Introduction

Marine mammals have evolved to use sound as their primary sensory modality—both actively (sound production) and passively (sound reception). That is because sound travels much better (i.e. with less loss and over longer ranges) underwater than does light. Sound also travels faster and over longer ranges underwater than it does in air. Marine mammals produce sound prolifically for communication [1]. Odontocetes (toothed whales) also emit echolocation clicks during foraging and navigation [2]. While marine mammal vocalisations are produced in the larynx, and echolocation clicks in nasal air sacs, some marine mammals also produce mechanical sounds during behaviours at the surface of the water such as breaching or fluke slapping [3].

Recording and listening to these sounds for investigating the presence, distribution, migration, relative abundance and behaviour of marine mammals has proven to be a powerful tool. While traditional visual detection methods of surveying marine mammals continue to be relevant for many applications, passive acoustic methods are increasingly employed due to their high success rates independent of time of day and weather, and their cost-effectiveness in remote, inaccessible locations where long-term data sets are required [4]. Such long data sets are becoming publicly available. For example, there are many years of recordings from Australia’s Integrated Marine Observing System’s (IMOS) passive acoustic stations located off the WA, SA, VIC and NSW coasts. All of the data can be downloaded or requested from the Australian Ocean Data Network (http://imos.org.au/aodn.html).

Underpinning all applications of passive acoustics is the ability to attribute recorded sounds to their sources, including marine mammals. Few Australian studies have comprised simultaneous visual and passive acoustic surveys, compared to studies off North America or Europe. While the underwater biophony in Australia is rich and complex, specifically in Australia’s tropical and subtropical regions (e.g. [5, 6]), we often do not know what species make what sounds. This is a major problem in particular for marine environmental impact assessments. Australia’s waters and continental shelf are rich in resources ranging from minerals and hydrocarbons to commercial fish. Industries such as fisheries, tourism, oil and gas exploration and production, mining, construction and defence abide by environmental health and safety guidelines which may include underwater soundscape analyses, marine mammal surveying and noise impact assessments.

To aid these studies and assessments, we present a summary of sounds emitted by marine mammals occurring in both inshore and offshore waters around the Australian continent and the Australian Antarctic Territory. Example sound files are available at http://cmst.curtin.edu.au/research/marine-mammal-bioacoustics/. Our article is not intended to be a complete summary of all of the literature published on the sounds made by these species, but rather a guide to the literature that best describes the species-specific sound characteristics. We envisage this overview to be used by readers tasked with analysing passive acoustic recordings from Australia.

Methods

A list of Australian marine mammal species and information on each species’ physical appearance, distribution, ecology and threats can be found on the website of the Australian Government Department of the Environment and Energy (http://www.environment.gov.au/marine/marine-species). Based on this list of species occurring in Australia, a literature search was undertaken of sounds recorded worldwide from each species. While some species have been shown to emit slightly different sounds in different geographic regions, information on the sounds of Australian mammals is typically so limited, that recordings from outside of Australia were included to give a broad overview of the types of sounds these animals emit. Articles that reported simultaneous visual and acoustic detections, rather than articles from autonomous passive acoustic recorders in the absence of visual observers, were preferred. The latter were considered only in cases where the sound-producing species could be identified without doubt based on the sound characteristics and where the study added new information important for passive acoustic monitoring (e.g. source levels). We further focussed on articles that described the spectrographic characteristics of sounds, showed example spectrograms and presented measurements such as bandwidth, duration and source level. In addition to the literature review, we provide spectrograms of vocalisations recorded in Australian waters and tables summarising their spectrographic features.

This article is organised by species, using the taxonomy of the Society for Marine Mammalogy (https://www.marinemammalscience.org/species-information/list-marine-mammal-species-subspecies/). Species distribution within Australia and the Commonwealth marine area is based on the maps and information available at the Department of the Environment and Energy website (http://www.environment.gov.au/cgi-bin/sprat/public/mapcaveat.pl), augmented with geographically farther-reaching information from maps on the Convention on the Conservation of Migratory Species of Wild Animals website (http://www.cms.int) and the Society for Marine Mammalogy website.

Marine Mammal Species of Australia

A total of 57 marine mammal species may be found around Australia, including 10 mysticetes (baleen whales), 36 odontocetes (toothed whales), one sirenian (dugong) and 10 pinnipeds (seals and sea lions). These species are listed in Table 1, also indicating whether they are most commonly encountered inshore versus offshore. Many of the species undertake annual migrations. For example, polar feeding mysticetes migrate between their polar feeding grounds in the summer and temperate to tropical breeding grounds in the winter. The “dwarf” mysticete species follow a similar annual migration pattern but shifted towards the equator, with most not reaching polar regions. Some odontocetes migrate, others do not and instead shift their distribution on more local scales typically driven by prey. Some of the pinnipeds stay within the vicinity of their breeding grounds, while others travel over extended distances for food. Indicative ranges of species distribution are given in Fig. 1.

Table 1 List of marine mammal species occurring around Australia and Antarctica and whether they are mostly encountered inshore or offshore

Types of Sounds

In the literature, animal sounds are sometimes given onomatopoeic names (phonetically imitating the sound) and grouped based on what they sound like to the human ear. The same sounds are often given different names by different authors. In this article, we group sounds based on their spectrographic features. However, these can depend on some of the recording and analysis settings (e.g. sampling frequency, filters and analysis window used during Fourier transform). A series of broadband pulses is sometimes called a bout of pulses in the literature, other times an amplitude-modulated (AM) sound, and, if the analysis window is longer than the pulses, the sound looks tonal in spectrograms. To make this even more interesting, some animals are able to produce sounds along a continuum from tonal to pulsed [7,8,9]. As we did not have access to the raw data underlying the various published spectrograms, we based our grouping on published spectral features.

All sounds were grouped into three classes based on their spectrographic features: 1) constant-wave (CW) tones, 2) frequency-modulated (FM) sounds and 3) broadband pulses. CW sounds appear as straight, horizontal lines in spectrograms. FM sounds include upsweeps, downsweeps or sinusoidal contours in spectrograms. Both CW and FM sounds may have harmonic overtones. Pulses are broadband and lack tonal characteristics.

Mysticetes make sounds of all three categories. Pulses may be emitted as fast pulse trains and have often been described as AM sounds (e.g. “moans”). Blue, fin and humpback whales may arrange their sounds into “song” that lasts for hours to days.

Odontocete sounds are typically classified as whistles, burst-pulse sounds and clicks in the literature though not all odontocetes whistle. Whistles are tonal sounds with CW or FM characteristics and may or may not have harmonic overtones. Most studies on whistles have focussed on describing the contour of the fundamental. Burst-pulse sounds are series of rapid pulses, typically reported for some delphinid species. Depending on the settings during Fourier analysis, burst-pulse sounds may appear as FM sounds with multiple overtones and non-harmonic sidebands [10]; however, they were classified as pulses in this article. Clicks are very short (typically <1 ms) and broadband pulses with peak energy at high frequencies (typically tens of kHz). Clicks can be heard as single signals or “slow” repetitive signals with inter-click intervals (ICI) of a few seconds, as in the case of sperm whales, or as rapid trains with millisecond ICI as in the case of odontocete echolocation. The ICI in click trains is typically larger than the inter-pulse interval (IPI) in burst-pulse sounds. The separation into FM sounds and pulses can be confusing. For example, the brief pulses of sei whales are FM downsweeps [11, 12]. And even the very brief clicks of odontocetes can be FM sweeps, like those of many beaked whales [13].

Fig. 1
figure 1

Indicative distribution of a baleen whales and some beaked whales, b, c odontocetes and dugong, d beaked whales and e pinnipeds. The grey circle is around Macquarie Island

Pinnipeds may emit sound in both media, air and water. Most studies have recorded their aerial vocalisations. Some species, such as the Ross seal, produce very similar sounds in both media [14]. The majority of pinniped sounds were classified as pulses in this article (including AM “moans” and “growls”); however, each pulse might be a brief frequency sweep or FM sound as in the case of Weddell seal chugs [15].

Despite these inconsistencies in naming sounds, for each species, we tabulate the sound types, with some notes on common vocalisation names, the frequency range, the duration, the source level (SL) as peak-to-peak sound pressure level (pp), root-mean-square sound pressure level (rms) or sound exposure level (SEL), the location where the sounds were recorded and the references identifying the source of the information. We have tried to keep the localities mentioned in the original literature, but sometimes summarise these to “Mediterranean” or “Northeast Atlantic” when there were several sites, studies or publications.

It is difficult to compile such summary tables from the literature, as each article lists different quantities in different formats. For frequency, duration and source level measurements, sometimes absolute ranges are given, other times means ± standard deviation (even though the measurements were often not Gaussian distributed), or medians and quartiles, or \(5{\mathrm{th}}\) and \(95{\mathrm{th}}\) percentiles, or \(10{\mathrm{th}}\) and \(90{\mathrm{th}}\) percentiles. These quantities and statistics cannot be accurately combined without access to the raw data. In an attempt to provide an overview of the available information, we decided to report the ranges (i.e. the broadest ranges). Therefore, the lowest frequency or duration or level we report is the minimum of all values published (e.g. the lowest percentile, the lowest range or the mean minus the standard deviation), and the highest value we list corresponds to the maximum of all values published (including the highest percentile, the highest range and the mean plus the standard deviation). If only one publication existed, we kept the original metrics as there was no need to combine data from multiple studies. Sometimes, measurements were not given, so we attempted to read data off spectrograms. All quantities, and in particular the observed bandwidth, can strongly be affected by the sound propagation environment and range of source to receiver; however, this is mostly not corrected in the literature.

Fig. 2
figure 2

Duration and fundamental frequency range of tonal (CW and FM) sounds of mysticetes, odontocetes, dugong and pinnipeds under water

Fig. 3
figure 3

Duration and frequency range of mysticete, odontocete, dugong and pinniped underwater pulses and burst-pulse sounds

In the case of whistles, the frequency range reported in this document corresponds to the range in primary frequency fp (i.e. the maximum frequency minus minimum frequency of the fundamental contour) and excludes harmonics. Some whistles are emitted with harmonics, some without. In fact, the same fundamental contour may sometimes be recorded with harmonics and other times without (e.g. in the case of killer whales [16]). This is different for FM sounds produced by mysticetes, where the frequency range reported in the literature mostly includes harmonics. If measurements were reported for only the fundamental frequency and not for harmonics present, we indicate this as “+ harmonics” in the following mysticete table.

In the case of clicks, listing only the range from minimum to maximum reported frequency resulted in all animals apparently producing hugely broadband clicks. Odontocete click emission is directional, and spectra change as a function of angle in both the horizontal and vertical planes, leading to variability in frequency measurements. Furthermore, clicks are commonly bimodal, exhibiting not just one but two peak frequencies, which is missed by studies that report mean or median frequencies. In addition to uncertainty in the angular aspect and range to the animal, the lack of a low-pass anti-aliasing filter during recordings at low sampling frequencies can affect the spectrum. If the click spectrum extends beyond the Nyquist frequency and no anti-aliasing filter is used, the spectral energy “folds” about the Nyquist frequency to lower frequencies. We list multiple frequency measurements for clicks if available: fp, the frequency of maximum spectral power; fc, the centroid frequency that divides the power in a spectrum into two equal parts; and the bandwidth at the \(-\)3 dB or \(-\)10 dB points, or the rms bandwidth, i.e. the standard deviation of the spectrum about fc.

Figures 2, 3 and 4 are an attempt to visualise the frequency ranges and durations of the various animal sounds based on the literature and the authors’ recordings.

Fig. 4
figure 4

Duration and frequency range of odontocete clicks

Fig. 5
figure 5

Southern right whale FM sounds recorded in Fowlers Bay (SA), August 2013 [26] [fs = 12 kHz, NFFT = 1200, 50% overlap, Hanning window]

The Sounds of Mysticetes—Baleen Whales

Eubalaena australis—Southern Right Whale

Southern right whales are found in the southern hemisphere, along the southern coasts of South America, Africa, Australia and New Zealand. They spend summers on their feeding grounds close to Antarctica and migrate to their northern breeding grounds in winter. Their seasonal Australian range (May–November) extends from Hervey Bay (QLD) on the east coast, along the southern coastline, to Exmouth (WA) on the west coast [17]. Close to shore, southern right whales seasonally form distinct groups in about a dozen coastal aggregation areas [18,19,20].

Descriptions of the sounds of southern right whales have been published from the South Atlantic, off Argentina and Brazil [21,22,23,24] and from the South Pacific, around New Zealand [25]. CMST has been recording southern right whales in Fowlers Bay (SA) during the austral winter [26].

All of the populations produce FM sounds, with fundamentals ranging from 30 to 500 Hz, harmonics up to 5 kHz, and durations of 0.2–3.8 s (Fig. 5a; [21,22,23,24,25]). The most frequently reported FM sound is a 0.2–2.7 s upsweep from 30 to 450 Hz plus harmonics (referred to as the “upcall” in the right whale literature; Fig. 5b; [21, 24, 25]). It seems to be the primary contact call, with variability in the call structure potentially coding caller identity [21]. Right whales off Argentina have shifted their upcall frequency upwards by about 10–20 Hz from 1977 to 2000, possibly due to an increase in low-frequency anthropogenic noise [27].

All populations also emit CW sounds, with fundamentals ranging from 30 to 500 Hz, and harmonics up to 3.5 kHz, and durations of 0.2–6 s [21,22,23,24,25]. An exception is a low-frequency CW sound of up to 15.5-s duration, recorded off New Zealand [25]. Pulsed sounds last 0.2–6.3 s with energy in the band 30–3300 Hz [21, 22, 24, 25]. Finally, southern right whales also emit brief, broadband (30–8400 Hz) pulses, called “gunshots” with source levels of 172–187 dB re 1 \(\upmu \)Pa rms @ 1 m, and 0.1–0.4-s duration [21, 22, 25].

Caperea marginata—Pygmy Right Whale

Pygmy right whales have been seen around the southernmost coasts of South America, Africa, Australia [28, 29] and New Zealand suggesting a circumpolar distribution in temperate waters of the southern hemisphere [30]. The rarity of sightings means that very little is known about the distribution of this species. Currently, there is only one published paper on pygmy right whale acoustics, describing an FM downsweep from 200 to 50 Hz, lasting 0.3 s (Fig. 6; [31])

Fig. 6
figure 6

Pygmy right whale downsweep [31] [fs = 12 kHz, NFFT = 1200, 50% overlap, Hanning window]

Balaenoptera acutorostrata—Common Minke Whale

B. acutorostrata is found globally and typically referred to as the common minke whale. There is a North Atlantic subspecies (B.a. acutorostrata), a North Pacific subspecies (B.a. scammoni) and an unnamed subspecies that occurs in the southern hemisphere and that is referred to as the dwarf minke whale. This subspecies is typically sighted and recorded in warm, tropical to subtropical waters, rather than at high latitude. Its range extends from \(11^{\circ }\hbox {S}\) to about \(41^{\circ }\hbox {S}\). These animals are more likely distributed closer to shore [32]. Like other baleen whales, their breeding grounds are likely in lower latitudes (e.g. the Great Barrier Reef), with mature animals migrating there during the winter months.

Common minke whales have been recorded in the east of Canada [33, 34], in the Gulf of Maine [35, 36], in the Caribbean [37, 38], in the North Pacific including Hawaii [39,40,41] and the Mariana Islands [42], in the Great Barrier Reef, Australia [43], and along the Western Australian coast to as far north as \(14^{\circ }\hbox {S}\) [44].

Common minke whales primarily produce pulsed sounds which are arranged into trains that either have a constant pulse repetition rate, or are sped up, or slowed down. The frequency band is 10–800 Hz. Pulses have a duration of 0.04–0.3 s; 1.5–4.5 pulses occur per second, in trains of 10–70-s duration, and with source levels of 159–176 dB re 1 \(\upmu \)Pa rms @ 1 m or 173–189 dB re 1 \(\upmu \)Pa pp @ 1 m [35,36,37]. Very-high-frequency clicks (4–7.5 kHz; 6.75 ± 1.02 clicks/s, 166–173 dB re 1 \(\upmu \)Pa pp @ 1 m) were only reported once from Newfoundland [33].

North Pacific minke whales were first reported to produce the famous “boing” sound. The “boing” is an AM sound, often described as a train of rapid pulses, and which appears in spectrograms (with analysis windows longer than the pulse duration) as a broadband FM sound with many sidebands. The “boing” covers a bandwidth of 1–5 kHz, lasts 1.4–4.2 s and exhibits about 112–118 pulses/s, at a source level of about 150 dB re 1 \(\upmu \)Pa rms @ 1 m [39,40,41]. These “boing” sounds are commonly recorded by CMST on Australia’s Northwest Shelf as well (e.g. [44]; Fig. 7a, b). At the Australian Great Barrier Reef, dwarf minke whales produce the pulsed, metallic-sounding “star-wars call” (50–9400 Hz, 1–3 s, 150–165 dB re 1 \(\upmu \)Pa rms @ 1 m; [43]; Fig. 7c).

FM downsweeps (50–250 Hz; 0.2–0.4 s) were reported from the St Lawrence Estuary, Canada [34], and the Great Barrier Reef, Australia [43].

Fig. 7
figure 7

a Spectrogram of a minke whale “boing” recorded by CMST on Australia’s Northwest Shelf, May 2009 [fs = 6 kHz, NFFT = 600, 50% overlap, Hanning window], and b corresponding waveform. c Spectrogram of the “star-wars call” recorded near the Great Barrier Reef, July 1998 [43] [fs = 44 kHz, NFFT = 2048, 50% overlap, Hanning window]

Balaenoptera bonaerensis—Antarctic Minke Whale, Dark-shoulder Minke Whale

There are two species of minke whales: B. acutorostrata and B. bonaerensis. The latter only occurs in the southern hemisphere from Antarctica to near the equator (\(10^{\circ }\hbox {S}\)) and migrates seasonally as do other baleen species. This species’ Australian range extends along the whole of the eastern and western coastlines (though less is known about their distribution on the western coastline), and its southern range extends to the ice edge well within the Australian Antarctic Territory [45]. Antarctic minke whales are considered an offshore and migratory species. In the winter breeding grounds at low latitudes in other regions, these animals appear to be distributed off the continental shelf edge [46], suggesting a similar winter distribution for Australian waters during the breeding season. Although some individuals might remain in Antarctica over the winter [47].

In Antarctica, FM downsweeps with a fundamental of 60–140 Hz, occurring with and without harmonics, with a duration of 0.2 ± 0.1 s and a source level of 147.3 ± 5.3 dB re 1 \(\upmu \)Pa rms @ 1 m have been recorded [48,49,50,51]. Antarctic mine whales also emit bouts of rapid broadband pulses (which are brief FM downsweeps), sounding like a duck’s quack (named the “bioduck” sound). This unique “bioduck” sound has been recorded in Antarctica and off Australia for decades, yet it was not until Risch et al. [49], that this sound was positively attributed to Antarctic minke whales. The “bioduck” sound contains pulses of 0.1–0.4-s duration, in bouts of 5–12 pulses, with a 1–3-s interval between bouts, and source levels of 140.2 ± 3.6 dB re 1 \(\upmu \)Pa rms @ 1 m [49, 51]. The “bioduck” (Fig. 8) is also commonly heard in the Perth Canyon in the austral winter, where it covers 100–500 Hz in frequency with about 0.2 s/pulse and 5 pulses/train [6] and occurs as far north as Exmouth, WA [44].

Fig. 8
figure 8

“Bioduck” sound recorded by CMST a south of Exmouth (WA), August 2011, and b, c in the Perth Canyon (WA), September 2003 [fs = 10 kHz, NFFT = 128, 50% overlap, Hanning window]

Balaenoptera borealis—Sei Whale

The sei whale is a global species and occurs in both hemispheres. The northern hemisphere subspecies is B.b. borealis, and the southern hemisphere subspecies is B.b. schlegelii. This species migrates once a year between its polar feeding grounds (summer) and tropical breeding grounds (winter). Sei whales tend to stay well offshore [52], though their tendency to be confused with Bryde’s whales has resulted in confusion about their distribution. The extent of occurrence, their geographic habitat, seasonal presence and absolute abundance around Australia are still poorly understood [53].

Acoustic recordings exist from the North Atlantic, off the US east coast [54,55,56] and Nova Scotia, Canada [11, 12], off the Azores [57], from Hawaii [58], south of New Zealand [59] and Antarctica [60]. There are no recordings of sei whales in Australian waters with simultaneous visual species identification.

Sei whales produce 0.5–2-s CW sounds at 200–500 Hz, without or with higher-frequency harmonics, including multi-part frequency stepping tonals [60]. The most frequently reported sounds are simple FM sounds: upsweeps and downsweeps, lasting 0.7–2.2 s, with the fundamental covering a low-frequency band (20–220 Hz; [54,55,56,57,58,59]) or mid-frequency band (200–600 Hz; [60]). These FM sounds have been recorded with and without harmonics. Source levels of FM and CW sounds are 147–183 dB re 1 \(\upmu \)Pa rms @ 1 m [55,56,57, 60]. Very brief (0.03–0.04 s) high-frequency (1.5–3.5 kHz) downsweeps occurring in bouts of 10–20 have been reported from Nova Scotia [11] and also been described as a series of pulses with peak energy at 3 kHz and 7–10 pulses over 0.7 s [12]. Broadband (100–800 Hz) pulsed sounds of 1–3-s duration have also been documented [60].

Balaenoptera edeni—Bryde’s Whale

Bryde’s whales inhabit tropical to subtropical waters in both the northern and the southern hemispheres. Two subspecies have been identified: the larger Bryde’s whale (B.e. brydei) that occurs worldwide in offshore waters and the smaller Eden’s whale (B.e. edeni) that may only occur in the Indo-Pacific. Offshore and coastal forms have been found in South Africa [61] and Japan [62], and similar forms likely exist in Australian waters. Bryde’s whales have been sighted in offshore and inshore waters all around the Australian mainland. The coastal form is considered resident, with no evidence of large-scale movements in Australia [63]. The offshore form may migrate seasonally, as do other baleen whale species, though more data are required.

Bryde’s whales have been recorded in the Gulf of Mexico [64, 65], in the Gulf of California [66,67,68], in the Caribbean [69], off Brazil [70], off Japan [69, 71], in the Eastern Tropical Pacific [69, 72] and off Western Australia.

Bryde’s whales emit 0.3–7-s CW tones at 8–80 Hz with and without higher harmonics [69, 70, 72]. They also make FM sounds, primarily downsweeps (50–1000 Hz, 0.1–5 s, 141–174 dB re 1 \(\upmu \)Pa rms @ 1 m) with and without harmonics [64,65,66, 68,69,70,71], sometimes occurring in sequences of 2–25 downsweeps [65]. FM sounds recorded by CWR in the presence of Bryde’s whales are shown in Fig. 9. Bryde’s whales further make AM sounds, which have been described as “moans” or rapid pulse trains (83–950 Hz, 0.3–51 s, 0.03–0.1 s/pulse, 10–130 pulses/s [64, 67, 70].

Fig. 9
figure 9

Sounds recorded by CWR in the presence of Bryde’s whales off Broome (WA), July 2013 [fs = 48 kHz, NFFT = 12,000, 50% overlap, Hanning window]

Balaenoptera musculus—Blue Whale

Blue whales (B. musculus) come in currently five recognised subspecies: the northern blue whale (B.m. musculus), which is only seen in the northern hemisphere, the Antarctic blue whale (B.m. intermedia), the northern Indian Ocean blue whale (B.m. indica), the pygmy blue whale (B.m. brevicauda) and the Chilean blue whale (unnamed). Based on acoustic analyses of blue whale sounds, size differences and genetics, there could be additional distinct populations or subspecies [73, 74]. Australian waters are home to at least two of these subspecies: the Antarctic blue whale and the pygmy blue whale. The Antarctic blue whale occurs at high latitudes into the Antarctic during the austral summer and migrates to lower latitudes around Australia, southern Africa and South America in the winter [75], although Antarctic blue whale vocalisations have been recorded in Antarctica all year-round, indicating that some animals remain at high latitudes throughout the year [76]. Pygmy blue whales are considered subantarctic, not occurring as far south as Antarctic blue whales. They also migrate annually between higher latitudes in the summer and lower latitudes in the winter and are believed to roam as far north as the northern Indian Ocean and Banda Sea [77]. The Perth Canyon (WA; [78, 79]) and the Bonney upwelling (SA; [80]) are two notable feeding aggregations of pygmy blue whales.

Southern hemisphere blue whales have been acoustically recorded in the southeast Pacific, off Chile [73, 81], in the southwest Pacific, off New Zealand [82, 83], in the Indian Ocean [84, 85], off Antarctica [76, 86, 87] and Australia [88,89,90]. They emit patterned and repetitive sequences of vocalisations, called “song”, lasting up to hours at a time. It appears only males sing, possibly as a form of reproductive display [91]. On a worldwide basis, song units can be CW, FM or pulsed, covering a frequency band of 14–300 Hz. Three- and four-unit phrases are common in blue whale song [73, 81,82,83,84]. Unit precursors in the form of 0.7–1-s pulses with peak energy between 347 and 430 Hz have been recorded in Chilean blue whale song [73, 81] and are similar to precursors recorded from blue whales in the North Pacific [92]. In addition to song, blue whales emit non-patterned sounds with FM sweeps being very common (e.g. 1–3 s, 80–40 Hz downsweeps called “D-calls” [91, 92]).

Pygmy blue whale song has been recorded off WA and along the southern Australian coast [6, 79, 85, 93,94,95,96,97,98]. Song units are FM type, with a fundamental frequency of 15–25 Hz and harmonics up to 100 Hz, although this can be higher for nearby animals. Song units are 15–40 s long, and several units (The classic form is three units.) are arranged into phrases of 60–120-s duration (Fig. 10a, b). Source levels of 168–181 dB re 1 \(\upmu \)Pa rms @ 1 m have been reported. It is interesting to note that the frequency of the song units has been decreasing steadily by about 0.35 ± 0.3 Hz/year over nine years [93] and is still decreasing at the same rate (A. Gavrilov, personal communication 2017). Non-song sounds can be 0.9–4.4 s long, are usually FM and are up to several hundred Hertz in frequency (Fig. 10c; [95]).

Blue whale song around New Zealand differs from the Western Australian pygmy blue whale song [83, 98, 99]. The typical sequence of units comprises two consecutive AM tones with associated harmonics, followed by an upsweep. A weak AM sound sometimes precedes this sequence. This phrase is of the order of 40–60 s, and the fundamental frequency is in the range 15–30 Hz, although harmonics can reach several hundred Hz (Fig. 10d). The New Zealand song type is also heard off the southeast coast of Australia with the dividing line between the New Zealand and Western Australian song types appearing to approximate the longitude of Bass Strait.

It is not always certain which subspecies of blue whale was recorded, except in the case of the “Z-call”, which has been attributed to the Antarctic blue whale (Fig. 10e). The “Z-call” appears as a Z in spectrograms, with the first part being a CW component at about 28 Hz lasting a few seconds, followed by a rapid downsweep to about 18 Hz, which is the last CW component, also lasting a few seconds [85,86,87, 89, 100]. In total, the “Z-call” has a duration of 16–30 s and can occur arranged into song of several hours. It is not necessarily emitted as a complete Z, but also occurs as just the higher tone, or the higher tone plus the rapid downsweep [100]. It is interesting to note that the frequency of the first tonal of the “Z-call” of Antarctic blue whales decreased by 1 Hz in 9 years [88] and is still decreasing at the same rate (A. Gavrilov, personal communication 2017). In addition to the “Z-call”, FM sweeps have been recorded from Antarctic blue whales, at 13–107 Hz, 0.8–7.5-s duration [100].

Blue whales produce some of the loudest biological sounds in the ocean (up to 189 ± 3 dB re 1 \(\upmu \)Pa rms @ 1 m; [81, 85, 86]), at the lowest frequencies, potentially yielding very long sound propagation ranges via ducting in the deep sound channel.

Fig. 10
figure 10

a Pygmy blue whale song units, b song sequence, c non-song sounds and d D-calls recorded by CMST in Geographe Bay (WA), November 2011. e Antarctic blue whale “Z-call” recorded by CMST in Bass Strait, May 2004. f New Zealand song type recorded by CMST in Bass Strait, March 2016. [fs = 12 kHz, NFFT = 8192, 50% overlap, Hanning window]

Balaenoptera omurai—Omura’s Whale

Omura’s whales were only described as a separate species from Bryde’s whales that have a similar appearance, in 2003 [101, 102]. Most confirmed records of the species are from specimens taken during research whaling and strandings. Based on these records, the distribution of the species is thought to be in tropical to subtropical latitudes of the Indo-Pacific region [103]. In Australia, validated specimens have been recorded offshore from the Cocos Islands [101], near Exmouth (WA; [103]) and in the Gulf of St. Vincent (SA; [104]). However, there is no information available on the spatial and temporal distribution, behaviour and abundance of this species in Australia. In fact, except for recent work off northwest Madagascar, there is a complete absence of ecological and behavioural data for the species throughout its distribution [105]. Only one published study to date has attributed vocalisations to Omura’s whale [105]. Low-frequency (15–50 Hz) AM vocalisations with an average duration of 9.2 ± 0.92 s produced every 134–177 s were described from boat-based recordings in the presence of Omura’s whales. Similar sounds (20–60 Hz, 12–15 s, repeated every 160–210 s; Fig. 11) are frequently recorded by CMST and CWR on Australia’s Northwest Shelf and simultaneous visual observations and photographs confirmed the presence of Omura’s whales.

Fig. 11
figure 11

Omura’s whale sound recorded by CMST on Australia’s Northwest Shelf, May 2006 [fs = 10 kHz, NFFT = 5000, 50% overlap, Hanning window]

Balaenoptera physalus—Fin Whale

The fin whale is a global species, inhabiting all the world’s oceans, migrating from the polar regions (feeding grounds) in summer to the equatorial regions (breeding grounds) in winter. However, there are exceptions, and non-migratory fin whales have been reported in the northern hemisphere [106]. There are three recognised subspecies: pygmy fin whales (B.p. patachonica), northern fin whales (B.p. physalus) and southern fin whales (B.p. quoyi). Both pygmy and southern fin whales occur in the southern hemisphere. Fin whale occurrence and migration along the Australian coasts are still poorly understood and known primarily from stranding events and whaling records. Strandings have been reported in WA, SA, VIC and TAS, suggesting that they are present around the southern coasts of the Australian continent. The migration routes and location of breeding grounds are uncertain [107].

Fin whales have been acoustically recorded off north-eastern Canada [108], in the St Lawrence Estuary [109], in the Bering Sea [110], in the North Pacific and North Atlantic [111,112,113], in the Gulf of California [66, 110, 114, 115], in Antarctica [76, 86] and in the Perth Canyon [6].

Fin whales emit characteristic 1-s FM downsweeps from about 30 to 15 Hz, commonly called the “20-Hz signal” [109, 114, 116]. These can be emitted in singlets, doublets or triplets and arranged into “song” lasting many hours, with IPI of 5–26 s [112, 115]. It is believed only males sing in order to attract females [117]. The IPI varies with geographic region, season and behaviour [118]. There sometimes is a higher-frequency component at 90–140 Hz (Fig. 12; [6, 76, 86, 108]). Many simultaneously singing fin whales can significantly raise ambient noise levels around 20 Hz by creating a continuous 20 Hz band in underwater acoustic recordings [6]. Source levels of up to 195 dB re 1 \(\upmu \)Pa rms @ 1 m have been reported [86, 111, 113] and are among the highest source levels of biological sounds underwater.

They also emit higher-frequency FM sounds of up to 310 Hz and up to 5-s duration [109, 110, 115], an AM “rumble” at 10–30 Hz and up to 30-s duration [109], AM “moans” up to 100 Hz, 2-s duration and 159–183 dB re 1 \(\upmu \)Pa rms @ 1 m source level [66], CW tones between 50 and 150 Hz and 0.3–0.7-s duration [109], and 34 and 68 Hz CW units arranged into 2-unit phrases of 3-s duration with 1.6–2.2 phrases/minute [66].

Fig. 12
figure 12

Fin whale song recorded by CMST in the Perth Canyon (WA), May 2003 [fs = 10 kHz, NFFT = 5000, 50% overlap, Hanning window]

Megaptera novaeangliae—Humpback Whale

Humpback whales occur globally as three subspecies: North Atlantic humpback whales (M.n. novaeangliae), North Pacific humpback whales (M.n. kuzira) and southern humpback whales (M.n. australis). All migrate annually between their winter breeding grounds in tropical waters and their summer feeding grounds in polar waters. The Australian migration follows two main routes along the QLD and WA coasts resulting in two different migratory populations [119]. Some animals skip a migration and over-winter at high latitudes [120].

The sounds of humpback whales have been studied extensively in Alaska [121,122,123], off the eastern USA [124], at low latitudes [125,126,127,128,129], off eastern Australia [130,131,132,133,134,135,136] and off Antarctica [137].

Male humpback whales use “song” likely as a means of sexual selection aimed at females or to mediate male–male interactions [128, 138, 139]. A song session can last for hours or days. All males of one population sing approximately the same song at any one time. Song normally changes slowly throughout the season and from season to season, although there can be abrupt song changes between years [140]. Males sing during the winter breeding season and during migration to and from the breeding grounds. Only rarely has singing been heard on the summer feeding grounds [137, 141]. Exceptions where humpback whales remained and sang at high latitudes during winter (breeding season) have been reported from Iceland [142]. Australian humpback whales that over-winter on their summer feeding grounds have not been heard singing [120].

There is a lot of literature on humpback whale song, geographic differences and changes over time (e.g. [121]). In Australia, the song differs between the west coast and the east coast populations of humpback whales; however, a rapid adaptation to the west coast song was observed by the east coast animals in 1 year, perhaps due to a few individuals moving from west to east [140] or due to song exchange on the summer feeding grounds in Antarctica where the Australian populations might converge in summer, and where some singing has been heard during the summer feeding season [137].

Song consists of 0.2–4 s units arranged into phrases, which are arranged into themes [127, 128, 139]. Units can be of CW, FM and AM type, typically with harmonics. The bandwidth of song with overtones can be as broad as from 20 Hz to more than 24 kHz; source levels are 151–189 dB re 1 \(\upmu \)Pa rms @ 1 m [125, 126] (Fig. 13).

Both male and female humpback whales produce social sounds. Social sounds include CW, FM, AM sounds and discrete pulses, as well as sounds from the blowhole, and sounds from breaching or slapping the pectoral or tail fins [3]. Social sounds can also include song units, termed “song-unit social sounds” [131], which are used during social interactions but are not emitted as continuous patterned song. Source levels of all types of social sounds range from 123 to 192 dB re 1 \(\upmu \)Pa rms @ 1 m [122, 130]. A 3-year study off the east coast of Australia found 34 different sound types [131], either as single sounds or in bouts [133]. However, only a small subset of these sound types remain in the repertoire over time [134], meaning that the sound catalogue is highly variable and constantly changing. All three of these studies recorded whales during their southern migration only; therefore, it is possible they produce many more sound types while feeding or on the breeding grounds.

Click trains termed “megapclicks” were recorded with DTAGs attached to two animals, during night-time foraging [143]. Clicks were broadband, with most energy below 2 kHz. The click interval decreased towards the end of a click train, ending in a buzz similar to feeding buzzes in odontocetes [144]. As further proof of these click trains being associated with feeding, click trains were emitted at the bottom of dives and the DTAGs recorded sharp body rolls towards the end of a click train [143]. Paired burst bouts of 0.25-s duration, with peak energy below 1 kHz, occurring in series of 2–120 paired bursts were also recorded during bottom feeding [145].

Fig. 13
figure 13

Examples of humpback whale song recorded by CMST on Australia’s Northwest Shelf, August 2013 [fs = 10 kHz, NFFT = 5000, 50% overlap, Hanning window]

The Sounds of Odontocetes—Toothed Whales

Delphinus delphis—Common Dolphin, Short-beaked Common Dolphin

This is a global species with currently four recognised subspecies: the common dolphin (D.d. delphis), the eastern North Pacific long-beaked common dolphin (D.d. bairdii), the Black Sea common dolphin (D.d. ponticus) and the Indo-Pacific common dolphin (D.d. tropicalis). Altogether, this species can be found in inshore and offshore tropical to temperate waters. It has been sighted all around Australia, albeit rarely in the north [146, 147]. There is some indication of migration and seasonal shifts in distribution [148]. There have been no publications of this species’ sounds in Australia previously. Instead, this species’ vocalisations have been described from New Zealand [149, 150], the Eastern Tropical Pacific [151, 152], the UK [153,154,155], the Mediterranean [156,157,158] and the Atlantic Ocean [158].

Whistles have fundamentals from 3 to 24 kHz and last 0.01–4 s [149,150,151,152,153,154,155,156,157,158,159]. Burst-pulse sounds have been reported at 2–14 kHz, lasting 0.5–0.75 s [157]; however, recordings from Australia appear much more broadband with energy up to 160 kHz (Fig. 14). Clicks peak between 23 and 67 kHz and have a 17–45-kHz bandwidth. They last 50–150 \(\upmu \)s (Dziedzic, 1978 quoted in [2]).

Fig. 14
figure 14

Common dolphin whistles, burst-pulse sounds and clicks; a recorded by CWR off SA, February 2013 [fs = 48 kHz, NFFT = 4800, 50% overlap, Hanning window]; b, c recorded by Sue Mason in Port Phillip Bay (VIC), June 2013 [fs = 384 kHz, NFFT = 19,200, 50% overlap, Hanning window]

Feresa attenuata—Pygmy Killer Whale

Pygmy killer whales are mainly tropical, but can be found in subtropical waters. They occupy offshore and inshore waters globally, though are thought to only come close to shore near oceanic islands. Sightings—without simultaneous recordings—and strandings have occurred off NSW and WA, and they have been sighted off the northeast of Australia [147, 160]. The spectral characteristics of their clicks (10–150 kHz, 25 \(\upmu \)s, 197–223 dB re 1 \(\upmu \)Pa pp @ 1 m) were published from the northern Indian Ocean [161]. Clicks, burst-pulse sounds and whistles were also recorded by CWR at Broken Ridge, Indian Ocean (Fig. 15). Whistles had fundamentals at frequencies above 5 kHz, often peaking around 10 kHz, with energy up to the Nyquist frequency, and lasted 0.3–0.8 s. Burst-pulse sounds had most energy above 3 kHz up to the Nyquist frequency, and commonly lasted 1–3.5 s.

Fig. 15
figure 15

Pygmy killer whale whistles, burst-pulse sounds and clicks recorded by CWR at Broken Ridge, Indian Ocean, November 2015 [fs = 48 kHz, NFFT = 4800, 50% overlap, Hanning window]

Globicephala macrorhynchus—Short-Finned Pilot Whale

Short-finned pilot whales occur globally in tropical to temperate waters both offshore and inshore; however, they are considered a deep-water species. Though thought abundant, only a few strandings have occurred in Australia [147]. Recordings from Australia have not been published yet.

In the Caribbean, Eastern Tropical Pacific, Gulf of Mexico, south-eastern USA, Canary Islands and Bahamas, this species emits whistles with fundamental frequencies ranging from 1 to 20 kHz and 0.1–2.5-s duration [8, 151, 152, 162,163,164,165,166]. Burst-pulse sounds cover a frequency band from 1 to greater than 30 kHz, lasting up to 1.5 s [8, 165]. Short-finned pilot whales exhibit the same graded structure of burst-pulse sounds as false killer whales, characteristic of a smooth burst-pulse to whistle transition [8]. Short-finned pilot whales further make biphonations, i.e. the simultaneous emission of two vocalisations by the same individual [8, 166, 167]. Foraging clicks and buzzes during deep dives have been recorded in the Canary Islands and off Hawaii showing a peak frequency of 8–39 kHz and 0.2–1.2 ms per click [168, 169]. Feeding buzzes last 0.7–6.8 s with 4–10-ms ICI; rasps are shorter (0.2–0.6 s) with longer ICI (11–50 ms) [166]. Finally, this species has been reported to copy tonal sounds of anthropogenic sources (e.g. mid-frequency sonar [170]). Australian recordings were obtained during a Coral Sea cruise by some of the authors (Fig. 16; [171]).

Fig. 16
figure 16

Short-finned pilot whale whistles (a) and clicks (b) recorded in the Coral Sea, July 2009 [171] [fs = 88 kHz, NFFT = 4400, 50% overlap, Hanning window]

Globicephala melas—Long-Finned Pilot Whale

There are two subspecies of long-finned pilot whales: G.m. melas in the North Atlantic and G.m. edwardii circum-Antarctic in the southern hemisphere. Little is known about the distribution of long-finned pilot whales in Australia, except that they prefer the colder water of the southern states and the Antarctic Territory [147].

Recordings from Australia have not been previously published. In the Mediterranean, eastern USA and Canada, and the northwest Atlantic, this species whistles between 1 and 11 kHz (fundamental) lasting 0.1–3 s per whistle [156, 157, 172,173,174]. CW tones have been reported at 1–8 kHz, 0.7–3-s duration off Newfoundland [173]. Burst-pulses cover a broader band (100–22,000 Hz) and last 0.1–2.2 s [157, 175]. Biphonations have been reported as well as a graded structure of burst-pulse to whistle transitions [157, 175]. Rhythmic repeated call sequences are common during social behaviours [176]. Long-finned pilot whales have also been reported to imitate mid-frequency sonar sounds in controlled-exposure experiments [177]. CMST and CWR recorded long-finned pilot whales on the shelf break south of Esperance (WA) and found CW and FM sounds with fundamentals between 1 and 6 kHz lasting 0.5–1.1 s (Fig. 17).

Fig. 17
figure 17

Long-finned pilot whale whistles, burst-pulse sounds and clicks recorded by CMST and CWR south of Esperance (WA), January 2016 [fs = 48 kHz, NFFT = 1200, 50% overlap, Hanning window]

Grampus griseus—Risso’s Dolphin, Grampus

Risso’s dolphins occur globally in tropical to subantarctic waters, both offshore and inshore. They have been reported in all Australian states except TAS and the NT [147]. They are thought to be an abundant species. Fraser Island has the only suspected “resident” population in Australia [178].

Recordings of low-frequency FM sounds with overtones and sidebands, whistles, burst-pulse sounds, clicks and biphonations (whistles and simultaneous burst-pulse sounds) were obtained off Newcastle (NSW; [179]). FM sounds with overtones and sidebands covered a bandwidth of 100–1300 Hz and lasted 4 ± 1.2 s. Such low-frequency sounds have not been reported from other parts of the world. Whistles recorded off Newcastle had fundamental frequencies from 4 to greater than 22 kHz (the recording bandwidth was limited to 22 kHz) and lasted 1–13 s [179]. While the frequency range matched that of whistles recorded off California, Egypt, Gran Canaria, the Azores, Scotland and in the Mediterranean [156, 180, 181], the duration of whistles from Australia was longer. Burst-pulse sounds off NSW covered a frequency band from 1 to above 22 kHz and lasted 2–13 s [179]. Animals recorded in the northern Indian Ocean, off California and off Gran Canaria, as well as captive animals in Hawaii and Taiwan emitted clicks with spectra extending from 10 to 140 kHz, durations of 26–118 \(\upmu \)s, ICI of 147–292 ms, source levels of 171–222 dB re 1 \(\upmu \)P pp @ 1 m, 163–210 dB re 1 \(\upmu \)Pa rms @ 1 m and 147–166 dB re 1 \(\upmu \hbox {Pa}^{2}\)s @ 1 m, in trains of 2-s duration [181,182,183,184,185]. Off NSW, click trains were longer, up to 24 s [179]. Click trains come in a variety of ICI patterns, e.g. as speed-up, slow-down and constant-ICI trains. Some click trains end in (feeding) buzzes [181]. An example of clicks recorded by CWR in Jervis Bay (NSW) is shown in Fig. 18.

Fig. 18
figure 18

Risso’s dolphin clicks recorded by CWR in Jervis Bay (NSW), February 2013 [fs = 32 kHz, NFFT = 320, 50% overlap, Hanning window]

Lagenodelphis hosei—Fraser’s Dolphin, Sarawak Dolphin

This is a global, tropical to subtropical, offshore species, also occurring along the outer continental shelf and slope in deep water (>1 km deep). Strandings have been reported along Australia’s northern, western and eastern coasts [147]. Recorded acoustically in the Caribbean and the Eastern Tropical Pacific, these animals make whistles with 4–24 kHz fundamentals, lasting 0.06–2 s [186, 187]. Their clicks range to greater than 40 kHz [187]. On Australia’s northwest coast, whistles had fundamentals from 1.5 to 7 kHz and lasted 0.7–3.5 s (Fig. 19).

Fig. 19
figure 19

Fraser’s dolphin whistles and clicks recorded by CWR on Australia’s northwest coast, October 2012 [fs = 32 kHz, NFFT = 1600, 50% overlap, Hanning window]

Lagenorhynchus cruciger—Hourglass Dolphin

The hourglass dolphin is a southern hemisphere, circumpolar species, found from close to the Antarctic ice to mostly \(45^{\circ }\hbox {S}\), but occasionally as far north as \(36^{\circ }\hbox {S}\). They are rarely seen near land and usually sighted in the Antarctic away from pack ice [147]. Their echolocation clicks cover 110–140 kHz, are 79–176 \(\upmu \)s long and have source levels of 190–203 dB re 1 \(\upmu \)Pa pp @ 1 m or 179–193 dB re 1 \(\upmu \)Pa rms @ 1 m [188, 189].

Lagenorhynchus obscurus—Dusky Dolphin

Dusky dolphins can be found off the coasts of South America (Fitzroy’s dolphin, L.o. fitzroyi; and Peruvian/Chilean dusky dolphin, L.o. posidonia), southern Africa (African dusky dolphin, L.o. obscurus), New Zealand (unnamed subspecies) and southern Australia and TAS, as well as some oceanic islands. The subspecific status of the Australian animals is unknown as they are rarely sighted in Australia but are thought to occur across southern Australia [190]. They have only been recorded acoustically in New Zealand. They emit whistles with fundamentals of 7 to greater than 16 kHz, lasting 0.3–0.8 s. Their burst-pulse sounds range from 1 to greater than 16 kHz and last 0.1–0.6 s. Clicks cover 40–110 kHz with bimodal structure and have a source level of 210 dB re 1 \(\upmu \)Pa pp @ 1 m [191, 192].

Lissodelphis peronii—Southern Right Whale Dolphin

The distribution range of this species is subtropical to subantarctic oceans of the southern hemisphere. The range and total population have not been estimated or closely studied, but the species is abundant off the coast of New Zealand, often recorded with other cetaceans, such as the dusky dolphin. In Australia, they are thought to occur off the southern continental shelf [147]. No records of vocalisations have been found in the literature.

Orcaella heinsohni—Australian Snubfin Dolphin

The Australian snubfin dolphin is a relatively “new” species. These animals were previously classified as Orcaella brevirostris, or Irrawaddy dolphin, common in southeast Asia. Yet consistent and statistically significant differences in anatomical skull characteristics, dorsal fin height, colouration and mtDNA supported the classification of the Australian populations as a new species: O. heinsohni [193]. Snubfin dolphins stay in tropical, shallow, protected coastal and estuarine waters, as far south as Broome (WA) and Brisbane (QLD).

Two publications are available on the acoustic repertoire of this species [194, 195]. Whistles have fundamentals ranging from 600 Hz to 13 kHz and last 0.1–0.45 s. CW sounds were not found. Burst-pulse sounds range from 1 to greater than 22 kHz and last 4 ± 2 s. Clicks trains span 300 Hz to greater than 22 kHz, last 0.1–20 s and have 10–46 clicks/s. Buzzes have up to 116 clicks/s [194]. Example whistles, burst-pulses and clicks are shown in Fig. 20, demonstrating energy of burst-pulse sounds and clicks up to 90 kHz when animals are at close range.

Fig. 20
figure 20

Snubfin dolphin whistles, burst-pulse sounds and clicks recorded by Sarah Marley in Roebuck Bay (WA), July 2014 [196, 197] [fs = 192 kHz, NFFT = 3840, 50% overlap, Hanning window]

Orcinus orca—Killer Whale

Killer whales are a global species, occurring both offshore and inshore, from the equator to the poles [198]. They are the most cosmopolitan and wide-ranging marine mammal. They have been sighted in all Australian states with frequent sightings around TAS, SA, VIC, as well as Ningaloo Reef and the Bremer Canyon in WA [199, 200].

Their sounds have been studied in coastal British Columbia, Canada, and Washington State, USA [201,202,203,204,205,206,207,208,209,210], in California [211], in other areas of the North Pacific, including Russia [212,213,214,215,216], in the North Atlantic off Norway and Iceland [217, 218], in Antarctica [219, 220] and south-western Australia [16]. A geographic comparison of North Atlantic and North Pacific killer whale sound frequencies was recently published [221].

Killer whales produce FM sounds (whistles) of variable duration (0.06–18.3 s) with a fundamental frequency range of 1–30 kHz (Fig. 21; [16, 208,209,210, 219]). Uncommon, low-frequency FM sounds with a fundamental of 50–270 Hz lasting 0.1–2.8 s were reported from Iceland [222]. Killer whales also produce high-frequency FM sounds with a fundamental of 16–75 kHz and source levels of 185–193 dB re 1 \(\upmu \)Pa pp @ 1 m [212, 215, 217]. CW tones are rare (3–5 kHz, 0.37 ± 0.08 s; [208]).

Fig. 21
figure 21

Killer whale sounds recorded off Bremer Bay (WA), February 2014 ([16]; data from [226]) [fs = 96 kHz, NFFT = 2400, 50% overlap, Hanning window]

Their burst-pulse sounds have been studied extensively and cover a frequency range of 500 Hz–25 kHz and last 0.1–1.5 s, at source levels of 131–176 dB re 1 \(\upmu \)Pa rms @ 1 m, or 146–158 dB re 1 \(\upmu \)Pa pp @ 1 m [16, 203,204,205,206,207, 211, 213, 214, 219].

Their echolocation clicks span 10–110 kHz, 100–250 \(\upmu \)s, in trains of 3–7 s, at source levels of 107–224 dB re 1 \(\upmu \)Pa pp @ 1 m [201, 202, 211, 218]. In Australia, slow click trains with an ICI of 0.1 s, as well as speed-up trains ending in a “buzz” sound with an ICI <2.5 ms were recorded. Clicks had a negative Gabor waveform and lasted 100 \(\upmu \)s. Peak energy lay between 12 and 24 kHz [16].

Biphonations are common, specifically in the form of lower-frequency burst-pulse sounds together with higher-frequency, independent FM components [203, 213, 214, 220, 223]. Populations along the Canadian British Columbia coast have been studied extensively using passive acoustics. These and other populations exhibit dialects, i.e. slight differences in vocalisations among pods sharing the same habitat [224, 225].

Peponocephala electra—Melon-Headed Whale

Melon-headed whales are offshore animals inhabiting tropical to temperate waters around the globe. They have been sighted at sea along the Australian coast, including WA and specifically the Browse Basin. Stranding records suggest they occur along the entire northern half of Australia [227]. Acoustic recordings from Australia do not exist. Off Hawaii and in the Caribbean, whistles with fundamentals of 890–24,500 Hz and 0.1–1.4-s duration have been recorded [7, 228, 229]. Burst-pulse sounds appear more broadband (0.5–40 kHz) with 0.1-s duration, 2.47-ms IPI, 47 pulses/train, and a source level of about 165 dB re 1 \(\upmu \)Pa rms @ 1 m [228, 229]. The spectro-temporal structure of vocalisations is graded and falls along a continuum of burst-pulses to whistles [7]. Click spectra cover 6–50 kHz with 0.02–0.45 ms per click, 86–150-ms ICI, up to 1200 clicks/s and source levels of 132–140 dB re 1 \(\upmu \)Pa pp @ 1 m [228,229,230,231].

Pseudorca crassidens—False Killer Whale

False killer whales are found worldwide in tropical to temperate waters, mostly offshore but also inshore, in particular when the continental shelf is narrow. They have stranded in all Australian states [232]. Acoustic recordings of wild animals exist from the northern Indian Ocean [182], Eastern Tropical Pacific [151] and Caribbean [180], in addition to recordings of captive individuals.

False killer whales produce whistles at 5–8 kHz (range of the fundamental) of 0.2–0.8-s duration [151, 180]. Their burst-pulse sounds consist of series of rapid pulses of sinusoidal signals usually in a speed-up pulse train. Burst-pulse sounds cover a band of 0.5–22 kHz and last 0.05–1 s [9]. As the IPI decreases and the damping of sinusoidal signals into pulses decreases, burst-pulse sounds gradually turn into tonals with harmonics and sidebands [9]. False killer whales sometimes emit bimodal clicks, which have two frequency peaks, one at about 40 kHz, the other at about 100 kHz [233]. Peaks at lower frequency (20 kHz) have also been reported [169]. Click duration ranges from 18 to 530 \(\upmu \)s [169, 182, 233,234,235,236,237]. Reported source levels are 195–225 dB re 1 \(\upmu \)Pa pp @ 1 m, 190–215 dB re 1 \(\upmu \)Pa rms @ 1 m and 145–168 dB re 1 \(\upmu \hbox {Pa}^{2}\)s @ 1 m [182, 233]. This species has been shown to imitate anthropogenic tonal sounds; false killer whales increased their whistle rate and copied the spectral features of mid-frequency sonar during controlled-exposure experiments [170]. False killer whale whistles, burst-pulse sounds and clicks were recorded by some of the authors in the Coral Sea, Australia (Fig. 22; [171]).

Fig. 22
figure 22

False killer whale whistles recorded in the Coral Sea, July 2009 [171] [fs = 300 kHz, NFFT = 2000, 50% overlap, Hanning window]

Sousa sahulensis—Australian Humpback Dolphin

This is a coastal species, occurring in shallow water along Australia’s northern coast with resident populations in Moreton Bay (QLD) and Shark Bay (WA) and potentially others [238]. The Australian humpback dolphin is slightly different in morphology from the other humpback dolphins in southern Asia and India, and it was declared a separate species in 2014 [239]. It had previously been classified as Indo-Pacific humpback dolphin (S. chinensis).

Whistles, burst-pulse sounds and clicks have been recorded off QLD and WA (Fig. 23; [195, 240, 241]). Signature whistles, identifying individuals, were also confirmed [242]. Whistles of S. sahulensis from QLD and of S. chinensis in southeast Asia have fundamentals ranging from 1 to 33 kHz with durations of 0.1–9 s; note that durations longer than 3 s were only reported from QLD [195, 240, 242,243,244,245,246,247]. CW tones were reported from S. chinensis in southern China at 500 Hz–22 kHz and 0.03–2-s duration [245]. The differences in whistle characteristics of S. sahulensis and S. chinensis have been detailed [248]. Burst-pulse sounds cover a frequency band of 600 Hz–25 kHz and last 0.1–8 s [195, 240, 247]. Clicks are very broadband with 10 dB bandwidths of \(116\pm 20\) kHz, peaking at 114 kHz [241]. In China, a bimodal spectrum was found peaking at about 100 kHz and 180 kHz [249, 250]. Durations of clicks are 10–100 \(\upmu \)s with ICIs of 0.01–0.25 s [241, 247, 249,250,251]. Source levels of 199 ± 3 dB re 1 \(\upmu \)Pa pp @ 1 m, 189 ± 3 dB re 1 \(\upmu \)Pa rms @ 1 m and 141 ± 3 dB re 1 \(\upmu \hbox {Pa}^{2}\)s @ 1 m have been reported [241, 250, 251].

Fig. 23
figure 23

Australian humpback dolphin whistles [195] [fs = 44 kHz, NFFT = 2200, 50% overlap, Hanning window]

Stenella attenuata—Spotted Dolphin, Pantropical Spotted Dolphin

Two species of spotted dolphin exist: Stenella attenuata in all the major tropical oceans (Atlantic, Pacific, Indian) and S. frontalis only in the Atlantic Ocean. S. attenuata has two subspecies: the offshore pantropical spotted dolphin (S.a. attenuata) and the coastal pantropical spotted dolphin (S.a. graffmani). In Australia, S. attenuata occupies the northern coasts and is not seen along the Great Australian Bight or Tasmania [147]. There are no recordings from around Australia. The only recordings of S. attenuata globally are from Hawaii [252], the Eastern Tropical Pacific [151, 152] and the Gulf of Mexico [253]. These animals whistle at 2–23 kHz (fundamental), lasting 0.3–1.3 s with a source level of 115–163 dB re 1 \(\upmu \)Pa rms @ 1 m.

Stenella coeruleoalba—Striped Dolphin, Euphrosyne Dolphin

This species exists globally in tropical to temperate offshore waters. Strandings along Australia’s coastline are infrequent but have occurred on the northern, western and eastern coasts [254]. Acoustic recordings exist from the Eastern Tropical Pacific [151, 152], the Azores [255], the Canary Islands [255] and the Mediterranean [255], where these animals whistle with fundamentals of 1–31 kHz and durations of 0.04–3 s. CWR saw striped dolphins in the Great Australian Bight, south of WA, in Dover Canyon and recorded whistles with 4–22 kHz fundamentals lasting 0.3–1 s (Fig. 24).

Fig. 24
figure 24

Striped dolphin whistles recorded by CWR in Dover Canyon (WA), February 2016 [fs = 48 kHz, NFFT = 2400, 50% overlap, Hanning window]

Stenella longirostris—Long-Snouted Spinner Dolphin

Spinner dolphins are found globally in tropical to temperate offshore waters, with some occurrence on the continental shelf in shallow water near islands and reefs. There are currently four recognised subspecies: the central American spinner dolphin (S.l. centroamericana), the eastern spinner dolphin (S.l. orientalis), the dwarf spinner dolphin (S.l. roseiventris) and Gray’s spinner dolphin (S.l. longirostris), which is found in Australian waters. Spinner dolphins have been sighted in the Indian Ocean hundreds of km from the nearest land. They have been sighted off northern, western and eastern Australia with numerous bycatch records from the Arafura and Timor Seas. They are also known to occur in the Great Barrier Reef [147].

There are no published sound recordings from Australia. Their sounds have instead been document from Hawaii [230, 256,257,258,259,260,261,262], the Eastern Tropical Pacific [151, 152], the north-western Atlantic [174], Brazil [263,264,265] and the western Indian Ocean off Madagascar [266].

This species emits CW sounds, but they have consistently been lumped with FM sounds during analysis. Whistles range from 2 to 23 kHz (fundamental) and last from 0.1 to 1.8 s [151, 152, 174, 256,257,258,259, 264, 266]. Examples of whistles recorded by CWR at Scott Reef (WA) are shown in Fig. 25. Burst-pulse sounds span a frequency band of 220 Hz-130 kHz, lasting 0.05–2 s. There can be 30 pulses/train with an IPI of 3.85 ms [259, 263]. Clicks cover 20–130 kHz, last 0.2–0.4 ms with an ICI of 121–235 ms [230, 260].

Fig. 25
figure 25

Spinner dolphin whistles recorded by CWR at Scott Reef (WA), August 2012 [fs = 48 kHz, NFFT = 4800, 50% overlap, Hanning window]

Steno bredanensis—Rough-Toothed Dolphin

Rough-toothed dolphins are a deep-water offshore species inhabiting tropical to temperate seas globally. In Australian waters, they have been sighted off Barrow Island (WA), NT, QLD and NSW suggesting tropical–subtropical distribution [267]. They have been acoustically recorded in the Mediterranean [268], Central and Eastern Tropical Pacific and off Brazil [151, 152, 269,270,271]. They emit FM sounds between 1.6 and 28.2 kHz (fundamental) of 0.03–2.2-s duration. Off Brazil, their FM sounds were very weakly frequency-modulated, with about 34% CW tones and 29% FM upsweeps [269]. Clicks ranging from 10 to 50 kHz, lasting 0.14–0.35 ms with 1.6–354-ms ICI were recorded in the Tropical Pacific [271].

Tursiops aduncus—Indian Ocean Bottlenose Dolphin, Spotted Bottlenose Dolphin

This species occurs in shallow, subtropical coastal and estuarine waters along the Indian Ocean, Indo-Pacific region and the western Pacific Ocean. In Australia, these animals are found in estuarine and costal waters along the west, north and east coasts [272]. Along Australia’s south coast, there might be a newly defined species endemic to SA, the Burrunan dolphin (Tursiops australis) [273, 274]; however, this species is currently not recognised by the Society for Marine Mammalogy.

Acoustic recordings exist from Australia’s east and west coasts, as well as Japan, Tanzania and South Africa. Comparative studies of vocal repertoires have been undertaken by many researchers [275,276,277,278,279]. Most research has focussed on whistles, which range in frequency from 1 to 22 kHz (fundamental) and in duration from 0.1 to 5.9 s (Fig. 26). CW sounds are mostly lumped with FM sounds and their parameters are therefore not always listed separately [280]. T. aduncus have been shown to produce signature whistles [281]. Echolocation clicks cover a frequency band from 45 to 109 kHz with a peak frequency of 124 ± 13 kHz, last 18 ± 6 \(\upmu \)s, with an ICI of 63 ± 45 ms and source levels of 205 ± 7 dB re 1 \(\upmu \)Pa pp @ 1 m, 195 ± 4 dB re 1 \(\upmu \)Pa rms @ 1 m and 146 ± 7 dB re 1 \(\upmu \hbox {Pa}^{2}\)s @ 1 m [241, 282]. There is no discussion of burst-pulse sounds in the literature. Acoustic recordings of Burrunan dolphins have not been published yet [283].

Fig. 26
figure 26

Whistles of Indian Ocean bottlenose dolphins recorded near Perth [279] [fs = 96 kHz, NFFT = 4800, 50% overlap, Hanning window]; see also [501]

Tursiops truncatus—Bottlenose Dolphin

There are two subspecies of T. truncatus: the Black Sea bottlenose dolphin (T.t. ponticus) and the common bottlenose dolphin (T.t. truncatus). Common bottlenose dolphins inhabit tropical to temperate waters globally and occur around Australia in coastal as well as offshore waters [147, 272]. Bottlenose dolphins have been acoustically recorded in QLD, Australia [284], New Zealand [285], the Eastern Tropical Pacific [151, 152], Hawaii [230], California [276, 286], the Gulf of Mexico [163, 253, 276, 287], the Caribbean [288], Central America [286, 289], Brazil [270, 290], Uruguay [291], Argentina [276], Namibia [292], the Northwest Atlantic [163, 174], the Northeast Atlantic [282, 293, 294], Scotland and the UK [282, 295, 296], Mediterranean [156, 293, 297,298,299], and Portugal [300]. Many recordings have also been taken of bottlenose dolphins in captivity (e.g. [2, 301,302,303]).

Much of the literature focuses on whistles and geographic differences [151, 152, 156, 163, 174, 253, 270, 276, 286, 289,290,291,292,293, 295,296,297,298, 300, 302]. Biphonic whistles have been reported [299]. CW sounds are typically lumped with whistles and called “flat” whistles. Together, CW and FM sounds range from 300 Hz to 39 kHz in fundamental frequency and from 40 ms to 4 s in duration (Fig. 27), with source levels of 114–163 dB re 1 \(\upmu \)Pa rms @ 1 m. FM sounds at lower frequency have sometimes been described separately and called “low-frequency narrow-band” sounds with a 30 Hz–1 kHz fundamental frequency and 10-ms–8.7-s duration [284, 287, 292, 303].

Bottlenose dolphins develop signature whistles during the first few months of their lives and emit these when meeting conspecifics; they also copy signature whistles of others interpreted as “calling others” [294, 296, 299, 302, 304,305,306,307,308,309,310,311,312,313].

Burst-pulse sounds range from a few kHz to more than 150 kHz, last 0.1–4 s, with 150–1050 pulses/s and source levels of about 195 dB re 1 \(\upmu \)Pa pp @ 1 m [292, 301, 314]. Echolocation clicks range from 20 to 150 kHz, peaking between 50 and 110 kHz. Clicks last 10–80 \(\upmu \)s with 20–180-ms ICI, 11.4 ± 1.6 clicks/train and 1–4 s/train, and source levels of 193–228 dB re 1 \(\upmu \)Pa pp @ 1 m have been reported [2, 230, 282, 291, 314].

The signals emitted by T. truncatus are directional (clicks, burst-pulse sounds and whistles). Directionality increases with frequency. Higher frequencies and amplitudes are emitted forward (i.e. on-axis; [301]). The spectra of extreme off-axis clicks were also published [315]. While click rates are listed in many publications, these can depend on the depth of the animal during a foraging dive [316].

Fig. 27
figure 27

Bottlenose dolphin whistles recorded by CWR off Albany (WA), March 2013 [fs = 200 kHz, NFFT = 10,000, 50% overlap, Hanning window]

Kogia breviceps—Pygmy Sperm Whale

Pygmy sperm whales, a global species, tend to stay in deeper water off the continental shelf ranging from tropical to warm temperate zones [317]. Strandings have been reported in all Australian states, yet sightings at sea are rare. There are no acoustic recordings of pygmy sperm whales in the wild at the time of concurrent visual sightings. Pygmy sperm whales that stranded in Monterey Bay and New Jersey and were then transported to an aquarium produced clicks in captivity between 60 and 200 kHz, with a spectral peak at 125 kHz, 120–600 \(\upmu \)s/click, in trains of 20 clicks/s, including speed-up click trains [318,319,320]. A captive animal that had stranded in Hawaii further produced 0.42-s FM sounds at about 1.4–1.5 kHz [321]. The earliest recordings of pygmy sperm whale sounds were done with a contact microphone in air, while the animal was out of the water [322], likely leading to unrealistically low-frequency content [319]. No recordings from around Australia exist.

Kogia sima—Dwarf Sperm Whale

Dwarf sperm whales are global, offshore animals, likely preferring tropical and warm temperate water [317]. Strandings have been reported in all Australian states. The dwarf sperm whale is an inconspicuous animal that is rarely seen at sea. No records of its acoustic signals were found in the literature. CWR obtained acoustic recordings simultaneously with visual species identification on the Rowley Shoals (WA) (Fig. 28). Whistles had a lot of steps (discontinuities in frequency).

Fig. 28
figure 28

Dwarf sperm whale sounds recorded by CWR on the Rowley Shoals (WA), October 2012: a CW, b FM, c burst-pulse sound and d clicks [fs = 32 kHz, NFFT = 1600 (640 for c), 50% overlap, Hanning window]

Phocoena dioptrica—Spectacled Porpoise

The spectacled porpoise occurs only in the southern hemisphere at subantarctic latitudes. Its distribution is circumpolar [147]. It is rarely seen at sea. Strandings have been reported from the south-eastern coast of South America, and various offshore islands. No records of its vocalisations were found in the literature.

Physeter macrocephalus—Sperm Whale

Sperm whales, a global species, tend to stay along the shelf edge and offshore. They have been sighted off all Australian states. Females and young tend to stay in warmer waters (N of \(45^{\circ }\hbox {S}\)), while males travel to Antarctica and back [323].

Sperm whales have been recorded off the US east coast [324,325,326,327,328], off Nova Scotia [329, 330], around the Bahamas [331,332,333], around Bermuda [334], in the Caribbean [335,336,337,338,339], in the Gulf of Mexico [340,341,342,343], in Alaska [344,345,346,347,348], in the northeast Pacific [349], off Japan [350, 351] around Hawaii [352], at the Galapagos Islands [173, 353, 354], at the Azores [355], off Scotland [356], off Norway [357,358,359,360,361], in the Mediterranean [362,363,364,365,366], in the South Pacific [367], in New Zealand [368,369,370,371,372] and Antarctica [373].

The only sounds that sperm whales produce are clicks, first reported in 1957 [374]. Apart from a geographic variation in repertoire, group-specific dialects have been found in interacting groups with overlapping geographic range [367, 375]. It has been suggested that clicks serve communication, individual identification and echolocation purposes and may be used to debilitate prey [144, 326, 336, 342, 354, 357, 370, 376,377,378,379,380,381].

Sperm whale clicks are now commonly categorised into five classes according to their temporal pattern [330]:

  1. 1.

    “Squeals” with up to 1600 clicks/s [382],

  2. 2.

    “Creaks” with up to 220 clicks/s,

  3. 3.

    “Usual clicks” with an ICI of about 0.5–1 s, and

  4. 4.

    “Slow clicks” (sometimes called “single clicks”) with a long ICI of 3–8 s.

  5. 5.

    Socialising groups emit stereotyped click sequences (patterns) called “codas” over periods lasting up to several hours [328].

Clicks have a bandwidth of 100 Hz–30 kHz, with peak energy around 15 kHz. Click duration is 0.2–3 ms. Each click can consist of a handful of pulses, and the time between pulses correlates with the size of the sperm whale (head) [383]. Source levels of usual clicks are 186 ± 0.9 dB re 1 \(\upmu \)Pa rms @ 1 m, 170 ± 0.7 dB re 1 \(\upmu \hbox {Pa}^{2}\)s @ 1 m and 200–205 dB re 1 \(\upmu \)Pa pp @ 1 m [346], with extremes of up to 196 dB re 1 \(\upmu \hbox {Pa}^{2}\)s @ 1 m and 236 dB re 1 \(\upmu \)Pa rms @ 1 m [358] and are the loudest levels reported for a biological source. Source levels of creaks have been reported as 166–205 dB re 1 \(\upmu \)Pa rms @ 1 m, 180 ± 6 dB re 1 \(\upmu \)Pa pp @ 1 m and 145–161 dB re 1 \(\upmu \hbox {Pa}^{2}\)s @ 1 m [346, 357]. Source levels of slow clicks are 175–190 dB re 1 \(\upmu \)Pa rms @ 1 m and 156–166 re 1 \(\upmu \hbox {Pa}^{2}\)s @ 1 m [357]. The click rate recorded from a group of sperm whales has been reported to increase with the number of animals in the group at 1.22 clicks/s/animal, depending on behaviour [384]. Two neonate sperm whales in rehabilitation produced clicks of low directionality, longer duration and lower frequency than did adults in the wild [385].

It is interesting to note that with such high click repetition rates of up to 1600 clicks/s, depending on sampling frequency and Fourier parameters, squeals will appear as tonal (CW and FM) sounds with sidebands in spectrograms and have been classified as typical odontocete burst-pulse sounds [382]. Squeals were only recorded during social behaviour and hence might serve a social communication role.

Click emission is directional (directivity index 27 dB, [358, 359]), i.e. the angle of measurement from the head affects the click spectrum. Computed source levels are higher on-axis than off-axis. Furthermore, usual clicks appear mono-pulsed on-axis (with 40 dB more energy in the main pulse than following pulses) and multi-pulsed off-axis [358, 386, 387]. With the general uncertainty about the orientation of the whale at the time of recording, differences in published source levels and spectral characteristics of usual clicks can be explained. Example sperm whale clicks recorded by CMST are shown in Fig. 29.

Fig. 29
figure 29

Sperm whale clicks recorded by CMST on Australia’s Northwest Shelf, April 2011 [fs = 192 kHz, NFFT = 768, 50% overlap, Hanning window]

Berardius arnuxii—Arnoux’s Beaked Whale

Arnoux’s beaked whale is circumpolar in the southern hemisphere, occurring in cold waters from the ice edge to its northernmost stranding reported at \(34^{\circ }\hbox {S}\). A small number of strandings have occurred in SA, WA, TAS and the subantarctic. There have been possible sightings near shore in SA and NSW and confirmed sightings in the Australian Antarctic Territory [388].

Dolphin-like clicks and squeaks from individuals with their heads exposed from the water were reported, but recordings were not taken at the time [389]. Click trains (12–19 kHz, 1.2 s/train, 25 clicks/train), burst-pulse sounds (3–11 kHz, 0.53 s), AM sounds (1–8.5 kHz, 0.77 s) and whistles (4.3–5.2 kHz plus higher-frequency harmonics, 0.65 s) were recorded off Antarctica (Fig. 30; [390]).

Fig. 30
figure 30

Arnoux’s beaked whale sounds: a, b whistles and c burst-pulse sounds [390] [fs = 192 kHz, NFFT = 768, 50% overlap, Hanning window]

Hyperoodon planifrons—Southern Bottlenose Whale

The southern bottlenose whale occurs circumpolar in deep water, with recorded sightings and strandings south of \(30^{\circ }\hbox {S}\). Strandings in Australia have occurred in WA, SA, VIC, TAS and NSW [147]. Click trains with peak energy at 18 kHz and click repetition rates of 17–200 clicks/s were attributed to this species based on nearby visuals in the Southern Ocean [391, 392].

Indopacetus pacificus—Longman’s Beaked Whale

Longman’s beaked whales have been reported from the Indo-Pacific region [393, 394]. Sightings and strandings in Australia are rare with only one record in QLD. This species is expected to occur in the warmer waters of Australia. In Hawaii, clicks (10–60 kHz, 90–340 \(\upmu \)s, 0.1–0.9-s ICI), an FM upsweep click (10–60 kHz, 100–500 \(\upmu \)s, 0.3–0.4-s ICI) and burst-pulse sounds (9–25 kHz, 0.5-s duration, 240 pulses/s) were recorded [13, 395].

Mesoplodon bowdoini—Andrew’s Beaked Whale

Distribution of Andrew’s beaked whales is thought to be circumpolar, though that is unconfirmed. Strandings have occurred off Australia in WA, VIC, TAS and NSW, as well as off New Zealand, Macquarie Island, the Falkland Islands, Argentina, Uruguay and Tristan da Cunha [396]. No reports of vocalisations by this species have been found in the literature.

Mesoplodon densirostris—Blainville’s Beaked Whale, Dense-beaked Whale

Blainville’s beaked whales are oceanic animals, occurring globally in tropical to temperate waters. They are very rarely seen at sea, yet strandings have been reported in all Australian states [147]. These animals embark on deep foraging dives of up to 45-min duration to 1250 m depth followed by a sequence of short and shallow resting dives [397], hence spending little time at the surface where they could otherwise be easily observed.

Blainville’s beaked whales have been recorded off the Canary Islands [398,399,400,401], in the Bahamas [402, 403], off Hawaii [13, 404] and off the Northern Mariana Islands [13].

They emit high-frequency FM clicks, where each click of 0.3–1 ms sweeps upwards from about 24 to 51 kHz [13]. During foraging, FM echolocation clicks are produced in long trains, including speed-up click trains (0.2–0.4-s ICI, 15–60 s/train), terminating in buzzes (400 clicks/buzz, 3 s/buzz) indicating a prey-capture attempt. Source levels are 200–220 dB re 1 \(\upmu \)Pa pp @ 1 m [399,400,401, 403, 405, 406]. Blainville’s beaked whales mostly do not echolocate during shallow dives or during deep descents and ascents, clicking vividly only at deep foraging depths. They were reported silent for 80% of time spent in water shallower than 170 m [398].

Blainville’s beaked whales also emit tonal CW and FM sounds (8–18 kHz fundamental, 0.2–2-s duration), as well as AM or rapidly pulsed sounds (2–80 kHz, 0.1–0.6 s) with source levels of 123–149 dB re 1 \(\upmu \)Pa rms @ 1 m, likely for communication [398, 402, 404]. There are no recordings from Australia with positive species identification.

Mesoplodon ginkgodens—Ginkgo-Toothed Beaked Whale, Ginkgo-toothed Whale, Ginkgo Beaked Whale

Ginkgo-toothed whales are rarely seen at sea; their distribution can only be inferred from strandings that have occurred along the tropical to temperate coasts of the Indian and Pacific Oceans. There have been a few stranding events in Australia, in southern NSW and VIC [407]. No records of sounds produced by Ginkgo-toothed beaked whales were found in the literature.

Mesoplodon grayi—Gray’s Beaked Whale

Gray’s beaked whales are distributed around the southern hemisphere in temperature waters with most sightings reported from south of \(30^{\circ }\hbox {S}\). Many stranding records have occurred along the coastlines of WA, SA, VIC, TAS and NSW as well as New Zealand, South Africa, Argentina, Chile and Peru [408]. Waters between New Zealand’s south island and the Chatham Islands are considered to be of high abundance. No reports of vocalisations by this species were found in the literature.

Mesoplodon hectori—Hector’s Beaked Whale

This species has a potentially circumpolar distribution in cool temperate waters of the southern hemisphere between \(35^{\circ }\) and \(55^{\circ }\hbox {S}\). There are no records from the northern hemisphere. Sightings are rare (with a small number recorded in WA, SA and TAS; [147]), and there are no reports of vocalisations.

Mesoplodon layardii—Strap-Toothed Beaked Whale

Strap-toothed beaked whales are distributed in cool temperate waters of the southern hemisphere between \(30^{\circ }\hbox {S}\) and the Antarctic Convergence. They may occur south of \(38^{\circ }\hbox {S}\) year-round moving north of \(38^{\circ }\hbox {S}\) seasonally, and they tend to occur in deep waters off the continental shelf [408]. This species is the most commonly stranded beaked whale species in Australia with stranding events recorded in WA, SA, VIC, TAS, NSW, QLD, Macquarie Island and Heard Island. There is only one mention of vocalisations from this species. This came from a stranded individual that reportedly “squealed” when attempts were made to drag it back to the ocean, but no acoustic characteristics were specified [409].

Mesoplodon mirus—True’s Beaked Whale

In the southern hemisphere, this species has stranded in the southern Indian Ocean, in South Africa, New Zealand, WA, VIC and TAS (though in very low numbers). In the northern hemisphere, records only exist from the North Atlantic. The species has not been recorded in the North Pacific, South Atlantic or northern Indian Ocean and appears to avoid tropical waters [408]. No reports of vocalisations of this species were available.

Tasmacetus shepherdi—Shepherd’s Beaked Whale

Few sightings and strandings of Shepherd’s beaked whale have been reported, with some located off New Zealand, southeast Australia (TAS and VIC), Argentina and islands around Tristan da Cunha. The species is thought to prefer subantarctic and temperate deep oceanic waters [408, 410]. No records of vocalisations were found.

Ziphius cavirostris—Cuvier’s Beaked Whale, Goose-Beaked Whale

Cuvier’s beaked whales occur in all the world’s oceans except for high-latitude polar waters, and they have stranded on all Australian coasts. Globally, they are considered to be one of the most abundant beaked whale species with the most extensive range [147, 411].

They are an oceanic species and have been acoustically recorded in the Mediterranean and Ligurian Seas [399, 412, 413] as well as various locations across the North Pacific [13]. They hold the marine mammal record for deep diving down to 3 km depth over 140 minutes [414]. They echolocate only when foraging at depth (>475 m; [399]). Their clicks are FM upsweeps from 13 to 64 kHz over 0.2–1.6 ms with an ICI of 0.1–0.5 s and source levels of 214 dB re 1 \(\upmu \)Pa pp @ 1 m or 164 dB re 1 \(\upmu \hbox {Pa}^{2}\)s @ 1 m [13, 399, 413]. Click trains last 16–45 s [412]. Buzzes are rapid clicks (up to 250 clicks/s) and are heard only at the end of foraging click trains, marking the prey-capture event [399]. There are no confirmed recordings from Australia.

The Sounds of Sirenians—Sea Cows

Dugong dugon—Dugong

Sirenians are found in shallow, inshore waters of tropical and subtropical climate worldwide. The dugong is the only species of Sirenia found in Australian waters. Distributed across the northern, tropical, inshore waters (up to 40 m depth) from Shark Bay (WA), along Australia’s northern coast, to Moreton Bay (QLD), dugongs are also occasional visitors to NSW coastal and estuarine waters [415, 416].

Dugong sounds (Fig. 31) have been recorded on both east and west coasts of Australia. They make FM sounds with 2–4.7 kHz fundamentals, lasting 0.1–2.2 s with source levels of 131–161 dB re 1 \(\upmu \)Pa rms @ 1 m, 145–180 dB re 1 \(\upmu \)Pa pp @ 1 m and 110–146 dB re 1 \(\upmu \hbox {Pa}^{2}\)s @ 1 m [417,418,419]. Their pulsed or AM sounds cover 300 Hz–22 kHz, last 0.1–1.2 s and have source levels of 129–163 dB re 1 \(\upmu \)Pa rms @ 1 m, 148–172 dB re 1 \(\upmu \)Pa pp @ 1 m and 110–135 dB re 1 \(\upmu \hbox {Pa}^{2}\)s @ 1 m [417, 418]. Broadband feeding (mastication) sounds ranging from 1 to over 22 kHz were recorded in Thailand [420].

Fig. 31
figure 31

Dugong sounds: a) whistle and b) burst-pulse sound [418] [fs = 48 kHz, NFFT = 512, 50% overlap, Hanning window]

The Sounds of Pinnipeds—Seals and Sea Lions

Arctocephalus forsteri—New Zealand Fur Seal

New Zealand fur seals, also known as long-nosed fur seals or South Australian fur seals, are found in subantarctic waters. They occur around both the south and north islands of New Zealand (up to 72 m depth [421]) and the coastal waters of SA and WA, with small populations in VIC, NSW and TAS [422, 423].

New Zealand fur seals have not been recorded underwater. In-air vocalisations are mostly of pulsed nature, and even FM sounds are typically amplitude-modulated (Fig. 32). Sounds have been classified as pup-attraction calls produced by females (burst-pulse sounds, 100–8000 Hz, 0.8–2.2 s), female-attraction calls produced by pups (burst-pulse sounds, 140–4600 Hz, 0.5–2.0 s), as well as guttural threats, growls preceding guttural threats, barks and a full-threat call, all of which were produced by males (pulsed sounds, 85–7800 Hz, 0.1–2 s) [424,425,426].

New Zealand fur seals breed sympatrically and hybridise with other fur seals (A. gazella and A. tropicalis) at subantarctic islands such as Macquarie Island. These hybrid fur seals produce vocalisations that have mixed traits and are intermediate to their parental species’ calls [425,426,427].

Fig. 32
figure 32

New Zealand fur seal sounds in air: a, b FM, c burst-pulse sounds recorded by UNSW on the Otago Peninsula [fs = 44 kHz, NFFT = 4410, 50% overlap, Hanning window]

Arctocephalus gazella—Antarctic Fur Seal

Antarctic fur seals are found in subantarctic waters. The majority of the Antarctic fur seal population is located on South Georgia Island. Other smaller populations are found on islands off South America and in the southern Indian Ocean. They are seen as vagrants in the Antarctic and along the coastlines of Australia, Argentina, Brazil, Chile and South Africa [428].

Antarctic fur seal calls have only been recorded in air. Sounds are of FM or pulsed nature (Fig. 33). Barks (100–7700 Hz) last 0.1–1.1 s/pulse and 5.6 s/bout. Guttural threat sounds cover a similar frequency band (85–6000 Hz) and last 0.5–4 s. AM sounds such as growls, pup-attraction calls and female-attraction calls cover 100–9000 Hz and last 0.2–2.1 s [424,425,426, 429]. The latter two call types are individually distinctive [426, 430].

The Antarctic fur seal breeds sympatrically and hybridises with other fur seals (A. forsteri and A. tropicalis) at subantarctic islands such as Macquarie Island. These hybrids produce vocalisations that are intermediate to their parental species’ vocalisations [425, 427].

Fig. 33
figure 33

Antarctic fur seal sounds in air recorded by UNSW in South Georgia in 2008: a, b FM, c burst-pulses [fs = 44 kHz, NFFT = 2205 (a, b), NFFT = 882 (c), 50% overlap, Hanning window]

Arctocephalus pusillus doriferus—Australian Fur Seal

There are two species of A. pusillus: the Cape fur seal (A.p. pusillus) and the Australian fur seal (A.p. doriferus). Australian fur seals are found in coastal waters of eastern SA, VIC, TAS and southern NSW. Breeding colonies exist at islands of Bass Strait, between TAS and VIC [431,432,433].

There is no record of Australian fur seal calls under water. In air, these animals mostly produce mostly pulsed vocalisations (Fig. 34). Male and female barks have most energy between 100 and 2400 Hz with each pulse of the bark lasting 0.1–0.2 s. Male and female guttural threat pulses are longer (0.14–0.3 s) but span a similar frequency band (40–3500 Hz). Burst-pulse sounds, including AM growls, pup-attraction and female-attraction calls, cover 100–5500 Hz in frequency and 0.2–1.8 s in duration [424, 434, 435]. Australian fur seal pups produce individually distinct calls to attract their mothers (female-attraction call) throughout the maternal dependence period [436].

Fig. 34
figure 34

Australian fur seal barks in air recorded by Joy Tripovich in VIC [fs = 22 kHz, NFFT = 1100, 50% overlap, Hanning window]

Arctocephalus tropicalis—Subantarctic Fur Seal

Subantarctic fur seals are widely distributed throughout the southern hemisphere and found in the southern waters of the Indian, Pacific and Atlantic Oceans. They breed on subantarctic islands such as Macquarie Island. Vagrants have been recorded on the coasts of Antarctica, Australia (from WA to QLD), New Zealand, southern South America and southern Africa [437, 438].

There are no records of subantarctic fur seals vocalising under water. On land, they produce pulsed calls: barks (120–3020 Hz, 0.07–0.2 s/pulse, 5–8 s/bout), guttural challenge calls (250–1930 Hz, 0.13–1.56 s) and burst-pulse sounds including pup-attraction calls, female-attraction calls and full-threat calls (100–8000 Hz, 0.2–3.5 s) [425, 426, 429, 439,440,441]. Individual identity has been demonstrated in mothers and pups [439, 440]. On Macquarie Island, subantarctic fur seals breed sympatrically and hybridise with other fur seals (A. forsteri and A. gazella). The vocalisations of hybrids have mixed traits from both their parental species [425, 427].

Neophoca cinerea—Australian Sea Lion

Australian sea lions are found in shallow waters along the southern and south-western coasts of Australia. They are endemic to Australia, distributed from the Houtman Abrolhos Islands (WA) to The Pages Island (SA; [442]), in waters up to 105 m depth [443]. The species has been known to venture to the southern coast of eastern Australia. Sixty-six Australian sea lion colonies have been recorded (38 in SA, 28 in WA). The four largest colonies are located on The Pages Island, at Seal Bay on Kangaroo Island and at Dangerous Reef [444].

There are no recordings of Australian sea lion sounds underwater. Their in-air sounds were first described in the early 1970s [445, 446]. In air, on their haul-out locations, males produce loud barks consisting of pulses with peak energy between 350 and 2600 Hz, lasting 0.02–0.1 s [441, 447, 448] (Fig. 35). Bark series that show accelerated rhythmicity and higher formants elicited stronger responses from other males [449]. A bleating burst-pulse sound (400–4000 Hz, 0.5 s) was also reported [441]. Mothers and pups produce FM sounds peaking between 400 and 3400 Hz, lasting 0.55–1.4 s [450]. Vocalisations of females and males are individually distinctive [441, 450,451,452,453]. Acoustic mother-offspring recognition is well developed in pinnipeds, with Australian sea lions being one example. Geographic variation along the southern WA and SA coasts was investigated by Ahonen et al. [448].

Fig. 35
figure 35

Australian sea lion barks in air recorded by Joy Tripovich in SA [fs = 22 kHz, NFFT = 1100, 50% overlap, Hanning window]

Fig. 36
figure 36

Leopard seal underwater burst-pulse sounds recorded by UNSW in the Davis Sea (Eastern Antarctica) in October 1996 [fs = 11 kHz, NFFT = 2000, 50% overlap, Hanning window]

Hydrurga leptonyx—Leopard Seal

Leopard seals are found in the Antarctic pack ice and on subantarctic islands. They are difficult to survey in the Antarctic, because they are widely dispersed at low densities throughout the Antarctic pack ice [454,455,456,457,458]. Vagrants are sighted in the austral winter on southern continents as far north as the coasts of QLD and WA [459].

Leopard seals produce vocalisations underwater particularly during their breeding season in the austral summer [455, 460]. In-air vocalisations have not been reported. Most of the underwater vocalisations consist of rapid pulse trains, commonly referred to as trills that are typically given in two parts (double trills; Fig. 36). Trills cover a frequency range of 150 Hz to 6 kHz and last 2–8 s with source levels of 156–175 dB re 1 \(\upmu \)Pa rms @ 1 m [459, 461,462,463]. These seals also produce lower-frequency burst-pulse sounds including AM growls at 50 Hz–6 kHz lasting 0.5–4.3 s [459, 462]. Ultrasonic (64–160 kHz) pulse trains consisting of rapid (9–40 ms) downsweeps were recorded in captivity and are the only account of such high-frequency sounds in pinnipeds [464]. Leopard seal vocalisations have individual [465] and age-related [466] differences, as well as geographic variation [462, 467, 468].

Leptonychotes weddellii—Weddell Seal

The Weddell seal is the world’s most southerly breeding mammal. It breeds on the fast ice around Antarctica where it occurs right up to the shoreline of the Antarctic continent. It also occurs offshore in the pack ice zone north to the seasonally shifting limits of the Antarctic Convergence. Vagrants have been recorded in South America, New Zealand and southern Australia [469,470,471].

Weddell seals produce a wide array of both in-air and underwater vocalisations. The majority of calling occurs during the breeding period, with vocalisation numbers dropping afterwards [472, 473]. Weddell seal calls were first described crudely phonetically by Weddell in 1825, then by Lindsey [474] and by Schevill and Watkins [475]. Different studies have used different descriptors and terminology making comparisons difficult. Weddell seals underwater make CW sounds at frequencies ranging from 100 to 1600 Hz, lasting 1–90 s [15, 476,477,478]. They make a variety of FM sounds, commonly downsweeps, between 100 Hz and 14 kHz, lasting 0.2–67 s, with source levels of 153–193 dB re 1 \(\upmu \)Pa rms @ 1 m (Fig. 37; [15, 476,477,478,479]). Pulse trains and AM sounds cover a band from 400 Hz to 15 kHz and last 1–54 s with source levels of 171–188 dB re 1 \(\upmu \)Pa rms @ 1 m [15, 476,477,478,479]. Pulsed sounds have distinct rhythmic patterns and can be emitted in constant, increasing as well as decreasing IPI [480]. In-air, FM calls between 100 and 500 Hz, lasting 0.25–2 s, commonly of convex spectrographic shape, have been recorded of mothers and pups [481, 482]. Individual [481,482,483] and geographic [477, 479, 484] variability of aerial and underwater vocalisations has been noted.

Fig. 37
figure 37

Weddell seal underwater FM sounds recorded by UNSW in the Davis Sea (Eastern Antarctica) in October 1996 [fs = 11 kHz, NFFT = 200 (a, b), NFFT = 500 (c), 50% overlap, Hanning window]

Lobodon carcinophaga—Crabeater Seal

Crabeater seals are found year-round in Antarctic circumpolar pack ice up to 528 m depth [485], and migrants are found as far north as Australia, South Africa and New Zealand.

Crabeater seal underwater vocalisations are mostly pulsed and include moans and groans (Fig. 38). There are lower-frequency calls with peak energy around a few hundred Hertz and higher-frequency calls with peak energy above 1 kHz. Together, these sounds cover a frequency range from 10 Hz to 6 kHz and last about 2 s with a range of 0.1–6 s [461, 486, 487]. Whistles have also been reported at 900–1200 Hz lasting about 1 s [486].

Fig. 38
figure 38

Crabeater seal underwater burst-pulse sound recorded by UNSW in the Davis Sea (Eastern Antarctica) in October 1996 [fs = 11 kHz, NFFT = 550, 50% overlap, Hanning window]

Mirounga leonina—Southern Elephant Seal

Elephant seals are found in subpolar waters worldwide. There are two species of elephant seal, but the southern elephant seal (M. leonina) is the only species in the southern hemisphere. Southern elephant seals are distributed across subantarctic waters north of the pack ice up to 1500 m depth [488] from Macquarie Island in Australia to the tip of the Antarctic continent [489].

To date, underwater vocalisations have not been recorded. Their aerial vocalisations were recorded on Sea Lion Island, one of the Falkland Islands [490,491,492]. Sounds consisted of broadband (10 Hz to greater than 4 kHz) pulses emitted in bouts with variable IPI (Fig. 39). Most energy lay between 200 and 800 Hz, and vocalisations lasted between 1 and 50 s with source levels of 103–120 dB re 20 \(\upmu \)Pa rms @ 1 m.

Fig. 39
figure 39

Southern elephant seal aerial pulsed call recorded by UNSW in the Western Antarctic Peninsula in 2008 [fs = 44 kHz, NFFT = 900, 50% overlap, Hanning window]

Ommatophoca rossii—Ross Seal

Ross seals are found in circumpolar pack ice about Antarctica. They are distributed in the Antarctic pack ice during the moult and breeding season and migrate north of the pack ice after moulting where they establish a pelagic lifestyle through the year [493]. Vagrant animals have been seen along the coastlines of the southern hemisphere including South Australia.

Ross seals vocalise both in air and under water with first recordings described by Ray [494]. Their vocalisations include pulses that are very brief downsweeps from 1000 to 100 Hz, lasting 0.05–0.1 s/pulse with 5–12 pulses in sequence—both in air and under water. They also produce U-shaped FM calls in both media, lasting 1–1.5 s. The frequency range in air is 100–800 Hz and higher (1000–4500 Hz) under water (Fig. 40). Underwater tones are amplitude-modulated creating sidebands at 650 Hz spacing [14].

Fig. 40
figure 40

Ross seal underwater siren recorded by UNSW in the Davis Sea (Eastern Antarctica) in October 1996 [fs = 22 kHz, NFFT = 440, 50% overlap, Hanning window]

Discussion

We intend this article to be used by researchers tasked with acoustically identifying species recorded in Australian waters. The first step should be to determine which species could have been present at the time of recording. A good starting place is the Department of the Environment and Energy website, which gives distribution and migration patterns for most species and was one of the main resources used for this study. Figure 1 summarises the geographic regions where the various species might be encountered. In addition, Table 1 indicates which species mostly stay offshore versus inshore. Figures 2, 3 and 4 can then be used to compare the bandwidth and duration of the signals recorded to the summary data for each species. More details on the sounds emitted by the resulting candidate species can be found in Tables 2, 3, 4 and Sects. 5 through 8, potentially leading to positive species identification.

There is an increasing amount of literature on the sounds of marine mammals in the world’s oceans. As hardware and software tools have advanced, the certainty of measured quantities has also improved. For example, in the case of echolocation clicks, sampling frequencies are now high enough to instil confidence that the full spectrum was captured in many recordings. In addition, the use of hydrophone arrays allows the determination of range and sometimes angular aspect to the animal so that researchers know whether spectra were recorded on- versus off-axis. What is missing though is a unified approach to analysing and reporting data, or the development of standards, so that results from different studies can be combined or compared [495].

Table 2 Mysticetes
Table 3 Odontocetes and dugong

Perhaps related to this is a lack of agreement on how to classify animal sounds. The biggest hurdle in completing this manuscript was the need to summarise information from hundreds of publications and the corresponding dilemma of whether to group two sounds as “the same” or to separate them. Onomatopoeic names have different connotations and imply different features for different people. Spectrographic representations are affected by various recording and analysis settings. The effect of sound propagation on altering received call structure from that transmitted is almost always ignored but hardly ever negligible. Rather than classifying sounds by a variety of characteristics that are recorded by specific gear at some range and angle from the animal, it might be less arbitrary to group them according to how they are produced. This requires more research into actual sound production, an understanding of resonances and filters and sound propagation through animal tissues. Until such time, it would at least be helpful to have consensus about minimum standards for the recording, analysis and reporting of animal sounds.

Symbols and Abbreviations

Sound types

CW

Constant-wave, flat tonal sound

FM

Frequency-modulated

AM

Amplitude-modulated

Pulses

Broadband sounds; typically brief; rapid series of pulses are often classified as AM sounds

Sound measurements

fp

Frequency of fundamental contour of whistles; frequency of maximum spectral power in clicks

fc

Centroid frequency that divides the power in a spectrum into two equal parts

bw3

\(-\)3 dB bandwidth

bw10

\(-\)10 dB bandwidth

bwrms

rms bandwidth, i.e. the standard deviation of the spectrum about the centroid frequency

SL

Source level

pp

Peak-to-peak

rms

Root-mean-square

IPI

Inter-pulse interval

ICI

Inter-click interval

Recording and analysis parameters

fs

Sampling frequency

NFFT

Number of Fourier components

Specific sounds

PAC

Pup-attraction call made by female pinnipeds

FAC

Female-attraction call made by pinniped pups

Geographic references

N

North

S

South

E

East

W

West

WA

Western Australia

SA

South Australia

VIC

Victoria

TAS

Tasmania

NSW

New South Wales

QLD

Queensland

NT

Northern Territory

Isl.

Island

Institutions

CMST

Centre for Marine Science & Technology, Curtin University

CWR

Centre for Whale Research

UNSW

University of New South Wales

Summary Tables of Marine Mammal Sounds

Table 4 Pinnipeds