wetland crossings, is tightly regulated, but, as a rule, the impact of noise for non-human animals associated with the presence of the road is only addressed in special cases, such as when species at risk are known to be highly sensitive to disturbance. Noise concerns as they relate to humans, on the other hand, are often considered to be fairly tightly regulated (but see abstract for emerging human health concerns). Such regulations only contribute to animal conservation in areas where humans are also likely to be affected.

In this section we discuss management of marine noise and then terrestrial noise, in part because management approaches have different histories and patterns of application that would make an integrated approach unwieldy. We include all anthropogenic noise, whether or not it is likely to have an effect through communication. We shall then consider whether communication effects need separate additional management and legislative treatment.

6.1 Management of Marine Noise

6.1.1 Background and Assessments

Since concerns about the potential impacts of underwater noise on marine life were raised in the early 1970s (e.g. Payne and Webb 1971), the issue has been debated (with some accompanying controversy) between scientists, the public, industry and other stakeholders including non-governmental organisations. This is especially during the past decade (e.g. OGP/IAGC 2007; Weilgart 2007; Parson et al. 2008). A milestone in scientifically driven debate was the formation of national fora such as the UK Working Group on Underwater Sound (now Underwater Sound Forum, see Defra 2010, working group report see IACMST 2006) and the US Joint Subcommittee on Ocean Science and Technology (JSOST, Southall et al. 2009). Also worth mentioning are information and guidance papers compiled by large industry platforms such as CEDA (2011) and WODA 2013. Internationally, the Intersessional Correspondence Group on Underwater Sound within OSPAR (Convention for the Protection of the Marine Environment of the North East Atlantic former Oslo-Paris Commission) was formed. This group has published two reports, one providing a background on impacts of man-made sound in the environment (OSPAR 2009a), the other assessing current (as of 2009) pressures due to underwater noise in the North East Atlantic (OSPAR 2009b). Both reports have been accepted by all 14 OSPAR member states and therefore carry some weight in informing policy. In this context, the OSPAR quality status report (QSR) 2010 (OSPAR 2010) is of particular relevance. The QSR provides a holistic assessment of the status of the North East Atlantic including for the first time underwater noise impacts. Although OSPAR (2010) notes the scarcity of information on noise-related effects it points out that OSPAR regions II (Greater North Sea) and III (Celtic Seas) seem to be most affected by noise and also calls for developing guidance on options for mitigation of noise and its effects. The importance of these sometimes time consuming and complex assessments should not be underestimated; they significantly inform policy makers on current status and, even more importantly, future research and management needs.

6.1.2 Legal Instruments

There are existing regulatory frameworks such as the US Marine Mammal Protection Act (1972) and the EU Habitats Directive (1992) which protect a variety of marine mammals and fish species that are sensitive to sound. Similarly, the EU-EIA Directive requires member states to perform an Environmental Impact Assessment (EIA) for projects likely to have significant impacts on the environment. EIAs can involve methods for assessing impacts of sound and can lead to mitigation measures such as ‘soft-start’ procedures during pile driving for offshore wind farms (e.g. Cefas 2004; JNCC 2009b). EIAs are also undertaken in many other parts of the world (for further information, see http://www.iaia.org).

The legal instruments mentioned above all go some way towards managing noisy activities. Yet until recently there was no regulatory framework specifically addressing underwater noise. This changed in Europe with the publication in June 2008 of the EU Marine Strategy Framework Directive (MSFD). The purpose of the MSFD is ‘establishing a framework for community action in the field of marine environmental policy’. The MSFD aim is to protect, conserve, and where possible, restore the marine environment in order to maintain biodiversity and provide diverse and dynamic oceans and seas which are clean, healthy and productive. The Directive requires Member States to achieve ‘Good Environmental Status’ (GES) in their marine environment by 2020 at the latest. Annex one of the MSFD lists the 11 qualitative descriptors for GES, one of which states that ‘the introduction of energy, including underwater noise, is at levels that do not adversely affect the marine environment’. Based on advice from an expert group (see Tasker et al. 2010) the EU has decided on two indicators that further specify GES. Indicator one addresses the distribution in time and place of loud, low- and mid-frequency impulsive sounds. The second indicator deals with continuous low-frequency sound (details in EU 2010). Whereas indicator one will perhaps require only an annual desk based assessment of activities generating low-frequency pulses, such as pile driving and seismic surveys, indicator two will most likely involve measuring ambient noise, perhaps at a regional level which would represent huge progress in identifying trends in existing pressures such as those from shipping (see Tasker et al. 2010; van der Graaf et al. 2012). Details of requirements for such monitoring are currently being investigated by an EU expert group and were planned to emerge in 2013 to keep up to speed with the very ambitious timeline of the Directive. The issue of ship noise has also been picked up by a corresponding group of the International Maritime Organisation (IMO, see IMO 2009)

6.1.3 Marine Mitigation Measures

The EU Marine Strategy Framework Directive will most likely lead to management measures that have to be undertaken in order to achieve Good Environmental Status. In this section, we shall discuss some of the existing and emerging measures to mitigate effects of underwater noise. However, before we do this we shall make a more general remark. It should have been clear from the previous section that our knowledge on sound-related effects has made huge progress in recent decades; yet, the overall picture remains incomplete, especially looking at the population level consequences of noise exposure. This calls for more controlled and replicable impact studies, especially looking at behavioural disturbance due to the potentially large impact ranges. Such studies are important because the information provided can be used to assess the cost to society of resulting mitigation measures. This is even more important as some of the structures now being considered by regulators, such as offshore wind farms, result from efforts to reduce other adverse impacts on the environment.

Geographical and seasonal restrictions Noise impacts can be mitigated effectively through geographical and seasonal restrictions on sound production, thereby protecting times and locations critical to mating, breeding, feeding or migration. An example is the moratorium that the Spanish Ministry of Defence has maintained since 2004 on the use of sonar within 50 nautical miles of the Canary Islands, following stranding events involving beaked whales Ziphiidae (see also Weilgart 2007; OSPAR 2009a). Protection zones could be designated under the EU Habitat Directive (Natura 2000 sites). Yet we have to remember that marine species are highly mobile and that distributional shifts (see for example harbour porpoises in the North Sea between 1994 and 2005; Hammond 2006) might lead to incongruity between protected areas and their originally postulated conservation objective. It is also likely that noise produced in the vicinity of a protected area can impact receivers therein; in Europe this has to be addressed in specialised Appropriate Assessments.

Another form of spatial restriction is the application of safety zones to avoid ensonification of receivers at distances thought to be critical, e.g. causing injury. For example, the Joint Nature Conservation Committee of the UK (JNCC) advises an exclusion zone of 500 m for seismic survey operations (JNCC 2009a). Specially trained marine mammal observers are required to detect marine mammals within the safety zone and passive acoustic monitoring (PAM) can additionally be used to detect marine mammals at night or during averse sighting conditions (see JNCC 2009a). PAM technology will very likely become more advanced in identifying senders and also in supporting real-time monitoring; however, it can only monitor safety zones if the species of interest produces sound most of the time. This may not be the case (see for example Barrett-Lennard et al. 1996; see Weilgart 2007; Compton et al. 2008 for other issues with regards to soft-start).

Noise exposure criteria Criteria for noise exposure were set by the US National Marine Fisheries Service in 2003 at 180 dB re 1 μPa (rms) for cetaceans and 190 dB re 1μPa for pinnipeds. More recently a US group of experts have suggested modified criteria for three functional hearing groups of cetaceans (low-, mid- and high-frequency; see Sect. 14.2.2) and pinnipeds in air and underwater both for injury (PTS) and behavioural response (using a TTS criterion). Both pulses (single and multiple) and non-pulses were considered (Southall et al. 2007). For fish, Popper et al. (2006) and Carlson et al. (2007) proposed interim criteria for injury and TTS for pile driving for different hearing groups. We have to bear in mind that all of these criteria are provisional as no hearing studies have been undertaken for most species (i.e. they are based on extrapolation from one species in which hearing abilities have been measured). Recently, Lucke et al. (2009) found TTS in a harbour porpoise at much lower received levels than those postulated by Southall et al. (2007) (see also Kastelein et al. 2012).

The exposure criteria are set for received sound pressure levels; however, these will be very difficult to establish as measurements might not be feasible in every case and thus modelling has to address site-specific transmission loss characteristics (see Madsen et al. 2006a for some values in the North Sea and Baltic). To avoid this issue Tasker et al. (2010) proposed criteria based on source rather than received levels. Most of the criteria address a limited set of sound types (e.g. one category ‘pulse’ or activity ‘pile driving’) and extrapolating criteria across sound types might not be appropriate due to differences in sound characteristics. We should also bear in mind that for many fish species particle motion rather than pressure is the appropriate stimulus (see Popper and Fay 2011). Multiple exposures to sounds below threshold can lead to injuries or TTS depending on the duty cycle and the overall dose received over time (Kastelein et al. 2012). Cumulative exposure criteria have to be considered that are very different from those for single strikes (Carlson et al. 2007; Southall et al. 2007). It is therefore evident that current suggestions will have to be revised when new data becomes available.

‘Soft-start’ methods (in which strike amplitude is slowly increased over several strikes in order to provide receivers with an opportunity to leave the area before adverse levels are reached) are applied before seismic surveys and pile driving operations. It is not yet known if this method achieves the desired effect (see Madsen et al. 2006b; Miller et al. 2009). Tools which emit an aversive signal so that receivers move out of the potential injury zone (acoustic management devices or ‘pingers’) have been used inter alia during offshore wind farm construction activities in Denmark (Tougaard et al. 2006). In general, these devices are quite effective in displacing receivers out of the immediate zone of danger (e.g. Culik et al. 2001). However, this of course raises the issue that one ‘evil’ (the effects of exposure to pile driving and other sounds) is replaced by another (the effects of exposure to pingers).

Technical mitigation measures Engineering solutions focus on the reduction of sound at the source. Examples are cofferdams, bubble curtains or plastic sleeves around pile drivers (see for example Nehls et al. 2007; Lucke et al. 2011). There are also attempts to develop alternative ship propeller designs, and methods to improve wake flows into ship propellers (Renilson 2009).

6.2 Management of Terrestrial Noise

6.2.1 Legal Instruments

The protection that exists for non-human animals generally relies on regulations that protect habitat and prevent disturbance, rather than regulations that protect against noise per se. These regulations fall under a wide range of legislation and policies, such as those applying to protected areas, critical habitat and environmental assessments. Extending these tools to incorporate the effects of noise requires broader interpretations of habitat and disturbance than are usually applied. Even then, such broader interpretations usually come to the fore only for the most intense effects of noise that result in obvious disturbance, such as escape behaviour or interruptions of foraging bouts (e.g. Pepper et al. 2003). For example, the Migratory Bird Convention Act, one of the most powerful tools for wildlife protection in North America, prohibits the disturbance of birds or their nests. Thus the prohibitions of the act can be applied to regulate noise that startles or disturbs birds, and can result in recommended set-backs based on flushing distances or the point where activities are interrupted. Even these blunt tools for measuring disturbance are generally applied only to species that are considered to be particularly sensitive to noise, such as waders and raptors (e.g. Rodgers and Smith 1995; Bautista et al. 2004; Wright et al. 2010). The Act offers no protection for the subtler, but spatially and temporally more extensive, effects of noise such as those that could affect communication, unless one takes a very broad view of the meaning of “disturbance”. Similarly, Canada’s Species at Risk Act, like much such legislation around the world, protects the critical habitat of endangered species, i.e. the habitat essential for the species’ survival or recovery. Again, in a broad interpretation, preventing disturbance may include establishing buffer areas around nest sites (Pepper et al. 2003), but would not apply to noise at the levels that are high enough to mask signals but too low to interrupt other activities. Even these gains in protection against more extreme noise levels are hard-won, because the sound environment that animals depend on is, somewhat understandably, not considered as integral to an animal’s habitat as are the more localised and tangible resources of food and shelter.

Given the lack of regulations and policies that target effects on non-human animals, it is worthwhile considering whether the many measures protecting humans from harmful noise are useful for protecting other animals. Setting aside regulations of extreme levels that can cause hearing impairment, anthropogenic noise is regulated in most countries based on two criteria: whether it interrupts human activities, such as conversation, and whether it is annoying to a certain percentage of the population (OECD 1995). Although annoyance seems to be a subjective way to gauge impacts on humans, it has been adopted because it is a readily obtained metric of objective impacts that underlie it, which include a wide range of cognitive and health impairments (Guski et al. 1999; Ouis 2001; Moudon 2009). Where these regulations apply, such as in residential areas, they might well be adequate for protecting most animals from deleterious effects of signal masking. Regulations protecting people from bothersome noise vary, but generally apply at levels around 60 dB(A) or less. Lab studies show signal masking at these levels (Chaps. 3, 6 and 8), but in the field receivers can readily overcome such masking (e.g. by turning their heads or moving slightly—Dooling and Popper, 2007). This result suggests that higher levels of noise are needed to mask signals. Of course, there is considerable doubt about this conclusion, especially because masking effects should vary considerably across species. Also, the regulations in place for human health and well-being offer no protection in areas where people are absent, which is often where species of conservation concern occur (Blickley and Patricelli 2010). As a remedy to these shortcomings, Blickley and Patricelli (2010) suggest that, because the effects of noise vary among species, a list of standards be developed that is specific to given species or groups of species. While it seems far too early to come up with such a list, given all the unknowns raised in this chapter and throughout this book, some such attempts have been made, mostly for local applications (e.g. Reijnen et al. 1995, 1997; Barber et al. 2011; Patón et al. 2012). As noted above, the measures in place for terrestrial species lag far behind those that are now routine for marine species, where much less is known but the potential harm is so severe that regulators have been more inclined to take a precautionary approach (see below).

6.2.2 Terrestrial Mitigation Measures

What can be done to mitigate the effects of noise on terrestrial animal communication? Of course, the most effective mitigation is to simply keep noise sources away from animals, a measure taken frequently for humans but rarely for non-human animals. Simply changing the timing of noise-related activities would be effective for species that signal preferentially at a certain time of day. This is analogous to the geographical and time restrictions discussed for the marine environment.

Other options for reducing noise are now available. For example, better noise reduction for car engines and tyres, quieter asphalt for roads, improved noise barriers and road siting will all contribute to reduce traffic noise (Makarewicz and Kokowski 2007; van Langevelde and Jaarsma 2009). However, most of these measures are directed at human complaints about interference with conversation (<3000 Hz) or disturbance from low-frequency rumbles (<100 Hz) and as such are more likely to fit the frequency range of large terrestrial mammals rather than birds or insects. (An exception might be high-pitched whines that humans also find annoying.) There are several engineering measures that could be employed to reduce noise that could affect animal signals, including traffic speed and flow control, quieter aircraft engines and more effective silencing of construction machinery and plant. We know of no instances in which any such engineering measures have been deployed to address conservation concerns. As Slabbekoorn and Ripmeester (2007) point out, however, small changes to current mitigation methods could have a disproportionately beneficial effect for the frequency range of most bird song (see also Halfwerk et al. 2011). Slight increases in barrier height can block road noise from the tree canopy where birds communicate, and angles or absorbent material on barrier top edges can absorb traffic noise (Slabbekoorn and Ripmeester 2007). One of the most effective noise absorption barriers is vegetation, which also provides additional habitat for the birds (Slabbekoorn and Ripmeester 2007) and an aesthetically pleasing driving experience for people.

7 Conclusion

We consider that the potential for anthropogenic noise to adversely impact conservation is amply demonstrated in both marine and terrestrial environments. The key challenges are to establish whether this potential is realised and if so, to assess its effects relative to more commonly considered conservation issues such as habitat loss. A further challenge (in terms of the topic of this book) is to establish which effects of anthropogenic noise have relevance to conservation through impacts on communication.

There are several reasons why establishing the link between anthropogenic noise and conservation impacts will be difficult:

  • The evidence base that feeds into the management process is limited. Although arguably there is more information for marine than terrestrial animals, it is apparent that knowledge of hearing abilities is still limited. Similarly, although we have some good information on amplitude and frequency characteristics of marine sound sources, we have far less information on the complexity of the emitted sound field. This leads to huge uncertainties when calculating sound transmission and ultimately the levels and characteristics of sound at the receiver which determine noise impacts.

  • Behavioural effects are likely to be underestimated. Although traditionally seen as less significant impacts in risk management approaches, behavioural effects can be pervasive and linked to fitness consequences. The potential for adverse behavioural reactions to have significant conservation impacts can be seen in certain circumstances such as the strandings observed in beaked whales in response to active sonar (see Sect. 14.6.1.3).

  • A relatively poor understanding of how animals trade off costs and benefits. For example the benefits of staying in an area to gain access to food or mates might outweigh the costs of exposure to noise such as signal masking and threshold shifts. This approach has been more fully developed in the welfare literature where it is termed adaptive cost gauging (Barnard and Hurst 1996; Barnard 2007) and it has been suggested as a complicating factor when interpreting response to playback (McGregor 2008).

  • There is little information on adverse fitness consequences of noise through effects on communication networks, information networks and soundscapes. However, such networks are likely to be as important as resource networks such as food webs.

The case for gathering more information on effects of noise on conservation is clear. We require more controlled studies looking at the nature, extent and persistence of behavioural responses to noise to assess likely population level consequences of acoustic disturbance. We also require more studies of masking and the efficacy of mitigation measures. However, the increase in anthropogenic noise means that advice is needed now, before such extra information has been gathered. One response to this need is to develop rules of thumb guidelines to be used in the absence of behavioural studies of the species of concern and in the absence of masking studies for the site at risk. A response to the absence of acoustic mapping tools (e.g. Barber et al. 2011) for non-human animals is to apply, as an interim measure, tools for acoustic mapping in humans (e.g. residential layouts that optimise disturbance zones, Theobald et al. 1997; and distance-based identification of open country quiet areas, Votsi et al. 2012; see Nega et al. 2013 for an application to urban park planning).

A continuing challenge will be that regulators rarely have the psychological or behavioural ecology background to understand the effects noise has on fitness through communication. Similarly, scientists who understand how signal masking can incur psychological costs such as distraction and physiological costs such as elevated heart rate that can translate into fitness costs, are unlikely to understand issues of legislation implementation and enforcement.

To conclude, the current perception of anthropogenic noise in biological conservation is similar to that of infectious diseases (Smith et al. 2009). In particular, neither noise nor disease has been cited as the principle cause of species extinction, but both can contribute to the effects of other drivers such as habitat loss. Also our current state of knowledge makes their relative contribution to species extinction hard to assess. It is encouraging that the role of infectious disease in conservation is beginning to be recognised. The same may be true for anthropogenic noise, where a recent report of the impact of low-frequency sound on cephalopods (André et al. 2011) led to an editorial in New Scientist (2011). We hope that this chapter further raises the profile of anthropogenic noise in conservation.