Collecting, Documenting, and Archiving Bioacoustical Data and Metadata

Abstract

The best practices for conducting bioacoustical research are described in this chapter. Ethical considerations, including the need for research integrity, animal ethics (e.g., in the USA, the Institutional Animal Care and Use Committee, IACUC) approvals, and data management and archiving are discussed. Suggestions for conducting a bioacoustical study in the laboratory, field, and captivity are described, including designing a data sheet. Recommendations for carrying out playback studies, a special type of bioacoustical study, are also provided. Finally, the importance of historical data, properly logging the details of a recording, documenting equipment specifications, and correctly storing analog and digital acoustic data are summarized.

Jeanette A. Thomas (deceased) contributed to this chapter while at the Department of Biological Sciences, Western Illinois University-Quad Cities, Moline, IL, USA

3.1 Introduction

Over the last 100 years, bioacoustical research has led to many important discoveries about the role of sounds in animal behavior. Over time, best practices have evolved in bioacoustical research; often through trial and error. In this chapter, these best practices, based on the literature and the co-authors’ experiences and opinions, are summarized. We recommend methods to properly collect and conserve data, use appropriate equipment, save time, and perhaps even make a study more affordable. It is advised, of course, that researchers conduct a current literature review before beginning their work, as developments in technique and technology are moving at a fast pace.

Although methods in bioacoustical studies are typically non-invasive, research should be conducted in an ethical way and any necessary permits obtained. Bioacoustical research should be able to be repeated reliably, where another investigator should be able to understand the circumstances of the recordings, replicate and apply the results, and be reassured the methods were appropriate for the goals of the study. Detailed logs of recordings are important and should include names of researchers; date and time; location; ambient conditions; equipment specifications; species, age, and sex; and behavioral context of the animal during the recording. Details of data collection and signal analysis should accompany any results, such as frequency range, sampling rate, bit-resolution, analysis bandwidth and interval, amplitude range, and any filtering or weightings used.

Here, we also discuss special considerations, or adaptation of methods, for acoustic studies in aquatic versus terrestrial field environments, as well as considerations for studies on captive animals. The “playback” technique, where a sound is played back to an animal and response noted, is a common method used in bioacoustical studies and this chapter provides recommendations for designing a robust playback study. Finally, methods for data archival, and current repositories for bioacoustical data, are provided as a resource for those interested in examining existing data or preserving their own recordings.

3.2 Ethical Research

As with all scientific endeavors, bioacousticians work to answer questions and address hypotheses by observing or manipulating the natural world. There is an ethical obligation to document procedures and methods, so that reported results are understandable and reproducible by other researchers. A reliable way for understanding data, and how they were collected, is by documenting metadata associated with a recording. Metadata are the description of basic information collected at the time of the recording, such as the recordist; date and time; specific location (GPS coordinates); equipment and settings; water depth or altitude; water or air medium; water or air temperature/humidity; weather conditions; and species, sex, age, and behavior of the animals. Knowing the who, what, when, and where, of acoustic recordings makes acoustic data more useful and allows a review of methods by other researchers to validate or supplement data.

Although bioacoustical studies are usually non-invasive, investigators need to consider and minimize any potential effects of their work on animals (e.g., avoid playbacks of extremely loud or injurious sounds that could disturb animals in critical breeding and feeding areas). In many cases, animal ethics permits and/or research permits are needed from the country, state, county, or any other political entity in which the study will be conducted. If the species is endangered, additional permits may be required. Most research institutions receiving funding from the USA government require investigators to submit an animal research protocol to an Institutional Animal Care and Use Committee (IACUC) for approval before conducting research involving any animals. Ethical conduct of research goes beyond satisfying the requirements of the IACUC and includes responsible data collection and management, appropriate statistical analyses, thorough presentation and archival of data, and a study that is reproducible. Additionally, research should be reported, peer-reviewed, and published ethically. This falls under research ethics principles and studies that are conducted with scientific integrity (Fig. 3.1). Most researchers consider their work with animals to be harmless and therefore ethical. However, the process of thinking through how animals could be affected, and proposing research methods during the preparation of an IACUC protocol can be very instructive. In some cases, preparing a protocol for review can save a project from mistakes (such as low statistical power, inadequate or illegal animal housing or handling methods, unnecessary duplication, unnecessary expense, or unrecognized alternative hypotheses). In fact, developing a research protocol can serve to make the research more robust.

Fig. 3.1

A collage of common reference materials and journals that are used to advise on the responsible conduct of research with animals. Considerations of the integrity of the scientific process and the ethics of how a study is conducted undoubtedly produce better science

Gannon (2014) provided two examples that illustrate a potentially unethical study and posed the question of whether a research permit was needed. In 1991, a rare migrant yellow-green vireo (Vireo flavoviridis) was spotted at protected parklands in Rattlesnake Springs, New Mexico, USA. The sighting was announced on the rare-bird hotline and a number of people went to the area to view the bird and to add it to their “life list.” During this time, a PhD student was collecting goldfinches (Spinus tristis). Knowing that genetic material and voucher specimens are important to taxonomic and conservation research, he decided to collect the rare bird for a museum research collection. To entice the bird to an unprotected area for easy and legal collection, he recorded calls of the vireo and then played them back where he could legally collect the bird. The birding community became incredulous and angry. Was it ethical to record and use playbacks of this species’ calls to lure the bird to an unprotected area for collection (see Gluck 1998)?

More recently, as characterized in Fig. 3.2, a smartphone birding application was used to lure a male common yellowthroat (Geothlypis trichas) into view. White (2013) described that broadcasting calls, using a smartphone application, generally elicits a quick response from a normally concealed bird. Possibly thinking the sounds were from another male of his species and threatening his territory, the male yellowthroat swooped down right in front of a birding tour and was photographed. Is it ethical to lure a bird to impress a tour group or does the playback burden the bird with unnecessary stress, perhaps reducing his fitness? Should acoustic luring be prohibited for all bird species or for only endangered animals? Conversely, should these techniques be encouraged in order to raise awareness of wild things to a public who are increasingly alienated from nature?

Fig. 3.2

Caricature of an ornithologist luring a bird by playback of bird calls (with permission of the illustrator Rohan Chakravarty)

Ethical treatment of animals serves to make a research project rigorous and results stronger. Given the personnel time to design experiments, obtain permits, and conduct bioacoustical research, and given the expense and potential disturbance to animals, is the project worth doing? If it is worth doing, it is worth doing well.

3.3 Good Practices in Bioacoustical Studies

Once research questions have been developed and equipment has been selected (see Chap. 2 on equipment choices), recording can begin! Animals can be recorded in a controlled laboratory or in the field. Bioacousticians often need to be innovative when collecting acoustic data in field situations because additional equipment, AC-power, and access to repairs are not always available. Below is a summary of some recommendations for beginning bioacousticians. All suggestions are relevant to both terrestrial and aquatic environments unless identified otherwise.

3.3.1 Recording Sounds

It is best to work toward making the cleanest recording possible for accurate acoustic analysis. Be sure that you have a solid understanding of the gain and level controls on your recorder. The gain and level meter work in concert and the person making the recording needs to be comfortable with these settings before serious acoustic research begins. Ideally the entire recording chain should be calibrated. Calibration generally refers to correlating the readings of an instrument with those of a standard for the purpose of checking the instrument’s accuracy. When recording sound, a calibration signal (a pure tone) of known frequency and amplitude should be placed at the beginning of all recordings. Some recorders have a built-in calibration tone. The tone also can be used to mark an important section of the recording. Having a calibration tone on a recording allows measurement of absolute amplitude, rather than just relative amplitude. This step is necessary if the researcher wants to report source-levels of animal or environmental sounds. Calibrating recording equipment is referred to in Chap. 2 of this volume. Ideally the distance to the sound source (vocalizing animals in our case) should be known. A common “trick” is dropping a colored poker chip at the point where the recording is started and then as moving toward the sound source, dropping additional chips until the point where the animal who had been calling has presumable run off. The distance can then easily be measured between chips. Absolute distance and calibration of the recording system is difficult in field studies.

If more than one channel is available on a recorder, use one channel to narrate metadata and the animals’ behaviors with the second channel dedicated to recording animal sounds. This allows all details and conditions of the situation to be documented in real-time and synchronized with the animals’ sounds and behaviors. After each session, the researcher should listen to the recordings to make sure signals were recorded and the equipment was working properly. We recommend making a copy of each recording and storing the backup and the original in different places.

When possible, use battery-power or direct-current (DC), rather than alternating-current (AC) wall- or shore-power. Using batteries eliminates background electronic noise and provides portability of the equipment. AC-power can create a 50-Hz (European power) or 60-Hz (North American power) hum or background noise on a recording. This frequency-specific noise is easy to recognize and filter-out, preferably during the recording. However, if the animal produces low-frequency signals (e.g., 20-Hz calls from some baleen whales, low-frequency knocks and grunts from fish, rumbles by elephants) the recordings should not be filtered. Note that in extremely cold locations, battery-life will be shorter and any type of mechanical components such as belts, gears, toggles, reels, or digital equipment can cease to operate correctly. We recommend that backup batteries be available or on-charge for quick battery exchange.

3.3.2 Environmental Conditions

Equipment should be selected based on environmental conditions at the field site including ambient temperature and humidity, prevalence of wind and waves, amount and type of precipitation, and frequency and amplitude of the target species (Fig. 3.3; see Chap. 2 on equipment choices). Before commencing field work, check the weather forecast. Recording animal sounds during precipitation, high wind, or a high sea-state often is futile because incoming signals will be masked. In addition, animals sometimes do not call during these conditions. In terrestrial environments, noise from wind, weather, moving vegetation, or other animal sounds can mask recordings of the target species (see Chap. 5 on the source-path-receiver model for airborne sound). In aquatic habitats, wind, sea-state, breaking waves, precipitation, and other animal sounds can create a noisy background. In both terrestrial and aquatic environments, anthropogenic noise (from vehicles and vessels, industrial operations, military activities, etc.) essentially is omnipresent (see Chap. 7 on soundscapes). If using a remote recording system, protect the unit from the weather and secure it as best possible. Be aware that even in remote locations, theft of field equipment occurs.

Fig. 3.3

Conditions in the field often contrast sharply from those in a controlled laboratory environment. Working to exclude bats (Townsend’s big-eared bat, Corynorhinus townsendii) from gold mining operations in Nevada, USA (top left). Recording assures animals are excluded prior to destroying the tunnel system for mineral extraction. Mitigation sites are identified (top right) which are gated and protected for bats to inhabit safely. Occasional sampling is completed by live-capture (bottom left) and acoustic monitoring (bottom right). All photos by authors except bottom left (MNH field biologists collect bat specimens, by Florante A. Cruz; https://www.wikiwand.com/en/UPLB_Museum_of_Natural_History; licensed under CC BY-SA 4.0; https://creativecommons.org/licenses/by-sa/4.0/

Documenting the ambient temperature and humidity is especially important when studying ectothermic terrestrial animals, such as reptiles, frogs, toads, insects, or other invertebrates. At low ambient temperatures, ectothermic animals are less active and sounds are lower in frequency than during higher ambient temperatures. For example, studies by Kissner et al. (1997) demonstrated that sounds from ectothermic animals, such as rattlesnakes (Crotalus viridis), change with ambient temperature and humidity.

3.3.3 Animal Considerations

The transducer should be positioned so target-animal sounds are recorded but the animal does not damage the equipment. An aggressive or curious animal can quickly demolish a recording system (Fig. 3.4). Equipment used in playback studies can be particularly susceptible to an animal attack. The goal of recording is to document sounds from natural circumstances and not from a charging or frightened animal. Captive animals often are curious about a hydrophone or a microphone in their enclosure and can need time to habituate to equipment before undisturbed sounds are produced. Placing the transducer in a protected area or in a protective mesh cage may be necessary.

Fig. 3.4

Photographs of researchers in Antarctica recording a killer whale (Orcinus orca; left) and Weddell seal (Leptonychotes weddellii; right). Equipment is both protected from being molested by the animal but also not prominent so as to not draw the subject’s attention. Note the researcher on the right maintains a distance from the seal so as not to disturb it

Researchers should not disturb animals while recording (Fig. 3.5). If possible, the recordist should hide in a blind spot or use an automated recording system with no observer present. Note that sometimes narrating observations of the animal’s behavior during the recording is useful which means that the researcher should decide between using a remote setup and a setup where they are nearby. To concurrently monitor animal behavior, a video camera on a tripod can be used, with minimal disturbance to the animal. However, the researcher should be aware that the audio track of a video camera has a limited frequency response and an auto-adaptive level control, meaning these sound recordings should not be relied upon for acoustical analysis. Closed Circuit Television (CCTV), synchronized with omnidirectional microphones on an ultrasonic detector, and coordinated using a mobile phone and speaking clock, has been used to document new vocalizations and activities patterns for barbastelle bats (Barbastella barbastellus; Young et al. 2018). With a little ingenuity, a researcher can create a robust recording system.

Fig. 3.5

What could go wrong? In the field, equipment failure is certain. Over-planning, backups, duplicate systems, checklists, and more will help avoid data collection failures

To save time and expense, it is important to know whether a species has a preferred time of day or season for producing sounds. Many species are most vocal during the breeding season. Some birds and amphibians are most soniferous at dawn and dusk whereas many chorusing insects primarily produce sounds at dusk. For example, Thomas and DeMaster (1982) showed that Antarctic crabeater seals (Lobodon carcinophaga) preferred to call under water between 2100 h and 0500 h and were hauled-out on the ice at other times. If the number of vocalizations was used as a population count, a census of crabeater seals at 1200 h would have yielded a much lower population estimate than a census at 2400 h. Bats, obviously, are active at night. However, there is usually a notable peak of activity approximately 30 minutes after dusk (Kunz and Parsons 2009). Some species (many in the genus Myotis and Tadarida) are more likely to be recorded during the first four hours of night, while others emerge past midnight (Euderma, Artibeus). Some bats have multimodal activity patterns (Sherwin et al. 2000) and many sciurids (e.g., Marmota and Neotamias) actively vocalize in the morning and then again in late afternoon (Gannon 1999). Some species (e.g., prairie dogs, Cynomys and pikas, Ochotona) are seasonally soniferous all day (Slobodchikoff et al. 1998; Smith et al. 2016).

It is important to know the effects of both time of day and month to interpret the behavioral context of a recording. For example, breeding data from the North American male rufous-sided towhee (Pipilo erythrophthalmus) showed that males reached breeding condition around mid-April. Testes were in regression by 20 July and had become inactive by mid- to late-September (Davis 1958). So, if a researcher desires to record sounds of this species associated with breeding, the study should be conducted from mid-April to mid-July. In addition, this species shifts their song to an earlier start time in relation to civil twilight. As day length increases between the spring equinox and the summer solstice, civil twilight occurs earlier in relation to sunrise, causing the dawn calling period to lengthen.

3.3.4 Documentation and Data Sheets

Documentation is very important. A logbook should accompany each recording to provide metadata on the recordist; the recording system and equipment settings (e.g., any filter or gain settings); the location, date and time; environmental conditions; types of sounds recorded; the animals’ behavior (e.g., breeding, feeding, or socializing); a specific animal number (if marked); and any other circumstances which could be valuable for analysis.

Many devices may record some of the metadata automatically. For instance, the Echo Meter Touch 2 PRO Ultrasonic Module using Kaleidoscope Pro software¹ (Wildlife Acoustics, Maynard, MA, USA) records calls to an iPhone or other device and collects metadata about each recording. Metadata can then be displayed with Kaleidoscope software or exported to a spreadsheet. Recording directly to a computer allows time-stamped (and often GPS-stamped) files.

If a datasheet (spreadsheet) is used, put metadata headers as the first column and fill the rows with your observations (Table 3.1). Each sound or bout of sounds should be assigned a unique number for easy reference later, and a variety of variables can then be noted for each sound (Table 3.2). Spreadsheets can be imported directly into a variety of statistical and graphing software products for analyses (see Chap. 9 on analytical approaches). Note that datasheets for playback studies usually include additional variables on animal behavior (Table 3.3).

Table 3.1 Sample logbook showing important metadata to be noted. Examples from author (JAT) notes for Weddell seal (Leptonychotes weddellii) and sea otter (Enhydra lutris)

Table 3.2 Sample data sheet for acoustic measurements on airborne vocalizations of California sea lions (Zalophus californianus); na not applicable. Courtesy of Schwalm (2012)

Table 3.3 Sample data sheet for playback study with white-cheeked gibbons (Nomascus leucogenys); na not applicable. Courtesy of Yegge (2012)

3.3.5 Trouble-shooting Equipment Problems

Often field work is conducted in remote locations, sometimes without easy access to the Internet, electricity, or equipment repairs. Consider all possible equipment problems and always have backups—of everything. A good motto for field work is to “bring one to use and one to lose” (Fig. 3.5). Studies usually are costly and time-consuming—in particular in remote locations. There is nothing worse than a missed field opportunity caused by the lack of a cable or battery.

Bring proper tools to the field site to make repairs: soldering iron, solder, electrical wire, heat-shrink tubing, electrical ties, electrical tape, extra cables and connectors, batteries (preferably rechargeable, with charger), multi-meter, etc. If possible, pack replacement equipment: anemometer, thermometer, laptop with extra charger, external speakers, software for data entry, backup hydrophone or microphone, headset, walkie-talkie, smartphone, microphone for narration onto a PC, and data storage devices (SD-cards, thumb-drive, external hard-drive). Why are duplicates necessary? If you cannot repair something, then use backups so the research effort is not wasted.

Moving or shipping equipment often creates problems with loose connections or fittings. If equipment is not operating properly, tighten fasteners on the equipment housing, make sure circuit boards are seated properly, check that batteries are fully charged, and make sure all cables are connected and working. To check for cable malfunction, use an ohm-meter to make sure the resistance of a cable is zero. If new equipment is used in a study, always unpack it and check its operation in the laboratory before going to the field. Bring manuals for all equipment to the field site or know where to reliably access them.

3.4 Playback Methods and Controls

Projections of sounds to animals (or playbacks) are common methods of study in bioacoustics (Fig. 3.6). Several authors have used playbacks to determine the function of a specific animal sound by measuring the animal’s behavioral response (Morton and Morton 1998).

Fig. 3.6

Playback studies are those by which an animal or group of animals is played their calls (or calls of their conspecifics) back to them and then their response is recorded. Research using playbacks has been used commonly in mammals (such as squirrels, prairie dogs, pika, carnivores, and primates), birds, reptiles, fish, and many others. Painting “His Master’s Voice” by Francis Barraud (1856–1924). Source: Victor Talking Machine Company. Public domain; https://commons.wikimedia.org/wiki/File:His_Master%27s_Voice.jpg

Playback studies on fish have been used to determine species recognition from a particular sound, to classify different call types, to identify effects of sound on fish behavior, to study how a call was coded, and to measure acoustic parameters of the call relevant to communication (Zelick et al. 1999). For example, Myrberg and Riggio (1985), studying bicolor damselfish (Stegastes partitus), found that males produced sounds more often in response to playbacks of conspecific sounds than to sounds of other species’, and responded more readily to sounds from non-resident fish than sounds from their nearest neighbor. Playbacks of male Lake Malawi cichlid fish (Pseudotropheus zebra) sounds to female cichlids caused them to lay eggs earlier than control female fish of another Lake Malawi cichlid species (Pseudotropheus emmiltos; Amorim et al. 2008). Simpson et al. (2011) played-back ambient sounds of different reefs to coral reef fish and showed that fish approached the sounds of their native coral reef versus sounds from a foreign reef. Hawkins et al. (2014) played back recordings of impulsive pile driving sounds attracting European sprat (Sprattus sprattus) in mid-water in the sea (Fig. 3.7).

Fig. 3.7

Responses of sprat (Sprattus sprattus) schools to sound exposure. Vertical lines indicate the beginning and end of each sound sequence. (a) Echogram of a medium-sized sprat school, cut off abruptly after the beginning of the sound, and reappearing a few seconds later as a denser school slightly closer to the seabed. (b) A medium-sized sprat school cut off at the onset of the sound and reappearing seconds later slightly closer to the seabed. (c) A large sprat school cut off at the onset of the sound and reappearing at a greater depth at lower density. (d) A small sprat school increasing in density in response to sound exposure. From Hawkins et al. 2014. © Acoustical Society of America, 2014. All rights reserved

Many birds respond to playbacks of their own or other animal sounds by approaching the projector and sometimes even attacking the speaker (Fig. 3.8). Emlen (1972) investigated how information is encoded in bird song by altering components of Indigo bunting (Passerina cyanea) song and playing-back the modified songs to male territory holders. He quantified the intensity of responses to modified songs and thus inferred the importance of temporal, structural, and syntactical features for both individual- and species-recognition.

Fig. 3.8

Diagram of a playback experiment with two different bird songs. The recording and the speakers should match the frequency range and levels of the original signals. Courtesy of G Pavan

Beecher and Burt (2004) played-back territorial sounds from male song sparrows (Melospiza melodia) that were in neighboring territories versus distant territories. The males were slower and less likely to fly over and explore the sounds from a neighbor than calls from a distant male. When a song from a distant territorial male was played, the subject almost always matched or replicated the song and approached the speaker as if looking for an intruder. In contrast, when the song of a neighbor male was played, 85% of the time the subject sang a different song, but one familiar to the neighbor. By responding with a different, but shared song, the subject sparrow indicated it recognized that the sounds were from a neighbor.

Much of the work in determining the function of alarm calls in ground squirrels and prairie dogs (Spermophilus and Cynomys, respectively) was determined or confirmed by playing-back previously recorded calls to an attentive colony of these rodents in the field and observing their responses (e.g., Slobodchikoff et al. 2009). Prat et al. (2016) used playback techniques of calls recorded from the Egyptian fruit bat (Rousettus aegyptiacus) to show that 16 sounds recorded and played-back from this bat provided enough information to identify who was calling, where they were calling from, what they were calling about, and what sort of response the receiver made to the vocalization.

Yegge (2012) and Thomas et al. (2016) reported using playbacks of duets to restore a pair-bond in yellow-cheeked gibbons (Nomascus gabriellae). A breeding pair of captive gibbons stopped duetting when construction occurred near their exhibit lasting for about 6 months. Afterwards, the authors played-back sounds of the pair’s previous duet, along with a silent- and music-controls. The pair slowly resumed their duet, established a pair-bond, and continued to duet, some 5 years later.

Playback experiments with marine mammals are less common due to the logistical challenges of undertaking these experiments at sea. However, there are a few examples. Weddell seals (Leptonychotes weddellii) produced geographically different vocal repertoires that has potential for identifying discrete breeding stocks of Antarctic seals (Thomas et al. 1983). Charrier et al. (2013) used playback methods to confirm that bearded seals (Erignathus barbatus) recognized vocalizations of their species from different regions. Male harbor seals (Phoca vitulina) that are territorial, use roars given by intruding seals to locate and challenge those intruders (Hayes et al. 2004). Deecke (2003) used playbacks to examine whether captive harbor seals could distinguish sounds from killer whales (Orcinus orca) that eat seals versus killer whales that eat fish; the seals exhibited fearful responses when sounds by the former were broadcast. Wild killer whales either approached or ignored playbacks of sounds from another killer whale pod, but did not call in response. However, when their own calls were played, most killer whales approached the source and the entire pod started calling in response (Filatova et al. 2011). Clark and Clark (1980) described right whale (Balaena australis) behavior from playback experiments where right whales can differentiate between conspecific sounds and other sounds. Playbacks of their own song or social sounds to wild humpback whales (Megaptera novaeangliae) resulted in some animals approaching, some charging the source, and others moving away (Mobley et al. 1988; Tyack 1983).

Before a playback session, the researcher should always check the projected sound near the animal to make sure the sound is not distorted and is of sufficient amplitude to mimic the intended sound. Ideally, playback experiments should be carried out on wild animals that are free to move within their natural habitats. Captive animals often are de-sensitized to reoccurring sounds, and confinement within a small space can greatly alter their behaviors and vocalizations. It is especially important to ensure that playback experiments are carried out under appropriate acoustic conditions, where the transmitted sounds are free from distortion, and reflection and reverberation are minimal. This is a particular problem with playback experiments on fish, where sounds can be greatly altered by the acoustic environment, especially in small aquarium tanks (Parvulescu 1964; Grey et al. 2016; Rogers et al. 2016).

Playback studies require controls to ensure the animal is responding to the projected sound and not to the noise/hum of equipment or the novelty of a new sound. Current sound analysis and sound-generation software allows the manipulation of many sound characteristics that could be used as a control. There are several types of controls used by investigators: 1) Merely turn on the equipment to replicate the electronic/background noise. 2) Play the animal’s own sound, but backwards. This projects the same frequency, amplitude, and time relationships of the actual sound, but in a different order. 3) Play the animal’s sound at a higher or lower speed. This transforms the projected sound into a different frequency range. 4) Play a call with parts filtered out. 5) Play something totally novel to the animal, such as sounds from another species it has never encountered, music, machinery noise, or human speech. 6) Play sounds typical of the animal’s natural environment.

3.5 Considerations for Terrestrial Field Studies

If recording on land, from a vehicle (such as during a truck survey for bat sounds), ground-generated noise can be a problem. In fact, Borkin et al. (2019) reported a negative relationship between bat activity and night-time traffic volume on New Zealand highways; when traffic increased, probability of detecting bats decreased. These researchers used stationary automatic bat detectors to avoid their own road noise. Some solutions include: stopping and turning the vehicle off and recording in silence; using a recently paved asphalt track rather than an older and noisier road or a dirt track; and carrying out vehicle transects using electric vehicles. Road surveys are valuable, but reducing non-biotic noise would make these transects even more valuable. Terrestrial recordings can be contaminated with nearby traffic noise. It is therefore advisable to make a sample recording, check it for ambient noise, and select an optimal quiet area.

Air temperature can be a problem. Thomas, Zinnel, and Ferm (1983), when recording Weddell seal breeding colonies, used water-activated chemical heat packs placed next to recording equipment and batteries in an insulated box to keep equipment warm in the Antarctic for 24-hour periods. In extremely warm locations with high humidity, moisture can collect on recorders or microphones. Placing recording equipment inside an insulated box with desiccants can minimize moisture problems. In rain forests, equipment must be totally waterproof. During periods of heavy rain, sounds from animals will either not be heard or masked by the rain.

A common problem in bioacoustical studies in terrestrial environments is the presence of acoustically-active non-target animals. If a non-target species calls in a specific frequency band, their sounds can perhaps be filtered out, but in many cases, this is not possible. Some analysis software allows to define the frequency and amplitude of a target species’ calls and automatically identifies only them in a recording. However, in many cases, finding locations and times when only an individual animal is vocalizing provides the best opportunity to make quality recordings.

A good solution for animals such as bats is to use units which are self-contained and weather resistant (see Chap. 2, section on bat detectors). Each unit can include a receiving transducer, storage device, or laptop programmed to record at intervals and can be powered by rechargeable battery packs or solar panels. Data can be recovered daily, weekly, monthly, or even uploaded in the proximity of Wi-Fi for automated data retrieval. Arrays of bat detectors have been used to record ultrasonic calls of bats, as well as to sample the acoustic landscape, estimate biodiversity, and estimate species density (Carles et al. 2007; Sherwin et al. 2000).

3.6 Considerations for Aquatic Field Studies

Studies in freshwater are easier on the equipment than in saltwater environments; saltwater’s corrosive properties require that underwater equipment be rinsed with freshwater after use and recorders and hydrophones be wiped down to remove saltwater deposited from the air. It is, of course, good practice to wipe down and dry all equipment, whether it was deployed in saltwater, in freshwater, or on land, after use to avoid any rusting or build-up of deposits.

Maintenance and calibration of equipment such as hydrophones has been shown to be important for long-term monitoring studies and data integrity. This includes considerations such as the pressure rating on the hydrophone and the length of cable that is waterproofed; the longer the cable, the higher the impedance and the greater the signal attenuation. Some plastic-coated cables, if deployed for long periods, are vulnerable to damage by marine organisms, shark bites, and even sea urchins. Polytetrafluoroethylene (PTFE) coated cables are less susceptible to damage of this kind. In addition, acoustic-release mechanisms (to allow equipment to surface) can malfunction when encrusted by marine creatures. In a review of underwater soundscape ecology to monitor habitat health in general, and fish spawning in particular, Lindseth and Lobel (2018) summarized current recording and sampling methods including metrics commonly used in analyses of aquatic acoustic data. They point out that there have been significant technological advances in equipment, especially hydrophones.

In aquatic situations, there can be electronic interference from improper grounding on the vessel, depending on the types of electronic equipment running onboard (e.g., lights, radios, freezers, generators, winches, fans, air conditioners, or furnaces). A quick-fix to grounding problems on a ship is to drop a bare wire into the water with the other end attached to the recording equipment. However, a trial-and-error approach may be needed to resolve this.

Flow noise is a problem that causes artifacts in the recordings. Noise from water flow over the hydrophone and its mooring can create turbulence and small eddies (vortex shedding). These lead to fluctuating pressure around the hydrophone, which is sensed by the hydrophone and appears as noise in recordings. But this “noise” is not due to a traveling acoustic wave and hence not due to sound in the environment. It is an artifact. Flow noise is often a problem in rivers but also offshore (see flow noise marked in the spectrograms in Fig. 3.3 in Erbe et al. 2015). It can require the use of a shield or deflector, or placement of the hydrophone in a sheltered area.

Sound-recording acoustic tags are attached to marine animals to record their vocalizations and examine the effects of anthropogenic noise in the marine environment relative to animal generated sound. Flow noise (generated simply by water flowing around the tag) can be useful in this instance, as it can measure whale speed (von Benda-Beckmann et al. 2016; Fig. 3.9). However, interference by background noise is also a common problem. Unfortunately, survey vessels produce noise while operating. Therefore, to avoid unnecessary mechanical background noise during recordings, turn off any non-essential equipment (such as engines, pumps, filters, fans, generators, lights, refrigerators, winches, etc.). However, fishing, military, research, and whale-watching boat operators often are reluctant to do this. Alternatively, these vessel sounds can be filtered out during recording or analysis.

Fig. 3.9

Non-animal generated noise can affect aquatic recordings adversely unless the research has a system in place that accounts for noise versus animal generated calls. Simply attaching a hydrophone or tag to a marine mammal can cause flow noise from water rushing around the attached object

In rivers or shallow coastal areas, currents and tides transport sediment which may create noise. It may come as quite a shock when an entire recording is ruined by nonstop sand swishing back and forth over the hydrophone, creating noise between 10 Hz and 2 kHz (Erbe 2009). Perhaps more amusing shallow-water “mooring noise” occurred when a group of teenage girls swam over to the mooring, held on to the floats and sang ABBA songs for 20 minutes—very clearly recorded. The entire recording session had to be discarded (Erbe 2013).

Similarly, a hydrophone fixed to a ship, boat, buoy, or dock will bob up-and-down and produce spurious signals such as flow noise as the water passes the hydrophone and artifacts from hydrostatic pressure changes as the hydrophone changes its depth. The recording can be saturated with such signals. This noise can be reduced by suspending the hydrophone with a bungee cord, decoupling the floating hydrophone from the surface through a catenary line, or mounting the hydrophone on the seafloor (Fig. 3.10; also see Chap. 2, section on PAM systems). Another solution to reduce flow noise is to use a sonobuoy or an anti-heave buoy (see photograph in Chap. 4, section on sonobuoys). The long cable of the sonobuoy acts as a bungee cord to dampen vertical oscillations of the hydrophone. The sonobuoy is isolated from self-noise of the vessel, but will detect sounds from the vessel until it moves out of range.

Fig. 3.10

Mooring options to avoid noise artifacts: (a) recorder on the seafloor, (b) recorder suspended from a float via a bungee cord and drogue, and (c) recorder suspended via a catenary line (Erbe et al. 2019). © Erbe et al.; https://www.frontiersin.org/articles/10.3389/fmars.2019.00606/full. Published under a Creative Commons Attribution License (CC BY); https://creativecommons.org/licenses/by/4.0/

Local sound propagation conditions will affect the recording (see Chap. 6 on sound propagation under water). It is important to measure and understand the sound speed profile in the study area to know the propagation pattern and range of a signal, which influence the recorded sound. For years, navies of the world measured sound speed profiles using disposable, battery-operated CTD (conductivity, temperature, depth) units, which were tossed into the ocean and data sent back to the ship as the unit fell in the water and unspooled a long copper wire. The units were not retrieved. Today, retrievable, digital CTD units are used. The sound speed profile may change over the course of a day—within the upper few meters below the sea surface. Turl and Thomas (1992) documented that a false killer whale (Pseudorca crassidens) echolocating during target-detection distance experiments in Kaneohe Bay, Hawaii, USA, consistently performed better during the morning than afternoon; i.e., the whale could detect the target at a greater distance during the morning. After taking CTD measurements prior to the morning and afternoon sessions, the researchers realized the water column, and thus sound speed profile, were very different between the two periods because or prevailing midday rains.

Sound propagation is particularly complicated in shallow water because of the close proximity of boundaries formed by the sea surface and seabed (Rogers and Cox 1988). Sound is reflected, scattered, and absorbed at these boundaries. There is far more attenuation of low-frequency sounds in shallow water compared to deep water. Rogers and Cox (1988) suggested that the lowest frequency that could propagate in water less than 1 m deep was about 300 Hz, but this was strongly dependent on the nature of the seabed (sand, rock, or mud).

Ambient noise is an omnipresent issue and may mask the signals desired for recording (see Chap. 7 on soundscapes). Wind and precipitation create noise underwater from coastal to offshore regions. In polar regions, ice popping and cracking may dominate the soundscape. When a hydrophone was dropped in the ice-covered water next to a group of Antarctic Weddell seals (JAT, personal observations), music was heard from the radio-station at the New Zealand Research Base in Antarctica about 2 km away! Organisms from tiny snapping shrimp to enormous singing whales may also mask recordings of a target species. Ship noise is almost omnipresent in the world’s oceans, so it can be difficult to obtain recordings of a target species in a quiet aquatic environment.

3.7 Considerations for Studies on Captive Animals

Because there are regulations on the housing and care of captive animals, research permit and IACUC requirements can be more detailed for research on captive species. However, often those regulations were written for laboratory animals used in medical research (mostly Rattus and Mus) and are not specified or applicable for wild animal research. For example, one of us (WLG) had to convince the university veterinarian to allow kangaroo rats (Heteromyidae, Dipodomys) to be housed using sandy desert soils instead of rat bedding so that these wild animals could properly sand-bathe and tunnel.

Zoos and aquaria support bioacoustical studies on a wide variety of species, including endangered species. Some benefits of studying captive animals in a zoo are that their history is usually known (i.e., wild caught vs. captive born, sex, age, reproductive history, relatedness to other animals, and health). Care should be taken to study healthy animals, as opposed to ill or rehabilitating animals, to best represent the acoustic abilities of their wild counterparts. However, burgeoning research by Therrien et al. (2012) indicated that changes in vocal behavior of bottlenose dolphins (Tursiops truncatus) and California sea lions (Zalophus californianus) actually could be used to indicate a health problem (Schwalm 2012). Moreover, captive animals, especially those that have been hand-reared or raised in a hatchery (such as salmon or sea bass) can show some degree of genetic selection, de-sensitization, and habituation to the presence of high levels of ambient sound. They can be much less responsive to sounds than wild animals.

Most zoos have noise created by loudspeaker announcements, music, shows, rides, or facility vehicles. Key events, such as hearing music for a show, or a vehicle delivering food, may affect animal behavior; therefore, studies should not be conducted during those times. Reminiscent of Ivan Pavlov in the 1890s experiment that dogs were being conditioned behaviorally (drooled) in response to being fed at the sound of a bell (conditioned response), researchers need to be aware of regular triggers to animal behavior. Of course, a common source of noise in captive studies is from visitors, keepers, and maintenance workers. If at all possible, it is best to conduct research before or after humans are near the study location (i.e., before or after the zoo is open). If possible, operation of air conditioners, furnaces, air-filters, and lights should be stopped, or minimized, to reduce or eliminate background sounds in recordings. Some facilities isolate their mechanical equipment in a separate building from the animals’ environment; this greatly reduces noise exposure for the animals. A preliminary survey of noise in the animals’ enclosure, using a sound pressure level meter, helps identify any particularly noisy or quiet areas.

Sometimes, ultrasonic noise or underwater noise can be present unbeknownst to zoo or aquarium staff. One of us (JAT, personal observations) provided two examples. In an underwater hearing study on a Pacific white-sided dolphin (Lagenorhynchus obliquidens) by Tremel et al. (1998), the test animal consistently reported hearing a 32-kHz signal at two different thresholds on different days. Spectrum analysis of the ambient noise in the pool revealed an intermittent noise near 32 kHz. So, on test days when the noise was present, the animal’s threshold at this frequency was much lower than on test days when the noise was absent. Because the noise was ultrasonic, it was not known by staff or researchers. In another study by Therrien et al. (2012), 24-hour recordings of bottlenose dolphins detected an almost continuous banging noise in the water. Zoo staff were unaware of the noise and upon a diver’s inspection of the pool, found a metal gate hinge that was broken and causing the banging sound. In both these examples, staff did not know about the noise, which could have been annoying to the animals and disturb bioacoustical research.

Researchers should understand the possible effects of the exhibit environment on the acoustic behavior of animals. For example, dolphins living in highly reverberant concrete pools echolocate less and at lower amplitudes than in the wild (Fig. 3.11) (Au 2000).

Fig. 3.11

Waveforms and spectra of echolocation clicks of bottlenose dolphins in open ocean (Kaneohe Bay, Hawaii, USA) and in a tank. The spectrum of the click from the tank had a lower frequency peak at 40 kHz and a lower source level of 170–185 dB re 1 μPa m. Reprinted by permission from Springer Nature. Hearing by Whales and Dolphins, edited by W. W. L. Au, A. N. Popper, and R. R. Fay, pp. 364–408, Echolocation in dolphins, W. W. L. Au; https://doi.org/10.1007/978-1-4612-1150-1_9. © Springer Nature, 2000. All rights reserved

Today, exhibit designers incorporate irregular wall and floor surfaces in pools, indoor enclosures, and outdoor exhibits to minimize reverberations. Projecting a signal into a regularly shaped (e.g., round or square) pool with a flat bottom (e.g., during a hearing test) can set up standing waves, which result in a sound-field that dramatically changes with receiver location and frequency. A resonant pool amplifies sound at its resonance frequencies and dampens others, essentially distorting the signal desired by the researcher. While concrete walls in a zoo or aquarium are easy to construct and clean, they provide a reflective surface that often causes annoying, cave-like reverberations.

Particular issues are encountered when trying to perform hearing tests and sound exposure experiments with fish or invertebrates in water-filled tanks that are only a few meters in dimensions, or even smaller. The complexities of the sound-field in small tanks were first pointed out by Parvulescu (1964) and recently discussed by Duncan et al. (2016), Grey et al. (2016), Rogers et al. (2016), and Popper and Hawkins (2018). Even in quite large tanks, the sound-field generated by even a simple sound source is transformed by interactions with boundaries (i.e., walls, floor of pool, and water surface) and can vary rapidly as a function of both space and frequency. The resulting sound-field can be difficult to model, or even characterize, and the sound-level can be very different from the natural environment. In particular, the levels of the particle motion components of the sounds (to which fish are sensitive) can be very high. Attempts at dampening reverberation by adding materials such as “horse hair” or bubble-wrap can be effective at high frequencies, but have little effect at the low frequencies to which fish are sensitive and where the sound wavelength often exceeds the dimensions of the tank (Popper and Hawkins 2018). In contrast, experiments performed in deep and open water allow the establishment of a relatively simple, well-controlled, and predictable sound-field (Hawkins 2014).

Grey et al. (2016) measured the sound-field in several large laboratory tanks and came to the following conclusions: 1) Tanks, even large ones, are not appropriate surrogates for open-water environments. 2) Tank wall-thickness is largely irrelevant. Walls backed by air essentially present a low impedance, and walls in contact with a solid foundation or ground present finite (non-rigid) impedance defined by the substrate materials. 3) Resonance of the tank walls can dominate underwater sound-field characteristics. 4) Lining the walls of a tank with acoustic absorbent material is futile, because the thicknesses required at low frequencies would leave no room for the fish. 5) Both the sound pressure and the particle motion of a sound need to be measured and checked for mutual validation by calculating the particle motion from pressure gradients. Special hydrophone systems, based on seismic accelerometers, are required to measure particle motion (see Chap. 2).

3.8 Digital File Format

Several file formats are available to save digital recordings. Digital file extensions include WAV, PCM, MP3, au, ram, MIDI, ogg, as well as others. It is best to record using uncompressed or WAV or PCM (Pulse Code Modulation) formats for faithful spectrum analysis.

MP3 is a digital audio-encoding format which uses data compression to reduce file size. It is a common audio-format for consumer audio and a de facto standard of digital audio-compression used for the transfer and playback of music. However, MP3 files and other compression methods are poor for spectrum analysis because compression only retains signals in a frequency band up to 16 kHz (i.e., the human hearing range). As a result, spectrum analysis using MP3 files is not trustworthy above 16 kHz. The psychoacoustic-based compression algorithms, in addition to limiting frequencies to below 16 kHz (and even less at higher compression ratios), discards fine details that cannot be heard by humans. Cuts introduced by compression appear as unpleasant “holes” in the spectrogram and can destroy details that could have meaning. However, MP3 files can be valuable for ecological monitoring of temporal and spatial patterns of well-known sounds.

A few digital recorders offer the Free Lossless Audio Codec (FLAC) format, which has less compression and reduces the storage space up to 50% without loss of detail. In addition, a few digital recorders employ a Direct Stream Digital (DSD) format; a proprietary system of digitally recreating audible signals for the Super Audio CD, using delta-sigma 1-bit A/D-converters at 2.8 or 5.6 MHz. Because of the intrinsic properties of the delta-sigma conversion made by the 1-bit A/D-converter, these recorders have the potential to record frequencies well beyond 100 kHz, but with increased noise at high frequencies. Spectrum analysis of recordings made in the DSD format is appropriate.

Waveform sound files (WAV; created by Microsoft) are perhaps the simplest of the common formats for storing audio samples. Unlike MPEG and other compressed formats, WAV files and their derivatives (like the Broadcast Wave File, BWF) store samples “in the raw” where no pre-processing is used, other than formatting of data. When there is a choice of a recording file format, the WAV (or BWF) format should be selected, rather than the MP3 format.

With continuous recording, WAV files can become quite large and subsequently be difficult to handle with sound analysis software. For example, WAV recordings sampling at 96 kHz and 24 bit for 1 hour will occupy approximately 1 GB of storage capacity (96,000 samples/s × 24 bits × 1 byte/8 bits × 60 minutes × 60 s/minute = 1.04 GB). If monitoring is required for long periods, it is therefore important to select the appropriate sampling rate to conserve storage space. For example, if mid-frequency fish sounds are the main features of interest, then it can be appropriately sampled at only 22 kHz, or at an even lower sampling frequency. Several possible sampling frequencies and sometimes a choice of bit depth (16 or 24 bit) are available, but not on all recorders. Some recorders enable a limit to be placed on the maximum size of each recorded file. Alternatively, a recording protocol can be adopted to limit the length of each recording.

3.9 Data Storage

All storage media should be carefully labeled with who, what, where, and when. Each recording period should have a unique number. Creating a master catalog of recording numbers allows researchers to cross-reference metadata from a logbook.

Magnetic media, including magnetic tape (e.g., reel-to-reel, cassette, or DAT tapes), and computer hard drives require storage in a dry, dark area away from any type of magnetic field. Exposure to a magnet could erase data. If tapes are not played often, the tightly packed tape could “bleed through” from one segment to another, thus contaminating data. Therefore, converting old recordings on magnetic tape to modern storage is becoming urgent for data on historic soundscapes and animals not be lost.

When converting analog to digital formats, usually using an A/D-converter, the sampling frequency must be at least twice the highest frequency recorded and the recordist needs to make sure that the parameters of the storage medium are adequate for the task. There are a number of free software applications for conversion of analog to digital formats.

Storage of digital recordings can be done on hard drives, optical drives, solid-state memory, or an Internet cloud. Bluetooth (a wireless technology standard) provides reliable exchange of data between fixed and mobile devices over short distances. Bluetooth uses UHF radio waves that are effective at a short distance.

3.10 Archiving Recordings

Properly curated recordings are critically important for assessing changes in soundscapes, ambient noise, and animal presence/absence and acoustic behavior over time. For example, underwater recordings made by the US Navy off the coast of California indicated a steady increase in background noise levels in the ocean in the last 60 years (from the 1960s). Marie Poland Fish, an oceanographer and marine biologist, recorded and analyzed the sounds of more than 300 species of marine life, from mammals to mussels. Her work (described and spectrograms provided in Fish and Mowbray 1970) helped the US Navy to distinguish fish and other animal sounds from the sounds made by submarines and remains a primary source for analysis of marine fish sounds.

Recordings of humpback whale songs date back to the 1970s and continue to document annual changes in their song within different populations. Williams et al. (2013) studied the changing songs of male savannah sparrows (Passerculus sandwichensis) recorded over three decades (1980–2011) on Kent Island, New Brunswick, in the Bay of Fundy. Life-long recordings of songs of white-crowned sparrows (Zonotrichia leucophrys) found they memorize syllables they hear at 10–50 days of age and sing the same song throughout their life. In contrast, life-long recordings of northern mockingbirds (Mimus polyglottos) found they add elements to their songs throughout their lives. Only long-term archival data could be used for analysis of these trends. In this time of global warming and accelerated ice melts, archived recordings from the polar regions might become instrumental in monitoring the rate of climate change (by quantifying ice-cracking noise) and the effects on soundscapes and ecology (Obrist et al. 2010). The take-home message here is that good research practices with solid documentation and data archiving allow for future knowledge generation.

3.11 Repositories of Bioacoustical Data

Hafner et al. (1997) noted that collections of animal recordings with ancillary data are rich sources of reference material for bioacoustical studies. Archiving analog data by converting to a digital format has played an essential role in preserving data for future use. Species-specific sounds from a variety of regions and times, with associated voucher specimens and metadata, are available for researchers at a number of organizations. All collections and their corresponding links were valid as of 13 June 2022.

In Europe, there is a long tradition of recording animal sounds, in particular bird songs, and many collections have been published on vinyl discs and CDs, mainly in France and the UK. In 1969, the British Library of Wildlife Sounds² established holdings of more than 160,000 well-documented field-recordings covering all classes of sound-producing animals from many regions. More than 10,000 species of invertebrates, insects, amphibians, reptiles, fishes, birds, and mammals, including many rare and threatened species. A large number of these recordings were made for radio by the BBC Natural History Unit. The British Library supported a citizen-science program to create a map of the UK coastal soundscape in 2015.³ Other European online sound libraries include: Tierstimmen Archiv⁴ (approximately 120,000 sound recordings; Museum für Naturkunde, Berlin, Germany) Xeno-Canto⁵ (595,000 recordings from approximately 10,250 bird species Naturalis Biodiversity Center, Leiden, Netherlands), and FonoZoo⁶ (11,657 recordings of 1621 animal species; Fonoteca Zoológica, Museo Nacional de Ciencias Naturales (CSIC), Madrid, Spain).

In the USA, the Macaulay Library⁷ (Cornell Lab of Ornithology, Ithaca, NY, USA) archived older analog, digital, and video recordings. To date, their holdings are approximately 24 million photos, 915,000 audio and 192,000 video recordings available for researchers. The K. Lisa Yang Center for Conservation Bioacoustics⁸ (Cornell Lab of Ornithology, Ithaca, NY, USA) is everything “bird” including citizen science and masterful guides and information in ornithology (including bird vocalization identification apps and bird cams). The Museum of Southwestern Biology⁹ (University of New Mexico, Albuquerque, NM, USA) and Museum of Vertebrate Zoology¹⁰ (University of California, Berkeley, CA, USA) have hundreds of thousands of cataloged natural history journals and voucher specimens and began to associate avian vocalizations with voucher specimens in the 2000s. These museum collections have shown a desire to include bat call libraries before 2023. The Watkins Sound Library¹¹ (Woods Hole Oceanographic Institution, Woods Hole, MA, USA) provides particularly good collections of marine mammal sounds with a highlighted “Best of” cuts section that contains 1694 sound cuts deemed to be of higher sound quality and lower noise from 32 different marine mammal species.

Several commercial companies market LPs and CDs of nature sounds. Bernie Krause¹² (Wild Sanctuary, Glen Ellen, CA, USA; Fig. 3.12) is unique among researchers, commercial ventures, and artists. From the Wild Sanctuary website, “The Wild Sanctuary Audio Archive represents a vast and important collection of whole-habitat field recordings and precise metadata dating from the late 1960s. This unique bioacoustic resource contains marine and terrestrial soundscapes representing the voices of living organisms from larvae to large mammals and the numerous tropical, temperate and Arctic biomes from which they come. The catalog currently contains over 4500 hours of wild soundscapes and in excess of 15,000 identified life forms.” The acoustic world is not only at our finger tips, but the world is becoming available for all to hear.

Fig. 3.12

Commercial companies and others market sounds of animals and soundscapes recorded by researchers such as Bernie Krause. Recording and analyzing natural sound is fulfilling and insightful, and can be a profound source for generating knowledge. Left photo by the authors; right photo, “Capturing the sounds of the lake” by S. Shiller; https://www.flickr.com/photos/12289718@N00/9454414945; licensed under CC BY 2.0; https://creativecommons.org/licenses/by/2.0/

3.12 Summary

As with other areas of science, good practices for bioacoustical research, as well as an awareness of the ethical implications of that research, should be employed. This chapter provides a list of considerations for terrestrial, aquatic, and captive studies—a list that will doubtlessly be improved as technology and access to the acoustic world improves. No longer is large, heavy, and expensive equipment necessary to make high-quality, meaningful acoustic recordings. Acoustic data are important beyond the immediate scope of a project, but data must be well documented with metadata (including field notes and ancillary information) and stored in a way that they are preserved and accessible for future research. The importance of a well-designed data sheet for easy data entry and analysis is also discussed along with special considerations for study design. Playbacks of sounds to animals are commonly used by bioacousticians and procedures for playbacks and controls are recommended.

Several sound libraries are publicly available for research. These facilities have invested a great deal of time in transferring analog recordings to digital formats for more permanent preservation. CDs of animal and nature sounds are now commercially available. Archives are useful for education and research. As we evaluate current hypotheses related to global warming, perhaps we can hear the world change.

3.13 Additional Resources

• Sound recording tips from eBird: https://www.macaulaylibrary.org/how-to/recording-techniques/

• Bioacoustics equipment and field techniques, Centro Interdisciplinare di Bioacustica e Ricerche Ambientali, Università degli Studi di Pavia: http://www.unipv.it/cibra/edu_equipment_uk.html

• Manual on Field Recording Techniques and Protocols for All Taxa Biodiversity Inventories and Monitoring (Eymann et al. 2010): https://issuu.com/ysamyn/docs/abctaxa_vol_8_part1_lr

All web resources were last accessed 13 June 2022.