NOTE TO READER: The text below was developed by the CIBRT and reproduces information readily available in other reports. In Sections II.B.6 and III.A.3 of this document, we provided information sufficient to justify the recovery criteria and actions addressing noise. Additional CI beluga hearing, vocalization, and noise information follows.
Having evolved from land based mammalian ancestors, cetaceans have inherited an ear that was first adapted to hearing sounds through air, which then readapted to receiving sounds through water (Thewissen and Hussain 1993). Cetaceans have retained the ear drum, ossicles, Eustachian tube, and middle ear structures, including an air-filled cavity within the temporal bone or bulla, connected by the Eustachian tube to the nasal cavity for equalization of pressure during dives (Gingerich et al. 1983; Thewissen and Hussain 1993; Ridgway et al. 2001). As a consequence, it was hypothesized that cetacean hearing might attenuate at depth due to the increased air pressure and density of air in the middle ear, which might make them less susceptible to the impacts of loud underwater sounds. This has been shown not to be the case in belugas, as their hearing was determined to be as good at 300 m (984 ft) depth as at the surface (Ridgway et al. 2001). This is consistent with the theory that sound may be received through odontocetes’ lower jaw “acoustic window” and transmitted directly to the ear (Norris 1968; Cranford et al. 2008). In fact, a study conducted with a captive beluga showed that the most efficient hearing pathway is from the rostrum tip (Figure E1), and may indicate that there are acoustic fat channels that begin at the beluga rostrum tip that effectively guide sound to the inner ear (Mooney et al. 2008). To date, belugas are the only odontocetes known to hear from the rostrum tip, although a similar pathway has been recently proposed for Cuvier’s beaked whale (Cranford et al. 2008). This feature probably gives belugas greater directional hearing abilities than other odontocetes. It is possible that belugas’ unfused vertebrae, which allows for a highly movable head, facilitates increased hearing directionality.
Beluga Echolocations and Vocalizations
Belugas utilize an alternative echolocation strategy compared with the bottlenose dolphin when performing identical echolocation tasks (Turl and Penner 1989; Rutenko and Vishnyakov, 2006). Bottlenose dolphins will emit a click and wait until the echo returns before emitting the next signal (i.e., the inter-click interval is always greater than the two-way time travel). Belugas appear to be able to transmit, receive, and process signal packets simultaneously, with the first click about two dB higher than the other clicks that follow, which may serve to identify the beginning of each signal packet (Turl and Penner 1989).
The first vocal repertoire description of belugas was made in the Canadian high Arctic by Sjare and Smith (1986a). They classified a total of 807 tonal calls (whistles) into 16 contour types and some 436 pulsed calls into three major categories that they describe as “click series,” “pulsed tones,” and “noisy vocalizations.” Subsequent studies have obtained varied results. The vocalizations of adult male beluga groups in Svalbard, Norway, were subjectively classified into 21 call types, which were dominated by a variety of whistles (Karlsen et al. 2002). Karlsen et al. (2002) highlighted the highly graded nature of these beluga calls, as one “call type” can merge into another type with very subtle changes, making the classification very challenging. A reproductive gathering of belugas in the White Sea, Russia, has been the subject of several repertoire studies (Belikov and Bel’kovich 2001, 2003; Bel’kovich and Kreichi 2004; Belikov and Bel’kovich 2007, 2008). Whistle-like signals were found to comprise approximately 10% of the total vocal production of this whale group. Of these, 750 signals were divided into 43 classes (Belikov and Bel’kovich 2001) with at least 16 whistle types (Belikov and Bel’kovich, 2007) and vowel-like signals and pulsed signals (Bel’kovich and Kreichi 2004; Belikov and Bel’kovich 2008).
The response of a decrease or cessation in acoustic activity has been observed in both captive and free-ranging belugas (Morgan 1979; Lesage et al. 1999; Karlsen 2002; Van Parijs et al. 2003; Castellote and Fossa 2006) and free-ranging narwhals (Finley 1990); the response has been associated with threat, startle, fright, alarm, or stress contexts and interpreted as a survival strategy to avoid detection by predators (Schevill 1964; Fish and Vania 1971; Morgan 1979; Finley 1990; Lesage et al. 1999). A broad band pulsed call labelled “Type A” (Vergara and Barrett-Lennard 2008) was identified as a contact call between mothers and their calves in a captive environment. It is thought that these calls, both in captivity and in the wild, function to maintain group cohesion, and the variants shared by related animals are used for mother-calf recognition (Vergara et al. 2010). The only study on vocal development in belugas suggests that neonates only produce pulse trains before they acquire rudimentary whistles at two weeks of age (Vergara and Barrett-Lennard 2008), although this is based on observations of one captive male beluga calf. Similarly, sound production of another neonate captive beluga consisted exclusively of low-frequency, short duration pulse trains that were not part of the adult’s repertoire (Castellote et al. 2007). Despite differences in populations of origin, captive facilities, health, and in acoustic context, the sound production observed in these two neonate belugas suggests a species-specific pattern of developmental stages in sound acquisition. Whether these observed captive neonate vocalization characteristics may prove useful in detecting the presence of wild neonates is still to be determined.
The most recent study on beluga social signals (Vergara et al. 2010) emphasized the two persistent problems commonly encountered in the study of animal communication: first, the great variability in the physical features of the sounds, with general call types grading into each other (Recchia 1994), introduces great uncertainty in the categorization schemes; second, the inherent difficulty in categorizing sounds that are biologically meaningful without testing how belugas themselves perceive or use them (Tyack and Clark 2000). Despite the challenges, some progress has been made in the attempt to correlate vocalization rate and call type with specific beluga behavioral states.
Effects on Beluga Hearing and Behavior from Anthropogenic Noise
There is an extensive body of literature regarding the effect of anthropogenic noise on marine mammal behavior. Most of the studies addressing this problem have used behavioral attributes such as changes in site fidelity, dive patterns, swimming speed, orientation of travel, herd cohesiveness, and dive synchrony to indicate possible disturbance or stress caused by noise (Richardson et al. 1995). However the current knowledge of the effects of anthropogenic noise to marine mammal acoustic behavior is more limited, and only a few studies have focused on belugas.
Their high auditory sensitivity, wide frequency bandwidth, and dependence upon sound to navigate, communicate, and find prey make belugas vulnerable to noise pollution. Noise pollution may mask beluga signals, or if intense, may lead to temporary or permanent hearing impairment (Awbrey et al. 1988; Finley 1990; Green et al. 1994; Richardson et al. 1995, 1988). Exposure to intense sound can produce an elevated hearing threshold, referred to as a threshold shift (TS). If the threshold later returns to normal it is considered a temporary threshold shift (TTS), but if not, it is considered a permanent threshold shift (PTS). Studies of TTS and PTS have helped to establish noise exposure limits in humans. There are no PTS data for cetaceans, yet a few studies have attempted to establish the TTS for belugas (Finneran et al. 2000, 2002a; Schlundt et al. 2000). Finneran et al. (2000) simulated sounds resembling signatures of underwater explosions from 5 or 500 kg HBX-1 charges at ranges from 1.5 to 55.6 km (0.9–34.5 mi), and while the simulated sounds were not intense enough to affect the beluga hearing capabilities, sound levels simulating explosions of 500 kg (1,102 lb) at 1.9 km (1.2 mi) and closer did disrupt the behavior of the belugas. However, they found no TTS after exposure to the highest level the underwater sound projector could produce. Finneran et al. (2002a) reported behaviorally measured TTS in a bottlenose dolphin and a beluga exposed to single pulses from a seismic water gun. Also, Schlundt et al. (2000) performed a study exposing five bottlenose dolphins and two belugas (same individuals as Finneran’s studies) to intense 1 second tones at different frequencies. The resulting levels of fatiguing stimuli necessary to induce 6 dB or larger masked TTSs were generally between 192 and 201 dB re 1 microPascal (µPa). Dolphins began to exhibit altered behavior at levels of 178–193 dB re 1µPa and above; belugas displayed altered behavior at 180–196 dB re 1 µPa and above. At the conclusion of the study, all thresholds were at baseline values. Results of this study indicate that at least these two odontocetes species are susceptible to TTS, but that they seem to recover from at least small levels of TTS.
A number of studies have examined other characteristics of beluga hearing. Johnson (1991) analyzed hearing thresholds, bandwidths, and integration times (basic descriptive parameters of the cetacean sonar system) for single pulsed tones and multiple pulsed tones of 60 kHz in the presence of noise. He found negative correlations between hearing thresholds and pulse repetition rate with abrupt 5–6 dB steps, and linear correlations between pulse repetition rate and integration times. The author related the abrupt hearing steps to a change in the echolocation strategy based on target distance, as has been described in some beluga echolocation studies, and is discussed in the next section. This result, together with a variable integration time and a constant system bandwidth of 1,000 Hz (much lower than previously reported), led the author to suggest that beluga sonar systems could not be fully described by a single filter model. In essence, this conclusion was a technical appreciation of the complexity of the beluga biosonar system. Finneran et al. (2002c) analyzed beluga sensitivity to acoustic particle motion, which is one of the two physically linked components of sound in water (together with pressure waves), and the main feature detected by all fish species (Fay and Popper 1975). Results suggested that the two belugas tested responded to changes in the acoustic pressure alone and were not able to use acoustic particle motion cues.
The possibility that noise conditions might mask the ability of animals to hear and decipher specific sounds has been studied in belugas in order to understand the potential impacts of anthropogenic noise on belugas. When a tonal signal is played in a broad spectrum of white noise (noise with equal loudness across all frequencies), only the noise energy in a relatively narrow band on either side of the tone frequency is effective in masking the signal, and the rest of the noise spectrum contributes little or nothing to the masking effect. Johnson et al. (1989) analyzed this feature in the hearing of a beluga in a wide frequency range (40–115 kHz) and found that the whale’s ability to detect the signal in noise was slightly better than results previously reported for bottlenose dolphins. Erbe et al. (1999) and Erbe (2000) analyzed the effect of masking of beluga calls by exposing a trained beluga to icebreaker propeller noise, an icebreaker’s bubbler system, and ambient Arctic ice cracking noise, and found that the latter was the least problematic for the whale detecting the calls. Finneran et al. (2002b) analyzed the ability of a beluga to detect acoustic signals in noise. A primary feature of the auditory system in these animals is the ability to resolve a complex sound into its individual frequency components by a set of auditory filters, and the filter shape and size affect the loudness and detectability of complex sounds and broadband signals (Scharf 1970). The authors analyzed 20 and 30 kHz pure- tone underwater hearing thresholds in one beluga and two bottlenose dolphins in the presence of broadband noise at two intensities: 90 and 105 dB re 1 µPa2/Hz. The filter shapes obtained for the dolphins and beluga were similar, but the filter width was consistently smaller for the beluga, conferring better ability to detect acoustic signals in noise.
Sheifele et al. (2005) studied a population of belugas in the SLE to determine whether beluga vocalizations showed intensity changes in response to shipping noise. This type of behavior has been observed in humans and is known as the Lombard vocal response (Lombard 1911). Sheifele et al. (2005) demonstrated that shipping noise did cause belugas to vocalize louder (Figure E2). The acoustic behavior of this same population of belugas was studied in the presence of ferry and small boat noise. Lesage et al. (1999) described more persistent vocal responses when whales were exposed to the ferry than to the small-boat noise. These included a progressive reduction in calling rate while vessels were approaching, an increase in the repetition of specific calls, and a shift to higher frequency bands used by vocalizing animals when vessels were close to the whales. The authors concluded that these changes, and the reduction in calling rate to almost silence, may reduce communication efficiency, which is critical for a species of a gregarious nature. However, the authors also stated that, because of the gregarious nature of belugas, this “would not pose a serious problem for intraherd communication” of belugas given the short distance between group members; the authors further concluded a noise source would have to be very close to potentially limit any communication within the beluga group (Lesage et al. 1999).
The fact that SLE belugas alter their vocal behavior by increasing the intensity or repetition rate, or by shifting to higher frequencies when exposed to shipping noise (from merchant, whale- watching, ferry and small boats), is indicative of an increase of energy costs (Bradbury and Vehrencamp 1998). If noise exposure is chronic, long-term adverse energetic consequences could occur for belugas, as it has been shown for birds (Oberweger and Goller, 2001). Chronic noise exposure could also increase stress levels for CI belugas, as has been shown in North Atlantic right whales (Rolland et al. 2012). Definitively linking adverse energetic consequences and chronic stress responses to detrimental health effects in belugas or other cetaceans is extremely difficult because of the logistics of studying free-swimming whales and the inability to conduct a controlled study. However, a large body of literature has demonstrated that chronic stress can lead to detrimental effects on health and reproduction across a variety of vertebrate taxa (Rolland et al. 2012). Both the degradation of the beluga acoustic communication and echolocation space, as well as the noise-induced chronic increase of signaling costs and stress, could lead to negative biological consequences at the population level. Even if these consequences are not yet well understood, there is sufficient evidence to suggest that the reproductive success and survival of cetaceans can be negatively impacted by noise (NRC 2000, 2003, 2005; Cox et al. 2006; Southall et al. 2007; Clark et al. 2009; Payne and Webb 1971; Tyack and Clark 2000).
While exhibiting a Lombard response provides a mechanism for animals to cope with varying levels of noise, the need for and use of this response suggests that the animal is attempting to cope with noise levels that are near a point where masking will occur. The effect of shipping noise in the acoustic environment of the endangered SLE beluga was studied recently by Gervaise et al. (2012) in the lower SLE. Noise from a car ferry line as well as a seasonal whale watching fleet were analyzed. The study found both beluga communication and echolocation bands were dramatically affected by these noise sources. Based on the background noise levels, spectra, and periodicity reported and based on the assumption of no behavioral or auditory compensation, beluga communication and echolocation signals could be masked 50% of the time with a reduction of potential communication ranges to less than 30% of their values under natural ambient noise conditions. Similarly, echolocation could be reduced to 80% of their range under natural ambient noise conditions. The study concludes that noise from these sources could easily limit long-range communication (in the order of 1–2 mi [1.6–3.2 km]) among scattered individuals or pods and affect echolocation efficiency in all exposed belugas.
There are some documented beluga spatial displacements caused by loud sources of noise. Two different research teams and data from several years showed that belugas typically avoided icebreakers at distances of 35–50 km (22–31 mi), at the point where they could probably just detect them. They travelled up to 80 km (50 mi) from the ship track and usually remained away for 1–2 days (Finley et al. 1990, Cosens and Dueck 1993). When drilling sounds were played to belugas in industry-free areas, the belugas only showed a behavioral reaction when received levels were high (Richardson et al. 1997). Belugas have been observed to show startle responses when drilling noises were played with a received level greater than or equal to 153 dB re 1 μPa. Considerable displacements have also been suggested for noise from air guns typically used during seismic surveys. One seismic survey in the Canadian Beaufort Sea determined behavioral reactions of belugas occurred when two 24 gun arrays of 2,250 in3 were operating (Miller et al. 2005). Results of the analysis of the differences between vessel-based and aerial-based beluga sighting distributions provided evidence of reactions of belugas to seismic operations at distances above 20 km (12.4 mi), beyond the effective visual range of the MMOs on the seismic vessel (Miller et al. 2005). Aerial surveys conducted in the southeastern Beaufort Sea in summer found that sighting rates of belugas were significantly lower at distances of 10–20 km compared with 20–30 km from an operating airgun array (Miller et al. 2005). The low number of beluga sightings by marine mammal observers on the vessel seemed to confirm there was a strong avoidance response to the 2250 in3 airgun array; however, it is unclear if the observed movement of the belugas was a direct consequence of the seismic surveys or related to the natural offshore migration at that time of year. More recent seismic monitoring studies in the same area seem to confirm that the apparent displacement effect on belugas extends farther than has been shown for other small odontocetes exposed to airgun pulses (e.g., Harris et al. 2007).
Similarly, aerial survey results from another seismic (array specifications unknown) and exploratory drilling activity conducted in the same area and same season in 2007 to 2008 showed belugas widely distributed offshore during the operation period, yet rarely sighted from seismic ships. This was interpreted as a tendency to temporarily avoid areas of seismic activity by greater distances than the range covered by MMOs on board seismic vessels (Harwood et al. 2010). However, the authors highlighted the temporary nature of these displacements, as belugas were observed back in the seismic operation area within days after the end of the seismic operations.
Belugas have been shown to have greater displacement in response to a moving sound source (e.g., air gun activity on a moving vessel) and less displacement or behavioral change in response to a stationary sound source. The presence of belugas has been documented within ensonified zones of industrial sites near platforms and stationary dredges, and the belugas did not seem to be disturbed by the activity (Richardson et al. 1995).