Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Soundscape of green turtle foraging habitats in Fiji, South Pacific

  • Elena Papale ,

    Roles Conceptualization, Data curation, Formal analysis, Methodology, Resources, Writing – original draft, Writing – review & editing (EP); (SP)

    Affiliations BioacousticsLab, IAS Capo Granitola, National Research Council, Torretta Granitola, Italy, Department of Life Science and Systems Biology, University of Torino, Torino, Italy

  • Shritika Prakash,

    Roles Data curation, Methodology, Project administration, Writing – review & editing

    Affiliation School of Marine Studies, The University of the South Pacific, Suva, Fiji

  • Shubha Singh,

    Roles Data curation, Methodology, Writing – review & editing

    Affiliation School of Marine Studies, The University of the South Pacific, Suva, Fiji

  • Aisake Batibasaga,

    Roles Funding acquisition, Project administration, Resources, Writing – review & editing

    Affiliation Ministry of Fisheries, Suva, Fiji

  • Giuseppa Buscaino,

    Roles Conceptualization, Formal analysis, Methodology, Resources, Writing – original draft, Writing – review & editing

    Affiliation BioacousticsLab, IAS Capo Granitola, National Research Council, Torretta Granitola, Italy

  • Susanna Piovano

    Roles Conceptualization, Formal analysis, Funding acquisition, Methodology, Project administration, Writing – original draft, Writing – review & editing (EP); (SP)

    Affiliation School of Marine Studies, The University of the South Pacific, Suva, Fiji


The soundscape features of the marine environment provide crucial information about ecosystem health for many species, and they are defined by the local biological, geophysical, and anthropogenic components. In this study, we investigated the soundscape at green turtle neritic foraging habitats in Fiji, South Pacific, with the aims of characterizing the contribution of each component and of comparing the levels of acoustic pressure among sites with different abundances of sea turtles. Four sites were selected at two islands, and one hydrophone was deployed at each site. Generalized additive models highlighted that sound pressure levels (SPLs) at low frequencies (125–250 Hz) were especially affected by wind conditions, while at higher frequencies (>250 Hz) SPLs were mostly influenced by fish and crustacean acoustic activity. Higher abundances of green turtles were found at sites with the highest levels of SPLs and the highest number of acoustic emissions by fishes and crustaceans but were not related to maximum seagrass and macroalgae coverage, or the highest number of fish. The selected coastal habitats have negligible anthropogenic noise, thus this study informs physiological and behavioral studies of the acoustic signatures that sea turtles might target and provides a baseline against which potential impact of soundscape changes on sea turtle spatial abundance and distribution can be evaluated.


The ecosystem soundscape can be considered a function of ecological variables, such as natural physical processes and meteorological events (geophonies), voluntary or involuntary biological sounds (biophonies), and noise generated by human activities (anthrophonies) [1]. Geophonies, biophonies, and anthrophonies are influenced by environmental and biological rhythms driven by diel, lunar, tidal, monthly, and seasonal cycles [2,3] and, as they interact and combine, they shape the temporal patterns and spectral composition of the soundscape [4,5]. Therefore, each ecosystem can be defined by a characteristic acoustic signature [6], determined by its unique combination of geophysical features, biological communities, and human interference [711], that is subject to temporal and spectral variability [4,12,13].

The contribution of sound components, as well as acoustic temporal variability, have been described for coral-dominated reefs [1416], macroalgae-dominated reefs [6], and oyster reefs [17,18], in addition to sandy bottoms [6,19], estuarine habitats [13,20,21], maerl beds [10], coastal embayments [12,22], Posidonia-dominated seagrass meadows [19], and oceanic Arctic and Antarctic areas [11,23,24]. Together, acoustic temporal and spatial patterns reflect the ecological conditions of habitats, as well as the acoustic adaptation of soniferous species to the environment [2527]. As a result, many marine vertebrates and invertebrates, at multiple life-history stages, rely on an integrated sensory response incorporating acoustic and other stimuli for orientation during navigation and selection of suitable habitat [16, 2830 but see also 31]. However, different sets of stimuli are likely used at different spatial scales [29]. For example, visual and mechanical cues are reliable over short (from centimeters to tens of meters) distances [32,33] while, depending on their intensity, acoustic cues can support habitat selection at short or long spatial scales [29,34,35].

Acoustic cues and their spatial and temporal variability likely provide marine animals with an ‘acoustic image’ of the underwater habitat structure and its community composition [36]. The underlying assumption, known as the ‘soundscape paradigm’, is that animals can perceive and interpret the acoustic stimuli [1,4,37]. However, little is known about the cognitive capacities of many marine species [37]. For example, the use of acoustic cues in habitat selection has been largely investigated in fishes and mammals that produce sounds; however, species-specific audiograms are available for <1% of marine fishes and ca. 25% of the marine mammals (see review in [38]), resulting in uncertainty regarding specific hearing thresholds.

In the marine environment, acoustic cues can be represented by the sound produced from high-energy waters, such as tidal currents, breaking waves, and wind, which commonly extend below 2 kHz [39], as well as by biological signals such as those generated by vertebrates and invertebrates. Fish signals, for example, are made of pulsations often below 1 kHz [40] and show high diel and seasonal variability [12]. In seagrass meadows and coral reef habitats, the loudest biological contribution that dominates the soundscape is provided by invertebrates, mainly crustaceans such as snapping shrimps (genera Alpheus and Synalpheus) [18,4144]. In tropical marine habitats, their short broadband pulses (<0.1 s, in [45]) can drown out all the other components at frequencies over 2 kHz [46] and contributes to habitat selection by pelagic larvae of corals, fishes and crustaceans [14,30,35,47,48]. Different acoustic communities create spatial variability. For example, at coral reefs in Hawaii, the frequencies of sounds produced by fishes range from <100 Hz to 1 kHz during non-feeding behavior (with peaks at 100–300 Hz), to 2–6 kHz during feeding behavior [49]. In Australia, they range from 100 Hz to 2 kHz [50], while in Panama and Florida they dominate the 400 Hz and 25–200 Hz bands, respectively [2].

Similarly, sea turtles likely use acoustic stimuli together with a variety of other potential cues such as magnetism, light, colors, and water chemicals to orient over long and short spatial scales during their different life stages, i.e., post-hatchling, pelagic juvenile, neritic juvenile and adult [5154]. The life cycle of hard-shell sea turtles is characterized by a terrestrial and a marine phase. In general, the terrestrial phase is relatively short, accounting for egg incubation in sandy beaches, hatchling emergence from the egg chamber, and subsequent crawling activity to reach the sea. In addition, mature females cyclically return to their natal beach to nest. The marine phase accounts for most of the sea turtles’ lives and is characterized by complex migration patterns that include periodic returns to specific geographic areas, such as coastal foraging grounds [52]. After an initial oceanic stage, sea turtles settle in a neritic environment and often stay in small home ranges for several years [55] and select, at a finer scale, specific resources available within neritic habitats [5659]. However, sea turtles in the neritic stage can travel across oceanic waters when moving between different foraging grounds, and adult individuals migrate across oceanic waters to reach (and return from) their mating and nesting areas.

Hard-shell sea turtles are capable of detecting and responding to acoustic stimuli within a frequency range of 50 Hz to 2 kHz in both aerial and seawater media (Table 1). The green turtle (Chelonia mydas) and the loggerhead turtle (Caretta caretta) have been the focus of the majority of the acoustic-detection studies. The green turtle reportedly has the broadest underwater hearing range, from 50 Hz (95 dB re: 1 μPa-rms) to 1.6 kHz (157 dB re: 1 μPa-rms) [54]. The loggerhead turtle appears to have a narrower underwater hearing range, from 50 Hz (101 dB re: 1 μPa-rms) [60] to 1.13 kHz (148 dB re: 1 μPa-rms) [61]. The loggerhead turtle is also the only sea turtle species in which hearing capabilities have been investigated at different life stages. Studies show consistency in the frequencies detected from hatchlings to juveniles and adults loggerhead turtle, despite the different environments where they live. Lavender et al. [60] suggested that loggerhead turtles have a basic hearing sense or, alternatively, hear low frequencies, which is useful in the coastal environment where hatchlings are born and neritic-stage animals (e.g. large juveniles and adults) live, and is simply maintained during the initial oceanic stage and the migration performed by large juveniles and adults.

Table 1. Hearing capabilities in sea turtles, by species.

Unless otherwise stated, the intensity is expressed as dB re: 1 μPa.

Both green and loggerhead turtles respond to underwater acoustic stimuli that overlap in frequency with fish-generated sounds and with common geophonies. In addition, both sea turtle species are likely capable of perceiving invertebrate-generated sounds at low frequencies: although this ability greatly depends on the sound source level, the active space of the sound, and the distance between the source and the receiver. For the same reasons, both sea turtle species likely miss out on the highest frequencies reached by the snapping shrimps [66]. The suggested ecological advantage of being able to detect low-frequency sounds, which travel farthest in the marine environment when produced at high intensities and dominate the soundscape [67], is associated with the possible use of acoustic stimuli as environmental cues [54]. Furthermore, recent studies proved the social role of vocalizations in freshwater turtles. For example, the giant South American river turtle (Podocnemis expansa) uses acoustic communication during the nesting season to facilitate synchronization of group migration to nesting beaches [68], as well as post-hatching parental care [69], while the northern snake-necked turtle (Chelodina oblonga) uses acoustic communication during the breeding season as a display to attract a mate in water with poor visibility [70].

This study aims to: 1) investigate the relationship between green turtle abundance and habitat quality data at tropical green turtle neritic foraging habitats, 2) characterize the levels of acoustic pressure present at each site, 3) assess the contribution of the acoustic components to the soundscape, and 4) infer the relationship between local soundscapes and green turtle abundance. Characterization of the acoustic stimuli present in a sea turtle coastal habitat with negligible anthropogenic noise has a twofold advantage. First, it will provide a better understanding of the biological significance of sea turtle hearing, as it could inform physiological and behavioral studies of the acoustic signatures that sea turtles might target [60]. Second, it will form a baseline against which changes in habitat soundscape (potentially associated with altered ecological communities in response to climate change or anthropogenic acoustic pollution) and their impact on turtle distribution can be evaluated.

Materials and methods

Ethics statement

All sea turtle handling procedures were approved by the Animal Ethics Committee of The University of the South Pacific and performed under a Ministry of Fisheries Exemption Permit (Ref. FI/G/12-5).

Study area

Fieldwork was carried out in 2018 at Makogai Island and Yadua Island in Fiji (Fig 1A), tropical South Pacific, where juvenile green turtles from different stocks form foraging aggregations in shallow coastal waters [71,72]. The surface areas of Makogai and Yadua Islands are about 8 km2 and 15 km2, respectively. Several back-reef embayments (hereafter “sites”) are present in both islands and used by green turtles for feeding. Based on previous data on green turtle distributions, green turtles move among sites within the same island, but not between islands (S. Piovano unpublished data). This study focused on four sites on each island (Table 2, Fig 1B and 1C). The sites ranged in area from 14,600 m2 to 68,300 m2, and consisted of neritic habitats extending from the intertidal zone (semidiurnal tidal cycle) to the subtidal zone (max depth: mean ± SD = 5.0 ± 1.9 m). All sites are delimited by coral reefs on the oceanic side and are mostly fringed by sandy beaches on the terrestrial side.

Fig 1. Study area.

(A) Main islands of Fiji (South Pacific Ocean) and relative position of Yadua Island and Makogai Island. (B) The four sites of hydrophone deployment at Yadua Island. (C) The four sites of hydrophone deployment at Makogai Island. Image prepared in R [73] with package “leaflet” [74], geodata from OpenStreetMap (data are available under the Open Database License 1.0, cartography is licensed under the Creative Commons Attribution-ShareAlike 2.0 license). Image retouched with GIMP [75].

Table 2. Description of the eight green turtle foraging sites at Makogai Island and Yadua Island, Fiji, South Pacific.

Data collection

Local soundscape and abundances of green turtles were recorded at each of the eight sites during the same periods: from 19 to 22 September 2018 at Makogai Island, and from 30 September to 3 October 2018 at Yadua Island. To minimize interference, fish counts and benthic surveys for habitat characterization were performed after completion of the sound recording period.

Acoustic data

Four underwater autonomous acoustic recorders (hereafter “UAARs”) were deployed at each island in the ratio of one UAAR per site. Deployment and collection of UAARs were done from a small boat at high tide. Each instrument was synchronized before deployment to simultaneously record uninterrupted sequences of 2-min files for four consecutive days. Unfortunately, one UAAR malfunctioned and recorded for less than 24 hours. An average 64.7 hr per site (SD = 28.2 hr per site) was recorded, at a sample rate of 48 kHz at 16 bits. The UAARs were located 1 m from the soft sea bottom, anchored with a 30 kg weight at a depth ranging from 3 m to 8 m, and kept vertical by a submerged buoy (S1 Fig). The instruments deployed were omnidirectional calibrated hydrophones with a flat sensitivity response of -174.5 (± 2) dB re: V/μPa from 100 Hz to 100 kHz (Benthowave Low Noise Broadband Hydrophone with Preamplifier, model BII-7016 T6) and a digital signal processor (Texas Instruments, model TMS320C5535), coupled with a stereo audio codec (Texas Instruments, model AIC3204).

Habitat characterization and sea turtle abundance

Fish counts were performed at each site by counting the number of fishes (by species or by genus, whatever possible) visible in underwater videos recorded with pocket cameras (GoPro Hero 4) along three belt transects at high tide. Each transect had a fixed width of 5 m (whose sides and middle lines were marked with measuring tapes) and an average length of 89 m (SD = 16.4 m) resulting from a maximum length of 100 m or until reaching the coral reef crest, whichever was encountered first. Fish count data were standardized as the number of fishes per 10 m, counted as an average of two swims. Each video lasted an average of 26.6 min (SD = 13.0 min), depending on transect length. To allow disturbed fishes to resume their normal activity, the two underwater operators undertook five minutes of no activity after laying the transect lines. Quadrat surveys to quantify habitat characteristics [76] were conducted after underwater video recording for fish counts. The biodiversity, distribution and percent coverage of seagrasses and macroalgae at each site were recorded within standard 100 × 100 cm quadrats placed every 10 m along the midline of the same belt transects used for the fish count. Quadrats were also used to assess substrate composition. Tide heights were estimated from the Fiji Meteorological Service recording at Levuka station (17° 40’ 00” S, 178° 48’ 00” E) by using the software WXTide32 4.6. Wind intensity and direction were obtained from Fiji Meteorological Service at Ratu Kadavulevu School Lodoni station (17° 44’ 27” S, 178° 33’ 04” E) for Makogai Island, and Seaqaqa station (16° 32’ 23” S, 179° 8’ 23” E) for Yadua Island.

The abundance of sea turtles at each site was recorded as the sum of sighted and captured sea turtles during a total of 16 hours of surveys performed during the days of UAARs deployment. Each site was monitored for 2 hr (30 min × 4 days) by boat at high tide. The capture of sea turtles was performed via “rodeo”, a standard technique that consists of the hand-capture of sea turtles by diving onto them out of a small boat [7780]. This method might have caused some disturbance, but it was the best possible approach. Before being released back into the water, each turtle was tagged individually with an Inconel flipper tag (National Band and Tag Company, Style 681) attached to the trailing edge of the right front flipper [81], and species and minimum curved carapace length were recorded. In addition, during each survey, all turtles sighted but not captured were noted. The relative abundance of sea turtles observed during this study was used to categorize each site as low abundance (<2 turtles/10,000 m2), moderate abundance (2–3.9 turtles/10,000 m2), and high abundance (≥4 turtles/10,000 m2).

Acoustic analyses

Acoustic analyses were conducted on each 2-min file, except for those recorded during turtle rodeos. To describe the general acoustic pressure for each site, sound pressure level values (SPL dB re: 1 μPa-rms, hereafter “SPL”) were measured through SPL analysis at 1/3 octave bands, centered from 16 Hz to 16 kHz [82] using the software Avisoft-SASLab Pro, calibrated to provide absolute dB values. SPL values at each frequency band were transformed from dB to Pa for calculation of mean and standard deviation, which are reported after back-transforming Pa into dB.

The dial cycle of the SPLs at the 1/3 octave bands, centered at 31.5 Hz, 125 Hz, 250 Hz, 500 Hz, 1000 Hz, and 2000 Hz, was examined to study temporal rhythms in sound levels and the daily abundance of the environmental and biological variables in the soundscape. Selected bands had a low correlation among them.

Twelve small boats, less than 10 m in length, are present at the islands (ten at Yadua Island and two at Makogai Island, including the one used during this research). The boats are equipped with outboard engines, and used to go fishing around the island and, when needed, to travel from the village to mainland (and back). To assess anthrophonies, the passages of the motorboats were counted through visual and aural inspection of the spectrograms.

The contribution of geophonies was evaluated using tide height and wind intensity as a proxy for high-energy waters [9,22]. Tide height and wind intensity were compared to SPL values at the different bands to infer their influence on background noise intensity.

The contribution of biophonies was evaluated using the number of acoustic events attributed to fishes and crustaceans as biological indicators of acoustic activity. Sounds of fishes and snapping shrimps were detected automatically with a custom-made code in MatlabTM 2019, irrespective of the different types of sounds produced by the different species (only the band pass filter was changed between fishes and snapping shrimp). Most fish sounds consisted of pulses at frequencies between 100 Hz and 800 Hz (see Fig 2), so, each 2-min file was down-sampled at 2 kHz and then filtered with a band pass at 100–800 Hz (equiripple, order 20). To count shrimp snaps in each 2-min file, the code filtered the signal (without down-sampling) with a band pass filter at 4–9 kHz (equiripple, order 20). This band pass filter was used to improve the identification of snapping shrimp signals by excluding low-frequency pulses produced by fishes and by photosynthetic activity [83]. Then, the code applied the Teager Kaiser filter and counted all the pulses that overcame an empirical threshold [84,85], one for fish and another one for snapping shrimps; these two thresholds were the same for all of the eight sites. Before applying the code to all the files, the accuracy of the detection was validated by manually counting the number of clicks from a randomly selected subsample (5%) of the audio files, and resulted in 95% correct identification of impulsive signals, i.e., fish pulses and snapping shrimp snaps.

Fig 2. Spectrogram (frequency vs time) of a portion of a recording at Vagabia at 18:00.

Squares highlight details of crustacean snaps (dashed line) and, at a lower frequency, fish impulsive signals (continuous line).

Statistical analysis

Statistical differences in curved carapace length of captured green turtles at the two islands were explored with a Wilcoxon-Mann-Whitney test. The correlation among SPL values at the frequency bands selected wind conditions, tide height, and biological acoustic signals were assessed using a Spearman correlation test (rs). Furthermore, since frequency bands from 100 Hz could be influenced by several variables, we tested which predictors better explained the variation of SPL values by using a generalized additive model (“GAM”) (R “mgcv” package version 1.8–2; [86]). SPL values were chosen as the response variable and tested as a function of the predictors hours (both daytime and nighttime, discrete variable ranging from 0 to 23), wind condition (mean Beaufort wind condition for each 2-min file), fish (number of fish acoustic emissions recorded in each 2-min file) and crustaceans (number of snapping shrimp clicks recorded in each 2-min file). A Gaussian distribution and identity link function were chosen. A penalized cubic regression spline was selected for the explanatory variables, except for hours, for which a cyclic cubic regression spline was used. SPLs at the frequency bands considered were compared among sites by using a Kruskal-Wallis test. Sites with less than 24 hours of recording were excluded from circadian and biological analyses. All analyses were performed in R [87].


Habitat characterization and sea turtle abundance

The eight sites presented several differences (Table 3). The dominant substrate was often sand, but mixed substrates and rock substrates were also recorded. In general, seagrass coverage was primarily composed of Halodule pinifolia, H. uninervis, Halophila ovalis, and Syringodium isoetifolium, with a prevalence of H. pinifolia at Yadua Island (58% of seagrass coverage), and H. uninervis (71% of seagrass coverage) at Makogai Island. The average number of fishes counted every 10 m ranged from 2.0 to 15.8. A total of 83 sea turtles were observed, of which 75% were sighted and 25% were captured. All were green turtles, except for two hawksbills (Eretmochelys imbricata) that were captured at Yadua Island and excluded from the statistical analysis. Based on green turtle relative abundance during the study period, two sites were ranked as high abundance, three as moderate abundance, and three as low abundance (Table 3). The two sites with a higher relative abundance of green turtles (i.e., Nasau and Vagabia) had a moderate to low average coverage of seagrass, moderate to high coverage of macroalgae, and a moderate to high number of fish observed, compared to the other sites. The lower relative abundance of green turtles was recorded at three sites (Matauvia, Navalowara, and Takewa) that had null or high average coverages of seagrasses, low to moderate coverages of macroalgae, and a moderate to high number of fish observed compared to the other sites. There was no statistically significant difference in green turtle size between the two islands (curved carapace length: median = 53 cm, mean ± SD = 54.5 ± 8.1 cm at Makogai Island and median = 53 cm, mean ± SD = 55.9 ± 6.9 cm at Yadua Island; Z = -0.60, p = 0.567), confirming previous findings by Piovano et al. [59] in 2015–2016.

Table 3. Habitat characterization of each green turtle foraging site at Makogai Island and Yadua Island, Fiji, South Pacific.

Wind direction (S2 Fig) concentrated around 145° (i.e., S-SE) and 245° (i.e., W-SW) at Makogai Island, whereas it converged prevalently from 100° to 155° (i.e., SE) and rarely around 360° (i.e., N) at Yadua Island. Wind intensity consistently showed higher levels at Yadua Island (mean Beaufort level ± SD = 4.07 ± 1.02) compared to Makogai Island, where a larger range of wind intensity was recorded (mean Beaufort level ± SD = 1.89 = ± 1.10). A strong peak was recorded during daylight hours at both islands (S3 Fig). Tide height ranged from 0.43 m to 1.63 m, as per Fiji Meteorological Service data.

Characterization of levels of acoustic pressure

A total of 31,118 min of recordings were collected. The mean and standard error of SPLs values measured at each site for the frequency bands centered from 16 Hz to 16 kHz are reported in S1 Table. The circadian rhythm for the frequency bands 31.5 Hz, 125 Hz, 250 Hz, 500 Hz, 1000 Hz, and 2000 Hz is reported for seven sites (Fig 3).

Fig 3. Circadian rhythm of the sound pressure levels at 31.5 Hz, 125 Hz, 250 Hz, 500 Hz, 1000 Hz, and 2000 Hz at the seven sites at Makogai Island and Yadua Island, Fiji, South Pacific.

Sites are in descending order of turtle abundance, arranged horizontally from left to right. One site, Matauvia, is not presented, as the data collected were insufficient to evaluate the circadian rhythm.

Contribution of the acoustic components to the soundscape

Anthrophonies were identified in boat passages and detected in 0.25% of the 15,535 2-min files, which makes their contribution negligible. Therefore, no further analysis was conducted on this sound component. The SPLs centered at 31.5 Hz showed a cycle matching the daily tide increase for 86% of the sites, however, this was not significantly correlated at Talai (rs = 0.01, p = 0.48). The SPLs centered at 125 Hz and 250 Hz were positively and significantly correlated with wind intensity (rs = 0.37 to 0.65, p <0.001), except for Votua, for which very weak negative correlations were found (125 Hz: rs = –0.05, p <0.04; 250 Hz: rs = –0.06, p = 0.02). These SPLs presented a daily increase coincident with increasing wind intensity at both islands (S2 and S3 Figs). At the three subsequent band levels, i.e., 500 Hz, 1000 Hz, and 2000 Hz, a clear diel cycle emerged, with SPLs peaking at dusk and dawn (S3 Fig). These last three bands were especially influenced by the biological component, with fish impulsive signals at lower frequencies (band 500 Hz, positive Spearman correlation for all the sites: p <0.001), and crustacean snaps sounds at higher frequencies (1 kHz and 2 kHz, positive Spearman correlation for all the sites: p <0.001). Predictor variables hours, fish and crustacean explained from 23% to 92% of the GAM model deviance. In the best fitting GAM, the majority of the deviance was explained by fish sounds at 250 Hz and 500 Hz bands and crustacean sounds at 1 kHz and 2 kHz (S2 Table).

Comparison among the soundscapes at the different sites

Average SPLs recorded at the eight sites along all the days (S1 Table) differed significantly at all the bands (Kruskal-Wallis test: 4,418 < χ2 < 12,723, p <0.001) (S3 Table). On average, SPL values were higher at two sites (Nasau and Vagabia) and lower at three sites (Matauvia, Takewa, and Votua), while moderate SPL values were measured at the remaining sites.

At the frequency bands of 125–250 Hz, a difference of about 8–16 dB was recorded between the sites with the most and least exposure to wind stress. Regarding biophonies, acoustic emissions from fishes and crustaceans are graphically summarized in Fig 4. Overall, higher numbers of biological signals were recorded at three sites, two of which (Nasau and Vagabia) also showed higher SPL values and geophonies. Matauvia was excluded from the analysis, as the amount of data collected was insufficient to evaluate biological emissions.

Fig 4. Plot of the mean number of acoustic emissions of fish and crustaceans at the seven green turtle foraging sites at Makogai Island and Yadua Island, Fiji, South Pacific, presented in order of turtle abundance.


This study documents the levels of acoustic pressure and the contribution of local biophonic and geophonic components to the overall soundscape of tropical neritic foraging habitats used by juvenile green turtles in a South Pacific island country. The soundscape can provide information about water turbulence under wind stress, which could inform sea turtles in their movements, and can provide information about habitat health through the intensity of local biological activity, a parameter that opportunistic green turtles might use to choose a feeding site.

Frequency bands observed in this study (16 Hz to 16 kHz) include the frequencies of green turtle electrophysiological response to underwater stimuli (50 Hz at 95 dB re: 1 μPa-rms to 1.6 kHz at 157 dB re: 1 μPa-rms, in [54]). The lower frequencies (<50 Hz) of the local soundscape were probably not perceived by the resident turtles; however, the effective hearing threshold for the juveniles of this species might be lower than what is known from auditory evoked potential measurements. Behavioral assessments, which are more sensitive than the auditory evoked potential for auditory threshold determination [60], could help to clarify this lower limit. The average level of energy of the local soundscape (86–102 dB re: 1 μPa-rms) observed at frequencies from 200 Hz to 700 Hz matched the hearing sensitivity of juvenile green turtles. Consequently, it is likely that the juvenile green turtles from different rookeries of the South Pacific and aggregating at the two Fijian islands [71] are capable of perceiving both the geophonic and biophonic components locally present.

The local soundscape was dominated by the sound generated by water turbulence such as due to breaking surf, tide, and wind-induced waves, that can be high energy and pervade low-frequency bands (31.5 to 125–250 Hz), especially during daylight hours. Higher frequency bands from 500 Hz can be associated with sonic fish choruses, emitted while in spawning aggregations or eventually defending territory and feeding, and by snapping shrimp clicks produced during aggressive behavior [8890]. Their emissions show a clear diel pattern, as recorded elsewhere [12,44,43], with the highest peaks at dusk and dawn at all the seven sites.

Soundscape components are not static but change daily and seasonally, reflecting the temporal scale of the contribution from geophonies, biophonies, and anthrophonies. As these components ultimately depend on the condition of a particular habitat, they might be used by neritic-stage sea turtles for their daily navigation and orientation within coastal habitats that are relatively close. In particular, since physical forcings are among the main drivers of zooplankton dispersal [91], geophonic components like wind intensity might indirectly inform sea turtles about the presence of macro- and mega- zooplankton brought to the coast by higher wind intensity. The use of sensory stimuli by hard-shell sea turtles to orient towards their food choice has been proven chemically [9294] and visually [9597], and might extend to the use of acoustic stimuli. Comparison between geophonies, the abundance of planktonic food, and sea turtle spatial behavior recorded via satellite tracking would help clarify the effect of wind intensity on sea turtle behavior. Other geophonic components, such as those produced by the tides, might provide information towards site accessibility. In particular, they might inform about the best time to forage in coastal waters where the low tide would restrict turtle access to shallow subtidal areas, as well as guide turtle movements in the intertidal zone. Such tidal-dependent behavior, for example, was observed in green turtles in Australia and at the Comoro Islands, where the presence of turtles in seagrass beds and mangrove habitats was very low or null at low tides [98,99]. Similar behavior has been described for bottlenose dolphins (Tursiops truncatus), whose higher presence in shallow estuarine waters in Australia was recorded at high tide and possibly resulted from increased accessibility to prey items [100]. However, tidal currents can also be perceived mechanically, thus separating the influence of the two stimuli produced by the same event (tide) might be difficult. Playback experiments could help clarify the acoustic influence in absence of the mechanical stimulus.

The intensity of biophonic components like sounds produced by fishes and invertebrates, together with their diel cycle, might inform sea turtles about the overall quality of habitat, as well as likely abundance of possible benthic prey. The high number of fish and invertebrate signals recorded at some sites in this study provides a regular contribution to the local soundscape generating dusk and dawn choruses. This high-energy activity produces a marked increase in sound levels, especially at Vagabia and Nasau, the two sites of high turtle abundance. Accordingly, a lower number of biological emissions, such as recorded at Votua, Talai, and Takewa, contributed less to energy levels and therefore to lower SPL measures.

In this study, significant variation in the soundscape has been identified among the sites at the fine-scale level. Similar among-site variability was recorded on close coral reefs on the Great Barrier Reef, in Australia [34]. Our analyses revealed differences in temporal patterns and sound levels among bays, especially below 800 Hz. These differences resulted from the combination of biological signals and sounds generated by water turbulence. Green turtles occurred most frequently in the two sites with the highest SPLs, at all frequencies and throughout the day. The differences recorded among the average SPLs at the sites, and caused by biological and geophysics signals reached 18 dB re: 1 μPa. The two sites of high turtle abundance were characterized by higher acoustic energy.

Although variation was identified in the biotic and abiotic components observed at the different sites, no clear pattern explaining the different sea turtle abundances emerged, suggesting that several factors might have driven the turtle presence at a site. For example, one site of high turtle abundance (Vagabia) had relatively low marine plant coverage, while another site with relatively high seagrass coverage (Matauvia) had low turtle abundance. Seagrass meadows and algal beds have a patchy distribution in the two foraging islands, and given that green turtle diet in this area is largely made up of invertebrates [59], it is possible that green turtles were using acoustic cues and other stimuli to choose a specific feeding site within the broad foraging area. Consistent with this, very few turtles occurred where the soundscape indicated low water turbulence and low biological activity. Therefore, our results suggest that soundscape can contribute to informing sea turtles with critical details regarding each site, but whether the acoustic data are effectively used to actively select a site, or if they are used as proximate acoustic cues, lies outside the scope of this work. In addition, green turtles might use information acquired from their past experiences, such as the use of memory about long-term average conditions, as recently described for blue whales (Balaenoptera musculus) [101]. Playback experiments with the bands identified in this study should be used to investigate whether green turtle behavior is influenced by habitat soundscape while combining data recording of habitat soundscape with satellite tracking of sea turtles could provide evidence on the contribution of the habitat soundscape towards green turtle fine-scale movements among neritic habitats.

This work provides useful baseline soundscape information for habitats with relatively low anthropogenic acoustic impacts. An increase in ocean noise levels, especially along coastal areas, is a critical issue around the world [67]. High levels of noise spreading from anthropogenic activities have been recorded in a major sea turtle foraging area on the eastern coast of the USA [102]. The noise occupied the frequencies below 1 kHz, thus largely overlapping with the range of sea turtle hearing and, in particular, with the range of greater hearing sensitivity. Furthermore, green turtles displayed increased swimming speed and erratic behavior [103] when exposed to noise generated by pile driving, airguns, and sonar. Therefore, this study provides the first step in understanding the relevance of underwater natural soundscape on turtle abundance in a green turtle tropical neritic area and, as such, forms the basis for future conservation efforts in an age of increasing acoustic pollution in coastal waters.

Supporting information

S1 Fig. Underwater autonomous acoustic recorder.

UAAR at Vagabia, in Makogai Island, Fiji. An iron 30 kg weight was used to anchor the UAAR, while a submerged buoy kept the mooring vertical.


S2 Fig. Wind rose representing the distribution of wind directions and Beaufort wind intensity.

(A) Makogai Island and (B) Yadua Island green turtle neritic foraging sites. Note: graphs have different scales.


S3 Fig. Mean Beaufort wind intensity at Makogai Island and Yadua Island (Fiji, South Pacific) green turtle neritic foraging sites throughout the day.


S1 Table. Mean and standard error of the sound pressure levels at 1/3 octave bands centered from 16 Hz to 16 kHz dB re: 1 μPa-rms for each of the eight green turtle neritic foraging sites at Yadua and Makogai Islands, Fiji, South Pacific.


S2 Table. Generalized additive model deviance.


S3 Table. Results of post-hoc Tuckey test for average SPL analysis.



We thank Turaga ni Yavusa Ratu Jone Cakau for allowing this research project to be carried out in Denimanu village’s iqoliqoli; all the volunteers from Yadua and USP, and the Fiji Fisheries officers, that helped during the surveys and the hydrophones deployment; particularly Pita Qarau, Senirusi Sevutia, Barry Hill, Waisea Delainimati, Pauliasi Levunakana, Seremaia Batikawai and Babitu Rarawa; Malakai Tuiono for accurate field work, in addition to Cristina Giacoma and Maria Ceraulo for their support and suggestions as well as Irene Kelly and Jeffrey A. Seminoff for their valuable comments on this manuscript. Environmental data were kindly provided by Fiji Meteorological Service, Ministry of National Disaster Management & Meteorological Services (Ref. MET 49/2). SPREP kindly provided flipper tags and applicator.


  1. 1. Pijanowski BC, Farina A, Gage SH, Dumyahn SL, Krause BL. What is soundscape ecology? An introduction and overview of an emerging new science. Landsc Ecol. 2011;26: 1213–1232.
  2. 2. Staaterman E, Rice AN, Mann DA, Paris CB. Soundscapes from a tropical Eastern Pacific reef and a Caribbean Sea reef. Coral Reefs. 2013;32: 553–557.
  3. 3. Staaterman E, Paris C, DeFerrari H, Mann D, Rice A, D’Alessandro E. Celestial patterns in marine soundscapes. Mar Ecol Prog Ser. 2014;508: 17–32.
  4. 4. Pijanowski BC, Villanueva-Rivera LJ, Dumyahn SL, Farina A, Krause BL, Napoletano BM, et al. Soundscape ecology: the science of sound in the landscape. BioScience. 2011;61: 203–216.
  5. 5. Farina A. Soundscape ecology. Dordrecht: Springer Netherlands; 2014.
  6. 6. Radford C, Stanley J, Tindle C, Montgomery J, Jeffs A. Localised coastal habitats have distinct underwater sound signatures. Mar Ecol Prog Ser. 2010;401: 21–29.
  7. 7. Kennedy EV, Holderied MW, Mair JM, Guzman HM, Simpson SD. Spatial patterns in reef-generated noise relate to habitats and communities: evidence from a Panamanian case study. J Exp Biol. 2010;395: 85–92.
  8. 8. Piercy J, Codling E, Hill A, Smith D, Simpson S. Habitat quality affects sound production and likely distance of detection on coral reefs. Mar Ecol Prog Ser. 2014;516: 35–47.
  9. 9. Kaplan M, Mooney T, Partan J, Solow A. Coral reef species assemblages are associated with ambient soundscapes. Mar Ecol Prog Ser. 2015;533: 93–107.
  10. 10. Coquereau L, Lossent J, Grall J, Chauvaud L. Marine soundscape shaped by fishing activity. R Soc open sci. 2017;4: 160606. pmid:28280559
  11. 11. Haver SM, Klinck H, Nieukirk SL, Matsumoto H, Dziak RP, Miksis-Olds JL. The not-so-silent world: measuring Arctic, Equatorial, and Antarctic soundscapes in the Atlantic Ocean. Deep Sea Res I. 2017;122: 95–104.
  12. 12. Buscaino G, Ceraulo M, Pieretti N, Corrias V, Farina A, Filiciotto F, et al. Temporal patterns in the soundscape of the shallow waters of a Mediterranean marine protected area. Sci Rep. 2016;6. pmid:27677956
  13. 13. Ricci S, Eggleston D, Bohnenstiehl D, Lillis A. Temporal soundscape patterns and processes in an estuarine reserve. Mar Ecol Prog Ser. 2016;550: 25–38.
  14. 14. Simpson S, Meekan M, McCauley R, Jeffs A. Attraction of settlement-stage coral reef fishes to reef noise. Mar Ecol Prog Ser. 2004;276: 263–268.
  15. 15. Radford CA, Jeffs AG, Tindle CT, Montgomery JC. Temporal patterns in ambient noise of biological origin from a shallow water temperate reef. Oecologia. 2008;156: 921–929. pmid:18461369
  16. 16. Bertucci F, Parmentier E, Berten L, Brooker RM, Lecchini D. Temporal and spatial comparisons of underwater sound signatures of different reef habitats in Moorea Island, French Polynesia. PLOS ONE. 2015;10: e0135733. pmid:26352701
  17. 17. Lillis A, Eggleston DB, Bohnenstiehl DR. Oyster larvae settle in response to habitat-associated underwater sounds. PLOS ONE. 2013;8: e79337. pmid:24205381
  18. 18. Bohnenstiehl DR, Lillis A, Eggleston DB. The curious acoustic behavior of estuarine snapping shrimp: temporal patterns of snapping shrimp sound in sub-tidal oyster reef habitat. PLOS ONE. 2016;11: e0143691. pmid:26761645
  19. 19. Ceraulo M, Papale E, Caruso F, Filiciotto F, Grammauta R, Parisi I, et al. Acoustic comparison of a patchy Mediterranean shallow water seascape: Posidonia oceanica meadow and sandy bottom habitats. Ecol Indic. 2018;85: 1030–1043.
  20. 20. Lillis A, Eggleston D, Bohnenstiehl D. Soundscape variation from a larval perspective: the case for habitat-associated sound as a settlement cue for weakly swimming estuarine larvae. Mar Ecol Prog Ser. 2014;509: 57–70.
  21. 21. Monczak A, Mueller C, Miller M, Ji Y, Borgianini S, Montie E. Sound patterns of snapping shrimp, fish, and dolphins in an estuarine soundscape of the southeastern USA. Mar Ecol Prog Ser. 2019;609: 49–68.
  22. 22. Putland RL, Constantine R, Radford CA. Exploring spatial and temporal trends in the soundscape of an ecologically significant embayment. Sci Rep. 2017;7. pmid:28720760
  23. 23. Ahonen H, Stafford KM, de Steur L, Lydersen C, Wiig Ø, Kovacs KM. The underwater soundscape in western Fram Strait: breeding ground of Spitsbergen’s endangered bowhead whales. Mar Pollut Bull. 2017;123: 97–112. pmid:28938997
  24. 24. Menze S, Zitterbart DP, van Opzeeland I, Boebel O. The influence of sea ice, wind speed and marine mammals on Southern Ocean ambient sound. R Soc open sci. 2017;4: 160370. pmid:28280544
  25. 25. Hansen P. Vocal learning: its role in adapting sound structures to long-distance propagation, and a hypothesis on its evolution. Anim Behav. 1979;27: 1270–1271.
  26. 26. Krause BL. The niche hypothesis: a virtual symphony of animal sounds, the origins of musical expression and the health of habitats. The Soundscape Newsletter. 1993; 4–6.
  27. 27. Mullet TC, Farina A, Gage SH. The acoustic habitat hypothesis: an ecoacoustics perspective on species habitat selection. Biosemiotics. 2017;10: 319–336.
  28. 28. Montgomery JC, Tolimieri N, Haine OS. Active habitat selection by pre-settlement reef fishes. Fish Fish. 2001;2: 261–277.
  29. 29. Kingsford MJ, Leis JM, Shanks A, Lindeman KC, Morgan SG, Pineda J. Sensory environments, larval abilities and local self-recruitment. Bull Mar Sci. 2002;70 (Suppl): 309–340.
  30. 30. Vermeij MJA, Marhaver KL, Huijbers CM, Nagelkerken I, Simpson SD. Coral larvae move toward reef sounds. PLoS ONE. 2010;5: e10660. pmid:20498831
  31. 31. Kaplan MB, Mooney TA. Coral reef soundscapes may not be detectable far from the reef. Scientific Reports. 2016;6. pmid:27550394
  32. 32. Stanley JA, Radford CA, Jeffs AG. Location, location, location: finding a suitable home among the noise. Proc R Soc B. 2012;279: 3622–3631. pmid:22673354
  33. 33. McWilliam JN, Hawkins AD. A comparison of inshore marine soundscapes. J Exp Mar Biol Ecol. 2013;446: 166–176.
  34. 34. Radford C, Stanley J, Jeffs A. Adjacent coral reef habitats produce different underwater sound signatures. Mar Ecol Prog Ser. 2014;505: 19–28.
  35. 35. Parmentier E, Berten L, Rigo P, Aubrun F, Nedelec SL, Simpson SD, et al. The influence of various reef sounds on coral-fish larvae behaviour: reef-sound influence on fish larvae behaviour. J Fish Biol. 2015;86: 1507–1518. pmid:25786340
  36. 36. Fay R. Soundscapes and the sense of hearing of fishes. Integr Zool. 2009;4: 26–32. pmid:21392274
  37. 37. Farina A. Perspectives in ecoacoustics: A contribution to defining a discipline. Journal of Ecoacoustics. 2018;2: TRZD5I.
  38. 38. Lucke K, Popper AN, Hawkins AD, Akamatsu T, André M, Branstetter BK, et al. Auditory sensitivity in aquatic animals. J Acoust Soc Am. 2016;139: 3097–3101. pmid:27369131
  39. 39. Erbe C, Verma A, McCauley R, Gavrilov A, Parnum I. The marine soundscape of the Perth Canyon. Prog Oceanogr. 2015;137: 38–51.
  40. 40. Kasumyan AO. Sounds and sound production in fishes. J Ichthyol. 2008;48: 981–1030.
  41. 41. Lillis A, Bohnenstiehl D, Peters JW, Eggleston D. Variation in habitat soundscape characteristics influences settlement of a reef-building coral. PeerJ. 2016;4: e2557. pmid:27761342
  42. 42. Lillis A, Perelman JN, Panyi A, Aran Mooney T. Sound production patterns of big-clawed snapping shrimp (Alpheus spp.) are influenced by time-of-day and social context. J Acoust Soc Am. 2017;142: 3311–3320. pmid:29195442
  43. 43. Lillis A, Mooney TA. The snapping shrimp conundrum: spatial and temporal complexity of snapping sounds on coral reefs. J Acoust Soc Am. 2016;140: 3071–3071.
  44. 44. Lillis A, Mooney TA. Snapping shrimp sound production patterns on Caribbean coral reefs: relationships with celestial cycles and environmental variables. Coral Reefs. 2018;37: 597–607.
  45. 45. Johnson MW, Everest FA, Young RW. The role of snapping shrimp (Crangon and Synalpheus) in the production of underwater noise in the sea. Biol Bull. 1947;93: 122–138. pmid:20268140
  46. 46. Au WWL, Banks K. The acoustics of the snapping shrimp Synalpheus parneomeris in Kaneohe Bay. J Acoust Soc Am. 1998;103: 41–47.
  47. 47. Simpson SD, Meekan MG, Jeffs A, Montgomery JC, McCauley RD. Settlement-stage coral reef fish prefer the higher-frequency invertebrate-generated audible component of reef noise. Anim Behav. 2008;75: 1861–1868.
  48. 48. Stanley JA, Radford CA, Jeffs AG. Induction of settlement in crab megalopae by ambient underwater reef sound. Behav Ecol. 2010;21: 113–120.
  49. 49. Tricas T, Boyle K. Acoustic behaviors in Hawaiian coral reef fish communities. Mar Ecol Prog Ser. 2014;511: 1–16.
  50. 50. Cato DH. Some unusual sounds of apparent biological origin responsible for sustained background noise in the Timor Sea. J Acoust Soc Am. 1980;68: 1056–1060.
  51. 51. Bartol SM, Musick JA. Sensory biology of sea turtles. In: Lutz PL, Musick JA, Wyneken J, editors. Biology of sea turtles. Boca Raton, Fla: CRC Press; 2003. pp. 79–102.
  52. 52. Lohmann KJ, Luschi P, Hays GC. Goal navigation and island-finding in sea turtles. J Exp Mar Biol Ecol. 2008;356: 83–95.
  53. 53. Lohmann KJ, Lohmann CMF, Brothers J, Putman NF. Natal homing and imprinting in sea turtles. In: Wyneken J, Lohmann KJ, Musick JA, editors. Biology of sea turtles. Boca Raton, Fla: CRC Press; 2013. pp. 59–77.
  54. 54. Piniak WED, Mann DA, Harms CA, Jones TT, Eckert SA. Hearing in the juvenile green sea turtle (Chelonia mydas): a comparison of underwater and aerial hearing using auditory evoked potentials. PLOS ONE. 2016;11: e0159711. pmid:27741231
  55. 55. Musick J, Limpus CJ. Habitat utilization and migration in juvenile sea turtles. In: Lutz PL, Musick J, editors. The biology of sea turtles. Boca Raton, Fla: CRC Press; 1997. pp. 137–163.
  56. 56. Senko J, Koch V, Megill WM, Carthy RR, Templeton RP, Nichols WJ. Fine scale daily movements and habitat use of East Pacific green turtles at a shallow coastal lagoon in Baja California Sur, Mexico. J Exp Mar Biol Ecol. 2010;391: 92–100.
  57. 57. Thomson J, Heithaus M, Burkholder D, Vaudo J, Wirsing A, Dill L. Site specialists, diet generalists? Isotopic variation, site fidelity, and foraging by loggerhead turtles in Shark Bay, Western Australia. Mar Ecol Prog Ser. 2012;453: 213–226.
  58. 58. Hancock J, Vieira S, Jimenez V, Carvalho Rio J, Rebelo R. Stable isotopes reveal dietary differences and site fidelity in juvenile green turtles foraging around São Tomé Island, West Central Africa. Mar Ecol Prog Ser. 2018;600: 165–177.
  59. 59. Piovano S, Lemons GE, Ciriyawa A, Batibasaga A, Seminoff JA. Diet and recruitment of juvenile green turtles at two foraging grounds in Fiji, South Pacific inferred from in-water capture and stable isotope analysis. Mar Ecol Prog Ser. 2020.
  60. 60. Lavender AL, Bartol SM, Bartol IK. Ontogenetic investigation of underwater hearing capabilities in loggerhead sea turtles (Caretta caretta) using a dual testing approach. J Exp Biol. 2014;217: 2580–2589. pmid:24855679
  61. 61. Martin KJ, Alessi SC, Gaspard JC, Tucker AD, Bauer GB, Mann DA. Underwater hearing in the loggerhead turtle (Caretta caretta): a comparison of behavioral and auditory evoked potential audiograms. J Exp Biol. 2012;215: 3001–3009. pmid:22875768
  62. 62. Ketten D, Bartol SM. Functional measures of sea turtle hearing. Final report. Woods Hole, MA: Woods Hole Oceanographic Institution; 2006. Available:
  63. 63. Bartol SM, Ketten D. Turtle and tuna hearing. In: Swimmer Y, Brill R, editors. Sea turtle and pelagic fish sensory biology: developing techniques to reduce sea turtle bycatch in longline fisheries. US NOAA; 2006. pp. 98–15.
  64. 64. Bartol SM, Musick JA, Lenhardt ML. Auditory evoked potentials of the loggerhead sea turtle (Caretta caretta). Copeia. 1999;1999: 836.
  65. 65. Ridgway SH, Wever EG, McCormick JG, Palin J, Anderson JH. Hearing in the giant green turtle, Chelonia mydas. Proc Natl Acad Sci USA. 1969;64: 884–890. pmid:5264146
  66. 66. Piniak WED, Mann DA, Eckert SA, Harms CA. Amphibious hearing in sea turtles. In: Popper AN, Hawkins A, editors. The effects of noise on aquatic life. New York, NY: Springer New York; 2012. pp. 83–87.
  67. 67. Hildebrand J. Anthropogenic and natural sources of ambient noise in the ocean. Mar Ecol Prog Ser. 2009;395: 5–20.
  68. 68. Ferrara CR, Vogt RC, Sousa-Lima RS, Tardio BMR, Bernardes VCD. Sound communication and social behavior in an Amazonian river turtle (Podocnemis expansa). Herpetologica. 2014;70: 149–156.
  69. 69. Ferrara CR, Vogt RC, Sousa-Lima RS. Turtle vocalizations as the first evidence of posthatching parental care in chelonians. J Comp Psychol. 2013;127: 24–32. pmid:23088649
  70. 70. Giles JC, Davis JA, McCauley RD, Kuchling G. Voice of the turtle: the underwater acoustic repertoire of the long-necked freshwater turtle, Chelodina oblonga. J Acoust Soc Am. 2009;126: 434–443. pmid:19603900
  71. 71. Piovano S, Batibasaga A, Ciriyawa A, LaCasella EL, Dutton PH. Mixed stock analysis of juvenile green turtles aggregating at two foraging grounds in Fiji reveals major contribution from the American Samoa Management Unit. Sci Rep. 2019;9: 3150. pmid:30816199
  72. 72. Álvarez-Varas R, Heidemeyer M, Riginos C, Benítez HA, Reséndiz E, Lara-Uc M, et al. Integrating morphological and genetic data at different spatial scales in a cosmopolitan marine turtle species: challenges for management and conservation. Zool J Linn Soc. 2020.
  73. 73. R Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2019. Available:
  74. 74. Cheng J, Karambelkar B, Xie Y. Create interactive web maps with the JavaScript “leaflet” library. 2019. Available:
  75. 75. GIMP Development Team. GNU image manipulation program. 2019. Available:
  76. 76. Murphy HM, Jenkins GP. Observational methods used in marine spatial monitoring of fishes and associated habitats: a review. Mar Freshw Res. 2010;61: 236.
  77. 77. Limpus CJ, Walter DG. The growth of immature green turtles (Chelonia mydas) under natural conditions. Herpetologica. 1980;36: 162–165.
  78. 78. Ehrhart LM, Ogren LH. Studies in foraging habitats: capturing and handling turtles. In: Eckert KL, Bjorndal KA, Abreu-Grobois FA, Donnelly M, editors. Research and management techniques for the conservation of sea turtles. Washington, DC: IUCN/SSC Marine Turtle Specialist Group Publication; 1999. pp. 61–64.
  79. 79. Fuentes MMPB, Lawler IR, Gyuris E. Dietary preferences of juvenile green turtles (Chelonia mydas) on a tropical reef flat. Wildl Res. 2006;33: 671.
  80. 80. Wildermann NE, Sasso CR, Stokes LW, Snodgrass D, Fuentes MMPB. Habitat use and behavior of multiple species of marine turtles at a foraging area in the northeastern Gulf of Mexico. Front Mar Sci. 2019;6.
  81. 81. Eckert KL, Beggs J. Marine turtle tagging: a manual of recommended practice. Beaufort, N.C.; 2006 p. 40. Report No.: 2. Available:
  82. 82. Acoustical Society of America. ANSI S1.11–2004 Specification for octave, half-octave and third octave band filter sets. American National Standard. 2004. Available:
  83. 83. Freeman SE, Freeman LA, Giorli G, Haas AF. Photosynthesis by marine algae produces sound, contributing to the daytime soundscape on coral reefs. PLOS ONE. 2018;13: e0201766. pmid:30281593
  84. 84. Kaiser JF. On a simple algorithm to calculate the “energy” of a signal. International conference on acoustics, speech, and signal processing. Albuquerque, NM, USA: IEEE; 1990. pp. 381–384.
  85. 85. Buscaino G, Filiciotto F, Buffa G, Di Stefano V, Maccarrone V, Buscaino C, et al. The underwater acoustic activities of the red swamp crayfish Procambarus clarkii. J Acoust Soc Am. 2012;132: 1792–1798. pmid:22978906
  86. 86. Wood SN. Generalized additive models: an introduction with R. 2nd ed. Chapman and Hall/CRC; 2017.
  87. 87. R Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2017. Available:
  88. 88. Nolan B, Salmon M. The behaviour and ecology of snapping shrimp (Crustacea: Alpheus heterochelis and Alpheus normanni). Forma et Functio. 1970;2: 289–335.
  89. 89. Schein H. Aspects of the aggressive and sexual behaviour of Alpheus heterochaelis Say. Mar Behav Physiol. 1975;3: 83–96.
  90. 90. Herberholz J, Schmitz B. Signaling via water currents in behavioral interactions of snapping shrimp (Alpheus heterochaelis). Biol Bull. 2001;201: 6–16. pmid:11526058
  91. 91. Torri M, Corrado R, Falcini F, Cuttitta A, Palatella L, Lacorata G, et al. Planktonic stages of small pelagic fishes (Sardinella aurita and Engraulis encrasicolus) in the central Mediterranean Sea: the key role of physical forcings and implications for fisheries management. Prog Oceanogr. 2018;162: 25–39.
  92. 92. Piovano S, Balletto E, Di Marco S, Dominici A, Giacoma C, Zannetti A. Loggerhead turtle (Caretta caretta) by catches on long-lines: the importance of olfactory stimuli. Ital J Zool. 2004;71: 213–216.
  93. 93. Piovano S, Farcomeni A, Giacoma C. Effects of chemicals from longline baits on the biting behaviour of loggerhead sea turtles. Afr J Mar Sci. 2012;34: 283–287.
  94. 94. Endres CS, Putman NF, Lohmann KJ. Perception of airborne odors by loggerhead sea turtles. J Exp Biol. 2009;212: 3823–3827. pmid:19915124
  95. 95. Swimmer Y, Arauz R, Higgins B, McNaughton L, McCracken M, Ballestero J, et al. Food color and marine turtle feeding behavior: can blue bait reduce turtle bycatch in commercial fisheries? Mar Ecol Prog Ser. 2005;295: 273–278.
  96. 96. Piovano S, Farcomeni A, Giacoma C. Do colours affect biting behaviour in loggerhead sea turtles? Ethol Ecol Evol. 2013;25: 12–20.
  97. 97. Hall RJ, Robson SKA, Ariel E. Colour vision of green turtle (Chelonia mydas) hatchlings: do they still prefer blue under water? PeerJ. 2018;6: e5572. pmid:30258709
  98. 98. Limpus CJ, Limpus DJ. Mangroves in the diet of Chelonia mydas in Queensland, Australia. Mar Turtle Newsl. 2000;89: 13–15.
  99. 99. Taquet C, Taquet M, Dempster T, Soria M, Ciccione S, Roos D, et al. Foraging of the green sea turtle Chelonia mydas on seagrass beds at Mayotte Island (Indian Ocean), determined by acoustic transmitters. Mar Ecol Prog Ser. 2006;306: 295–302.
  100. 100. Fury CA, Harrison PL. Seasonal variation and tidal influences on estuarine use by bottlenose dolphins (Tursiops aduncus). Estuar Coast Shelf Sci. 2011;93: 389–395.
  101. 101. Abrahms B, Hazen EL, Aikens EO, Savoca MS, Goldbogen JA, Bograd SJ, et al. Memory and resource tracking drive blue whale migrations. Proc Natl Acad Sci USA. 2019;116: 5582–5587. pmid:30804188
  102. 102. Samuel Y, Morreale SJ, Clark CW, Greene CH, Richmond ME. Underwater, low-frequency noise in a coastal sea turtle habitat. J Acoust Soc Am. 2005;117: 1465–1472. pmid:15807034
  103. 103. McCauley RD, Fewtrell J, Duncan AJ, Jenner C, Jenner M-N, Penrose JD, et al. Marine seismic surveys—a study of environmental implications. The APPEA Journal. 2000;40: 692.