Hearing sensitivity in context: Conservation implications for a highly vocal endangered species

Global Ecology and Conservation 6 (2016) 121–131

Contents lists available at ScienceDirect

Global Ecology and Conservation journal homepage: www.elsevier.com/locate/gecco

Review paper

Hearing sensitivity in context: Conservation implications for a highly vocal endangered species Megan A. Owen a,∗ , Jennifer L. Keating a , Samuel K. Denes b , Kathy Hawk c , Angela Fiore c , Julie Thatcher c , Jennifer Becerra c , Suzanne Hall a , Ronald R. Swaisgood a a

Institute for Conservation Research, San Diego Zoo Global, 15600 San Pasqual Valley Road, Escondido, CA 92027-7000, USA

b

Department of Biology, Syracuse University, 107 College Pl., Syracuse, NY 13244, USA

c

Collections and Husbandry Science, San Diego Zoo Global, PO Box 120551, San Diego, CA 92112, USA

article

info

Article history: Received 3 November 2015 Received in revised form 22 February 2016 Accepted 23 February 2016 Available online 17 March 2016 Keywords: Acoustic ecology Giant panda Hearing sensitivity Noise Psychoacoustic

abstract Hearing sensitivity is a fundamental determinant of a species’ vulnerability to anthropogenic noise, however little is known about the hearing capacities of most conservation dependent species. When audiometric data are integrated with other aspects of species’ acoustic ecology, life history, and characteristic habitat topography and soundscape, predictions can be made regarding probable vulnerability to the negative impacts of different types of anthropogenic noise. Here we used an adaptive psychoacoustic technique to measure hearing thresholds in the endangered giant panda; a species that uses acoustic communication to coordinate reproduction. Our results suggest that giant pandas have functional hearing into the ultrasonic range, with good sensitivity between 10.0 and 16.0 kHz, and best sensitivity measured at 12.5–14.0 kHz. We estimated the lower and upper limits of functional hearing as 0.10 and 70.0 kHz respectively. While these results suggest that panda hearing is similar to that of some other terrestrial carnivores, panda hearing thresholds above 14.0 kHz were significantly lower (i.e., more sensitive) than those of the polar bear, the only other bear species for which data are available. We discuss the implications of this divergence, as well as the relationship between hearing sensitivity and the spectral parameters of panda vocalizations. We suggest that these data, placed in context, can be used towards the development of a sensory-based model of noise disturbance for the species. © 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Contents 1. 2.

∗

Introduction............................................................................................................................................................................................. 122 Methods ................................................................................................................................................................................................... 123 2.1. Study subjects ............................................................................................................................................................................. 123 2.2. Background noise measurements and test space ..................................................................................................................... 123 2.3. Stimuli presentation and data acquisition ................................................................................................................................ 124 2.4. Subject training ........................................................................................................................................................................... 124 2.5. Threshold determination and data analysis.............................................................................................................................. 124

Corresponding author. Tel.: +1 619 231 1515. E-mail address: [email protected] (M.A. Owen).

http://dx.doi.org/10.1016/j.gecco.2016.02.007 2351-9894/© 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

122

3.

4.

M.A. Owen et al. / Global Ecology and Conservation 6 (2016) 121–131

2.6. Digital audio recordings and analysis of spectral content of vocalizations ............................................................................ 125 Results...................................................................................................................................................................................................... 126 3.1. Hearing thresholds ..................................................................................................................................................................... 126 3.2. Comparison with polar bear audiogram ................................................................................................................................... 126 3.3. Relationship between hearing sensitivity and acoustic signals .............................................................................................. 127 Discussion ................................................................................................................................................................................................ 127 Acknowledgments .................................................................................................................................................................................. 130 References................................................................................................................................................................................................ 130

1. Introduction The large-scale transformation of the acoustic landscape by human activities may have significant consequences for wildlife (Francis, 2015; Francis and Barber, 2013; Shannon et al., 2015), as noise can readily permeate regulatory boundaries and can be detected in even the most remote protected areas (Barber et al., 2010). While the nature and severity of noise impacts is dependent upon the interaction of acoustic, biological, and abiotic environmental factors, this complexity should not preclude the development of predictive, mechanistic models that integrate key aspects of species biology and environmental parameters (Southall et al., 2007). Given the rapidly increasing footprint of noise-generating human activities, estimating impacts, predicting consequences, and developing targeted mitigation strategies may be essential to species and habitat conservation (Francis and Barber, 2013). Hearing is the selective filter through which animals integrate acoustic signals, cues and inadvertent environmental sounds that promote successful reproduction and survival (Blanchet et al., 2010). However, hearing can also be a nonselective entryway through which audible noise may influence or impede communication (Bee and Swanson, 2007), or act as a disturbance (Shannon et al., 2015); ultimately influencing successful reproduction or survival. Thus, hearing sensitivity is a fundamental aspect of an animal’s vulnerability to anthropogenic noise (Gerstein and Gerstein, 1999; Hastie et al., 2015). Because of this fundamental role in noise perception, data describing a species’ hearing capacity should be a key component of models that predict noise impacts (Erbe and King, 2009; Gerstein and Gerstein, 1999), and efforts should be made to integrate audiometric data with data describing other aspects of species’ acoustic ecology and the environmental soundscape (i.e., consider audiometric data ‘in context’, Ellison et al., 2011 and Southall et al., 2007). Here we model the integration of hearing sensitivity and the spectrographic characteristics of vocalizations of an endangered Ursid, the giant panda (Ailuropoda melanoleuca); exploring potential areas of acoustic vulnerability. The Ursidae share a number of ecological and life history characteristics (Stirling and Derocher, 1990), however, they also diverge widely in the topography of their habitat – a factor which can shape the evolution of both signal design and hearing sensitivity (Morton, 1975) – and the use of acoustic signals for communication and survival (Jackson et al., 2010). Together, these factors suggest that hearing capacity may be varied among bear species, especially for those exhibiting dramatic differences in these characteristics and in the relative evolutionary distance between them (Kutschera et al., 2014; Stirling and Derocher, 1990). From a conservation perspective, this is important because differential hearing capacity will impact both the susceptibility to noise disturbance (Delaney et al., 1999), and the validity of extrapolating confamiliar hearing capacity from one species to the next. To date, the polar bear (Ursus maritimus), a relatively recently-evolved species of bear, is the only other Ursid for which audiometric data are available (Owen and Bowles, 2011). The endangered giant panda (Ailuropoda melanoleuca) relies on acoustic communication for successful reproduction (Kleiman and Peters, 1990). Courtship and breeding are coordinated through multimodal signaling, however acoustic signals are prominent, especially during the peri-ovulatory period (Owen et al., 2013). The vocal repertoire of adult giant pandas ranges from intensity-graded agonistic vocalizations (Nie et al., 2012) to information-rich affiliative vocalizations (Kleiman and Peters, 1990). For example, characteristics of female chirps vary according to female reproductive stage (Charlton et al., 2010a), and characteristics of bleats contain information about male androgen levels and size (Charlton et al., 2011, 2012, 2009b), female age (Charlton et al., 2009b) and the identity of male and female callers (Charlton et al., 2009a,c). Male copulation calls convey information regarding mating success (Keating, 2011). Cubs emit a range of vocalizations that elicit discrete behavioral responses from mothers (Baotic et al., 2014) and may be essential for cub survival. Indeed, the vulnerability of highly altricial cubs during the denning period is well documented (Zhu et al., 2001), and den abandonment and cub mortality as a result of disturbance are of concern for all conservation dependent species of bear (Linnell et al., 2000). The majority of free-ranging pandas inhabit wildlife reserves (State Forestry Administration, 2015), however, human pressures on habitat are intense and increasing. Further, climate projections indicate that within 50 years, these reserves may no longer hold suitable bamboo stands for giant pandas and populations may shift outside reserve boundaries in search of appropriate habitat (Tuanmu et al., 2012). These range shifts will increase the potential for giant pandas to be in increased proximity to human activities, thus exposing them to a suite of potentially disturbing acoustic stimuli (Zhu et al., 2013). We used behavioral psychoacoustic techniques to measure hearing thresholds across frequencies and generate the first comprehensive audiogram for the species. We compare our findings to the audiogram of the polar bear (Owen and Bowles, 2011), to the relative spectral energy of adult and cub vocalizations, and to published data of the fundamental frequencies of cub and adult vocalizations (Baotic et al., 2014; Keating, 2011). We also discuss the implications of the panda’s hearing

M.A. Owen et al. / Global Ecology and Conservation 6 (2016) 121–131

123

Table 1 Study subjects and extent of testing. ID

371 415 596 634 794 a

Sex

F M F F M

Birth year

1991 1990a 2005 2007 2009

Provenance

Captive Wild Captive Captive Captive

Testing duration: Start date

End date

2009.06.26 2011.01.02 2009.06.29 2010.05.23 2011.10.19

2011.11.12 2012.03.13 2010.09.18 2010.09.19 2012.12.15

Freqs. tested

Thresholds measured

Presentations

19 20 20 11 8

194 200 184 49 40

1177 1622 3351 1140 831

Estimated birth year (±1 year) based on rescue history from Gansu province, PRC.

Fig. 1. Schematic diagram of sound-attenuated test space, data recording equipment and the relative positions of research staff and study subject during testing. *Attenuators were only used to lower tone amplitude at certain frequencies.

capacity relative to published findings that demonstrate the information bearing content of the resonance frequencies of the vocal tract, i.e., ‘‘formants’’ previously documented in adult bleats (Charlton et al., 2009b, 2010b) and that we describe here in growls. Formants do not appear to be present in the chirp vocalization. Further, we discuss the management implications of these data and the necessity of incorporating sensory data into predictive models of disturbance for this and other conservation dependent species. 2. Methods 2.1. Study subjects We measured hearing in 3 female (1 adult and 2 sub-adult) and 2 male (1 adult and 1 sub-adult) giant pandas. All subjects were housed at the San Diego Zoo, and research was conducted between 2009 and 2013. Because all bears were in good health at the time of testing, had never been exposed to ototoxic medication nor to exceptionally high noise levels, we assumed that each panda’s hearing capacity was species typical and normal for their sex and age class. Logistical constraints and subject tractability resulted in a variable number of both experimental presentations and thresholds measured between the five individuals in our study, as well as the duration of the testing period (Table 1). However, for the less cooperative subjects (e.g., #794), we strategically chose to measure thresholds at frequencies that would enable us to visualize the overall shape of the audiogram. Our experiment was carried out in accordance with the Animal Behaviour Society Guidelines for the Use of Animals in Research (2012). Experimental protocols were approved by the Zoological Society of San Diego’s Institutional Animal Care and Use Committee (IACUC). 2.2. Background noise measurements and test space Testing was conducted in a custom designed sound attenuated chamber. The test chamber was constructed in a modular fashion around a transport crate (Fig. 1). To reduce the amplitude of ambient noise in the test chamber we applied a layer of 7 cm thick acoustic foam panels that were attached to 2 cm thick sheets of plywood with acoustic glue. We assessed

124

M.A. Owen et al. / Global Ecology and Conservation 6 (2016) 121–131

50

40

30

20

10

0 100

1000

10000

Fig. 2. The degree of attenuation achieved by sound dampening of test space as demonstrated by ambient noise levels recorded across frequencies, both inside and outside of the test space.

the degree of sound attenuation achieved in the test chamber by measuring the background noise levels, using calibrated microphones and recording system, across frequencies, inside and outside the enclosure (Fig. 2). 2.3. Stimuli presentation and data acquisition Audiometric data were collected using an adaptive up–down staircase presentation order (Levitt, 1971). We delivered 500 ms shaped tones, and used a go/no-go response protocol to present test stimuli to the pandas. We gated presentations with a delay of 10 ms. We measured thresholds at 20 frequencies, ranging between 0.125 and 50 kHz for subjects 415 and 596, and we measured thresholds at 19 frequencies within this range for subject 371. For subjects 634 and 794 we measured thresholds at 8 and 7 frequencies respectively. All tones delivered during each experimental trial were recorded to ensure that no incidental noise was being generated by the system. We used an ACO-7013 1/2′′ microphone for frequencies between 0.125 and 31.5 kHz, and an ACO-7016 R 1/4′′ microphone for 40 and 50 kHz tones. Test tones were digitally generated using a National Instruments⃝ NI USB 6221 data acquisition unit (DAQ), and the system was calibrated using a Larson Davis CAL200 precision acoustic calibrator R AMP102 amplifier, which was in (pistophone). The DAQ was connected via the analog output jack to an AudioSource⃝ turn connected to a pair of speakers. Speakers were custom-built, and included full-range drivers and horn super tweeters to ensure simultaneous tone delivery. Speakers were positioned at ‘‘ear-level’’ in relation to the panda’s station position and were oriented towards the subject at a distance of 0.5 m. Tone presentations were controlled, and response data were R collected, using a custom-designed program written in LabView© software. We used a MacBook Pro⃝ laptop computer with R operating system partition, as an operator interface and for data storage (Fig. 1). a Windows⃝ 2.4. Subject training Because the psychoacoustic methodology we used is akin to a human behavioral hearing test, each subject required extensive operant conditioning training prior to participation in this study (Remington et al., 2012). While each panda was naïve to the training protocol prior to the initiation of our study, all bears had previously participated in other husbandry related operant conditioning, and had demonstrated good tractability. The ‘‘go/no-go’’ response protocol was achieved by training pandas to sit with their head facing forward and held stationary (‘‘no-go’’). From this stationary position, they were trained to break station (‘‘go’’) when a tone was broadcast and place their nose on a target located 40 cm to the right of the station. We measured thresholds by presenting pure tones and reducing tone-amplitude sequentially, in a stepwise fashion. We reduced tone amplitude in 6 dB increments for the first three presentations, and by 3 dB increments thereafter. To control for false positive responses (i.e., guessing), we included blank ‘‘catch’’ trials in our presentation order. Each trial had a 30% chance of being a catch trial, and if the rate of false positives (i.e., going to the target on a catch trial) was ≤30% during an experimental session, data collected during that session were retained for analysis. 2.5. Threshold determination and data analysis Behavioral thresholds were determined across frequencies by taking the average of the intensity level of the last correct detection, and the intensity level of the following missed detection (or the ‘‘reversal’’ point) (Szymanski et al., 1999).

M.A. Owen et al. / Global Ecology and Conservation 6 (2016) 121–131

125

Table 2 Number of recordings used in the analyses of spectral content and relative energy of vocalizations. Age class

Vocalization

Count

Cub Cub Cub Cub Cub Cub Adult Adult Adult

Birth croak Birth squall Birth squawk Croak Squall Squawk Bleat Chirp Growl

14 19 1 19 7 40 68 74 17

We used R statistical analysis software for all analyses (R Core Team, 2013). Data were examined for both normality using the Shapiro–Wilkes test and for homoscedasticity of variance using Bartlett’s K. When assumptions were not met, we used an appropriate non-parametric test. Estimates of upper and lower limits of functional hearing outside the range of testing were calculated via linear extrapolation (MacGillivray et al., 2013; Ramsier et al., 2012). To examine the consistency of hearing thresholds between male and female giant pandas, and adult and sub-adult giant pandas we fit linear mixed effects models, including sex, age and sex ∗ age as fixed factors (Pinheiro et al., 2015). We included panda ID as a random factor to avoid problems associated with temporal pseudoreplication due to the uneven repeated sampling on individual pandas. Correlations between hearing sensitivity and energy associated with giant panda vocalizations is demonstrated visually (sensu Ladich and Yan, 1998). We also visually inspected the shared frequency space between vocal formants (individually noted as ‘Fi’) present in adult vocalizations and hearing sensitivity. For the bleat vocalization, we use formant values published in Charlton et al. (2009a,b). For the growl vocalization, we measured frequency values (for a maximum of five formants) using PRAAT (version 6010). Parameter settings included a window length of 0.1 s and maximum frequency of 5.5 kHz. Formant frequencies were selected from the center of each call, and we report a mean formant value for all calls analyzed. We also examined and visually illustrate the overlap of spectral energy for 3 adult and 3 cub vocalizations relative to the measured region of good hearing sensitivity. We chose to assess hearing thresholds using comparable procedures to those outlined in Owen and Bowles (2011) in order to facilitate comparative evaluation of the thresholds we generated for the giant panda, with those generated for the polar bear (Reichmuth and Southall, 2012). We used a Wilcoxon Signed Ranks test to compare the mean thresholds for the giant panda presented here with published mean thresholds for the polar bear. We partitioned threshold data into relatively low, medium, and high frequency ranges.

2.6. Digital audio recordings and analysis of spectral content of vocalizations

Cub vocalizations were recorded from a single male giant panda cub born in 2009 at the San Diego Zoo. Cub vocalizations were captured using an ACO Pacific microphone installed in the ceiling of the birthing den, approximately 1.5 m from the cub’s typical location. We used a Marantz PMD660 solid-state digital recorder to capture recordings. Cub vocalizations were categorized according to qualitative acoustic parameters described in a widely used ethogram for giant pandas (Swaisgood & Owen, unpublished). We separately analyzed cub vocalizations recorded within 2 h of birth (birth squall, birth squawk and birth croak), and vocalizations recorded in the third month of life (squall, squawk and croak). Adult vocalizations (bleat, chirp and growl) were captured using a Sennheiser MKH 70 P48 shotgun microphone (sampling rate of 48 kHz, 16 bits of amplitude resolution) and Marantz PMD660 recorder at the China Conservation and Research Center for the Giant Panda (CCRCGP), Bi Feng Xia base, in Ya’an, China in 2009. Adult recordings included vocalizations from 13 individual giant pandas (Table 2). Spectral content of vocalizations was calculated for one-third octave bins using MATLAB (MathWorks, Torrance, CA, USA). The limits of frequency content were determined by identifying the first and last 1/3 octave band that had a signal to noise ratio (SNR) of at least 3 dB. One-third octave band levels were computed over the duration of a vocalization and relative levels were computed by subtracting the level of the maximum energy band from all other bands over the range of frequencies determined from SNR. Relative energy levels were linearly averaged over all recordings for each vocalization. We used PRAAT to visually inspect the adult growl and chirp vocalizations and identify formant frequencies. We used formants published in Charlton et al. (2009a,b) to compare the formants found in the bleat vocalization with the panda audiogram.

126

M.A. Owen et al. / Global Ecology and Conservation 6 (2016) 121–131

Table 3 Mean hearing thresholds (dB [re : 20 µPa]) (±SEM) measured for 5 giant pandas. Freq. (kHz)

Subject ID

Weighted grand mean

0.13 0.25 0.4 0.5 0.8 1 1.25 2 4 5 8 10 12.5 14 16 20 25 31.5 40 50

52.0 34.3 31.7 21.3 27.9 25.6

(0.79) (1.02) (0.92) (0.80) (1.25) (1.00)

8.0 25.2 20.5 19.4 7.5 3.5 5.3 8.3 16.6 18.9 19.5 36.2 32.5

(0.50) (0.83) (0.55) (2.51) (0.45) (0.45) (0.49) (0.66) (1.68) (0.40) (1.75) (0.80) (0.39)

371

415 42.3 37.3 33.3 18.3 27.6 25.0 25.1 14.3 21.6 19.5 15.4 4.2 3.8 7.1 10.7 12.1 26.2 25.6 36.4 27.1

596 (0.66) (0.92) (0.90) (0.80) (0.78) (0.81) (0.75) (1.02) (1.68) (1.10) (1.55) (1.64) (0.70) (0.60) (0.66) (1.47) (1.07) (0.46) (0.52) (1.03)

46.5 43.8 31.0 26.7 29.6 17.1 29.7 25.1 19.7 24.6 14.1 21.0 12.2 12.1 18.7 15.5 18.6 9.8 33.9 28.9

634 (0.00) (0.75) (1.50) (0.73) (3.25) (3.80) (2.04) (0.70) (2.48) (2.08) (1.57) (2.66) (1.13) (1.66) (1.68) (1.84) (1.19) (0.80) (1.77) (2.50)

41.5

794 (1.22)

24.2

(2.46)

11.3 29.8 18.0 14.7 14.9 5.0

(1.20) (2.56) (2.97) (4.37) (1.96) (0.56)

20.0 16.8

(1.50) (1.89)

28.6

(3.43)

46.0 42.3

(1.27) (2.13)

25.0

(0.93)

22.3

(2.23)

23.7

(2.33)

6.2

(0.75)

17.2

(1.11)

24.7

(1.42)

47.2 38.3 32.4 21.8 28.2 31.5 27.6 13.9 23.2 21.6 16.0 11.7 7.0 7.6 12.6 15.3 21.2 17.5 34.2 29.4

(1.14) (0.83) (0.62) (0.78) (0.97) (1.11) (1.26) (1.22) (1.06) (1.13) (1.12) (1.43) (1.03) (0.72) (1.03) (1.19) (0.85) (1.17) (0.74) (0.87)

Fig. 3. Comparison of giant panda and polar bear (Owen and Bowles, 2011) average hearing thresholds, including comparison between relatively low (≤1 kHz), medium (1.25–10 kHz) and high (>10 kHz) frequency thresholds (* <0.05).

3. Results 3.1. Hearing thresholds Our data show that pandas have good hearing sensitivity between 10.0 and 16.0 kHz, with best sensitivity documented between 12.5 and 14.0 kHz (Table 3). Low frequency hearing sensitivity declined below 0.25 kHz. Functional hearing, defined as a threshold of sensitivity of 60 dB or lower (Ketten, 2004), was documented into the ultrasonic range, with thresholds measured from 27.1 to 32.5 dB SPL at 50 kHz. Age did not influence hearing capacity, and we found good agreement among thresholds for both adult and sub-adult pandas (t = 0.64, P = 0.56). We also found that sex did not influence hearing capacity, and we determined that male and female thresholds were not significantly different (t = 0.63, P = 0.53). By extrapolation, we bound the range of functional hearing between 0.10 and 70.0 kHz. 3.2. Comparison with polar bear audiogram For the audiogram as a whole, giant panda hearing was similar to that of the polar bear (Owen and Bowles, 2011) (V = 66.5, p = 0.65). However, we found that there was a significant difference between the species above 10 kHz (V = 3, p = 0.04), but not in the low (V = 6, p = 0.81) and medium (V = 20, p = 0.36) frequency ranges (Fig. 3).

M.A. Owen et al. / Global Ecology and Conservation 6 (2016) 121–131

127

Fig. 4. Comparison of hearing thresholds and (a) frequency range and relative spectral energy of cub (red/blue) and adult (gray) vocalizations; (b) fundamental frequency (F0) range (horizontal bars) of cub vocalizations (V1: squall; V2: squawk; V3: croak) (Baotic et al., 2014), adult bleat (V4, from Charlton et al., 2009a,b) and adult copulation-call (V5) (Keating, 2011). Mean (± SE) amplitude of F0, for each vocalization (vertical bar). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

3.3. Relationship between hearing sensitivity and acoustic signals Graphic comparison of between hearing sensitivity and spectral energy from panda vocalizations (Baotic et al., 2014; Keating, 2011) illustrated that infant squawk, infant squall, and adult chirp, bleat, growl and copulation-call fell well within the hearing curve (i.e., were audible), however the fundamental frequency (F0) of infant croaks (as documented by Baotic et al., 2014) fell below the audiogram (Fig. 4). Characteristic formant frequencies documented in panda bleats (Charlton et al., 2009a,b) and growls showed some degree of overlap with relative increases in panda hearing sensitivity (Fig. 5). Specifically, the fourth and fifth formants of the panda bleat (F4 and F5), which encode information related to individual identity (Charlton et al., 2009a,b), correspond to increased hearing sensitivity at 2.0–3.2 kHz. Growl formants showed some similarly interesting overlap with hearing sensitivity in the same frequency range, with F1 and F3 overlapping with a range of relatively good hearing, between 0.5 and 3.2 kHz. However, it is important to note that the information content of growl formants is not known. The chirp vocalization does not appear to have formants, however we note that panda hearing sensitivity has a notable decrease in relative sensitivity at 1.0 kHz, corresponding the fundamental frequency (F0) of the chirp (Charlton et al., 2011). Graphic inspection of spectrograms for 3 adult and 3 cub vocalizations, highlighting the frequency range of good (10.0–16.0 kHz) and best (12.5–14.0 kHz) hearing sensitivity (Fig. 6), demonstrated that bleats, chirps and cub squawks, had discernable energy in this frequency range. Adult growl and cub croak did not appear to have much, if any, spectral energy in the range of good hearing. 4. Discussion Our data provide a first description of hearing sensitivity in the giant panda, and demonstrate that this most primitive of bear species retains functional hearing into the ultrasonic range (>20 kHz). While we demonstrated similar hearing capacity to that of the polar bear at 10 kHz and below, giant panda hearing is notably more sensitive than that of the polar bear at and above 16 kHz (Owen and Bowles, 2011). That pandas retain functional ultrasonic hearing and polar bears, a more recently evolved bear with a strikingly different habitat and relatively limited use of acoustic communication, do not, is consistent with hypotheses that suggest ultrasonic hearing is an ancestral trait in mammals (Sales and Pye, 1974), and that signaling systems would have been shaped to some degree by the habitat topography over evolutionary time (Morton, 1975). Specifically, the acoustic adaptation hypothesis predicts that species in forested habitats may retain better high frequency hearing than those inhabiting more flat and open spaces (Morton, 1975). The predictions from this hypothesis reflect our comparative findings for these bear species inhabiting dramatically divergent habitats. In the context

128

M.A. Owen et al. / Global Ecology and Conservation 6 (2016) 121–131

Fig. 5. Comparison of vocal formants documented in the (a) adult bleat (F1–F5, as published in Charlton et al., 2009a,b) and (b) the adult growl (F1–F4).

of conservation management, these divergent data provide further evidence that generalizing hearing capacity based on contemporary taxonomic affinity is not always appropriate. Giant pandas are unique among the Ursidae in that females ovulate spontaneously (Stirling and Derocher, 1990). The temporal constraints that this strategy puts on fertility means that accurate assessment of breeding opportunities via acoustic signals is essential for reproductive success (Charlton et al., 2010a, 2009b). Indeed, pandas appear to rely more heavily on acoustic signals than other bears to coordinate mating (Stirling and Derocher, 1990). While most energy from affiliative vocalizations is found between 0.4 and 1.1 kHz, the need to detect and discriminate salient features of conspecific vocalizations may have shaped the evolution of good hearing above this range (Bohn et al., 2006). In particular, formants identified in the bleat vocalization, have been shown to convey information regarding the caller’s identity (specifically the fifth formant (F5), Charlton et al., 2009a,b), sex and body size (in the case of male callers; Charlton et al., 2009b, 2010a,b), and our results show correspondingly good hearing sensitivity around the information bearing formants of the bleat (e.g., F4–F5). We also found relative increases in hearing sensitivity corresponding to the four formants we identified in the growl vocalization (Fig. 5). While, the information content of these formants is not known, it is likely that being able to identifying the quality and identify of same-sex competitors would be adaptive; as males typically engage in aggressive competitive interactions for breeding access to sexually receptive females (Nie et al., 2012). In contrast to bleats and growls, our inspection of chirps did not demonstrate the presence of formants per se. However, at the fundamental frequency (F0) of the chirp, where most of the spectral energy for this vocalization is found, there is a correspondingly marked decrease in hearing sensitivity. Of note, Charlton et al. (2010a,b) found that no individual information was encoded by F0. Further, comparison of vocalization spectrograms with the region of good hearing sensitivity demonstrates (i.e., 10.0–16.0 kHz) the overlap of some acoustic energy from bleats and chirps (Fig. 6(a) and (b)), emphasizing the potential for information content to be encoded this frequency range. These results suggest that more research be done to ascertain the potential information content of this higher frequency vocal energy. In light of our results, we also note that future studies of giant panda acoustic ecology should consider the hearing capacity to ensure that recording equipment can reliably capture ultrasonic sound. Infant vocalizations are emitted at lower center-frequencies than courtship vocalizations (Fig. 2). Indeed, there is evidence (Baotic et al., 2014) that the fundamental-frequency of the infant ‘croak’ falls outside the range of panda hearing, providing support for hypotheses that this vocalization may have evolved as a form of self-soothing or autocommunication, analogous to the Ursid ‘hum’ (Peters et al., 2007). Given the lower sensitivity of panda hearing within the range of cub vocalizations, increases in ambient noise may more readily mask acoustic signals or prompt energetically demanding compensatory adjustments on the part of the signaler (Parks et al., 2010). In parts of the panda’s range, cubs may be vulnerable to these potentially fitness-reducing effects as the availability of large, and acoustically-protective, old-growth tree dens has declined (Zhang et al., 2011). An understanding of a species’ hearing provides a foundation for developing estimates of noise disturbance (Gerstein and Gerstein, 1999) along with other acoustic ecological traits. Indeed, Francis (2015) found that in birds, hearing associated

M.A. Owen et al. / Global Ecology and Conservation 6 (2016) 121–131

a

0

0

b

3.1

20

0

0

20

c

0

0

d

0.9

20

0

0

20

e

0

129

0

f

0.3

0

0

Fig. 6. Range of good hearing (10.0–16.0 kHz, red lines) overlaid on exemplar spectrograms of adult and cub vocalizations (10.0–16.0 kHz), including: (a) bleat, (b) growl, (c) chirp, (d) cub croak, (e) cub squawk and (f) cub squall. Dashed red lines indicate the range of best hearing sensitivity (12.5–14.0 kHz). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

traits, such as the frequency range of vocalizations, are strong predictors of habitat avoidance due to anthropogenic noise exposure. Although the systematic measurement of behavioral and physiological responses to noise provides the most salient information regarding the impacts of noise (Southall et al., 2007), ethical and regulatory constraints preclude the implementation of controlled noise exposure experiments for most conservation dependent species. These limitations require that a bottom-up approach be used when estimating disturbance. For the giant panda, vocalizations are typically emitted in proximity to conspecifics, however the ability to discriminate between fine-scale differences in vocalizations is important for successful reproduction (Charlton et al., 2011, 2010a, 2009b) and so a thorough understanding of acoustic ecology is merited in order to estimate the potential for disturbance. Captive breeding programs are an important component of the overall conservation strategy for giant pandas and concerns regarding the impact of noise disturbance on reproductive success ex situ are prominent, as pandas are exposed to a wide range of human generated noise in captive facilities. As in the wild, the typically close proximity of individuals in breeding or mother–cub pair minimizes concerns regarding the masking of acoustic signals during courtship encounters or during the post-partum period. However noise disturbance, and any associated stress response, remains a concern during both the breeding season and during the post-partum period, when maternal energy stores are at their lowest and highly altricial and fragile cubs rely on vocalizations to convey needs to their mother. Indeed, in the captive setting, disturbance has lead to the mishandling and subsequent injury/mortality of a neonate by the mother (D. Lindburg, pers. comm.). Given these challenges, understanding the frequency ranges of good hearing sensitivity in the species should be applied to the

130

M.A. Owen et al. / Global Ecology and Conservation 6 (2016) 121–131

development of the strategic management of noise in the vicinity of panda housing, thus minimizing the potential of acoustic disturbance to compromise reproductive success (Owen et al., 2004). While noise disturbance has not been identified as a primary conservation threat to free-ranging giant pandas, this may change in the future as climate change may drive panda populations outside the boundaries of reserves (Fan et al., 2014), further reducing important access to the old growth trees preferred for maternal denning. These predicted range shifts may also drive giant pandas towards closer contact with human habitation and activities – including industrial scale resource extraction (including logging), road building and use – increasing the likelihood of noise exposure and the probability that noise will disturb pandas or disrupt communication. Acknowledgments We thank the generous support of private donors, and the logistical support of the Construction and Maintenance department at the San Diego Zoo. We thank Ann Bowles for helpful comments on design of the space, and we thank Nicki Boyd, Gaylene Thomas and Joanne Simerson for advising us on the development of the animal training strategy. We thank ACO Pacific Inc. for their in-kind support of this research. We also thank Ben Charlton for his insightful comments on an earlier version of this manuscript. Author contributions MAO co-designed experiment, collected and analyzed data, and wrote manuscript; JLK co-designed experiment, developed test-space, collected audiometric data and recorded adult panda vocalizations; SKD wrote programs for tone delivery and data collection, and processed and analyzed panda vocalizations; KH, AF, JT, JB trained subjects for data collection; SH collected data; RRS played an advisory role. Additional information The author(s) declare no competing financial interests. References Baotic, A., Stoeger, A.S., Li, D., Tang, C., Charlton, B.D., 2014. The vocal repertoire of infant giant pandas (Ailuropoda melanoleuca). Bioacoustics 23, 15–28. Barber, J.R., Crooks, K.R., Fristrup, K.M., 2010. The costs of chronic noise exposure for terrestrial organisms. Trends Ecol. Evolut. 25, 180–189. Bee, M.A., Swanson, E.M., 2007. Auditory masking of anuran advertisement calls by road traffic noise. Anim. Behav. 74, 1765–1776. http://dx.doi.org/10.1016/j.anbehav.2007.03.019. Blanchet, S., Clobert, J., Danchin, É., 2010. The role of public information in ecology and conservation: an emphasis on inadvertent social information. Ann. New York Acad. Sci. 1195, 149–168. http://dx.doi.org/10.1111/j.1749-6632.2010.05477.x. Bohn, K.M., Moss, C.F., Wilkinson, G.S., 2006. Correlated evolution between hearing sensitivity and social calls in bats. Biol. Lett. 2, 561–564. Charlton, B.D., Huang, Y., Swaisgood, R.R., 2009a. Vocal discrimination of potential mates by female giant pandas (Ailuropoda melanoleuca). Biol. Lett. 5, 597–599. http://dx.doi.org/10.1086/506971. Charlton, B.D., Keating, J.L., Kersey, D., Li, R., Huang, Y., Swaisgood, R.R., 2011. Vocal cues to male androgen levels in giant pandas. Biol. Lett. 84, 71–74. Charlton, B.D., Keating, J.L., Li, R., Huang, Y., Swaisgood, R.R., 2010a. Female giant panda (Ailuropoda melanoleuca) chirps advertise the caller’s fertile phase. Proc. R. Soc. B Biol. Sci. 277, 1101–1106. Charlton, B.D., Swaisgood, R.R., Zhang, Z., Snyder, R.J., 2012. Giant pandas attend to androgen-related variation in male bleats. Behav. Ecol. Sociobiol. 66, 969–974. http://dx.doi.org/10.1007/s00265-012-1345-0. Charlton, B.D., Zhihe, Z., Snyder, R.J., 2009b. The information content of giant panda, Ailuropoda melanoleuca, bleats: acoustic cues to sex, age and size. Anim. Behav. 78, 893–898. Charlton, B.D., Zhihe, Z., Snyder, R.J., 2009c. Vocal cues to identity and relatedness in giant pandas (Ailuropoda melanoleuca). J. Acoust. Soc. Am. 126, 2721. Charlton, B.D., Zhihe, Z., Snyder, R.J., 2010b. Giant pandas perceive and attend to formant frequency variation in male bleats. Anim. Behav. 79, 1221–1227. Delaney, D.K., Grubb, T.G., Beier, P., Pater, L.L., Reiser, M.H., 1999. Effects of helicopter noise on Mexican spotted owls. J. Wildl. Manage. 63, 60–76. Ellison, W.T., Southall, B.L., Clark, C.W., Frankel, A.S., 2011. A new context-based approach to assess marine mammal behavioral responses to anthropogenic sounds. Conserv. Biol. 26, 21–28. http://dx.doi.org/10.1111/j.1523-1739.2011.01803.x. Erbe, C., King, A.R., 2009. Modeling cumulative sound exposure around marine seismic surveys. J. Acoust. Soc. Am. 125, 2443. http://dx.doi.org/10.1121/1.3089588. Fan, J., Li, J., Xia, R., Hu, L., Wu, X., Li, G., 2014. Assessing the impact of climate change on the habitat distribution of the giant panda in the Qinling Mountains. Ecol. Modell. 274, 12–20. http://dx.doi.org/10.1016/j.ecolmodel.2013.11.023. Francis, C.D., 2015. Vocal traits and diet explain avian sensitivities to anthropogenic noise. Global Change Biol. 21, 1809–1820. http://dx.doi.org/10.1111/gcb.12862. Francis, C.D., Barber, J.R., 2013. A framework for understanding noise impacts on wildlife: an urgent conservation priority. Front. Ecol. Environ. 11, 305–313. Gerstein, E., Gerstein, L., 1999. The underwater audiogram of the West Indian manatee (Trichechus manatus). J. Acoust. Soc. Am. 105, 3575–3584. Hastie, G.D., Russell, D.J.F., McConnell, B., Moss, S., Thompson, D., Janik, V.M., 2015. Sound exposure in harbour seals during the installation of an offshore wind farm: predictions of auditory damage. J. Appl. Ecol. 52, 631–640. http://dx.doi.org/10.1111/1365-2664.12403. Jackson, M.E., Asi, N.S., Fullard, J.H., 2010. Auditory sensitivity and ecological relevance: the functional audiogram as modelled by the bat detecting moth ear. J. Comp. Physiol. 196, 453–462. Keating, J.L., 2011. Implications of Vocalizations During Giant Panda Breeding Encounters. California State University, San Marcos. Ketten, D.R., 2004. Marine mammal auditory systems: a summary of audiometric and anatomical data and implications for underwater acoustic impacts. Polarforschung 72, 79–92. Kleiman, D.G., Peters, G., 1990. Auditory communication in the giant panda: motivation and function. In: Proceedings of the 2nd International Symposium on the Giant Panda. Tokyo Zoological Society. Kutschera, V.E., Bidon, T., Hailer, F., Rodi, J.L., Fain, S.R., Janke, A., 2014. Bears in a forest of gene trees: Phylogenetic inference is complicated by incomplete lineage sorting and gene flow. Mol. Biol. Evol. 31, 2004–2017. http://dx.doi.org/10.1093/molbev/msu186. Ladich, F., Yan, H.Y., 1998. Correlation between auditory sensitivity and vocalizationin anabantoid fishes. J. Comp. Physiol. [A] 182, 737–746. Levitt, H., 1971. Transformed up-down methods in psychoacoustics. J. Acoust. Soc. Am. 49, 467–477. Linnell, J.D.C., Swenson, J.E., Andersen, R., Barnes, B., 2000. How vulnerable are denning bears to disturbance? pp. 400–413. MacGillivray, A.O., Racca, R., Li, Z., 2013. Marine mammal audibility of selected shallow-water sources. J. Acoust. Soc. Am. 135, EL35–EL40. http://dx.doi.org/10.1121/1.4838296.

M.A. Owen et al. / Global Ecology and Conservation 6 (2016) 121–131

131

Morton, E.S., 1975. Ecological sources of selection on avian sounds. Am. Nat. 109, 17–34. Nie, Y., Swaisgood, R.R., Zhang, Z., Liu, X., Wei, F., 2012. Reproductive competition and fecal testosterone in wild male giant pandas (Ailuropoda melanoleuca). Behav. Ecol. Sociobiol. 66, 721–730. Owen, M.A., Bowles, A.E., 2011. In-air auditory psychophysics and the management of a threatened carnivore, the polar bear (Ursus maritimus). Int. J. Comp. Psychol. 24, 244–254. Owen, M.A., Swaisgood, R.R., Czekala, N.M., Steinman, K., Lindburg, D.G., 2004. Monitoring stress in captive giant pandas (Ailuropoda melanoleuca): behavioral and hormonal responses to ambient noise. Zoo Biol. 23, 147–164. http://dx.doi.org/10.1002/zoo.10124. Owen, M.A., Swaisgood, R.R., McGeehan, L., Zhou, X., Lindburg, D.G., 2013. Dynamics of male–female multimodal signaling behavior across the estrous cycle in giant pandas (Ailuropoda melanoleuca). Ethology 119, 869–880. Parks, S.E., Johnson, M., Nowacek, D., Tyack, P.L., 2010. Short-and long-term changes in right whale calling behavior: The potential effects of noise on acoustic communication. Biol. Lett. 122, 51–54. http://dx.doi.org/10.1121/1.2799904. Peters, G., Owen, M.A., Rogers, L., 2007. Humming in bears: a peculiar sustained mammalian vocalization. Acta Theriol. 52, 379–389. Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., R Core Team, 2015. nlme:Linear and nonlinear mixed effects models. Ramsier, M.A., Cunningham, A.J., Moritz, G.L., Finneran, J.J., Williams, C.V., Ong, P.S., Gursky-Doyen, S.L., Dominy, N.J., 2012. Primate communication in the pure ultrasound. Biol. Lett. 8, 508–511. R Core Team, 2013. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Reichmuth, C., Southall, B.L., 2012. Underwater hearing in California sea lions (Zalophus californianus): Expansion and interpretation of existing data. Mar. Mam. Sci. 28, 358–363. http://dx.doi.org/10.1111/j.1748-7692.2011.00473.x. Remington, E.D., Osmanski, M.S., Wang, X., 2012. An operant conditioning method for studying auditory behaviors in marmoset monkeys. PLoS One 7, e47895. http://dx.doi.org/10.1371/journal.pone.0047895.g005. Sales, G.D., Pye, J.D., 1974. Ultrasonic Communication by Animals. Chapman and Hall, London. Shannon, G., McKenna, M.F., Angeloni, L.M., Crooks, K.R., Fristrup, K.M., Brown, E., Warner, K.A., Nelson, M.D., White, C., Briggs, J., McFarland, S., Wittemyer, G., 2015. A synthesis of two decades of research documenting the effects of noise on wildlife. Biol. Rev. http://dx.doi.org/10.1111/brv.12207. n/a–n/a. Southall, B.L., Bowles, A.E., Ellison, W.T., Finneran, J.J., Gentry, R.L., Greene Jr., C.R., Kastak, D., Ketten, D., Miller, J.H., Nachtigall, P.E., Richardson, W.J., Thomas, J.A., Tyack, P.L., 2007. Marine mammal noise exposure criteria: initial scientific recommendations. Aquatic Mammals 33, 411–532. State Forestry Administration, 2015. The Fourth National Survey Report on Giant Panda in China. Science Press, Beijing, China. Stirling, I., Derocher, A.E., 1990. Factors affecting the evolution and behavioral ecology of the modern bears. In: Bears: Their Biology and Management. pp. 189–204. Szymanski, M.D., Bain, D.E., Kiehl, K., Pennington, S., Wong, S., Henry, K.R., 1999. Killer whale (Orcinus orca) hearing: Auditory brainstem response and behavioral audiograms. J. Acoust. Soc. Am. 106, 1134–1141. Tuanmu, M.-N., Viña, A., Winkler, J.A., Li, Y., Xu, W., Ouyang, Z., Liu, J., 2012. Climate-change impacts on understorey bamboo species and giant pandas in China’s Qinling Mountains. Nature Clim. Change 3, 249–253. Zhang, Z., Swaisgood, R.R., Zhang, S., Nordstrom, L.A., Wang, H., Gu, X., Hu, J., Wei, F., 2011. Old-growth forest is what giant pandas really need. Biol. Lett. 7, 403–406. Zhu, L., Hu, Y., Zhang, Z., Wei, F., 2013. Effect of China’s rapid development on its iconic giant panda. Chinese Sci. Bull. 58, 2134–2139. Zhu, X., Lindburg, D.G., Pan, W., Forney, K.A., Wang, D., 2001. The reproductive strategy of giant pandas (Ailuropoda melanoleuca): infant growth and development and mother-infant relationships. J. Zool. Lond. 253, 141–155.