Effects of Anthropogenic Underwater Noise on Harbour Porpoise Hearing in Areas of High Ecological Importance

University of Veterinary Medicine Hannover, Foundation

Passive acoustic monitoring for evaluating potential impacts of anthropogenic noise on marine animals: Tools for ecological assessment and monitoring

Passiv akustisches Monitoring zur Bewertung möglicher Auswirkungen von anthropogenem Lärm auf die marine Fauna: Tools zur ökologischen Bewertung und zum Monitoring

INAUGURAL - DISSERTATION submitted in fulfilment of the requirements for the degree - Doctor rerum naturalium - (Dr. rer. nat.)

submitted by Johannes Baltzer Burg

Büsum 2020

Scientific supervision: Prof. Prof. h. c. Dr. Ursula Siebert Institute for Terrestrial and Aquatic Wildlife Research, University of Veterinary Medicine Hannover, Foundation.

Prof. Dr. Magnus Wahlberg Marine Biological Research Center, Kerteminde, University of Southern Denmark, Odense.

First supervisors: Prof. Prof. h. c. Dr. Ursula Siebert Institute for Terrestrial and Aquatic Wildlife Research, University of Veterinary Medicine Hannover, Foundation.

Prof. Dr. Magnus Wahlberg Marine Biological Research Center, Kerteminde, University of Southern Denmark, Odense.

Second examinor: Prof. emeritus Dr. Paul Nachtigall Institute of Marine Biology, University of Hawaii, Kailua Hawaii.

Date of oral examination: 13.11.2020

Parts of this study were published in a peer-reviewed journal: Baltzer, J., Maurer, N., Schaffeld, T., Ruser, A., Schnitzler, J.G., Siebert, U. (2020). Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea. Journal of Sea Research, Volume 162, 1-8, doi: 10.1016/j.seares.2020.101912.

Parts of this study are submitted in a peer-reviewed journal: Baltzer, J., Iwata, T., Akamatsu, T., Sato, K., Aoki, K., Lucke, K., Rasmussen, M.-H., Wahlberg, M., Schnitzler, J.G., Siebert, U. (2020). Blue whale off Iceland: Silent giants in a Northern feeding ground.

Table of Contents

Table of contents

List of abbreviations ...... I Introduction ...... 1 I. Passive acoustic monitoring of marine mammals ...... 1 II. Advantages of acoustic monitoring over other monitoring techniques ...... 2 III. Marine mammals are vulnerable towards anthropogenic effects ...... 4 IV. PAM for studying anthropogenic effect ...... 5 V. Sound propagation ...... 6 VI. Development in PAM equipment since the 1990s ...... 7 VII. Introducing the SoundTrap ...... 8 VIII. Methodological development of the analysis of recordings ...... 9 IX. Application of the concept of TOAD – Localisation ...... 10 X. Aims of the thesis ...... 11 Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea. .... 13 Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea ...... 14 Abstract ...... 14 I. Introduction ...... 16 II. Material and methods ...... 19 A. Study area and anchor pipes vibration embedment operations ...... 19 B. Sound propagation modelling and frequency analysis ...... 21 III. Results ...... 22 A. Sound propagation in the study area ...... 22 IV. Discussion ...... 26 A. Potential effect on marine mammals ...... 26 B. Potential effect on marine fish ...... 30 C. Marine Strategy Framework Directive ...... 31 D. Conclusion ...... 32 V. Acknowledgement ...... 33 VI. Appendix ...... 34 Blue whales off Iceland: Silent giants in a Northern feeding ground...... 35 Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground ...... 36 Abstract ...... 36 Author contributions ...... 37 I. Introduction ...... 38 II. Material and methods ...... 40 A. Study area and equipment...... 40 B. Analysis of the call characteristics ...... 42

Table of Contents

C. Tag data analysis ...... 45 III. Results ...... 47 A. Propagation loss and source level based on array data ...... 48 B. Acoustic and diving behaviour from tag data ...... 50 IV. Discussion ...... 56 V. Conclusion ...... 60 VI. Acknowledgement ...... 60 VII. Appendix ...... 61 The effect of marine vibrators on blue whale vocalisation behaviour ...... 62 Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour ...... 63 Abstract ...... 63 I. Introduction ...... 65 II. Material and methods ...... 67 A. Study area ...... 67 B. Sound recorders (SoundTraps) ...... 67 C. Sound source ‘Argotec SS-2’ ...... 67 D. Experimental design ...... 69 E. Data analysis ...... 70 III. Results ...... 73 IV. Discussion ...... 80 V. Conclusion ...... 82 VI. Acknowledgement ...... 83 VII. Appendix ...... 84 Overall discussion ...... 85 I. Anthropogenic noise affects marine fauna ...... 85 II. Anthropogenic noise and masking ...... 87 III. Anti-masking strategies by marine mammals ...... 90 Conclusion ...... 93 Summary ...... 95 Zusammenfassung ...... 97 Bibliography ...... 99 List of figures ...... 118 List of tables ...... 119 Acknowledgement ...... 120

List of abbreviations

In addition to the commonly used abbreviations following short forms were used:

ADD Acoustic Deterrent Device AVS Absolute Vertical Speed CEE Controlled Exposure Experiments FFT Fast Fourier Transform GES Good Environmental Status ITAW Institute for Terrestrial and Aquatic Wildlife Research MSFD Marine Strategy Framework Directive MV Marine Vibrator PAM Passive Acoustic Monitoring PL Propagation Loss (dB) RMS Root-Mean-Square RMSE Root-Mean-Square Error RNL Radiated Noise Level (dB re 1 µPa) SEL Sound Exposure Level (dB re 1 µPa2s) SL Source Level (dB re 1 µPa m) SMC Seed Mussel Collector SNR Signal-to-Noise Ratio SPL Sound Pressure Level (dB re 1 µPa) ST SoundTrap TOAD Time-Of-Arrival Difference TTS Temporary Threshold Shift

Introduction

Sound plays an important role in the lives of marine mammals. All marine mammal species produce sound, and sound production has been associated with a variety of behaviours including those related to mating, rearing of young, social interaction, group cohesion, and feeding (Erbe et al., 2016). Sounds emitted by cetaceans are very diverse. Some animals produce low- frequency moans (e.g., blue whales down to 9 Hz), whereas others emit high-frequency clicks or whistles. The harbour porpoise (Phocoena phocoena), for example produces only one specific vocalisation type for echolocation, a narrowband high-frequency click around 130 kHz (Au et al., 1999). These clicks are emitted in a certain pattern associated with foraging (DeRuiter et al., 2009; Verfuß et al., 2009). The same sound is also used in click bursts for communication (Amundin and Cranford, 1990; Clausen et al., 2010; Sørensen et al., 2018).

Underwater acoustic recordings have provided ecological, geographical, and behavioural data on a variety of species. Passive acoustic monitoring takes advantage of the fact that cetaceans like harbour porpoises echolocate (Akamatsu et al., 2007; Møhl and Andersen, 1973), and has therefore been used to investigate porpoise occurrence (Carlström et al., 2009; Dähne et al., 2013; Kyhn et al., 2012). Recorded calling patterns have also given clues to whales’ daily calling behaviour and seasonal presence (e.g., Širović et al., 2004).

I. Passive acoustic monitoring of marine mammals Passive acoustic monitoring (PAM) is a label for all techniques recording or logging the detections of acoustic signals. PAM is used to determine e.g., seasonal and diurnal patterns of sound production in animals and animal distribution. Initially, manual recording and analysis systems were used, but nowadays there exists a range of both automated data loggers that can collect animal sounds for months or even years, as well as sophisticated software packages to extract and analyse relevant sounds. These techniques have been especially applicable and useful for studies of marine mammals, where species-specific and far ranging underwater signals are frequently emitted from many species.

Species‑specific factors, such as vocal behaviour, frequency, directionality, and source level can have an influence on acoustic monitoring. The pitfall with PAM is that it will only work for some marine mammal species. Some species are more amenable to make use of their vocalisations when scientists are investigating their behaviour. Marine mammals have to

Introduction

breathe, but they do not necessarily have to vocalise. There are species that emit sounds more frequently than others depending on the sound production rate. Animals use social sounds for example for breeding or as contact calls. Here, we focus on sounds made by cetaceans. Vocalising behaviour varies with gender, age, and season. For instance, adult males of many baleen whale species vocalise regularly and loudly during the breeding season. The larger cetaceans, including baleen whales and sperm whales, produce intense vocalisations that can be detected at distances of several tens of kilometres on a single hydrophone (Barlow and Taylor, 2005; Stafford et al., 2007) and much further (hundreds of kilometres) on hydrophone arrays (Clark, 1995). Odontocetes produce even more sounds than baleen whales. However, due to their vocalisations being higher in frequency than those of most baleen whales, they are detected at shorter distances.

High-frequency clicks of some odontocetes are highly directional. For instance, bottlenose dolphins have a directivity index (which describes the sharpness of their emitting beam pattern) of about 26 dB (Au, 1993). The source level of sperm whale clicks vary around 35 dB between different directions (Møhl et al., 2000). In contrast, low-frequency baleen whale sounds are believed to be emitted omnidirectionally, because the long wavelengths compared to the size of the source make directional sound emission impossible.

Sounds below 1 kHz (typical for baleen whales) have significantly less seawater absorption loss than sounds above 10 kHz produced by odontocetes (François and Garrison, 1982), and thus can be detected at greater distances. Some cetaceans vocalise more frequently or more consistently and more intensely than others, making them better subjects for acoustic surveys. For instance, blue whale (Balaenoptera musculus) tonal calls have been measured at 179– 189 dB re 1 μPa m (rms) (Cummings and Thompson, 1971; McDonald et al., 2001; Samaran et al., 2010; Širović et al., 2007; Thode et al., 2000), while on-axis sperm whale clicks have been measured at instantaneous levels up to 223 dB re 1 μPa m (peak) (Møhl et al., 2000). In contrast, bottlenose dolphin (Tursiops truncatus) tonal sounds (whistles) have been measured at source levels up to 169 dB re 1 μPa m (rms) (Janik, 2000), while their clicks have been measured at 210–213 dB re 1 μPa m (rms) (Au, 2004).

II. Advantages of acoustic monitoring over other monitoring techniques Although there are many challenges with marine mammals, PAM is actually the best method for studying the behaviour for these less accessible animals. The sounds of baleen whales are

Introduction

some of the loudest in the ocean, and their low frequencies travel long distances (Richardson et al., 1995). This is one of many reasons that makes PAM an appropriate tool for monitoring distributions of whales over a large area (Stimpert et al., 2015). Acoustic monitoring has many advantages in terms of effectively recording animals for long time periods without needing much manpower.

In contrast, visual monitoring is a default method for detecting marine mammals as they are difficult to observe. Some make long dives and hence are unavailable to be seen at the surface for extended periods. The efficiency of sightings is also decreased when weather conditions are bad. Visual surveys often detect only a small portion of the marine mammals in an area, are limited to daylight hours, and are strongly influenced by animal behaviour and weather conditions (Mellinger and Barlow, 2003). Passive acoustic detection of marine mammals can complement and, in some cases, offer advantages over traditional visual surveys. Provided that their calls are distinctive and well characterised, the probability of detecting vocal species of marine mammals is increased by using sound recordings in addition to or, in some cases, instead of visual surveys (Mellinger et al., 2007).

Another consideration is that maintaining high levels of vigilance is demanding on observers. During visual surveys for example, many observers are needed to effectively cover the survey area (Gillespie et al., 2008). Visual detection is, of course, extremely limited at night, yet for economic reasons, operators may need to continue activities in a 24-hour-cycle. Technological advancements in automated long-term recording systems are increasing the importance of acoustic recordings as a passive monitoring tool for whale populations (Mellinger et al., 2007). Acoustic monitoring provides a complementary approach to conventional visual line transect surveys. Long-term acoustic recordings covering much of the world’s oceans are becoming freely available from a variety of sources, such as U.S. Navy’s Pacific Missile Range Facility (Helble et al., 2015). PAM can improve our knowledge of migration patterns, population structure, abundance and behaviour (McDonald et al., 2009).

Fixed passive acoustic methods can be performed year-round at relatively low cost (e.g., Thompson and Friedl, 1982). Further, they can be conducted in remote areas that are difficult to survey. These areas can be far from land (Clark, 1995; Nieukirk et al., 2004; Stafford et al., 1999) or polar regions (Mellinger et al., 2007; Moore et al., 2006; Munger et al., 2005; Širović

Introduction

et al., 2004, 2007; Stafford et al., 2007) exposed to harsh weather conditions making visual surveys impossible, in some seasons (Mellinger et al., 2007).

When combined with other data such as visual, environmental (i.e., water temperature, pressure and salinity), and satellite measurements, long-term acoustic monitoring can be an especially powerful observational tool for studying marine mammals. Acoustic monitoring can also provide information when analysing behavioural responses of vocalising animals towards acoustic events, either anthropogenic or natural. Regulators in many regions require mitigation involving the real-time detection of marine mammals during activities emitting intense sound, such as seismic surveys, pile driving and military sonar exercises (Gillespie et al., 2008).

III. Marine mammals are vulnerable towards anthropogenic effects Marine mammals live in an medium, where vision is limited due to a rapid light absorption but sound is propagating efficiently (Gordon and Tyack, 2002). Therefore, they evolved to actively and passively use sound for biologically important behaviours (Tyack, 2000). Marine mammals use sound ranging from a few Hz to over 100 kHz to navigate through their environment, communicate and search for prey (Gordon and Tyack, 2002). The fact that sound propagates very efficiently under water and marine mammals depend on sounds in many different ways, makes these animals potentially vulnerable to underwater noise caused by anthropogenic sound sources in the oceans (Gordon and Tyack, 2002).

Underwater noise can have effects on all aspects of their life, such as foraging, mating, nursing, resting, migrating. These effects can cause hearing impairment, masking of acoustic signals, behavioural responses, or physiological stress (e.g., Erbe et al., 2018; Nowacek et al., 2007; Richardson et al., 1995).

An impairment of hearing sensitivity has been documented in harbour porpoises after the exposure with pile driving strikes (Kastelein et al., 2016) resulting in a temporary threshold shift (TTS) (Lucke et al., 2009). Behavioural responses (Kastelein et al., 2013a, 2013c) and changes in spatial distribution (Brandt et al., 2016; Carstensen et al., 2006; Dähne et al., 2013; Scheidat et al., 2011; Teilmann and Carstensen, 2012; Tougaard et al., 2009) have been observed for the same species caused by impulsive noise. Other responses of porpoises can be stress or the interruption of their natural behaviour, such as feeding (Wisniewska et al., 2018).

Introduction

Other odontocetes, such as killer (Orcinus orca) and beluga whales (Delphinapterus leucas) as well as bottlenose dolphins are also affected especially by ship noise (reviewed in Erbe et al., 2019). Ship noise has the potential to mask relevant signals for these species and elicit behavioural responses resulting in adjustments of their vocal behaviour when ambient noise is elevated (Foote et al., 2004; Holt et al., 2009, 2011; Kragh et al., 2019; Scheifele et al., 2005). These behavioural responses are also seen in several baleen whales (mysticetes), like humpback (Megaptera novaeangliae), minke (Balaenoptera acutorostrata), bowhead (Balaena mysticetus) and North Atlantic right (Eubalaena glacialis) (Dunlop et al., 2014; Guazzo et al., 2020; Helble et al., 2020; Parks et al., 2007, 2009; Thode et al., 2020).

It is therefore important to know how marine mammals perceive their environment, how far their acoustic communication ranges reach, and how they adapt to ambient and anthropogenic noise (Erbe et al., 2018). Empirical methods are necessary to study the behavioural and acoustic responses of marine mammals towards exposures from sound sources, such as military sonar, seismic exploration, shipping vessels, or construction (Latusek-Nabholz et al., 2020).

IV. PAM for studying anthropogenic effect Tyack (2009) summarised methods on how effects of anthropogenic sounds on marine mammals can be assessed. These methods involve observational studies focussing on determining behavioural changes of animals near the anthropogenic sound source in situ or experimental studies testing animal responses with controlled exposure experiments (CEEs; Tyack et al., 2003). CEEs are a good method to directly measure acoustic exposure and behavioural responses of an animal with a specific sound stimulus of a known dose (Tyack, 2009).

DeRuiter et al. (2006) used data of CEEs on tagged sperm whales to measure sound characteristics and the propagation of airgun pulses. A study by Kyhn et al. (2015) used PAM together with CEE to investigate how harbour porpoises respond to pinger-sounds. Pingers are acoustic deterrent devices (ADD) invented to prevent porpoises from getting entangled in gill nets, however they were found to displace the animals temporarily from their habitat. Another study investigated the effects of pingers on the vocalisation behaviour of harbour porpoises for different distances (Kindt-Larsen et al., 2019). They documented that pinger sounds affected porpoise up to a range of 400 m. When comparing their findings with the study of Hardy et al. (2012), they observed that the effects of pingers were differently pronounced possibly due to

Introduction

varying background noise levels between regions (Kindt-Larsen et al., 2019). Areas with high background noise levels have a reduced signal-to-noise ratio (Urick, 1983) impeding to detect certain signals, e.g. from pingers. Differences in propagation could also be caused by differences in sea bed morphology and water depths as shallow waters can lead to multipath sound propagation (Shapiro et al., 2009).

V. Sound propagation Underwater acoustic propagation addresses how acoustic signals in water move from the source to a receiver (Lurton, 2010). The intensity of an acoustic signal decreases with distance through geometrical spreading and absorption depending on chemical properties of the seawater. The propagation loss (PL) is an important factor for PAM due to its effects on the receiver’s ability to detect and classify sound sources. Therefore, PL has to be evaluated when considering the performance of underwater acoustic systems (Lurton, 2010).

One way to estimate PL from absorption and attenuation are acoustic propagation models. Several parameters, such as source frequency band and configuration, sound speed profile, bathymetry, bottom properties as well as source and receiver geometry are often included in these models to reliably estimate PL (Küsel et al., 2009). Acoustic propagation modelling has been applied to military operations, marine seismology, and physical oceanography and is more recently being used to address questions in regards to marine ecology, physics, and conservation (Tennessen and Parks, 2016).

Propagation modeling is particularly important in PAM studies to assess marine mammal occurrence as well as ambient and anthropogenic noise effects on species and populations. For example, knowledge of acoustic propagation of seismic exploration signals is critical when predicting exposure levels and potential impacts to marine wildlife (Jochens et al., 2008). These types of anthropogenic noise propagation studies In addition to examining the characteristics of anthropogenic noise and potential impacts on marine mammals, propagation modeling is used to localise and track individual sound sources. These models are particularly important when assessing the vocalisation behaviour of different animals and species and trying to discern acoustic sounds of a specific individual (Sidorovskaia, 2005). For instance, the ability to discern spectral features of whale clicks from single hydrophone recordings based on surface- and bottom-reflected arrivals helps researchers develop algorithms for animal localisation and tracking (Sidorovskaia, 2005; Tiemann et al., 2006).

Introduction

VI. Development in PAM equipment since the 1990s Nowadays, passive acoustic devices are used in many different ways with a variety of types for investigating marine mammals. Tags can be attached to the animals to obtain data on the acoustic behaviour, diving behaviour, and acceleration for individuals (Johnson and Tyack, 2003). Acoustic monitoring systems were developed to be bottom-mounted on the sea floor or floating at the surface. Single hydrophones or sparse to dense arrays can be deployed. What devices are used and in which design depends on the species and the scientific questions.

Since the 1990s, several autonomous recording devices with various capabilities have been developed and used in different settings for PAM to record whale sounds throughout the world’s oceans (Au et al., 2000). Acoustic studies include shore cabled hydrophones, autonomous hydrophone recorders, towed hydrophones, drifting and moored sonobuoys. However, many systems have been limited in sample rate and record only low-frequency baleen whales (below 1 kHz) such as blue (Balaenoptera musculus), fin (B. physalus), humpback (Megaptera novaeangliae) and right (Eubalaena spp.) whales. These systems, however, are not capable of providing both long-term (months) and broadband (up to 100 kHz or more) recordings that are required for monitoring odontocetes (e.g., Wartzok and Ketten, 1999).

Therefore High-frequency Acoustic Recording Packages (HARPs) have been developed to acquire acoustic data with high-bandwidth for long-term marine mammal monitoring (Wiggins and Hildebrand, 2007). Another challenge to overcome in acoustic studies is to localise marine mammals to learn more about their distribution. One example for solving the localisation problem are the Directional Frequency Analysis and Recording (DIFAR) sonobuoys. These have been used by the U.S. Navy for many decades and provide magnetic bearings to low frequency (<4 kHz) sound sources from a single sensor (McDonald, 2004). DIFAR sonobuoys are well suited for localising not only baleen whales, but also various other sound source, e.g., ships. Marine Acoustic Recording Units (MARU; Cornell Bioacoustics Research Program) have been used to study effects of seismic activities on blue whales (Di Iorio and Clark, 2010).

Real-time Acoustic Tracking Systems (RATS) enable to track large marine predators accurately over small spatial scales to facilitate proximate environmental sampling. These systems consist of an array of four free-floating buoys capable of detecting 36-kHz pings emitted by an animal- borne acoustic transmitter. They are based on the detection and localisation of acoustic transmissions emitted from an animal-borne tag by a free-floating moveable array of buoys

Introduction

(Baumgartner et al., 2008). RATS provide location estimates in real time and does not depend on the target animal to produce sound for tracking. The intension of developing RATS was to monitor foraging behaviour of baleen whales.

As the need grows to conduct long-term studies on cetaceans and to learn about their acoustic behaviour and population dynamics as well as their response towards anthropogenic noise, autonomous recorders with enhanced capabilities are required (Wiggins and Hildebrand, 2007).

VII. Introducing the SoundTrap Here, a novel recording system is introduced, an autonomous GPS-linked receiver array consisting of four acoustic devices (SoundTrap 300 HF, OceanInstruments NZ, AcousticMonitoring Systems). SoundTraps (STs) are compact self-contained underwater sound recorders for ocean acoustic research and offer 20 Hz to 150 kHz bandwidth frequency applications (max. sampling rate: 576 kHz, resolution: 16 bit). The clipping level of the devices, i.e., the level with the maximum voltage that can be supplied when digitising the signal before it is getting distorted, is 171.2 to 173 dB re 1 µPa (peak). The SoundTraps have low self-noise (about 40 dB re 1 µPa2/Hz in the frequency band of 40–200 Hz, measured by placing the loggers in the sound-proof chamber recording for five minutes) ensuring recordings with an extremely good signal-to-noise ratio.

As mentioned above, recording systems are often restricted to limited space for data storage as well as battery capacity. When using hydrophones, they can either be moored on the sea floor or manually operated on-board ships. In both cases, large effort is needed for data acquisition. Depending on which region is used, the mooring system has to be secure ensuring that other ships see the measuring station or that anchors and ropes are robust enough to withstand rough weather conditions and hold the system in position. Therefore, big ships are needed to deploy these oftentimes large mooring systems. These fixed positions can only be used when it is guaranteed that animals are around and passing by the recording system. Otherwise, one has to get to them and start recording in the vicinity, while observing them visually. One way is to get close to the whales and operate the recording system manually from different boats in order to obtain for example source levels.

The system provides a new method in order to either monitor noise events or biological sounds caused by whales. The SoundTrap provides real time data continuously recording underwater

Introduction

sound for weeks, even months depending on the sampling rate. Using several of these sound recorders in combination with a connected GPS device is a powerful tool to estimate source levels of marine mammals as well as sound exposure levels of noise caused by anthropogenic activity. The recording system is easy to handle and does not require well equipped ships and several ship hours associated with costs. Along with the recording system we developed an algorithm for detecting blue whale down sweeps that can be adjusted to other species vocalisations or acoustic events as well. Furthermore, a custom-made localisation algorithm was developed further for the data derived from the new system.

VIII. Methodological development of the analysis of recordings Detection of the vocalisations of a target species can be done either by listening through the recorded sound files or by looking at spectrograms to find vocalisations of the focal species (Clark et al., 1996; Stafford et al., 1999, 2001). Due to the huge amount of data coming along with advanced technologies to store them, methods to automatically detect vocalisations are often required. Many different techniques have been applied for detecting and classifying cetacean sounds, either in the time domain or in a spectrogram. These techniques are matched filters (Stafford et al., 1998), energy summation in a certain band followed by statistical classification (Oswald et al., 2004), image-processing techniques in spectrograms (Gillespie, 2004), spectrogram correlation (Mellinger and Clark, 1997, 2000), neural networks (Kirsebom et al., 2020; Shiu et al., 2020), wavelet-based decomposition (Lopatka et al., 2005), band- limited amplitude in either the time series (Gillespie and Chappell, 2002) or spectrogram (Mellinger et al., 2004a) among others. Whatever method is used, two problems occur. First, determining the vocalisation types and the amount of variability in these vocalisations. Some species, such as fin whales (B. physalus), have highly stereotyped vocalisations. These are amenable to detection using one of the template-matching methods mentioned above. Other species, such as common dolphins (Delphinus delphis), produce highly variable tonal sounds (Oswald et al., 2004). These typically require band-limited energy summation for detection, possibly followed by statistical classification techniques for species classification. Other species produce sounds with intermediate levels of variability that can be detected using neural networks (Kirsebom et al., 2020; Shiu et al., 2020) and filter banks (Urazghildiiev and Clark, 2006). The second challenge is the desired accuracy of detection. In a perfect world, a detection method would find all instances of a certain call type, and nothing more. This ideal is never

Introduction

met, in part because there are inevitably faint calls that are difficult to classify, even by the best human specialists. The issue then becomes one of configuring the detector’s sensitivity, or threshold, to achieve a certain trade-off between missed calls (false negatives) and wrong detections (false positives). For a survey of a relatively rare species such as right whales (Eubalaena spp.), for which one wishes to miss no calls, detection can be configured at a relatively sensitive level so that there are no or few missed calls, but a large number of false detections; the resulting detections can be checked manually to determine which really were from the desired species (Mellinger et al., 2004b; Munger et al., 2005). For a survey of a common species, such as fin whales, for which determining an accurate index of call occurrence is paramount, detection can be configured to be relatively insensitive, so that there are few wrong detections and a very high proportion of correct detections. For a survey using the cue- counting statistical methods discussed below, it may be important to have the number of missed calls be as equal as possible to the number of false detections, so an intermediate sensitivity is used.

IX. Application of the concept of TOAD – Localisation Studies of acoustic sound sources, both natural and anthropogenic, are often restricted by the problem of knowing where the sound source was located. This problem is usually best solved by recording the same source with several synchronised receivers at known locations. By measuring the time-of-arrivals at the various receivers, the sound source location can be determined by hyperbolic mathematics (Spiesberger, 2006). The more receivers are used, the better will the location estimate be (Wahlberg et al., 2001). Traditional localisation systems achieve synchronised recordings by cabling or VHF-linking the receivers to a multitrack tape recorder. This puts severe limits to how many receivers can be used, and how far they can be spaced. There are also signal to noise issues with radio links. By the introduction of GPS, both the receiver locations and their timing can be determined with great precision. A bioacoustic recording system that exploited this was first described by Møhl et al. (2001) for sperm whales (Møhl et al., 2003) The original system was manned, again making fieldwork cumbersome and difficult to handle.

Many studies have addressed the issue of localising cetaceans by means of recording and analysing their underwater vocalisations (Madsen and Wahlberg, 2007; Møhl et al., 2001; Roy et al., 2010; Simard et al., 2004; Simard and Roy, 2008). For this purpose, hydrophone arrays

Introduction

consisting of several hydrophones proved to be most efficient to determine the locations of individuals or groups of animals. Using multiple hydrophones provides the opportunity to have the same signal arriving on all devices with a slight time difference (time-of-arrival difference, TOAD), usually within the order of microseconds for short baseline systems. Given the sound speed profile along the transmission path is known, it is possible to calculate the position of the animal relative to the array. The number of hydrophones necessary for calculating an animal’s location depends on the environment in which the measurement is conducted. For the three dimensional marine environment a minimum of four hydrophones is needed to track an individual marine mammal, such as a blue whale (Madsen and Wahlberg, 2007).

By tracking different individuals, this approach allows collecting useful information about their intraspecific communication. Furthermore, information on the position and 3D-movement of a sound source provides the data necessary to quantify its acoustical properties such as source level, directionality and phonation rate. For marine animals, such information can be used to estimate their abundance and for anthropogenic sound sources (e.g., ships) to describe their trajectory and how much they contribute to the underwater soundscape (Barlow and Taylor, 2005). Furthermore, by revealing the underwater movements of marine mammals this kind of information provides insights into their natural behaviour and allows assessing responses to sound generated by human activity, such as from ships or seismic activities (Hildebrand, 2009).

X. Aims of the thesis The aim of my thesis is to develop methods to empirically measure the transmission loss for different signal types in different habitats as well as develop and test a novel GPS-linked receiver array to make it possible to understand in more detail the acoustic response of marine mammals to anthropogenic noise.

The thesis consists of three chapters:

 Chapter 1 focusses on the assessment of noise caused by the construction of seed mussel collectors in the Wadden Sea. We conducted empirical noise measurements during the construction of seven anchor pipes. We aimed at determining accurate estimates of received levels at a range of distances from the source with a propagation model and evaluate the potential impacts on marine fauna based on sound exposure level thresholds from the literature for marine mammals and fish.

Introduction

 In Chapter 2, the vocal behaviour of Icelandic blue whales was analysed with a GPS- linked receiver array. We recorded ambient noise and calling blue whales in Skjálfandi Bay, Northeast Iceland and determined call characteristics, such as frequency, duration, production rate, and especially source levels. The aim was to investigate whether Icelandic blue whales are quieter than conspecifics in other parts of the world. The motivation behind it was a large difference between the source levels measured for tagged blue whales in the same region and other regions of the world’s oceans.

 Chapter 3 was to clarify if blue whales show a response towards anthropogenic sounds due to a Lombard effect. Little is known about the Lombard effect on blue whales in particular but also on baleen whales in general.

In the conclusion we summarised the most important findings and developments presented in this thesis and offer perspectives for supplementary applications of the system in further acoustic aspects.

Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea.

Published: Reproduced from, Baltzer, J., Maurer, N., Schaffeld, T., Ruser, A., Schnitzler, J.G., Siebert, U. (2020). Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea. Journal of Sea Research, 162, 1-8, https:// doi.org/10.1016/j.seares.2020.101912. with the permission of the Journal of Sea Research.

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

J. Baltzer, N. Maurer, T. Schaffeld, A. Ruser, J.G. Schnitzler, U. Siebert.

Institute for Terrestrial and Aquatic Wildlife Research, University of Veterinary Medicine Hannover, Foundation, Werftstraße 6, 25761 Büsum, Germany

Abstract Anchor pipe vibration embedment operations during the construction of seed mussel collectors were performed in the Wadden Sea, a designated World Heritage Site by UNESCO in 2009. We recorded 200 min of underwater noise during the construction of seven anchor pipes. Underwater noise was recorded simultaneously at three positions with a water depth of 9 m with increasing distance to the construction site to assess the disturbance potential to the marine fauna. The recorded vibration embedment noise was a continuous sound with durations of 2– 55 s, with most energy below 1 kHz and peak frequencies around 900 Hz. Background noise level at a distance of approximately 1 km increased around 13 dB at frequencies between 800 and 1000 Hz. We estimated the sound propagation by a non-linear logarithmic regression by means of the intercept, slope and attenuation factor, which allowed us to evaluate the received sound levels that reach an animal in certain distances from the construction site. The estimated sound exposure level (SEL) of the source was 148.2 dB re 1 μPa2s and the median SEL ranged from 120 to 99 dB re 1 μPa2s at distances between 394 and 2288 m, respectively. Behavioural thresholds for indigenous species of marine mammals in the Wadden Sea as well as representative fish species were used to determine effect radii of vibration embedment noise. Our study showed that the detected anchor pipe vibration embedment noise might exert a behavioural reaction on a local scale. Marine mammals could be affected by the construction operations up to a distance of 375 m and fish up to a distance of 766 m. These zones of responsiveness for vibration embedment operations are relatively small, compared to pile driving, which is regularly used during construction operations. Our study shows that it is important to monitor and assess any kind of noise introduction to verify, whether a sustainable human use with respect to the complied guidelines is ensured without affecting the marine

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

fauna. That is the first step to maintain a good environmental status as implemented in the MSFD.

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

I. Introduction The Wadden Sea is one of the largest intertidal areas in the world, with extensive wetland areas characterised by large intertidal flats stretching from the Netherlands to Denmark (Hild, 1999). Indigenous marine mammal species in the Wadden Sea are the common seal (Phoca vitulina), the grey seal (Halichoerus grypus) and the harbour porpoise (Phocoena phocoena) (Jensen et al., 2017). The Dutch and German parts of the Wadden Sea Conservation Area have been designated as a World Heritage Site by UNESCO in June 2009, recognising the global importance of the Wadden Sea as a nature area (CWSS, 2017). The Wadden Sea region is an area where people work, but also come for leisure or recreational activities. About 3.7 million people live along the Wadden Sea coast interacting with the landscape, plants and wildlife. The Trilateral Wadden Sea Plan (2010) concedes that sustainable human use has to be continuously balanced in a harmonious relationship between the needs of society and ecological integrity (CWSS, 2010).

Activities at sea increased extensively over the last decades, among which shipping, fisheries, tourism, military activities, dredging and energy exploitation are the most concerning activities (CWSS, 2017). These activities contribute a lot to ambient underwater noise (Rako-Gospić and Picciulin, 2019). The introduction of noise into the oceans is getting more and more in focus when it comes to impact assessment of anthropogenic activities on the environment.

The North- and Baltic Seas are classified as two areas with excessive human exploitation (Halpern et al., 2015). In offshore areas of the North Sea many wind farms have already been constructed and a lot more are planned. In comparison, in shallow coastal waters noise a chronic and constant pollution due to urbanisation, shipping and expanding tourism. Along with those activities, studies have been conducted to figure out, to which extent marine life is affected and how severe potential effects might be.

Anthropogenic noise can cause behavioural responses of harbour porpoises (Kastelein et al., 2013a, 2013c) or lead to changes in spatial distribution (Brandt et al., 2016; Carstensen et al., 2006; Dähne et al., 2013; Scheidat et al., 2011; Teilmann and Carstensen, 2012; Tougaard et al., 2009). Even hearing impairment resulting in a temporary threshold shift (TTS) has been documented by Lucke et al. (2009) or Kastelein et al. (2016) after the exposure to impulsive noise, such as pile driving strikes. Other responses of porpoises can be stress or the interruption of their natural behaviour, such as feeding (Wisniewska et al., 2018). Behavioural responses

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

towards pile driving of a Dutch wind farm were documented for both, harbour (Heinis, 2013) and grey seals (Aarts et al.,2018). Russell et al. (2016) predicted a displacement of harbour seals in response to pile driving.

There are also studies of anthropogenic noise affecting different fish species by deteriorating body condition (Bruintjes et al., 2016; Buscaino et al., 2010; Casper et al., 2013a, 2013b), decreasing catch rates (Purser and Radford, 2011), inhibiting anti-predator defence (Simpson et al., 2016; Spiga et al., 2017; Voellmy et al., 2014) or changing school coordination (Hawkins et al., 2014; Herbert-Read et al., 2017) and cohesion (Kastelein et al., 2017; Neo et al., 2014). Reproduction of fish could also be affected, if anthropogenic noise causes masking and therefore disrupts the intra-specific communication. This masking can result in an increase in amplitude of communication signals to compensate for a decreased signal-to-noise ratio, also known as the Lombard effect (Holt and Johnston, 2014; Ladich, 2019; Luczkovich et al., 2016).

To assess and quantify environmental effects of anthropogenic noise, it is crucial to estimate the levels of sound generated by the sender (source level) and the rate at which the sound decays as it propagates to the receiver (transmission loss) (Rako-Gospić and Picciulin, 2019). All those studies show that underwater noise has a huge effect on the marine environment and is therefore of international concern. Thus, it is important to monitor and assess the introduction of noise. This is crucial to develop measures to keep the noise levels low in order to achieve a Good Environmental Status (GES) as implemented in the Marine Strategy Framework Directive (MSFD, Descriptor 11, European Union, 2008).

In addition to the three occurring marine mammal species, the shallow Wadden Sea is an important area for many fish species, which rely on the coastal area for at least one part of their life cycles (Tulp et al., 2017). Numerous species of marine fish (flatfish, other benthic and pelagic fish species) reach the Wadden Sea as post-larvae and spend their juvenile phase there benefitting from the high food availability and shelter from predators (marine juveniles e.g. plaice (Pleuronectes platessa), sole (Solea solea), dab (Limanda limanda), whiting (Merlangus merlangus), cod (Gadus morhua), sea bass (Dicentrarchus labrax) and herring (Clupea harengus) (Elliott et al., 2007; Van der Veer et al., 2000). Other species cross the region on their way to either marine or fresh water spawning sites. These species can be diadromous, such as eel (Anguilla anguilla), smelt (Osmerus eperlanus), twaite shad (Alosa fallax), river lamprey

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

(Lampetra fluviatilis) and sea lamprey (Petromyzon marinus). Others visit the area during certain times of the year like marine seasonal migrants, such as anchovy (Engraulis encrasicolus) and pilchard (Sardina pilchardus)) or only sporadically like marine adventitious species, such as mullets (Mugilidae) and sprat (Sprattus sprattus) (Elliott et al., 2007). Apart from the temporary visitors, the Wadden Sea is also inhabited by resident species that spend (almost) their entire life in the Wadden Sea (e.g. flounder (Platichthys flesus), eelpout (Zoacres viviparous), bullrout (Myoxocephalus scorpius), fivebeard rockling (Ciliata mustela), hooknose (Agonus cataphractus) and pipefishes (Syngnathus sp.)) (Tulp et al., 2017).

The Wadden Sea represents a rich food source for humans, offering large amounts of natural grown mussel beds. Nowadays, it is more profitable to rather cultivate mussels in the Wadden Sea than harvesting from natural sites. This is practised with blue mussel (Mytilus edulis) cultures by placing wild-caught mussel seeds (spat) at specific sites, i.e. on-bottom culture plots, where survival and growth is enhanced. As an alternative to catching the spat, artificial seed collection technologies, such as seed mussel Collectors (SMCs) are used. SMCs are fixed or hanging net constructions where mussel larvae can settle on and develop into young mussels that will be harvested. Therefore, anchor pipes are fixed to the ground, either through pile driving or drilling, which is accompanied by the introduction of high noise levels into the water (Brandt et al., 2018). An alternative form of SMC construction is to vibrate anchor pipes into the seabed.

The aim of this study was to assess the effects of underwater noise from the construction of seed mussel collectors on the marine mammals and fish in the Wadden Sea. Therefore, we conducted empirical noise measurements during the construction of seven anchor pipes. We aimed to determine accurate estimates of received levels at a range of distances from the source with a propagation model and evaluate the potential impacts on marine fauna based on sound exposure level thresholds from the literature for marine mammals and fish.

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

II. Material and methods A. Study area and anchor pipes vibration embedment operations Fieldwork was conducted in the German Wadden Sea in the tidal creek ‘Sueder Piep’ off the coast of Schleswig-Holstein (North Sea) at 25th of April in 2019 (Fig. 1). Underwater recordings during the construction of seven anchor pipes were conducted during 200 minutes simultaneously at three measuring positions in order to determine underwater noise attenuation by distance. Therefore, three recording buoys consisting of a PVC rod, a floating body and an 18 mm nylon rope were placed at distances between 400 and 2300 m from the construction site. Additionally, a small anchor stone was attached to the system to keep the buoy in position. On each buoy we fixed a SoundTrap (ST300 HF, Ocean Instruments NZ, Acoustic Monitoring Systems), a compact self-contained underwater sound recorder for ocean acoustic research. The water depth at the measuring positions was 9 m and all SoundTraps were deployed at a depth of 4.5 m. The devices were linked to GPS (Global Positioning System)-receivers placed at the PVC rod of the buoy via a cable to associate acoustic recordings (wave-files) to the specific position of the buoy and to synchronise recordings and distance to the construction activities. A sample rate of 576 kHz was used to record the vibration sounds.

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

Fig. 1: Map of the North Sea coastline including the entire Wadden Sea area with an enlarged section of the area of investigation (square therein).Recording buoys were bottom mounted at three measuring positions (red, green and yellow dots) around the construction site (blue dots) for culturing areas for seed mussels at distances between 400 and 2300 m to anchor pipe vibration embedment sites. The water depth at the measuring positions was 9 m and all recorders were deployed at a depth of 4.5 m. Grey dots show positions of additional piles without available underwater noise recordings, because they were already vibrated into the seabed.

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

B. Sound propagation modelling and frequency analysis Underwater recordings were analysed to determine the sound propagation in the construction area. The calculated sound propagation is essential to determine sound exposure levels (SEL) within the acoustic field of all anchor pipe vibration embedment events the surrounding marine fauna might be exposed to. To calculate the SEL, underwater recordings were loaded with the R package ‘tuneR’ (Ligges et al., 2016) and high pass filtered (Butterworth) with a 1st order filter to 100 Hz, to eliminate a possible offset with the R package ‘seewave’ (Sueur et al., 2008). Anchor pipe vibration embedment in the recordings was detected by visually screening the spectrograms (Fast Fourier Transform: 16384, Hanning window, 50% overlap.) and listening to the underwater recordings in Adobe Audition 3.0. The SEL was calculated according to the ISO 1996-1:2016 (International Organization for Standardization, 2016) as:

E SEL = 10 × log10 , E0

푇 where 퐸 = 푝(푡)²푑푡, for a 1 s time window. ∫0

The sound propagation was based on the median SEL50 and the SEL05, defined as the noise levels exceeded by 50% and 5% of all values, respectively. The sound propagation in relation to the distances to the anchor pipe vibration embedment site was estimated by a non-linear regression. The intercept and the logarithmic regression factor were estimated by a non-linear least squares (nls) approach, using the nls function in R (R Core Team, 2017). We further estimated an attenuation factor A accounting for absorption and further complicating factors, such as multipath propagation, refraction, diffraction and scattering of sound due to suspended particles in the water column (Urick, 1983) within the nls approach.

Background noise was determined for a 30 s fraction of underwater sound recordings prior to construction activities. Third octave spectra were calculated for centre frequencies ranging from 62.5 Hz to 128 kHz for each 100 Hz high pass filtered 1 s window within the fraction, for each anchor pipe at each measuring position. In total, background noise recordings prior to seven anchor pipe vibration events at three measuring positions were analysed, building a data base of eleven 30 s windows.

All analyses were performed and figures created using R (R Core Team, 2017).

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

III. Results A. Sound propagation in the study area Underwater recordings during the construction of seven anchor pipes were conducted during 200 min in order to determine underwater noise. The detected vibration embedment noise can be defined as continuous sound with durations from 2 to 55 s. Energy spectra show that most of the energy was found below 1 kHz with a peak around 900 Hz (see Fig. 2).

Fig. 2: Spectrogram of the first detected anchor pipe vibration embedment noise with a duration of 52 s recorded at around 1 km distance to the pile.Fast Fourier Transform: 16384,

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

Hanning window, 50% overlap. The dB scale is colour-coded with red as the highest and blue the lowest intensity.

Fig. 3 shows the median third octave sound exposure level (SEL in dB re 1 µPa2s) of the vibration embedment noise (red) seen in the spectrogram (Fig. 2) in relation to background noise (green). In the frequency range from 630 Hz to 25 kHz the noise generated by anchor pipe vibration embedment operations exceeded the background noise level, particularly at 800– 1000 Hz by around 13 dB (Fig. 3).

Fig. 3: Median third octave sound exposure level (SEL) for the first detected anchor pipe vibration embedment noise event (red dots) relative to the background noise (green triangles).The dashed lines display the 0.25 (lower) and 0.75 (upper) quantiles for vibration and background noise, respectively. The SEL of the vibration noise was calculated for each third octave and each second in a 52 s time window according to Fig. 2. The background noise SEL was calculated the same way for the same time window (52 s) right before the construction work began.

The sound propagation was estimated by a non-linear logarithmic regression, estimating the intercept, slope and attenuation factor based on the determined SEL per second. The best fit was determined at:

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

Received level: 148.2 − 10.05 × 푙표푔10(푅) + 0.0067(푅), where R accounts for the distance.

Thus, the estimated source level resulting from the intercept was 148.2 dB re 1 µPa2s.

The transmission loss is shown in Fig. 4.

Fig. 4: Calculated transmission loss for the median sound exposure level (SEL50, orange line) and the sound exposure level exceeded for 5% (SEL05, red line) of all detected vibration embedment noise events (black circles), respectively as well as for the background noise (grey dashed line).

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

2 The received median SEL50 ranged from 120 to 99 dB re 1 µPa s and the received median 2 SEL05 ranged from 125 to 103 dB re 1 µPa s at distances between 394 and 2288 m, respectively (Fig. 4).

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

IV. Discussion Our recordings quantified the introduced underwater noise into the Wadden Sea marine environment by the construction of the seed mussel collectors. The background noise was already slightly elevated due to the presence of three ships involved with construction activities (such as vibration embedment, painting and construction supervision) or passing ships due to the construction site being close to a fairway. The source level of anchor pipe vibration embedment operations was determined at 148.2 dB re 1 µPa2s and can be considered as relatively low compared to pile driving of small diameter (0.61 or 0.71 m) piles with a source SEL of 192 dB re 1 μPa2s at 1 m investigated by Leunisssen and Dawson (2018). Nevertheless, the signal exceeds background noise levels out to around 3000 m from the construction site and is therefore audible for the marine fauna. The construction for each anchor pile took 2 and 55 s and was interrupted by multiple breaks, most likely to readjust the anchor pile and vibration embedment equipment. The vibration embedment operations produced continuous noise, which was also stated by Branstetter et al. (2018) for vibratory pile driving. The vibration noise contained most of the energy below 1 kHz and did not contain distinct high-energy incidents in the construction periods.

Based on the calculated transmission loss, we were able to evaluate the received sound levels that reached an animal at a certain distance from the construction site. In terms of frequency content, the noise generated by anchor pipe vibration embedment operations was comparable to impulsive noise sources like airguns (Lucke et al., 2009) or pile driving noise (Dähne et al., 2017). These impulsive noise events lose their characteristics, when recorded at larger distances. The repetition of these impulsive noise events may become diffuse with distance and reverberation and become indistinguishable from continuous noise (van der Graaf et al., 2012).

A. Potential effect on marine mammals Noise exposure can induce hearing shifts, which can occur permanently or temporarily. In comparison to the estimated source levels of the construction work in this study, much higher source levels are needed to induce hearing impairment (Lucke et al., 2009). Our focus is on behavioural reaction thresholds that occur at larger distances from the source where received levels are lower compared to those causing a TTS. Thus, we can proceed from this assumption. Vessel noise, as a continuous noise source, contains also most of the energy below 1 kHz, but

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

can be very variable, depending on vessel type, propeller and speed (Kipple and Gabriele, 2003, 2004; Putland et al., 2017) and include also high frequency content to which animals might respond (Dyndo et al., 2015). Therefore, our impact assessment of the recorded vibration embedment noise was based on thresholds derived from studies on impulsive noise (Aarts et al., 2018; Kastelein et al., 2016; Lucke et al., 2009). Thresholds for behavioural reactions could be found in the literature for indigenous species of marine mammals in the Wadden Sea (harbour porpoise, harbour seal and grey seal) as well as representative fish species (herring, sole, cod). Based on the data from the literature and the calculated transmission loss we were able to determine effect radii for marine mammals inhabiting the German Wadden Sea (Fig. 5).

The figure displays effect radii based on the transmission loss for the median SEL50 (dashed line) and the SEL05 (solid line), defined as the noise levels exceeded by 50% and 5% of all detected vibration embedment noise events, respectively.

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

Fig. 5: Effect radius to the vibration embedment location for different fish and marine mammal species living in the German Wadden Sea: Cod (yellow), sole (blue), seal (red) and harbour porpoise (turquoise). The effect radii are calculated for the sound exposure level (SEL) of the vibration noise exceeded for 5% and for 50%.

Due to the relatively low determined source level of 148.2 dB re 1 µPa2s a TTS from the vibration embedment noise is unlikely. Although it could be shown for harbour porpoises, that multiple exposure to pile driving strikes with single sound exposure levels of 145 dB re 1 µPa2s has the potential to induce a TTS, if a cumulative energy of 175 dB re 1 µPa2s is reached (Kastelein et al., 2016). However, the short duration of construction work and the breaks in between are unlikely to induce hearing shifts, even for multiple exposures.

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

Behavioural reactions might be induced more likely. Thus, Lucke et al. (2009) found that a single harbour porpoise consistently showed aversive behavioural reactions to pulsed sound from an airgun at a received SEL of 145 dB re 1 μPa2s. This is the same threshold at which evasive actions had been predicted for harbour seals (Heinis, 2013), based on the hearing ability and response of captive animals (Kastelein et al., 2013b). Our determined SEL is also well in the range of the findings of Russell et al. (2016). They predicted a displacement of harbour seals as a response to pile driving at SELs between 142 and 151 dB re 1 µPa2s. A SEL of 2 145 dB re 1 μPa s can be reached at a distance of 2 m (up to 8 m for the SEL05) from the construction site, resulting from the estimated intercept, slope and attenuation factor given by the non-linear logarithmic regression.

A study on grey seals indicated that a behavioural response to pile driving occurred in response to a SEL of 133 dB re 1 μPa2s. At this threshold a change in descent speed could be observed. Moreover, exceeding a SEL of ~137 dB re 1 µPa2s leads to a significant behavioural response in any of the dive or movement variables (Aarts et al., 2018). This threshold could be reached around 31 m (up to 132 m for the SEL05) from the SMC construction site.

Harbour porpoises react to underwater noise in the most sensitive way. The lowest threshold has been found by Kastelein et al. (2013a), who showed that above a received broadband sound exposure level of 127 dB re 1 μPa2s the respiration rate of harbour porpoises increased in response to the pile driving sounds. This would correspond to a received SEL at a distance of

109 m (up to 375 m for the SEL05) from the construction site.

Estimated zones of responsiveness for marine mammals differ between harbour seals, grey seals and harbour porpoises for anchor pipe vibration embedment operations. This might be related to the sensitiveness of the measurement of the reaction (respiration rate vs. aversive behaviour or change in dive speed) or variability in the response of animals in the wild and in human care. The expected hazard zones, extending from 2 to 375 m from the anchor pipe vibration embedment operations can be considered as negligible for the indigenous species of marine mammals in the Wadden Sea (Kastak et al., 2005, 2007).

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

B. Potential effect on marine fish Although sound is of key importance for almost all vital functions among most marine fauna, sound perception has been studied only for a small percentage of fish species (Ladich and Fay, 2013). To date, around 100 fish and invertebrate species have been shown to be impacted by anthropogenic noise (Slabbekoorn et al., 2010), resulting in decreases in growth (Anderson et al., 2011), immune competency (Celi et al., 2015), productivity (Lagardère, 1982; Stanley et al., 2017), body condition (Bruintjes et al., 2016; Buscaino et al., 2010; Casper et al., 2013a, 2013b), catch rates (Purser and Radford, 2011), anti-predator defence (Spiga et al., 2017), school coordination (Hawkins et al., 2014; Herbert-Read et al., 2017) and cohesion (Kastelein et al., 2017; Neo et al., 2014).

Injuries and lethal effects could be shown for multiple fish species, such as spot, pinfish, lake sturgeon, Nile tilapia, hogchoker and Chinook salmon after exposure to impulsive sounds in the close vicinity of offshore wind farm construction site (Govoni et al., 2008; Halvorsen et al., 2012a, 2012b) and were more likely to affect the swim bladder and surrounding organs than the inner ears (Casper et al., 2013c). Such impairment of hearing or other organs is unlikely in the framework of anchor pipes vibration embedment operations.

Behavioural reactions might be more likely to be evoked. Two fish species displayed significant movement response to a pile driving stimulus at relatively low received sound pressure levels (SPL; sole: 144–156 dB re 1 µPa peak; cod: 140–161 dB re 1 µPa peak, particle motion between 6.51×10-3 and 8.62×10-4 ms-2 peak) (Mueller-Blenkle et al., 2010). Sole showed a significant increase in swimming speed during the playback period compared to before and after the playback. Cod exhibited a similar reaction, yet results were not significant. Cod showed a significant freezing response, i.e. decreasing the swimming speed, at the onset and cessation of playback. There were indications of directional movements away from the sound source in both species. Further, the results showed a high variability in behavioural reactions across individuals and a decrease of responses with multiple exposures (Mueller-Blenkle et al., 2010).

The study of Mueller-Blenkle et al. (2010) did not identify a single threshold, but a range over which behavioural responses occurred for sole and cod. The behavioural reaction threshold range for sole was found to be 126–142 dB re 1 μPa2s. Since this range is given as a SPL, it cannot be referenced to our data directly. The corresponding SEL range was determined by a

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

best linear fit between the recorded SPL vs SEL values. The reported range corresponds to a received SEL at a maximum distance of 132 m (up to 430 m for the SEL05) from the construction site. The behavioural reaction threshold range for cod was 121–149 dB re 1 μPa2s, which corresponds to a received SEL at a maximum distance of 314 m (up to 766 m for the

SEL05) from the construction site.

Our results show that these two fish species might be affected by the noise generated from the anchor pipe vibration embedment operations to a distance of 132 to 766 m from the construction site. Different fish species react very differently to noise and the reaction threshold must not inevitably follow hearing curves. Moreover reactions to anthropogenic sounds depend on the context, meaning for example animal’s age, school size, individual body size, temperature, location or physiological state, and generalisations should therefore be made with great caution (Kastelein et al., 2008).

An increasing amount of anthropogenic noise in the ocean can mask biologically relevant acoustic signals possibly leading to complete or partial loss or misinterpretation of signals. Masking is therefore considered as one of the main effects of noise pollution on marine animals and might alter acoustic communication, impact predator avoidance and prey detection and as a result might have a major effect on whole ecosystems (Slabbekoorn et al., 2010). Acoustic signals of Wadden Sea fish that might be affected by masking can be found in herring (Wahlberg and Westerberg, 2003), sand goby (Pomatoschistus minutus) (Lindström and Lugli, 2000), Atlantic cod (Hawkins and Rasmussen, 1978; Rowe and Hutchings, 2005), tub gurnard (Chelidonichthys lucerna, Trigla lucerna) (Amorim, 2006), pollack (Pollachius pollachius) and tadpole fish (Raniceps raninus) (Amorim, 2006; Hawkins and Rasmussen, 1978) (see Tab. 1). The sound communication in fish in the Wadden Sea might be affected in terms of calling activity (de Jong et al., 2016, 2018; Van Oosterom et al., 2016; Putland et al., 2017), alteration of sound characteristics (Lombard effect) (Holt and Johnston, 2014; Ladich, 2019; Luczkovich et al., 2016) and reduction in detection distance (Stanley et al., 2017).

C. Marine Strategy Framework Directive Germany, as European member state, is obliged to assess the current situation in their waters and to monitor changes in the future to implement the European Marine Strategy Framework

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

Directive (MSFD, Descriptor 11, European Union, 2008). In the course of this directive underwater noise is embedded as one of the descriptors (Descriptor 11) and is entitled ‘Introduction of energy, including underwater noise, is at levels that do not adversely affect the marine environment’. There are few studies that provide useful methods on how to assess the impact of underwater noise (Boyd et al., 2008; Faulkner et al., 2018). Compared to the study of Codarin and Picciulin (2015), we also investigated a coastal area exposed to anthropogenic noise. However, their study focused more on ship noise within two third-octave bands (63 Hz and 125 Hz), relevant for chronic exposure to low frequency ambient noise (van der Graaf et al., 2012). We on the other hand, considered the noise exposure within several third-octave bands to figure out in which frequency the sound exposure level was highest. The investigated sound source was an innovative method of fixing anchor pipes at the seabed, therefore the kind of noise had to be identified. Our study can be seen as the a first step of risk assessment, the hazard identification according to the research strategy of Boyd et al. (2008). We however went one step further and presented potential effects on different species in theory without measuring the exposure on the animals.

D. Conclusion Underwater noise is the most widespread and pervasive kind of anthropogenic energy which is introduced in the marine environment. Our study showed that the detected anchor pipe vibration embedment noise might induce a behavioural reaction in indigenous species from the Wadden Sea on a local scale. The construction of seed mussel collectors was done on certain conditions, for instance the anchor pipes for the seed mussel collectors were vibrated into the seabed as an alternative to pile driving. It can be concluded that the method of vibrating is much less harmful to the marine fauna than pile driving, at least for the species considered in our study. Our study gives an example that a sustainable human use in respect to the complied guidelines, can lead to a harmonious relationship between the needs of society and ecological integrity as conceded by the Trilateral Wadden Sea Plan (2010).

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

V. Acknowledgement We thank the mussel fishers for providing the information needed to do proper measurements. We also thank the WSA (Waterways and Shipping Office) Tönning for the permission to deploy our measuring devices and our colleague Patrick Stührk for creating the mooring system as well as for a safe deployment and retrieval of our measuring devices.

This study has been supported by the Ministry of Energy, Agriculture, the Environment and Rural Areas of Schleswig-Holstein / Schleswig-Holstein Agency for Coastal Defence, National Park and Marine Conservation (LKN.SH) Germany.

Chapter 1: Effect ranges of underwater noise from anchor vibration operations in the Wadden Sea

VI. Appendix Tab. 1: Acoustic signals of the Wadden Sea fish.

Blue whales off Iceland: Silent giants in a Northern feeding ground.

Manuscript: Baltzer, J., Iwata, T., Akamatsu, T., Sato, K., Aoki, K., Lucke, K., Rasmussen, M.H., Wahlberg, M., Schnitzler, J.G., Siebert, U. (submitted). Blue whales off Iceland: Silent giants in a Northern feeding ground.

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

J. Baltzer1, T. Iwata2,3, T. Akamatsu2, K. Sato3, K. Aoki3, K. Lucke4, M.-H. Rasmussen5, Magnus Wahlberg6, J.G. Schnitzler1, U. Siebert1

1Institute for Terrestrial and Aquatic Wildlife Research, University of Veterinary Medicine Hannover, Foundation, Germany

2Ocean Policy Research Institute (OPRI), Sasakawa Peace Foundation, Japan

3Atmosphere and Ocean Research Institute, The University of Tokyo, Japan

4Jasco Applied Sciences, Australia

5The University of Iceland’s Research Centre Húsavík, Iceland

6Department of Biology, University of Southern Denmark, Denmark

Abstract Whereas the acoustic behaviour of some blue whale (Balaenoptera musculus) populations is well-known, relatively little is known about their vocalisations in North Atlantic waters. We conducted fieldwork on blue whale vocalisation in Skjálfandi Bay, Northeast Iceland with two different passive acoustic monitoring (PAM) methods. The animals were tagged with animal- borne sound recorders and behavioural data loggers in order to achieve information on diving and vocalisation behaviour of individuals. A GPS-linked receiver array was used to record vocalising blue whales in the bay. We estimated a source level of 172±9 dB re 1 µPa m for 13 blue whale-down sweeps. The frequency of these specific calls ranged from maximum 107 Hz to minimum 28 Hz. Blue whales produced down sweeps on average 2.0 times per hour and rather infrequently during feeding events. The ambient noise level was 88 dB re 1 µPa within the frequency bandwidth of the calls. The source level of the recorded down sweeps was 7 to 17 dB lower than the ones reported in other studies from different regions of the world. This suggests that blue whales are able to vary their call intensity in order to adapt to their local environment.

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

Author contributions Participated in the fieldwork: J.B., T.I., T.A., K.L., M.-H. R., M.W. and J.S. Analysed array data: J.B., M.W. and J.S. Analysed tag data: T.I., T.A., K.S. and K.A. Wrote the paper: J.B. and T.I. Revised the paper: T.A., K.S., K.A., K.L., M.-H. R., M.W., J.S. and U.S.

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

I. Introduction Acoustic communication plays a major role in the life of many animals, for attracting mating partners, deterring rivals or to signal the presence of feeding areas or predators (Bradbury and Vehrencamp, 2011). Animals are known to adapt their call intensity to prevailing ambient noise conditions to ensure that their acoustic signals are perceived by conspecifics. The fact that animals increase their source level when ambient noise increases is known as the Lombard effect, first described in humans (Lombard, 1911) and later also in birds and mammals (Kragh et al., 2019). Several studies have documented this phenomenon in marine mammals. Toothed whales, such as killer whales (Orcinus orca), beluga whales (Delphinapterus leucas) and bottlenose dolphins (Tursiops truncatus) are known to adjust their signal intensity as well as other signal characteristics (e.g. duration) when ambient noise is elevated (Foote et al., 2004; Holt et al., 2009, 2011; Kragh et al., 2019; Scheifele et al., 2005). There is also evidence that baleen whales, like humpback (Megaptera novaeangliae), minke (Balaenoptera acutorostrata), bowhead (Balaena mysticetus) and North Atlantic right (Eubalaena glacialis) whales exhibit a Lombard effect towards natural and anthropogenic sound sources (Dunlop et al., 2014; Guazzo et al., 2020; Helble et al., 2020; Parks et al., 2007, 2009; Thode et al., 2020).

For blue whales Balaenoptera musculus (Linnaeus, 1758), the largest of the baleen whales and the largest animal on the planet, there are only a few studies on their response towards anthropogenic noise or increased ambient noise levels in general. Blue whales increase their call rate when exposed to seismic exploration (Di Iorio and Clark, 2010). They have also been observed to stop call production, while mid-frequency sonar was active, and increase the intensity of their calls in the presence of ships (Melcón et al., 2012). Blue whales produce a variety of sounds. Three major types of calls have been described: 1) low-frequency pulsed A and tonal B calls. These calls can occur together as rhythmic repetitive song sequences (McDonald et al., 2006) or as intermittent, singular calls (Oleson et al., 2007a). Blue whales are known to produce different songs in different areas, with the North Atlantic population having its own distinct song (McDonald et al., 2006; Mellinger and Clark, 2003). They also emit 2) highly variable amplitude or frequency-modulated calls and 3) D calls, also known as arch sound (Mellinger and Clark, 2003) and down sweep. Some of these vocalisations contain the lowest frequencies (down to 9 Hz) recorded from any marine mammal (Cummings and Thompson, 1971; Mellinger and Clark, 2003). The function of their acoustic signals is not really

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

clear, but singular A and B calls were recorded both on feeding grounds and while travelling (Oleson et al., 2007a), and also on presumably breeding grounds (Cummings and Thompson, 1971; McDonald et al., 2001). A and B calls as song, however are believed to play a role in reproduction and D calls have been documented among feeding blue whales (Oleson et al., 2007a).

Source levels (defined as the sound intensity 1 m in front of the sound source related to the source power (ISO 18406, 2017)) of blue whale calls have been estimated from feeding grounds in different parts of the world. Source levels ranged from 179 to 189 dB re 1 µPa m (Cummings and Thompson, 1971; McDonald et al., 2001; Samaran et al., 2010; Širović et al., 2007; Thode et al., 2000). From Icelandic blue whales, the measured source levels are lower, 159 to 169 dB re 1 µPa m (Akamatsu et al., 2014). Currently, it is not known whether the low source levels are due to specific adaptations in Icelandic blue whales or due to methodological differences as compared to other studies, since source levels of Icelandic blue whales were calculated from animal-borne acoustic tags rather than by free-field recordings.

To investigate whether Icelandic blue whales are indeed quieter than conspecifics in other parts of the world, we recorded ambient noise and calling blue whales in Skjálfandi Bay, Northeast Iceland (Fig. 6). In our study, we used two different approaches to get data on blue whale call behaviour via passive acoustic monitoring (PAM). Animal-borne sound recorders were attached to blue whales, and a GPS-linked receiver array was used to record ambient noise and calling blue whales in Skjálfandi Bay. Such an array has the advantage that we can localise the animals and determine their position in order to obtain source levels.

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

II. Material and methods A. Study area and equipment The fieldwork was conducted in Skjálfandi Bay in Northeast Iceland (Fig. 6) in June 2017 and 2018. The bay is approximately 30 km wide across its mouth and 20 km long and the water depth is up to 200 m.

Fig. 6: Area of investigation, Skjálfandi Bay in Northeast Iceland.

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

We used an autonomous GPS-linked receiver array consisting of four acoustic devices (SoundTrap 300 HF, OceanInstruments NZ, AcousticMonitoring Systems). SoundTraps (STs) are compact self-contained underwater sound recorders for ocean acoustic research and offer 20 Hz to 150 kHz bandwidth frequency applications (max. sampling rate: 576 kHz, resolution: 16 bit). The sampling rate for the recordings was set to 48 kHz. The clipping level of the devices is 171.2 to 173 dB re 1 µPa peak. The units have low self-noise (about 40 dB re 1 µPa2/Hz in the frequency band measured by placing the loggers in the sound-proof chamber recording for five minutes) ensuring recordings with a sufficient signal-to-noise ratio for our purposes. The sound recorders were linked to GPS-receivers attached to buoys at the water surface and connected via a cable lowered to a water depth of approx. 20 m.

All recordings were made in Sea State 3 or less. During field trials, the buoys were deployed in line 500 m up to 1 km apart from each other in East-West direction. Recordings were mostly conducted during night due to calm weather and absent boat traffic.

An Automatic Underwater Recording Systems-micro (AUSOMS-micro, Aqua Sound Co., Ltd., Kyoto, Japan) was used for tagging. The AUSOMS-micro is a cylindrical pressure-resistant recorder including a monaural hydrophone (sensitivity of -210 dB re V/1 µPa), with a length of 100 mm and a diameter of 22.5 mm. The AUSOMS-micro is capable of providing continuous, non-compressed recordings (sampling rate: 48 kHz, resolution 12 bit) of up to 48 h. The memory capacity of the AUSOMS-micro is 16 GB. The pressure-resistant housing is made of carbon fibre, which secures the electronics inside the device at depths of up to 300 m. The AUSOMS-micro was assembled with a suction cup to tag on whales, flotation material (copolymer, up to 400 m, Nichiyu Giken Co., Ltd., Tokyo, Japan) and a VHF transmitter (MM130, Advanced Telemetry Systems, Isanti, MN) to enable retrieval after detachment. Two types of behavioural data loggers, installed inside the flotation material, were used. A DST- compass magnetic tag (Star-Oddi, Garðabær, Iceland) which records the magnetic-field, depth, temperature and tilt was used in 2017 season. The resolution of the magnetic field, depth and temperature was 100 nT, 0.03 m and 0.032°C, respectively. We used 1 s or 4 s of sampling intervals for all parameters. An ORI400-D3GT tag (Little Leonardo, Tokyo, Japan) recording three-axis acceleration, depth and temperature was used in 2018 season. The sampling interval and acceleration resolution were 0.05 s and 0.1 m s-2, respectively. The sampling interval and resolution of depth and temperature were 1 s for both values, and 0.1 m and 0.1°C, respectively.

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

We used 0.05 s and 1 s as sampling intervals for acceleration, depth and temperature, respectively. In this study, data of magnetic field and acceleration were not used in data analysis, because the common behavioural parameter of the DST-compass magnetic and the ORI400-D3GT tag is the depth.

We used a Zodiac inflatable boat to approach whales for tagging. An experienced operator maneuvered the boat close to the right side of a target whale after the first blow was observed. The boat was maintained parallel to the whale and within approximately 6 m of the dorsal ridge. An 8-m carbon fibre pole (Villum Pole, Mikkel Villum Jensen, http://www.mikkelvillum.com, Horve, Denmark) was used for the tagging attempts. Tagged animals were photographed for individual identification when weather and other conditions permitted. After detachment, the tag floated vertically on the sea surface with the VHF antenna in the air. The detection range of the VHF signal was approximately 12 km from the Zodiac using a AR8200 receiver (AOR Co., Ltd., Tokyo, Japan) and a three-element Yagi antenna (RY-144M3, Radix, Chiba, Japan). For the safety of the operation, we used three boats simultaneously. One boat carried out the tagging, a second boat deployed and retrieved the GPS-linked receiver array and obtained photographic identification of each whale and a third boat was operating the loudspeaker for transmitting artificial upsweep signals.

B. Analysis of the call characteristics Detection and localisation of blue whale calls

The recordings were first screened aurally and visually in Adobe Audition 3.0 (Adobe Systems Inc., USA). Therefore, we adjusted a Hann window with 80% overlap and a window point size of 8192. Afterwards underwater sound recordings were loaded with the R package ‘tuneR’ (Ligges et al., 2018). For detecting blue whale down sweeps the R package ‘monitoR’ was used (Hafner and Katz, 2018). Feeding by a call template, the software searches similar calls in the sound files of all four receivers. The search is done using spectrogram cross-correlation (Mellinger and Clark, 1997). Once all calls fitting the template were detected, the timing of each call was used to match each call received by all four receivers using a nearest neighbour algorithm. Having the times for the calls received on all receivers, we obtained the GPS time

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

(in UTC) of the detected call in the original sound file. In this way, all recordings were synchronised by GPS.

After the detected calls were synchronised, they were cut 2-3 s before and after the synchronisation peak. The calls were Butterworth filtered individually (1st order bandpass filter with pass band minimum 20 to maximum 120 Hz) using the R package ‘seewave’ (Sueur et al., 2008) and subsequently cross-correlated with the one on the reference receiver, respectively. Thus, we got time shifts to the recorded call on the reference, time-of-arrival differences (TOADs). By means of the TOADs and the given speed of sound in the area of investigation we could obtain hyperbolic intersects resulting in possible positions of the calling blue whale (Møhl et al., 2001; Wahlberg et al., 2001).

Estimating the source levels

In order to determine the source level (SL) of a blue whale call, the received sound pressure level (SPL) was calculated in root-mean-square (rms) units. When calculating SPL, the length of the signal has to be defined carefully because the rms level is reduced if the averaging window surpasses the length of the signal being averaged (Madsen, 2005). The duration of the signal in our study was defined as including 90% of the energy of the entire signal waveform (Madsen, 2005; Southall et al., 2007), i.e. from the time where the signal includes 5% of the total energy to the time at which 95% of the energy was reached. The SPLrms in a time window incorporating 90% of the signal’s energy is calculated in units of dB re 1 µPa by

1 푇 푆푃퐿 = 20 푙표푔 √( ∫ 푝2(푡)푑푡), (1) 푟푚푠 10 푇 0

where p(t) is the instantaneous pressure derived from the output voltage (Urick, 1983).

Having the source position of the whale from the hyperbolic intersects, the range can be measured from the source (blue whale) to the each of the four receivers (SoundTraps) using the ‘distm’ function of the R package ‘geosphere’ (Hijmans, 2019).

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

The sound propagation was validated using playbacks of artificial upsweep signals that are in the same frequency band of the blue whale down sweeps. An underwater loudspeaker transmitted upsweeps with a constant source level and the position was tracked via GPS. Thus, we could estimate the propagation loss of the upsweep signals based on their SPLrms in relation to the distances to the sound source by a nonlinear regression. The resulting slope of the nonlinear regression fit gives an estimate of the propagation loss coefficient and can be applied to two different models for calculating the radiated noise level RNL as well as the source level SL.

Model 1 is based on spherical spreading and model 2 is based on cylindrical spreading, taking water depth and sea bed reflection into account.

Model 1: Spherical spreading

We determined RNL with the received SPLrms on each SoundTrap and the propagation loss PL 푅 of 20 푙표푔10 ( ) that fits to the upsweep signals of the loudspeaker over the measured distances 푅0 using the following equation:

푅 푅푁퐿 = 푆푃퐿푟푚푠 + 20 푙표푔10 ( ), (2) 푅0

where R is the distance from the blue whale to the SoundTrap and R0 is the reference range of 1 m. The value 20 is the propagation loss coefficient resulting from the nonlinear regression based on the upsweep signals and accounts for spherical spreading. Equation (2) corresponds to spherical spreading and is suitable for deep waters and closer distances. However, it is an established method and is used for comparison reasons with previous studies.

Model 2: Combining spherical and cylindrical spreading

Due to the fact that the area of investigation is on average 200 m deep and blue whales produce calls with a low frequency bandwidth, we additionally used another propagation model following methods in Ainslie et al. (2014). They combined spherical propagation up to bottom distance and cylindrical propagation and included the effect of sediment absorption in different

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

expressions using a seabed critical angle ψ (Ainslie et al., 2014). Thus, PL is calculated after equation (7) in Ainslie et al. (2014):

푅6 퐻⁄(2휓) 푃퐿 = 10 푙표푔10 ( ) + 10 푙표푔10 ( ), (3) 푅0 푅0

where R6 is the range up to 6 km, H is the water depth (200 m) and ψ is 0.1 for silt. This propagation fits well to the nonlinear regression of the upsweep signals considering a range of up to 6 km. The source level SL would then be calculated like that:

푆퐿 = 푆푃퐿푟푚푠 + 푃퐿 (4)

Source and radiated noise levels were calculated separately for each receiver and the average of these four values was used to calculate these levels for each call with a standard deviation.

In addition to source level measurements, the duration, frequency range (maximum to minimum frequency), centroid frequency, peak frequency were calculated for each blue whale down sweep as was done by Tervo et al. (2012) for bowhead whale song. Minimum and maximum frequencies of the signal were defined as the lowest and highest -10 dB points in the power spectrum. Peak frequency corresponds to the frequency in the signal with maximum energy and centroid frequency divides the signal into two parts of equal energy on a linear scale.

Ambient noise levels were calculated the same way as call source levels using a time window of 2 s before the cutting start of the detected call and a 1st order Butterworth filter for the same bandwidth of the call.

C. Tag data analysis Behavioural data

We analysed depth data using IGOR Pro version 8.04 (WaveMetrics, Inc). We defined the beginning and end of a dive as 1 m of depth and the behaviour diving, when maximum dive

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

depth is deeper than 10 m. We detected lunge feeding events by a change of the absolute vertical speed (AVS) which is calculated from depth data, because blue whales also perform vertical lunge feeding (also reported by Akiyama et al. (2019) for humpback whales). The maximum peaks of positive AVS more than 10 m in 4 s within a 40 s time window is detected. These maximum peaks are defined as lunge feeding events, when rapid acceleration or deceleration occurs, i.e. the AVS is faster or slower than 4 m in 4 s within an 8 s time window. Depth data of a 4 s interval were used, because we needed to make sampling interval of depth data all the same, which is 4 s. Behavioural data are categorised into feeding dive, non-feeding dive and surfacing. A feeding dive is a dive with lunge feeding events and a non-feeding dive is a dive without. Surfacing is a period except diving.

Acoustical data

Acoustical data were analysed with Audacity (Audacity®) which change from raw acoustic data to spectrogram. The sound events emitted by whales were detected by visually inspecting the spectrogram and aurally checking those sound events. Call sonograms were created with Audacity using the 8192, 16384 or 32768 Fast Fourier Transform (FFT) algorithm with a Hann window. For the power spectrum calculations, a 16384 FFT was used to enable a better resolution in the low-frequency range.

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

III. Results We recorded acoustical data with a GPS-linked receiver array during periods when no noise from anthropogenic activities apart from the tagging activities was present in the area. We detected 30 blue whale down sweeps, simultaneously recorded on all four receivers. Thirteen down sweeps could be localised. Fig. 7 shows the spectrogram of such a blue whale down sweep.

Fig. 7: Spectrogram of a detected blue whale down sweep at 4.6 km from the receiver. Fast Fourier Transform: 16384, Hann window, 95% overlap.

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

The detected call is in a frequency range from 70 to 45 Hz and includes highest energy in the lower frequencies (around 48 Hz) of that sweep. Tab. 2 shows the call characteristics of all blue whale down sweeps that were clearly distinguishable from ambient noise.

Tab. 2: Call characteristics of blue whale down sweeps in Skjálfandi Bay, Northeast Iceland derived from the array data.

Call characteristics Magnitude Centroid 65 ± 9 (SD) frequency (Hz) range: 58 - 72 Peak 63 ± 10 (SD) frequency (Hz) range: 56 - 70 Frequency 28 - 107 range (Hz) 1.2 ± 0.4 (SD) Duration E (s) 90 range: 0.9 - 1.3 Source level 172 ± 9 (dB re 1 µPa m) range: 151 - 185 2.0 Call rate (calls h-1) range: 1 - 8

The received sound pressure levels ranged from 125 to 87 dB re 1 µPa at distances between 258 m and 14 km, respectively (Fig. 8).

Received ambient noise levels were on average around 88±9 dB re 1 µPa in the frequency band of the down sweeps.

A. Propagation loss and source level based on array data The source level is the sound intensity measured 1 m from the sound source. To obtain the source level, the propagation loss from a range of 1 m to the recorded range needs to be added to the received sound pressure level (SPL). The propagation loss was estimated by a nonlinear logarithmic regression, estimating the intercept and slope based on the received SPL of artificial upsweep signals, assuming a constant source level and omnidirectional sound source. The regression was applied for two different models.

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

The best fit for model 1 (see Fig. 8) was:

Radiated noise level = 172 − 20푙표푔10(푅), where R is the entire range in which the sound source signal was measured. The root-mean-square error (RMSE) of the model as compared to 2 the sound source measurements was 2.8 dB and R was 0.83. Using 20 log10 R, or spherical spreading, as our propagation loss, we obtained an average blue whale source level of 174±7 dB re 1 µPa (Fig. 8, right).

Fig. 8: Propagation loss (left) and resulting radiated noise levels (RNL; right) for blue whale down sweeps after model 1 assuming spherical spreading. Blue circles represent the determined received sound pressure levels (SPL) of the localised down sweeps for each of the four receivers, respectively. The propagation loss (red line) is based on spherical spreading of SPL of artificial upsweep signals (red dots) for the entire range in which down sweeps were detected. Boxplot of the calculated RNL of down sweeps based on model 1 (right).

The best fit for model 2 (see Fig. 9) was:

Source level = 172 − 10푙표푔10(푅6) − 30, where R6 is the close range (up to 6 km) from sound source to the receiver and the value 30 accounts for sediment absorption and water depth. The RMSE was 1.5 dB and R2 was 0.83. This model resulted in a blue whale mean source level of 172±9 dB re 1 µPa for distances up to 6 km (Fig. 9).

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

Fig. 9: Propagation loss (left) and resulting source levels (SL; right) for blue whale down sweeps after model 2 assuming cylindrical spreading. Blue circles represent the determined received sound pressure levels (SPL) of the localised down sweeps for each of the four receivers, respectively. The propagation loss (red line) is based on cylindrical spreading of SPL of artificial upsweep signals (red dots) for a range of up to 6 km in which down sweeps were detected. Boxplot of the calculated SL of down sweeps based on model 2 (right).

B. Acoustic and diving behaviour from tag data We analysed data from four tags. Tag data of 59 h were obtained, including 1142 dives in total. Behavioural states were categorised as feeding dive and non-feeding dive according to depth (Fig. 10).

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

Fig. 10: Tag data (vertical speed and depth) of two behavioural categories. A Feeding B Non- feeding.

Dives of all animals were mainly (83-96% of all dives) of the feeding dive throughout the record duration. Average dive depth of feeding and non-feeding is 48 m ± 10 (SD, n = 1024 dives) and 51 m ± 9 (SD, n = 118 dives), respectively. Two types of sounds were recorded with the tag,

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

down sweeps and growls. Down sweeps were recorded 33 times (Tab. 3) and growls 27 times from three animals, respectively.

Tab. 3: Characteristics of blue whale down sweeps in Skjálfandi Bay, Northeast Iceland derived from tag data. Recording Number of Frequency Behavioural ID duration Depth (m) down sweeps (Hz) state (h) bm17_178a 3.8 0 - - - Feeding: 1 Upper: 90-120 Average: 8 ± 10 (SD) Non- bm17_179a 8.1 4 Lower: 50-55 Range: 0 – 21 feeding: 1 Surface: 2 Feeding: 4 Upper: 85-170 Average: 25 ± 10 (SD) Non- bm18_175a 20.1 24 Lower: 25-100 Range: 4 – 43 feeding: 20 Surface: 0 Feeding: 2 Upper: 45-180 Non- bm18_180a 23.7 5 Range: 3 – 17 Lower: 30-155 feeding: 1 Surface: 2

Down sweeps ranged from 45 to 180 Hz (upper frequency) and from 30 to 155 Hz (lower frequency). Growls showed a broadband frequency structure without any clear energy peak within the frequency band. Blue whales emitted only a few down sweeps during feeding dives (n = 7). Many down sweeps were detected during non-feeding dives (n = 22) and surfacing (n = 4) when the behavioural state changed. The diving and vocalisation behaviour of one tagged blue whale (bm18_180a) over a day is depicted in Fig. 11.

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

Fig. 11: Vertical speed and dive profile of one tagged blue whale (bm18_180a) with behavioural phases and vocalisation. Red dots on vertical speed indicate lunge feeding events. The arrows show vocalisation of two species (H = humpback whale, D = dolphin). The light blue, grey and red lines indicate surface, non-feeding and feeding behaviour, respectively. Blue circles on depth account for down sweeps and green triangles for growls.

The blue whale exhibits vocalisations like down sweeps and growls within different behavioural phases. Some down sweeps (blue circles) are emitted during non-feeding dives, whereas many growls (green triangles) occurred during feeding (Fig. 11). Growls showed a broadband frequency structure without any clear energy peak within the frequency band. Growls made by bm18_180a at 12:18 of 28th June 2018 occurred after a humpback whale call (sonogram; Fig. 12).

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

Fig. 12: Segment of Fig. 11 between 12:15 and 12:50 (top) and sonogram of 12:18 (bottom). The light blue, grey and red lines indicate surfacing, non-feeding and feeding behaviour, respectively in the top panel. A green triangle indicates a growl in the top panel. The dotted

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

arrow in the sonogram indicates a call from a humpback whale and the solid arrow a growl of the tagged blue whale.

Also 13 growls made by the same individual at a period between 17:13 and 17:20 of 28th June 2018 were detected, when many dolphin sounds were recorded (Fig. 11).

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

IV. Discussion We recorded vocalisations of blue whales in the Skjálfandi Bay in Northeast Iceland by both animal-borne tags and a GPS-linked receiver array. The recorded vocalisations consist nearly exclusively of down sweeps, also known as type D calls (McDonald et al., 2001; Thompson et al., 1996). Similar observations were made by other studies conducted in Iceland (Akamatsu et al., 2014; Iversen et al., 2010, 2011). We compared call characteristics of blue whale vocalisation with calls of the same species recorded in other regions of the world. Thompson et al. (1996) were the first who described down sweeps as a vocalisation type of blue whales in the Gulf of California and reported a frequency range of 57 to 11 Hz and a duration lasting for about 1 s. The calls, we recorded in Skjálfandi Bay ranged from maximum 107 to minimum 28 Hz with a duration of 0.9 to 1.3 s long. Akamatsu et al. (2014) performed acoustic measurements on individual tagged blue whales in the same bay and found similar frequency ranges from 105 to 48 Hz and durations from 1 to 2 s. A study by Oleson et al. (2007a) in the eastern North Pacific calculated sound characteristics of D calls with averaging start frequency of 75.7 Hz, end frequency of 39.3 Hz as well as a duration of 1.8 s, which is close to our findings

(Fmin=52 Hz; Fmax=79 Hz). Down sweep calls recorded off California had a duration of about 1 s and ranged from 60 to 45 Hz (McDonald et al., 2001). Blue whale down sweeps are thus comparable in frequency and duration but show slight variations across different regions of the world. Variations in vocalisation not only occur between regions, but also between seasons. Oleson et al. (2007a) documented that blue whale song in feeding season occurs temporally and geographically separate from breeding grounds. Also, B and D calls were found to be separated, both seasonally and daily, suggesting that different calls are used for different behaviours (Oleson et al., 2007b).

Another aspect to consider is the call rate. For example, A and B calls referred to as song show a stereotyped form and are repeated frequently (McDonald et al., 2006). Songs are hypothesised to attract females and used in a mating context (Oleson et al., 2007a). The presence of blue whales in Skjálfandi Bay during the summer months occurs in a completely different context in a foraging area. The call rate in this area was found to be relatively low, 2.1 times per hour ranging from minimum one call to maximum eight calls over a one-hour period. Low call rates are not unusual for feeding grounds. In the study of Akamatsu et al. (2014) the tagged blue whales were barely calling during feeding (four calls over a period of 21 hours). Icelandic blue

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

whales called less frequently during feeding than Pacific blue whales emitting singular A and B calls 3.9 times per hour and some of them were feeding (Oleson et al., 2007a).

D calls, also known as down sweeps are produced during foraging by both male and female blue whales (Oleson et al., 2007a). The findings of a few studies (McDonald et al., 2001; Oleson et al., 2007a; Thode et al., 2000) suggest that the function of D calls is likely to be related to social interaction. The reason for producing these down sweeps in a feeding area is not clear. They might be associated with group (McDonald et al., 2001) and foraging behaviour (Oleson et al., 2007a; Thompson et al., 1996). Oleson et al. (2007b) suggested that D calls attract other individuals to feeding grounds or maintain cohesion within the foraging group. We detected 30 blue whale down sweeps on the array data and 33 in total on the tag recordings (Tab. 3). An interesting point is that only seven down sweeps occurred during 1024 feeding dives, whereas 23 down sweeps were detected in 118 non-feeding dives. We hypothesise that the greater number of down sweeps in non-feeding dives compared to feeding dives could be a possible signal for disappointment in situations where no feeding occurred. Thus, conspecifics might understand that there is no prey available. However, a few down sweeps were also present in feeding dives. There is evidence that blue whales also emit growls along with sounds of other species like dolphins and a humpback whale (Fig. 11). It suggested that the animals use these growls for inter-specific communication.

The source level in our study is 7 to 17 dB lower than down sweeps and songs recorded in the Southern Hemisphere and the North Pacific (Tab. 4). Studies on blue whale vocalisation in the eastern Pacific documented source levels from 180 to 188 dB re 1 µPa (McDonald et al., 2001; Thode et al., 2000; Thompson et al., 1996). The highest source level of blue whale calls with 189 dB re 1 µPa was found by Širović et al. (2007) in the Southern Ocean. Our findings are comparable with the level of the z-call documented by Samaran et al. (2010), but 3 to 13 dB higher than the findings of Akamatsu et al. (2014) for the same region. The discrepancy in estimated source levels can be explained by three factors, due to different used methodologies of source level estimation, local geographic conditions and local acoustic conditions.

Propagation losses from source to receiver are subject to many factors and might be affected by a variety of source level estimation methods used. We estimated the propagation loss via a nonlinear regression based on the received SPL of the artificial upsweep signals. This

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

propagation was applied to the received SPL of the blue whale down sweeps. In our study, we used two different models to estimate both, radiated noise levels (RNL) and source levels (SL). Model 1 is based on spherical spreading and model 2 corresponds to cylindrical spreading accounting for sediment absorption and water depth. The resulting nonlinear regression showed that model 1 fits well to the received SPL of the blue whale down sweeps when considering the entire range in which these down sweeps were detected. However, model 2 fits better when looking at the SPL in the proximity of the sound source (up to 6 km). R2 was 0.83 for both models, i.e. the variability in the received SPL is well explained by both models. The change from spherical spreading at close ranges to the sound source to cylindrical spreading at further ranges was also described by Ainslie et al. (2014).

There are many different approaches to estimate a source level. Cummings and Thompson (1971) applied ‘normal spreading’ and ignored attenuation due to the close proximity of the whales. McDonald et al. (2001) recorded in deep waters and used spherical spreading to calculate the source levels. They did not consider a detailed propagation model due to a lack of data on sound speed, bathymetry, seafloor characteristics and depth of the calling animals. Several propagation loss models were tested resulting in values close to spherical spreading (McDonald et al., 2001). Thode et al. (2000) used multi-modal convolution techniques to calculate the source levels (in shallow waters). Samaran et al. (2010) performed propagation modelling and Akamatsu et al. (2014) assumed spherical sound propagation from the possible sound source location to the tag. Romagosa et al. (2015) and Guazzo et al. (2020) calculated source levels of sei and humpback whale sounds using a combination of both spherical and cylindrical spreading methods. For comparison reasons, not only of the source levels, but also the method of obtaining them, we used different models to show the complexity of sound propagation when interpreting source levels.

The recordings of our study were performed under rather quiet conditions. The ambient noise has been quantified at 88±9 dB re 1 µPa and is considered as low to moderate noise level (Woelfing et al., 2020). Therefore, blue whales do not have to overcome that noise by increasing their source levels, which might explain the lower source intensities from Icelandic blue whales as compared to other recording sites.

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

Finally, both our study and the one from Akamatsu et al. (2014) show lower source levels in blue whale down sweeps and were conducted in the same geographical area. The Skjálfandi Bay is approximately 30 km in width across its mouth and 20 km in length from its mouth to the southern end. The water depth at the mouth is more than 150 m. The lower SL in our study could be associated with the environmental conditions (bathymetry, water depth, substrate, geoacoustic profile etc.) and the size of the area of investigation. The audible range was estimated to be around 16 km given the ambient noise of 88±9 dB re 1 µPa and the parameters of the nonlinear regression for estimating the propagation loss. It is thus most likely that the whales do not have to put so much effort in producing very loud calls to reach nearby conspecifics by. Our findings suggest that the whales adapt their calls to the size of the bay.

Akamatsu et al. (2014) found that blue whales produced calls with low source level occurring very infrequently in their feeding grounds, especially during foraging. Our calculated source levels were 3 to 13 dB higher than the ones measured by Akamatsu et al. (2014), but the low call production rate was similar. However, Skjálfandi Bay is an important feeding area, where blue whales acquire energy and also a place where this wide-roaming highly dispersed population congregates to engage in social interactions (Sears and Perrin, 2017). If anthropogenic noise will be introduced, the whale’s ability to detect socially relevant signals could be reduced (Di Iorio and Clark, 2010). The low call production rate and source levels impede the passive acoustic monitoring in terms of for example density estimation. Knowing the call characteristics, such as the source level, is important to understand over which distances the whales communicate. Furthermore, they are useful for assessing how anthropogenic noise can potentially affect the animals.

The variability of 34 dB (range of source level, Tab. 3) in source level could be related to the changing in source level for adapting to fluctuations in ambient noise conditions. Blue whales do not necessarily have to produce calls with higher intensity, when staying in an area where the conditions are acoustically beneficial for the animals. Guazzo et al. (2020) argued that the location can have an effect on source levels, as they compared for intensities of humpback whale song from deep and shallow waters. However, if the ambient noise level is raised, it might have direct costs through higher energy consumption when producing calls. An indication for metabolic costs due to modifying sound was found by Noren et al. (2013) for bottlenose dolphins. Holt et al. (2015) showed evidence of metabolic costs of increased vocal

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

effort by bottlenose dolphins. The cumulative effects of these costs in relation to other anthropogenic pressures are an important aspect in terms of evaluating anthropogenic noise effects on population level (Wartzok et al., 2005).

V. Conclusion Blue whales in Skjálfandi Bay, Northeast Iceland infrequently produced down sweep calls during feeding. The source level of these calls were 7 to 17 dB lower than the ones documented in other regions of the world. Blue whales show the ability to vary the intensity of their calls showing adaptation to the local environment. An additional increase of ambient noise due to anthropogenic effects can result in a Lombard effect resulting in additional energy consumption to produce calls. Future studies are needed to clarify the meaning of these costs for the population of this species.

VI. Acknowledgement This research was supported by the Federal Ministry for the Environment, Nature Conservation, Building and Nuclear Safety, German Environment Agency, Project number: FKZ 3715 55 299 0. Further, we were supported by “Grant-in-Aid for Research Activity Start-up from the Japan Society for the Promotion of Science (JSPS) Grant JP19K24380”, “The JSPS Postdoctoral Fellowships Research Abroad”, “The Mitsui & Co. Environment Fund”, “Biologging Science, the University of Tokyo (UTBLS)”, and “The Research Grant against Global Warming of the Ichimura Foundation for New Technology”. We greatly thank Mikkel Villum for his huge effort and assistance during fieldwork.

Chapter 2: Blue whales off Iceland: Silent giants in a Northern feeding ground

VII. Appendix

Tab. 4: Comparison of estimated source levels of blue whale vocalisation from our study and those recorded in other regions of the world from different blue whale populations.

Population Location Vocalisation type Source level (dB re 1 µPa m North Skjálfandi Bay, Down sweep 172 ± 9a Atlantic NE Iceland North Skjálfandi Bay, Down sweep 159 to 169 dBb Atlantic NE Iceland Southern Southern Indian z-call 179 ± 5c Hemisphere Ocean off the Western Southern Southern Ocean Antarctic 189 ± 3d Hemisphere blue whale song Peninsula North Pacific off California B call 186e North Pacific off California B call 180f Southern off Chile Down sweep 188g Hemisphere aThis study, bAkamatsu et al. (2014), cSamaran et al. (2010), dŠirović et al. (2007), eMcDonald et al. (2001), fThode et al. (2000), gCummings and Thompson (1971)

The effect of marine vibrators on blue whale vocalisation behaviour

Manuscript: J. Baltzer, T. Iwata, T. Akamatsu, K. Lucke, M.-H. Rasmussen, M. Wahlberg, J.G. Schnitzler, U. Siebert (in prep.). The effect of marine vibrators on blue whale vocalisation behaviour

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

J. Baltzer1, T. Iwata2,3, T. Akamatsu2, K. Lucke4, M.-H. Rasmussen5, M. Wahlberg6, J.G. Schnitzler1, U. Siebert1

1Institute for Terrestrial and Aquatic Wildlife Research, University of Veterinary Medicine Hannover, Foundation, Germany

2Ocean Policy Research Institute (OPRI), Sasakawa Peace Foundation, Japan

3Atmosphere and Ocean Research Institute, The University of Tokyo, Japan

4Jasco Applied Sciences, Australia

5The University of Iceland’s Research Centre Húsavík, Iceland

6Department of Biology, University of Southern Denmark, Denmark

Abstract The effect of marine vibrator (MV) signals on blue whale vocalisations was investigated. We recorded blue whale down sweep signals with a GPS-linked receiver array. The detected down sweeps were analysed for different characteristics, such as source level (SL), call rate, duration and frequency (minimum, maximum, centroid and peak frequency) in three experimental phases with and without sound exposure. Additionally, we investigated the ambient noise levels (NLs) within these phases. Our findings are preliminary and give insights that elevated noise has an effect on blue whale vocalisation behaviour. We could see that the exposure phases showed highest NLs, probably caused by the MV signal. The post-exposure was the quietest phase, in which the highest call rates (6.2 down sweeps per hour) and the longest down sweeps (2.8 s) occurred. This increase was probably caused by the preceding elevated noise. However, we could also see a frequency shift that might be elicited specifically by the MV signals. In contrast, no changes in source level occurred. Additionally, changes in call characteristics towards the noise exposure have to be verified. We speculate, how the MV signals could affect

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

blue whale vocalisation, but more data are needed from both hydrophone arrays together and tags to confirm our findings.

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

I. Introduction The Arctic is under constant change due to human exploitation and anthropogenic as well as natural climatic alterations. Environmental alterations have implications for the animals living in the Arctic.

Hydrocarbon exploration activities are predicted to increase in the Arctic and sub-Arctic (for example, around Iceland) in the near future. The instruments in seismic exploration are seismic airguns, emitting impulsive signals into the water column and towed behind a slow-moving vessel. Airguns are aligned so that their acoustic signals are directed vertically towards the bottom. Even though the fact that the sound source is pointing downwards makes it less likely for animals being affected by high intensities from airguns, the frequency spectrum of the signal gives rise to concerns regarding the potential negative effects on the marine fauna. The sound generated by airguns may affect marine mammals over large areas (Richardson et al., 1995).

The effects on marine mammals from airguns can be physiological, perceptual, behavioural, or indirect (e.g., effect on prey availability) (Gordon et al., 2003). One perceptual effect of special interest here is masking, i.e., the fact that biologically relevant signals cannot be perceived due to noise interference (Erbe et al., 2016). Masking may lead to difficulties for the animals in finding prey, or reducing in their communication range (Clark et al., 2009). A variety of studies describe that high noise levels over a broad frequency band can mask relevant signals (Branstetter et al., 2013; Clark et al., 2009; Hatch et al., 2012; Parks et al., 2011a; Richardson et al., 1995). Marine mammals may compensate for increased ambient noise in changing their signal’s amplitude, duration, production rate, and/or frequency (Holt et al., 2009). This ability to overcome masking noise by adjusting their vocalisation may be crucial for communicating in a dynamic acoustical environment (Parks et al., 2011a).

The compensation mechanism towards elevated noise is called Lombard effect, a tendency to increase vocal effort when communicating in noisy environments to enhance the audibility of their vocalisations. Not only the intensity of the signal is affected but also other acoustic features such as pitch, rate, and duration. This compensation effect results in an increase in the auditory signal-to-noise ratio of the animal’s call. Studies have shown that seismic airguns cause a Lombard effect in baleen whales (Blackwell et al., 2015; Castellote et al., 2012; Cerchio et al., 2014; Di Iorio and Clark, 2010; Thode et al., 2020). Bowhead whales (Balaena mysticetus) have been observed to increase their call source level and decrease their calls when exposed to

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

seismic airgun pulses (Thode et al., 2020). Di Iorio and Clark (2010) documented that also blue whales (Balaenoptera musculus) respond to seismic surveys by increasing their vocalisation rates, presumably to compensate for elevated ambient noise caused by these surveys. It is currently unknown if other types of sound sources would affect blue whale signalling in a similar manner.

A new developed seismic sound source, the marine vibrator (MV) is expected to reduce environmental impacts (Duncan et al., 2017; Pramik, 2013). Next to their usage in environmental sensitive areas, they can be employed in very shallow waters where traditional sources can be difficult to deploy. MVs could be a good alternative to airguns as their emitted signals have a lower sound pressure level (SPL) and a smaller bandwidth than airgun pulses (Pramik, 2013; Tenghamn, 2006). Thus, they are believed to have potentially lower environmental impact than airgun signals. The MVs emit frequency sweeps below 100 Hz (Duncan et al., 2017; Tenghamn, 2006) that are longer in duration than airgun signals. The MV signals are within the same frequency range of blue or fin whale (Balaenoptera physalus) calls (Širović et al., 2007). The combination of longer durations within an overlapping frequency range give the signals the potential to mask biologically relevant signals for e.g., blue whales. Therefore, before MV are used in the marine environment, potential effects on e.g., blue whales have to be assessed.

The aim of this study was to investigate the response of blue whales towards artificial MV signals at their feeding grounds. Therefore, fieldwork was conducted in Skjálfandi Bay, Northeast Iceland where blue whales occur frequently throughout the summer months to feed on krill (Euphausia superba). A major aspect is the investigation of the sound exposures on the vocalisation behaviour of the whales. Documented behavioural adaptations to sound exposure show cessation of vocalisation (Melcón et al., 2012), prolonged or increased repetition of vocalisations or a change in frequency and source level (Di Iorio and Clark, 2010).

The behaviour of blue whales was measured by recording their down sweeps with a GPS-linked receiver array and analysed them for changes in call rate, duration and frequency between different experimental phases with and without exposure of MV signals. We hypothesise that we can reveal indications of a potential Lombard effect towards the elevated ambient noise.

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

II. Material and methods A. Study area The fieldwork was conducted in Skjálfandi Bay (Fig. 6) in Northeast Iceland in June 2017 and 2019. The bay is approximately 30 km wide across its mouth and 20 km long and the water depth is up to 200 m. The University of Iceland’s Research Centre is situated at this bay, thus providing easy access to the animals by boat as well as the logistical infrastructure necessary to conduct measurements on this species.

B. Sound recorders (SoundTraps) We used an autonomous GPS-linked receiver array consisting of four acoustic devices (SoundTrap 300 HF, OceanInstrumentsNZ, AcousticMonitoring Systems). SoundTraps (STs) are compact self-contained underwater sound recorders for ocean acoustic research and offer 20 Hz to 150 kHz bandwidth frequency applications (max. sampling rate: 576 kHz, resolution: 16 bit). The sampling rate for the recordings was set to 48 kHz. The clipping level of the devices was measured to be 171.2 to 173 dB re 1 µPa (peak) by relative calibration in a sound-proof chamber in the frequency range of 40–200 Hz. The units have a low self-noise level (about 40 dB re 1 µPa2/Hz in the same frequency range) ensuring recordings with a sufficient signal- to-noise ratio for our purposes. The sound recorders were lowered to 20 m and connected with a cable to GPS-receivers attached to buoys 1 m above the water surface. Every second the GPS location and timing pulse was recorded on the ST in synchrony with the acoustic recordings.

All recordings were made in Sea State 3 or less. During field trials, the buoys were deployed in an approximate line 500 m up to 1 km apart in East-West direction. Recordings were mostly conducted during the night due to calm weather and considerably reduced boat traffic. A speed boat was used for deployment and retrieval of the buoys.

C. Sound source ‘Argotec SS-2’ A high-power transducer, Argotech SS-2 (Fig. 13), was used to mimic sounds produced by MVs. Since the SS-2 is used underwater, the air chambers need to be pressurised to compensate their internal pressure for the ambient water pressure at operating depth. The air pressure in the air chamber should equal the pressure of the water on the depth of the piston face. Pressure

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

hoses connect the sound source to a compressed-air source at the surface for this pressure compensation.

Fig. 13: Sound source ‘Argotec SS-2’ for emiting marine vibrator sounds.

The SS-2 emits an upsweep signal in the frequency range from 20 to 100 Hz with a duration of 10 s and a 1 s on- and offset (Fig. 14). After the transmission of the signal there is a silent period of 8 s before the next signal. This results in a total signal length of 20 s with duty cycle of 60%.

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

Fig. 14: Upsweep signal imitating the marine vibrator like sound.

D. Experimental design In order to investigate possible effects of the marine vibrator (MV) signal, we defined three different phases: pre-exposure, exposure and post-exposure phase. After successfully deploying the tags, the buoys with the SoundTraps and GPS were deployed in a linear array of 2–4 km distance before the pre-exposure phase started. A period of 30 min was chosen for collecting pre-exposure data of the behaviour of the tagged animal. Following this, the approach by the source vessel was initiated. The ship’s engine and all ancillary machinery were stopped once

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

the source vessel was in position close enough to the animal. The sound source was deployed by crane to a depth of approximately 7 m without any active transmissions. Then the sound source emitted the upsweep signal for 1 h (exposure phase) to allow measurements of behavioural changes during several dive periods. After sound exposure, a 1 h post-exposure period was initiated, during which group and surface behaviour were observed. The study setup is depicted in Fig. 15.

Fig. 15: The experimental design cycles through different phases: a 30 minutes pre-exposure phase, a 60 minutes exposure phase, where simulated MV signals are emitted and a 60 minutes post-exposure phase.

E. Data analysis Source level

In order to determine the source level (SL) of a blue whale call, the received sound pressure level (SPL) was calculated in root-mean-square (rms) units. When calculating rms, the length of the signal has to be defined carefully because the rms level becomes lower if the averaging window surpasses the length of the signal being averaged (Madsen, 2005). The duration of the signal in our study was defined as including 90% of the energy of the entire signal waveform (Madsen, 2005; sensu Southall et al., 2007), i.e., from the time where the signal includes 5% of the total energy to the time at which 95% of the energy was reached. The SPLrms in a time window incorporating 90% of the signal’s energy is calculated in units of dB re 1 µPa by

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

1 푇 푆푃퐿 = 20 푙표푔 √( ∫ 푝2(푡)푑푡) (1) 푟푚푠 10 푇 0

where p(t) is the instantaneous pressure derived from the output voltage (Urick, 1983).

Having the source position of the whale from the hyperbolic intersects, the range could be measured from the source (blue whale) to the each of the four receivers (SoundTraps) using the ‘distm’ function of the R package ‘geosphere’ (Hijmans, 2019).

The sound propagation was validated using playbacks of the MV signals that are in the same frequency band of the blue whale down sweeps. An underwater loudspeaker transmitted these signals with a constant source level and the position was tracked via GPS. Thus, we could estimate the propagation loss of the MV signals based on their SPLrms in relation to the distances to the sound source by a nonlinear regression. The resulting slope of the nonlinear regression fit gives an estimate of the propagation coefficient for calculating an approximate source level SL.

We determined SL with the received SPLrms on each SoundTrap and the propagation loss PL of 푅 20 푙표푔10 ( ) over the measured distances using the following equation: 푅0

푅 푆퐿 = 푆푃퐿푟푚푠 + 20 푙표푔10 ( ), (2) 푅0

where R is the distance from the blue whale to the SoundTrap and R0 is the reference range of 1 m. The value 20 is the propagation loss coefficient resulting from the nonlinear regression based on the MV signals and accounts for spherical spreading. Source and radiated noise levels were calculated separately for each receiver and the average of these four values was used to calculate these levels for each call with a standard deviation.

Ambient noise levels (NL) were measured after the same principle of calculating SPL using a time window of 20 s before the cutting start of the detected down sweep and a first order Butterworth filter for the same bandwidth of the down sweep. By subtracting the NL from the SPL of the down sweep, we were able to calculate also the signal-to-noise ratio (SNR).

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

Additionally to source level measurements, further call characteristics, such as call rate as well as duration and frequency (maximum, minimum, centroid, and peak frequency) were determined for each blue whale down sweep similar to Tervo et al. (2012).

Call duration

In order to estimate the call duration, the cumulative energy function was used on the squared vector of the signal, as the signal squares are proportional to the sound intensity. Thus, ith element of the cumulative signal vector is proportional to the amount of energy passing from the first sample until the ith sample of the signal. We defined the duration of the signal as the interval containing 90% of the energy (starting at 5% and ending at 95% of the total energy.

Frequencies

Minimum and maximum frequency

Minimum and maximum frequencies of the signal were defined as the lowest and highest - 10 dB points relative to the peak of the envelope of the waveform. This has been applied to calculate the duration of narwhal clicks (Møhl et al., 1990). The frequencies were calculated by first performing a Fast Fourier Transform (FFT) on the signal and then divide it by the number of samples to obtain the true amplitude of each frequency component. Then, the absolute numbers of that FFT signal are used to create a frequency vector.

Centroid frequency

The centroid frequency divides the signal’s spectrum into two parts of equal energy on a linear scale. It was calculated using the absolute numbers of the FFT signal and a frequency vector based on the absolute signal.

Peak frequency

Peak frequency is the signal frequency with maximum energy. It was calculated using the absolute numbers of the FFT signal and a frequency vector based on the absolute signal. Within this vector, we searched for the frequency corresponding to the maximum of the absolute signal.

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

III. Results Over the entire study (14 fieldwork trips), we detected 118 blue whale down sweeps during 1936 minutes of recording. The number of recorded down sweeps fluctuated among our fieldworks, i.e. in some fieldworks we did not detect any down sweep, whereas in others we detected up to 57 down sweeps. Tab. 5 summarises all measured call characteristics, such as source level (SL), call rate, duration, and frequency (minimum, maximum, centroid, and peak frequency) for the down sweeps that were clearly distinguishable from the ambient noise (good signal-to-noise ratio (SNR)). The call characteristics were calculated for 72 down sweeps (approx. 61% of all detected ones). Source levels could be estimated for 20 down sweeps (17% of all detected ones), due to a reasonable hyperbolae intersection when localising these sweeps (Tab. 5).

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

Tab. 5: Call characteristics during the three experimental phases.

experimental phase Call Pre-exposure exposure Post-exposure characteristics Magnitude Total number of 30 26 62 calls Number of calls 21 (70%) 16 (62%) 35 (56%) analysed Source level 175 ± 12 176 ± 5 174 ± 4 (dB re 1 µPa m) (n = 5) (n = 8) (n = 7) 2.0 2.9 6.2 Call rate (calls h-1) range: 1 - 8 range: 1 - 10 range: 1 - 33 Minimum 52 ± 10 41 ± 9 37 ± 9 frequency (Hz) Maximum 79 ± 11 70 ±13 68 ± 12 frequency (Hz) Frequency range 28 - 107 24 - 110 24 - 128 (Hz) Centroid 65 ± 9 (SD) 55 ± 9 (SD) 53 ± 8 (SD) frequency (Hz) range: 58 - 72 range: 50 - 59 range: 49 - 60 Peak 63 ± 10 (SD) 52 ± 11 (SD) 51 ± 10 (SD) frequency (Hz) range: 56 - 70 range: 45 - 59 range: 45 - 60 1.2 ± 0.4 (SD) 1.2 ± 0.6 (SD) 2.8 ± 1.3 (SD) Duration E (s) 90 range: 0.9 - 1.3 range: 0.8 - 1.4 range: 1.8 - 3.4 Minutes of 634 530 772 recording

SLs showed no differences between the experimental phases and were estimated at 174 to 176 dB re 1 µPa m. However, there was a high variability of the SL in the pre-exposure phase (Tab. 5).

Ambient noise levels (NLs), measured for the same frequency range of blue whale down sweeps, were 11 to 22 dB higher in the exposure than in the pre-exposure and post-exposure phase (Fig. 16).

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

Fig. 16: Comparison of the ambient noise level between the three experimental phases. A Wilcoxon rank sum test was performed resulting in p-values with certain significances (* p < 0.05, ** p < 0.01, *** p < 0.001).

The SNR was significantly lower (p < 0.001) in the exposure phase than in the pre- and post- exposure phase.

The call activity was highly variable form one fieldwork day to another, and blue whale calls could be detected in 8 out of 14 fieldworks. We could observe a call rate of 2.0 calls per hour in the pre-exposure phase, 2.9 calls per hour during exposure, and 6.2 calls per hour were

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

emitted during post-exposure phases. Call rates per hour were two to three times higher in the post-exposure phase compared to the pre- and exposure phase, respectively. The call rate in the pre-exposure phase was significantly lower (p < 0.05, paired Wilcoxon rank sum test) compared to the post- and exposure phase. During exposure, 26 calls were detected (Tab. 5), 16 of which occurred in between upsweep signals. Fig. 17 shows an example of a blue whale arch sound between the artificial MV sounds.

Fig. 17: Blue whale D call (arch sound) among marine vibrator like upsweep signals played back via the SS-2.

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

However, there was no clear pattern if down sweeps rather occurred within or in between MV signals.

The call duration was significantly higher (p < 0.001) in the post-exposure compared to the pre- and exposure phase (Fig. 18). The average signal duration of the call in the post-exposure phase was twice as much in the other two phases.

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

Fig. 18: Comparison of the call duration (including 90% of the signal’s energy) between the three experimental phases. A Wilcoxon rank sum test was performed resulting in p-values with certain significances (* p < 0.05, ** p < 0.01, *** p < 0.001).

In Fig. 19, the centroid frequencies of the three different phases were compared. In order to investigate whether call characteristics changed, a Wilcoxon rank sum test was performed.

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

Fig. 19: Comparison of the centroid frequency between the three experimental phases. A Wilcoxon rank sum test was performed resulting in p-values with certain significances (* p < 0.05, ** p < 0.01, *** p < 0.001).

The test showed that centroid frequencies significantly differed between pre- and exposure (p < 0.01) as well as pre- and post-exposure phases (p < 0.001), respectively (Fig. 19). All frequencies (minimum, maximum, centroid and peak frequency) were significantly higher in the pre-exposure phase compared to the other two phases.

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

IV. Discussion In this study, we investigated whether marine vibrator (MV) signals had an effect on blue whale vocalisation behaviour. We determined the characteristics of blue whale down sweeps in different experimental phases with and without exposure of the MV signal.

The source levels (SLs) of blue whale down sweeps ranged from 174 to 176 dB re 1 µPa m (Tab. 5) over the course of the experiment. There were no significant differences when comparing SLs between the experimental phases.

We also calculated the ambient noise level (NL) to determine whether there was a relation between NL and source level (SL). Although NLs differed significantly among the experimental phases, no changes in SL occurred. Blue whales may not be able to counteract the MV signals, despite high variability in SL. However, changes in SL as an adaptation towards anthropogenic noise have been documented in several studies on cetaceans.

Animals can adapt their vocalisation according to the surrounding acoustic conditions and increase the amplitude of communication signals to compensate for a decreased SNR. This phenomenon is known as the Lombard effect (Lombard, 1911) and was first described in humans and later also in birds and mammals (Kragh et al., 2019). There are several studies documenting this phenomenon in marine mammals. Studies on toothed whales (Holt et al., 2009, 2011; Kragh et al., 2019; Scheifele et al., 2005) documented such an effect expressed in an increase of source level or call rate when the ambient noise level was raised. Killer whales (Orcinus orca) have been observed to increase their amplitude in calling when exposed to vessel noise (Holt et al., 2009, 2011). Scheifele et al. (2005) reported higher source levels of communication signals by beluga whales (Delphinapterus leucas) as a response towards boat noise in the St. Lawrence river estuary. Bottlenose dolphins (Tursiops truncatus) adjusted their signature whistles when ambient noise was elevated (Kragh et al., 2019). There is evidence that migrating humpback whales (Megaptera novaeangliae) exhibit a Lombard effect towards natural sound sources, such as sea surface motion due to the wind speed (Dunlop et al., 2014). Thode et al. (2020) found that bowhead whales (Balaena mysticetus) increase their source level when seismic airguns are active. Blue whales have been documented to show an increase in call rate when exposed to seismic exploration sounds (Di Iorio and Clark, 2010) and a decrease when mid-frequency sonar was active (Melcón et al., 2012). Unlike the findings of other

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

studies, we did not observe changes in SL, but we could however confirm changes in call rate. In our study, the call rate increased significantly in the post-exposure phase (Tab. 5).

Not only changes in call rate can occur, but also in the length of the emitted vocalisation signal. A study on killer whales documented that they adjust their call duration when boats were present as a form of ‘anti-masking’ behaviour (Foote et al., 2004). This anti-masking strategy was also observed in humpback whales that altered their songs during playbacks of low- frequency active sonar. Our findings further showed a similar pattern as the duration of blue whale down sweeps increased significantly in the post-exposure phase compared to the two previous phases (Fig. 18).

While call rate and duration of down sweeps increased after the noisier phases (pre- and exposure), the frequencies showed an opposite effect. Blue whale down sweeps differed significantly in frequency in the post- and exposure phase compared to the pre-exposure phase. We observed a decrease in frequency over the course of the experimental phases. It was described that bottlenose dolphins adjust frequencies to avoid an overlap with elevated ambient noise (Gospić and Picciulin, 2016; Heiler et al., 2016; Marley et al., 2017; Papale et al., 2015; Strahan et al., 2020). Morisaka et al. (2005) observed that Indo-Pacific bottlenose dolphins emit whistles of lower frequencies when ambient noise was increased. Cetaceans are thought to have different acoustical niches by either decreasing their communication sounds below (Dahlheim et al., 1984) or increase them above ambient noise (Wang et al., 1995). Our preliminary results also showed a frequency shift as a possible response towards the noise introduction. However, there is no difference in frequency between exposure and post-exposure phase. It seems likely that the MV signal itself elicited a shift towards lower frequencies, which lasted within in the post-exposure phase. A closer look at the frequencies in the post-exposure phase revealed that frequencies increased again after half an hour (Fig. 20).

In the post-exposure phase, we observed most down sweeps that were significantly longer in duration than in the experimental phases with noise (pre- and exposure). The longer down sweeps and the higher call rate can be an adjustment towards higher ambient noise in general, not specifically the MV signals. The frequency shift, however, could be caused by the MV signal. Nevertheless, whether and to what extent marine animals are able to compensate for environmental noise remains unclear (Papale et al., 2015). Jensen et al. (2009) suggested that

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

changes in acoustic behaviour could be costly, because of possible increases in energy expenditure or modifications of the original information content.

The primary focus of this chapter was on data recorded via a GPS-linked hydrophone array. Additionally, tagging was conducted to investigate individual reactions towards the MV signals. However, no vocalisation occurred on the tags during the experimental phases. We could observe a slight behavioural reaction in terms of diving behaviour towards the sound source. In the second chapter, we already confirmed that blue whales primarily produce down sweeps in an Icelandic feeding ground, which is also documented for feeding grounds in other regions of the world (see Oleson et al., 2007a). Given the tag data and the number of down sweeps in non-feeding dives compared to feeding dives, we hypothesised that the down sweep could function as a ‘call of disappointment’. That may be an important information for conspecifics that no food is available. Additional noise, such as MV signals potentially mask this communication signal and thus could have an indirect effect on feeding success. The findings of Akamatsu et al. (2014) and our study showed that blue whales produce down sweeps very infrequently and with low intensity in their feeding grounds, especially during foraging behaviour. Higher ambient noise would require double effort for the animals to ensure that their signals are perceived by conspecifics. The higher call rate at the end of the experiment may be related to the fact that the exposure to MV signals raises the blue whales’ call production. However, simultaneous tag recordings could not confirm this observation. Observations from boats could also not confirm whether the number of surrounding blue whales raised because of the sound exposure and therefore more calls per hour were detected.

V. Conclusion The findings presented in this chapter are preliminary, which enables insights that elevated noise has an effect on blue whale vocalisation behaviour. A change in call rate and duration might rather be a response towards the elevated noise in general, whereas the frequency shift could specifically be caused by the marine vibrator signal. Source levels, however, showed no changes at all. In order to increase the sample size, the estimation of source levels has to be investigated further. Additionally, changes in call characteristics towards the noise exposure have to be verified. We speculate, how the MV signals could affect the blue whales in the area,

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

but more data are needed from both, hydrophone arrays together with tags to confirm our findings. Thus, we plan for another fieldwork season in 2021.

Chapter 3: The effect of marine vibrators on blue whale vocalisation behaviour

VII. Appendix

Fig. 20: Example of frequency change within the postexposure phase. A Wilcoxon rank sum test was performed resulting in p-values with certain significances (* p < 0.05, ** p < 0.01, *** p < 0.001).

Overall discussion

The investigation of effects of noise on marine mammals is interdisciplinary. On the one hand there are the challenges of physics, such as sound generation, propagation, measurement, and modelling. On the other hand, biological challenges, like monitoring animals, determining impacts, and understanding biological significance have to be faced. Misinformation and miscommunication can lead to various issues with underwater acoustic quantities, units, recording and reporting, as well as experimental design, statistical analysis, and interpretation (Erbe et al., 2019).

I. Anthropogenic noise affects marine fauna Anthropogenic noise is one of the most hazardous pollutions and has become a major threat to many different species (Andrew et al., 2002; Kunc et al., 2016; World Health Organization, 2017). Noise can be defined as sound that is disturbing both, the terrestrial and the marine fauna. In the marine environment, underwater noise is caused by various anthropogenic activities, such as ship traffic (reviewed in Erbe et al., 2019), construction and operation of offshore wind farms (OWFs) (Brandt et al., 2018; Dähne et al., 2013; Graham et al., 2019), military sonar (Curé et al., 2016; Filadelfo et al., 2009; Isojunno et al., 2016) or seismic surveys for oil exploration (Bain and Williams, 2006; Sarnocińska et al., 2020).

Anthropogenic noise is known to negatively affect various marine animals (Slabbekoorn et al., 2010; Weilgart, 2007). The effects of underwater noise from anthropogenic activities have been documented in several studies including behavioural responses, acoustic interference (i.e., masking), temporary or permanent threshold shifts in hearing ability), and stress (e.g., Erbe et al., 2018; Nowacek et al., 2007; Richardson et al., 1995). Sound sources of high-intensity can have severe consequences, such as injuries or strandings (Frantzis, 1998; Parsons et al., 2008) or lead to temporarily or permanently elevated hearing thresholds (Kastak et al., 2005; Mooney et al., 2009; Smith, 2004). At lower exposure levels, anthropogenic noise may affect behavioural patterns (Nowacek et al., 2007; Samson et al., 2016), which can be associated with fitness consequences and eventually population-level effects (Nabe-Nielsen et al., 2018; New et al., 2014).

Overall discussion

In order to estimate potential noise effects, sound emissions from anthropogenic activities need to be either recorded in field or controlled exposure experiments using playbacks of certain sounds need to be conducted. For instance, during the construction of OWFs accompanying ecological research was done to investigate, whether pile driving sound has an effect on harbour porpoises. There are activities at sea with consequences that cannot be assessed in advance (Weilgart, 2007). An example is the construction of seed mussel collectors (SMCs) in the German Wadden Sea. SMCs were constructed at a World Heritage site with respect to complied guidelines that anchor pipes are vibrated into the seabed and not pile driven. To what extent vibration noise is affecting the environment was however completely unknown.

We assessed potential effects of noise generated by the construction of SMCs in the Wadden Sea (chapter 1). We conducted empirical noise measurements during the construction of seven anchor pipes. It is important to know how noise propagates through the environment. Thus, we determined accurate estimates of sound exposure levels (SELs) of manually detected vibration noise events at different distances to the construction site. Based on these SELs at a range of distances to the source, we developed a sound propagation model. The Wadden Sea forms a shallow body of water with tidal flats and wetlands, here sound travels on paths that interact with the seafloor and ocean waves at the sea surface (Moore et al., 2007). Therefore, we performed a nonlinear least-squares approach to our empirical data. The non-linear logarithmic regression resulting from the propagation model was applied to literature based SEL thresholds for effects on different marine mammals and fish. Thus, we could determine effect ranges for indigenous species.

Underwater noise is the most widespread and pervasive kind of anthropogenic energy which is introduced in the marine environment. Our results showed that the detected anchor pipe vibration embedment noise can elicit behavioural response in indigenous species from the Wadden Sea on a local scale. The construction of SMCs was done on certain conditions, for instance the anchor pipes for the seed mussel collectors were vibrated into the seabed as an alternative to pile driving. It can be concluded that the method of vibrating is much less harmful to the marine fauna than pile driving, at least for the species considered in our study. Our study gives an example that a sustainable human use in respect to the complied guidelines, can lead to a harmonious relationship between the needs of society and ecological integrity as conceded

Overall discussion

by the Trilateral Wadden Sea Plan (2010). This chapter 1 shows the importance of assessing the impact of noise for a sustainable management.

Germany, as European member state, is obliged to assess the current situation in their waters and to monitor changes in the future to implement the European marine strategy framework directive (MSFD, Descriptor 11, European Union, 2008). In the course of this directive underwater noise is embedded as one of the descriptors (Descriptor 11) and is entitled ‘Introduction of energy, including underwater noise, is at levels that do not adversely affect the marine environment’. There are few studies that provide useful methods on how to assess the impact of underwater noise (Boyd et al., 2008; Faulkner et al., 2018). The investigated sound source was an innovative method of fixing anchor pipes at the seabed, therefore the kind of noise had to be identified. Our study can be seen as a first step of risk assessment, the hazard identification according to the research strategy of Boyd et al. (2008). We however went one step further and predicted potential effect areas round the construction site on different species in theory without measuring the effects on the animals.

II. Anthropogenic noise and masking Anthropogenic noise may also interfere with the ability of marine animals to detect biologically relevant sounds. This phenomenon is also known as masking (Erbe et al., 2016). In order to analyse, whether biological sounds can be masked sound characteristics need to be measured to get an idea how animals use certain signals and face the challenges of their marine environment. We used a GPS-linked receiver array to study the vocal behaviour of Icelandic blue whales (chapter 2). Therefore, we recorded ambient noise and calling blue whales in Skjálfandi Bay, Northeast Iceland which is the focus in the second chapter. The aim was to investigate whether Icelandic blue whales are quieter than conspecifics in other parts of the world. The motivation behind it was a big difference between source levels documented for tagged blue whales in the same region compared to other areas of the world.

Customised tools were developed to analyse blue whale recordings using custom written scripts in R. Sound and GPS files downloaded from all receivers of the array had to be first synchronised and merged in preparation for data analysis. Furthermore, a call detector was applied in order to find down sweep calls of blue whales. This detector had to be evaluated by

Overall discussion

checking the files again manually and see how good the detector recognised the down sweeps. Having all down sweeps detected, we determined their received sound pressure level (SPL). The goal was to estimate source levels (SLs) of blue whale down sweeps. In order to calculate SLs, the SPL at the sound recorders needs to be known and the distance to the animal has to be measured (Wahlberg and Larsen, 2017). The estimation of source levels requires two important factors, a high signal-to-noise ratio (SNR) and the animal being close to the hydrophone array. If these requirements are not fulfilled, the estimation of the source level could be afflicted with errors, which makes it difficult to estimate TOADs, necessary to get the animal’s position. Our developed lightweight portable autonomous drifting passive acoustic array showed the advantages to be deployed quickly and repeatedly from a small boat very close to the animals of interest.

However, the capacity of passive acoustic monitoring (PAM) systems strongly relies on the characteristics of the whale calls, the local ambient noise levels, and the propagation conditions (Stafford et al., 2007). If whales are close to the water surface, it could happen that their emitted signal will be cancelled, due to reflections from the water surface resulting in a phase shift. Thus, calls might not be recorded on single receivers, which makes it difficult for the source level estimation. Because a minimum of four receivers is needed to triangulate the position of the emitting animal (Madsen and Wahlberg, 2007). However, surface reflection could also lead to the opposite, i.e., the signal being amplified due to interference of the signal’s waves as described by Charif et al. (2002). This phenomenon is called the Lloyd’s mirror effect and can lead to either an overestimation of the source level or an underestimation of the call rate. Calls produced in shallow depths may also result in shorter detection ranges due to increased attenuation of sounds propagating close to the surface (Parks et al., 2011b). In our study, we may have experienced this effect as the acoustic receivers were only lowered 20 m below the water surface. Understanding the calling behaviour, including call types, behavioural function of sounds, and call rates, is important for assessing the effectiveness of passive acoustic monitoring in detecting a particular species in a particular habitat area (Parks et al., 2011b).

The distance from the receiver (hydrophone) to the animal (sound source) is usually estimated by triangulation using several receivers at known positions and measuring the time-of-arrival differences (TOADs) between the same signal arriving on different receivers (Spiesberger and Fristrup, 1990; Wahlberg et al., 2001; Watkins and Schevill, 1972). By means of TOADs, we

Overall discussion

were able to adjust localisation algorithms using hyperbola intersects to get the position of the blue whale. The localisation, however, was challenging as down sweeps were degraded through background noise or cross-correlations of the same signals on different receivers resulted in vague TOADs, so that a hyperbola intersection was not possible.

For all localised down sweeps that were clearly distinguishable from background noise, we could estimate a propagation loss (PL) that is needed to obtain SLs. PL is mostly based on assumptions, for instance following spherical spreading combined with frequency-dependent sound absorption in deep waters (Wahlberg and Larsen, 2017). Several parameters, such as source frequency band, sound speed profile, bathymetry, bottom properties as well as source and receiver geometry have to be considered to reliably estimate PL as they all can have influence the propagation (Küsel et al., 2009). Therefore, it is always better to obtain empirical data in order to get clear insight on how sound radiates in a certain environment. To validate the sound propagation in Skjálfandi Bay we used playbacks of artificial upsweep signals in the same bandwidth of blue whale down sweeps with known source levels and positions. PL turned out to be a combination of cylindrical and spherical spreading depending on which distances from source to receiver are considered. Thus, two propagation models were applied to the SPL and distances measured for blue whale down sweeps. We were able to estimate an average SL at 172±9 dB re 1 µPa m for 13 down sweeps. Reported source levels from blue whales ranged from 179 to 189 dB re 1 µPa m (Cummings and Thompson, 1971; McDonald et al., 2001; Samaran et al., 2010; Širović et al., 2007; Thode et al., 2000). From Icelandic blue whales, the measured source levels are lower, 159 to 169 dB re 1 µPa (rms) (Akamatsu et al., 2014). We discussed the discrepancy in estimated source levels which can be explained by three factors, due to different used methodologies of source level estimation, local geographic conditions and local acoustic conditions.

Blue whales in Skjálfandi Bay, Northeast Iceland infrequently produced down sweep calls during feeding. The source level of these calls were 7 to 17 dB lower than the ones documented in other regions of the world. Blue whales show the ability to vary the intensity of their calls showing adaptation to the local environment and might make them more susceptible to show a Lombard effect.

Overall discussion

III. Anti-masking strategies by marine mammals Finally, we investigated if blue whales show a response towards anthropogenic sounds from marine vibrator (MV) signals (chapter 3). Little is known about this effect on baleen whales, especially on blue whales. Of all the ways in which noise can affect the lives of marine mammals, auditory masking is perhaps the most pervasive. Masking occurs when the ability to detect a sound of interest is obscured in the presence of another sound (Erbe et al., 2016). Marine mammals may alter the characteristics of their vocalisation sounds in the presence of noise, known as Lombard effect which includes not only changes in sound intensity but also other acoustic features such as changes in repetition rate, frequency shift, or increase of the signal duration (Lombard, 1911). Not only acoustic communication is at risk of masking (as in the present example), but also echolocation and the detection of environmental, predator and prey sounds (Erbe et al., 2016).

In order to investigate possible effects of the marine vibrator (MV) signal, we defined three different phases: pre-exposure, exposure and post-exposure phase. These controlled exposure experiments (CEEs) provide the best method of proving that a particular sound stimulus causes a response because a specific known dose of sound is broadcast to an animal, and the acoustic exposure and behavioural responses can be directly measured (Tyack, 2009). We developed tools to analyse call characteristics and compare parameters between experimental phases. Therefore, the number of blue whales down sweeps per hour was analysed. We also determined the signal duration and different frequency parameters such as minimum, maximum, centroid and peak frequency. The results of this chapter give evidence that elevated noise from MV signals has an effect on blue whale vocalisation behaviour. We observed changes in call rate and duration as well as changes in frequency that could be a consequence of the exposure from MV signals. But we did not observe changes in source levels.

Animals that communicate acoustically need to overcome the challenge of successfully transmitting signals in varying conditions of ambient noise (Kragh et al., 2019). Killer whales have been shown to raise the amplitude of their communication signals in the presence of ship noise (Holt et al., 2009, 2011); as have beluga whales (Scheifele et al., 2005). Humpback whales increased the source level of their songs proportionately to increases in wind-dependent ambient noise (Dunlop et al., 2014). Bottlenose dolphins were found to raise the amplitude of their outgoing echolocation clicks when masking noise was added to the environment (Au et

Overall discussion

al., 1974). A beluga whale adapted its echolocation clicks, shifting to higher intensities and higher frequencies after being moved to a different ambient noise environment (Au et al., 1985). Certain marine mammals have been reported to alter the frequency content of their calls in the presence of interfering noise, potentially as a way to minimise spectral overlap (reviewed in Erbe et al., 2016). Studies suggest that right whales have vocally adapted to environments with increased low-frequency noise through a shift in vocalisation frequency and duration (Parks et al., 2007, 2009, 2011a), which may have been a response to compensate for a loss in communication range (Clark et al., 2009). Right whales (Eubalaena glacialis) (Parks et al., 2007), beluga whales (Lesage et al., 1999), and common dolphins (Delphinus delphis) (Ansmann et al., 2007) apparently shifted the frequency of their calls away from the environmental frequencies containing the most noise energy. These observations, however, were not made in controlled conditions and more research is needed on the usage and effectiveness of this Lombard response in marine mammals. There is also evidence that odontocetes shift the main energy of their echolocation signals outside of the band of masking noise (Au et al., 1985, 1974; Moore and Pawloski, 1990), another observation that should lead to more systematic investigation of the extent to which frequency-shifting, as an anti-masking strategy, may occur in marine mammals. Increasing repetition rate or redundancy of calls reduces the potential for masking and improves the odds of detection by a listener. Many marine mammals display high redundancy in calling behaviour in typical conditions, presumably as a strategy for increased detectability by conspecifics. Increases in call repetition rates have been documented for bottlenose dolphins (Buckstaff, 2004) and beluga whales (Lesage et al., 1999) in the presence of ship noise. Furthermore, increases in call duration have been reported for killer whales in the presence of masking noise (Foote et al., 2004).

A hydrophone array has the advantage to record ambient noise and all marine mammals in the vicinity of the array to estimate source levels and other signal features received at the hydrophones. But with an array, however only a general statement on vocalisation behaviour and changes towards anthropogenic noise can be made. It is difficult to identify individuals and their possible reactions. Therefore, it is reasonable to also use animal-borne tags together with a hydrophone array. By means of tags, not only the vocalisation, but also diving behaviour and acceleration of an individual can be obtained to have clear proof of the animals respond to the

Overall discussion

introduced noise or not. However, no vocalisation of the tagged animals occurred during the whole experiment.

Conclusion

The aim of my thesis was to monitor different sound sources with a novel GPS-linked receiver array and to develop customised analytical solutions for targeted data processing of the recordings. The thesis consists of three chapters where we assessed the noise input caused by the construction of seed mussel collectors in the Wadden Sea, we studied the vocal behaviour of blue whales in Skjálfandi Bay, Northeast Iceland and verified if blue whales show a response towards anthropogenic sounds from marine vibrator (MV) signals.

The GPS-linked receiver array used for the investigations within the framework of this thesis can be used for many different marine mammals, from very low-frequency species like blue whales to high-frequency species like harbour porpoises (Phocoena phocoena) and dolphins. The customised R routines developed to detect and localise cetaceans as well as determine the characteristics of their vocalisation signals can be applied to analyse the vocal behaviour on every marine mammal species amenable.

Not only biological sound sources, but also anthropogenic sounds from any kind of offshore activities, such as drilling, dredging, pile-driving, seismic exploration and many more can be investigated. This study is an example for how an autonomous GPS-linked hydrophone array can be used as a tool to assess the effect of anthropogenic sound. Here, vibration sound caused by two different activities at sea was analysed. This type of sound can be either emitted by construction of anchor pipes or by seismic surveys for oil exploration, e.g., vibroseismic signals from marine vibrators. Our system can also be used to measure sound propagation from passing boats or other anthropogenic activities at sea.

We also provided simple propagation models useful and applicable for different sound types to analyse how these sounds radiate in certain areas of interest. Furthermore, it can be modified as a vertical array to measure the sound propagation with depth. Preliminary results revealed that in certain depth with a certain distance from a loudspeaker the sound is cancelled. This can happen due to constructive or destructive phase shift between direct path and surface reflection. This information might help to understand, if the whales might use these silent areas caused by the Lloyd’s mirror effect due to surface reflection.

One could use playbacks of sound to perform controlled measurements of sounds of interest from new or advanced techniques of industrial activities less harmful to the marine

Conclusion

environment. Thus, a sustainable use of the oceans could be ensured. Furthermore, a global use of PAM and the development of effective standardised statistical methods to analyse acoustical data on many cetacean species should increase our knowledge and help to understand their role in marine ecosystems (Mellinger et al., 2007).

Summary

Johannes Baltzer - Passive acoustic monitoring for evaluating potential impacts of anthropogenic noise on marine animals: Tools for ecological assessment and monitoring

In this thesis, we monitored different natural and anthropogenic sound sources with a novel GPS-linked receiver array and developed tailor made analytical solutions for targeted data processing of the recordings. We assessed the noise of anchor pipe vibration embedment operations during the construction of seed mussel collectors in the Wadden Sea (chapter 1). We studied the acoustic behaviour of blue whales (Balaenoptera musculus) in an Icelandic feeding ground (chapter 2) and we analysed how an anthropogenic sound source (marine vibrator signals) can affect the vocalisation of blue whales (chapter 3).

During this thesis we took advantage of an autonomous drifting lightweight portable passive acoustic array that can be deployed quickly and repeatedly from a small boat. We developed tools to estimate the sound propagation by a non-linear logarithmic regression by means of the intercept, slope and attenuation factor, which allowed us to evaluate the received sound levels that reach an animal in certain distances from the construction site of seed mussel collectors in the Wadden Sea. We provided tools for time-synchronisation, detection of blue whale vocalisations in Skjálfandi Bay, Northeast Iceland. We calculated general sound characteristics and compared them to recorded blue whale calls from other regions of the world. Eventually we used the GPS-linked receiver array to analyse how an anthropogenic sound source can affect the calling behaviour of blue whales. This was realised by a controlled exposure experiment. We looked at call characteristics of blue whale vocalisation, such as source level (SL), call rate, call duration and frequency parameters (minimum, maximum, centroid and peak frequency) as well as possible changes throughout the experimental phases. In addition, we determined ambient noise levels (NLs) within these phases.

Our study showed that the detected anchor pipe vibration embedment noise might exert a behavioural reaction on a local scale. Marine mammals could be affected by the construction operations up to a distance of 375 m and fish up to a distance of 766 m. These zones of responsiveness for vibration embedment operations are relatively small, compared to pile driving, which is regularly used during construction operations. Our study shows that it is important to monitor and assess any kind of noise introduction to verify, whether a sustainable

Summary

human use with respect to the complied guidelines is ensured without affecting the marine fauna. That is the first step to maintain a good environmental status as implemented in the MSFD.

We evaluated the call characteristics of blue whales in a feeding ground. The source level of the recorded down sweeps was 7 to 17 dB lower than the ones reported in other studies from different regions of the world. This finding suggests that blue whales show the ability to vary their call intensity in order to adapt to their local environment.

Our results showed that elevated noise from marine vibrator (MV) signals has an effect on blue whale vocalisation behaviour. We observed that the pre- and exposure phases showed highest NLs, probably caused by the MV signals and additional boat noise. The call rates increased from 2.0 to 6.2 down sweeps per hour and the down sweeps were lengthened two-fold (2.8 s) in the post-exposure phase, the quietest period of the experiment. This increase was probably caused by the preceding elevated noise in general. However, we could also see that the frequency parameters decreased as a result from the MV signals specifically. In contrast, no changes in source level occurred.

With the array we were able to make good recordings, coupled with advanced propagation estimates we were able to make predictions of impacts, evaluate the vocalisation behavior of blue whales and we could assess potential noise-induced reactions towards a controlled exposure to marine vibrator signals.

Zusammenfassung

Johannes Baltzer - Passiv akustisches Monitoring zur Bewertung möglicher Auswirkungen von anthropogenem Lärm auf die marine Fauna: Tools zur ökologischen Bewertung und zum Monitoring

In dieser Arbeit haben wir verschiedene natürliche und anthropogene Schallquellen mit einem neuartigen GPS-verbundenen Hydrophonarray aufgenommen und angepassten analytische Lösungen für die gezielte Datenverarbeitung der Aufnahmen entwickelt. Wir haben den Lärm beim Einvibrieren von Verankerungsrohren beim Bau von Saatmuschelgewinnungsanlagen im Wattenmeer untersucht (Kapitel 1). Wir haben das akustische Verhalten von Blauwalen (Balaenoptera musculus) in einem isländischen Nahrungsgebiet untersucht (Kapitel 2) und wir haben analysiert, wie eine anthropogene Schallquelle (Signale mariner Vibratoren) die Vokalisierung von Blauwalen beeinflussen kann (Kapitel 3).

Im Laufe dieser Thesis haben wir ein autonomes, leichtes, tragbares, passives, akustisches Array, das schnell und wiederholt von einem kleinen Boot aus eingesetzt werden kann, verwendet. Wir haben Tools entwickelt, um die Schallausbreitung durch eine nichtlineare logarithmische Regression anhand des Achsenabschnitts, der Steigung und des Dämpfungsfaktors abzuschätzen. So konnten wir die empfangenen Schallpegel, die ein Tier in bestimmten Entfernungen vom Baugebiet erreichen, bewerten. Wir haben Tools für die zeitliche Synchronisation und die Detektion von Blauwal-Vokalisierungen in der Skjálfandi Bucht, im Nordosten Islands, entwickelt. Wir haben die Charakteristika von Blauwallauten berechnet und sie mit aufgezeichneten Blauwallauten aus anderen Regionen der Welt verglichen. Das GPS-verbundene Hydrophonarray wurde weiterverwendet, um zu analysieren, wie eine anthropogene Schallquelle das akustische Verhalten von Blauwalen beeinflussen kann. Dies wurde durch ein kontrolliertes Expositionsexperiment umgesetzt. Wir haben die Charakteristika der Laute, wie z.B. Quellschallpegel, Dauer und Frequenzparameter des Signals bei der Vokalisierung von Blauwalen sowie mögliche Änderungen während der experimentellen Phasen untersucht. Zusätzlich haben wir den Umgebungslärmpegel innerhalb dieser Phasen bestimmt.

Unsere Studie zeigte, dass der detektierte Lärm beim Einvibrieren der Verankerungsrohre eine Verhaltensreaktion auf lokaler Ebene hervorrufen kann. Meeressäuger könnten bis zu einer

Zusammenfassung

Entfernung von 375 m und Fische bis zu einer Entfernung von 766 m von den Bauarbeiten betroffen sein. Die Bereiche, in denen eine Verhaltensreaktion durch das Einvibrieren hervorgerufen wurde, sind im Vergleich zu Rammverfahren, welches üblicherweise während Bauarbeiten verwendet werden, relativ klein. Unsere Studie zeigt, dass es wichtig ist, jede Art von Lärmbelastung zu kontrollieren und zu bewerten, um zu überprüfen, ob eine nachhaltige Nutzung durch den Menschen in Bezug auf die eingehaltenen Richtlinien gewährleistet ist, ohne die Meeresfauna zu beeinträchtigen. Dies ist der erste Schritt zur Aufrechterhaltung eines guten Umweltzustands, wie in der MSRL vorgesehen.

Wir haben die Charakteristika von Blauwallauten in einem Nahrungsgebiet. Der Quellschallpegel der aufgezeichneten Down-Sweeps war 7 bis 17 dB niedriger als in anderen Studien aus verschiedenen Regionen der Welt. Diese Erkenntnis legt nahe, dass Blauwale die Fähigkeit zeigen, die Intensität ihrer Laute zu variieren, um sich an ihre Umgebung anzupassen.

Unsere Ergebnisse zeigten, dass zunehmender Lärm, einen Einfluss auf das Vokalisierungsverhalten von Blauwalen hat. Wir stellten fest, dass die Pre- und Exposure- Phasen die höchsten Lärmpegel aufwiesen, was vermutlich auf die Signale des marinen Vibrators (MV) und zusätzlichen Schiffslärm zurückzuführen ist. Die Frequenz, in der Blauwale Laute ausgesendet haben, stieg von 2,0 auf 6,2 Down-Sweeps pro Stunde an. Down- Sweeps wurden in der Postexposure-Phase, der leisesten Phase des Experiments, um das Zweifache (2,8 s) verlängert. Dieser Anstieg wurde wahrscheinlich generell durch den erhöhten Lärmpegel verursacht. Wir konnten auch beobachten, wie sich die Frequenzen aufgrund der MV-Signale verringerten. Im Gegensatz dazu traten keine Änderungen des Quellschallpegels auf.

Mit dem Array konnten wir gute Aufzeichnungen machen, zusammen mit weiterentwickelten Schallausbreitungsmodellen konnten wir Auswirkungen von Lärm vorhersagen, das Vokalisierungsverhalten von Blauwalen bewerten und mögliche Reaktionen auf eine kontrollierte Exposition mit MV-Signalen bewerten.

Bibliography

Aarts, G., Brasseur, S., and Kirkwood, R. (2018). Behavioural response of grey seals to pile- driving Den Helder: Research report C006/18, 54 pages. Retrieved from https://library.wur.nl/WebQuery/wurpubs/fulltext/466039 Ainslie, M. A., Dahl, P. H., de Jong, C. A. F., and Laws, R. M. (2014). “Practical Spreading Laws : the Snakes and Ladders,” UA2014 - 2nd Int. Conf. Exhib. Underw. Acoust. (isl. Rhodes, Greece),. Akamatsu, T., Rasmussen, M. H., and Iversen, M. (2014). “Acoustically invisible feeding blue whales in Northern Icelandic waters.,” J. Acoust. Soc. Am., 136, 939–944. doi:10.1121/1.4887439 Akamatsu, T., Teilmann, J., Miller, L. A., Tougaard, J., Dietz, R., Wang, D., Wang, K., et al. (2007). “Comparison of echolocation behaviour between coastal and riverine porpoises,” Int. Symp. Underw. Technol. UT 2007 - Int. Work. Sci. Use Submar. Cables Relat. Technol. 2007, , doi: 10.1109/UT.2007.370755. doi:10.1109/UT.2007.370755 Akiyama, Y., Akamatsu, T., Rasmussen, M. H., Iversen, M. R., Iwata, T., Goto, Y., Aoki, K., et al. (2019). “Leave or stay? Video-logger revealed foraging efficiency of humpback whales under temporal change in prey density,” PLoS One, 14, 1–21. doi:10.1371/journal.pone.0211138 Amorim, M. C. P. (2006). “Diversity of Sound Production in Fish,” Commun. Fishes, 1, 71– 105. Retrieved from http://nova.wh.whoi.edu/palit/Amorim_2006_Communication in Fishes_Diversity of sound production in fish Amundin, M., and Cranford, T. (1990). “Forehead anatomy of Phocoena phocoena and Cephalorhynchus commersonii: 3-dimensional computer reconstructions with emphasis on the nasal diverticula,” In J. A. Thomas and R. A. Kastelein (Eds.), Sens. Abil. Cetaceans - Lab. F. Evid., pp. 1–18. Anderson, P. A., Berzins, I. K., Fogarty, F., Hamlin, H. J., and Guillette, L. J. (2011). “Sound, stress, and seahorses: The consequences of a noisy environment to animal health,” Aquaculture, 311, 129–138. doi:10.1016/j.aquaculture.2010.11.013 Andrew, R. K., Howe, B. M., Mercer, J. A., and Dzieciuch, M. A. (2002). “Ocean ambient sound: comparing the 1960s with the 1990s for a receiver off the California coast,” Acoust. Res. Lett. Online, 3, 65–70. doi:10.1121/1.1461915 Ansmann, I. C., Goold, J. C., Evans, P. G. H., Simmonds, M., and Keith, S. G. (2007). “Variation in the whistle characteristics of short-beaked common dolphins, Delphinus delphis, at two locations around the British Isles,” J. Mar. Biol. Assoc. United Kingdom, 87, 19–26. doi:10.1017/S0025315407054963 Au, W. W. L. (1993). The sonar of dolphins, Springer Verlag, New York. Au, W. W. L. (2004). “Dolphin sonar detection and discrimination capabilities,” J. Acoust. Soc. Am., 115, 2614–2614. doi:10.1121/1.4784793 Au, W. W. L., Floyd, R. W., Penner, R. H., and Murchison, A. E. (1974). “Measurement of echolocation signals of the Atlantic bottlenose dolphin, Tursiops truncatus Montagu, in

Bibliography

open waters,” J. Acoust. Soc. Am., 56, 1280–1290. Au, W. W. L., Kastelein, R. A., Rippe, T., and Schooneman, N. M. (1999). “Transmission beam pattern and echolocation signals of a harbor porpoise ( Phocoena phocoena ),” J. Acoust. Soc. Am., 106, 3699–3705. doi:10.1121/1.428221 Au, W. W. L., Mobley, J., Burgess, W. C., Lammers, M. O., and Nachtigall, P. E. (2000). “Seasonal and diurnal trends of chorusing humpback whales wintering in waters off western Maui,” Mar. Mammal Sci., 16, 530–544. doi:10.1111/j.1748- 7692.2000.tb00949.x Au, W. W., Penner, R. H., Carder, D. A., and Scronce, B. (1985). “Demonstration of adaptation in beluga whale echolocation signals,” J. Acoust. Soc. Am., 77, 726–730. doi:10.1121/1.392341 Bain, D. E., and Williams, R. (2006). Long-range effects of airgun noise on marine mammals: responses as a function of received sound level and distance Meet. Int. Whal. Comm. 26 May–6 June 2006, St. Kitts,SC58/E35,. Retrieved from https://tethys.pnnl.gov/sites/default/files/publications/Bain-and-Williams-2006.pdf Barlow, J., and Taylor, B. L. (2005). “Estimates of sperm whale abundance in the northeastern temperate Pacific from a combined acoustic and visual survey,” Mar. Mammal Sci., 21, 429–445. Baumgartner, M. F., Freitag, L., Partan, J., Ball, K. R., and Prada, K. E. (2008). “Tracking large marine predators in three dimensions: the real-time acoustic tracking system,” IEEE J. Ocean. Eng., 33, 146–157. doi:10.1109/JOE.2007.912496 Blackwell, S. B., Nations, C. S., McDonald, T. L., Thode, A. M., Mathias, D., Kim, K. H., Greene, C. R., et al. (2015). “Effects of airgun sounds on bowhead whale calling rates: Evidence for two behavioral thresholds,” PLoS One, 10, e0125720. doi:10.1371/journal.pone.0125720 Boyd, I., Brownell Jr, R. L., Cato, D. H., Clark, C. W., Costa, D. P., Evans, P. G. H., Gedamke, J., et al. (2008). “The effects of anthropogenic sound - A draft research strategy,” Eur. Sci. Found. Mar. Board, Position P, 96. Bradbury, J. W., and Vehrencamp, S. L. (2011). Principles of Animal Communication, Sinauer Associates, Sunderland, MA, 2nd ed., 697 pages pages. Brandt, M. J., Dragon, A.-C., Diederichs, A., Schubert, A., Kosarev, V., Nehls, G., Wahl, V., et al. (2016). Effects of offshore pile driving on harbour porpoise abundance in the German Bight Husum, 246 pages. Retrieved from http://bioconsult- sh.de/site/assets/files/1573/1573.pdf Brandt, M. J., Dragon, A. C., Diederichs, A., Bellmann, M. A., Wahl, V., Piper, W., Nabe- Nielsen, J., et al. (2018). “Disturbance of harbour porpoises during construction of the first seven offshore wind farms in Germany,” Mar. Ecol. Prog. Ser., 596, 213–232. doi:10.3354/meps12560 Branstetter, B. K., Bowman, V. F., Houser, D. S., Tormey, M., Banks, P., Finneran, J. J., and Jenkins, K. (2018). “Effects of vibratory pile driver noise on echolocation and vigilance in bottlenose dolphins (Tursiops truncatus),” J. Acoust. Soc. Am., 143, 429–439. doi:10.1121/1.5021555

100

Bibliography

Branstetter, B. K., Trickey, J. S., Bakhtiari, K., Black, A., Aihara, H., and Finneran, J. J. (2013). “Auditory masking patterns in bottlenose dolphins (Tursiops truncatus) with natural, anthropogenic, and synthesized noise.,” J. Acoust. Soc. Am., 133, 1811–1818. doi:10.1121/1.4789939 Bruintjes, R., Simpson, S. D., Harding, H., Bunce, T., Benson, T., Rossington, K., and Jones, D. (2016). “The impact of experimental impact pile driving on oxygen uptake in black seabream and plaice,” Proc. Meet. Acoust., 27, 010042. doi:10.1121/2.0000422 Buckstaff, K. C. (2004). “Effects of watercraft noise on the acoustic behavior of bottlenose dolphins, Tursiops truncatus, in Sarasota bay, Florida,” Mar. Mammal Sci., 20, 709–725. doi:10.1111/j.1748-7692.2004.tb01189.x Buscaino, G., Filiciotto, F., Buffa, G., Bellante, A., Stefano, V. Di, Assenza, A., Fazio, F., et al. (2010). “Impact of an acoustic stimulus on the motility and blood parameters of European sea bass (Dicentrarchus labrax L.) and gilthead sea bream (Sparus aurata L.),” Mar. Environ. Res., 69, 136–142. doi:10.1016/j.marenvres.2009.09.004 Carlström, J., Berggren, P., and Tregenza, N. J. C. (2009). “Spatial and temporal impact of pingers on porpoises,” Can. J. Fish. Aquat. Sci., 66, 72–82. doi:10.1139/F08-186 Carstensen, J., Henriksen, O. D., and Teilmann, J. (2006). “Impacts of offshore wind farm construction on harbour porpoises: Acoustic monitoring of echolocation activity using porpoise detectors (T-PODs),” Mar. Ecol. Prog. Ser., 321, 295–308. doi:10.3354/meps321295 Casper, B. M., Halvorsen, M. B., Matthews, F., Carlson, T. J., and Popper, A. N. (2013a). “Recovery of barotrauma injuries resulting from exposure to pile driving sound in two sizes of hybrid striped bass.,” PLoS One, 8, 1–12. doi:10.1371/journal.pone.0073844 Casper, B. M., Smith, M. E., Halvorsen, M. B., Sun, H., Carlson, T. J., and Popper, A. N. (2013b). “Effects of exposure to pile driving sounds on fish inner ear tissues,” Comp. Biochem. Physiol. - A Mol. Integr. Physiol., 166, 352–360. doi:10.1016/j.cbpa.2013.07.008 Casper, B. M., Smith, M. E., Halvorsen, M. B., Sun, H., Carlson, T. J., and Popper, A. N. (2013c). “Comparative Biochemistry and Physiology , Part A Effects of exposure to pile driving sounds on fish inner ear tissues,” Comp. Biochem. Physiol. Part A Mol. Integr. Physiol., 166, 352–360. doi:10.1016/j.cbpa.2013.07.008 Castellote, M., Clark, C. W., and Lammers, M. O. (2012). “Acoustic and behavioural changes by fin whales (Balaenoptera physalus) in response to shipping and airgun noise,” Biol. Conserv., 147, 115–122. doi:10.1016/j.biocon.2011.12.021 Celi, M., Filiciotto, F., Vazzana, M., Arizza, V., Maccarrone, V., Ceraulo, M., Mazzola, S., et al. (2015). “Shipping noise affecting immune responses of european spiny lobster (Palinurus elephas),” Can. J. Zool., 93, 113–121. doi:10.1139/cjz-2014-0219 Cerchio, S., Strindberg, S., Collins, T., Bennett, C., and Rosenbaum, H. (2014). “Seismic Surveys Negatively Affect Humpback Whale Singing Activity off Northern Angola.,” PLoS One, 9, e86464. doi:10.1371/journal.pone.0086464 Charif, R. A., Mellinger, D. K., Dunsmore, K. J., Fristrup, K. M., and Clark, C. W. (2002). “Estimated source levels of fin whale (Balaenoptera physalus) vocalizations: Adjustments

101

Bibliography

for surface interference,” Mar. Mammal Sci., 18, 81–98. Retrieved from http://onlinelibrary.wiley.com/doi/10.1111/j.1748-7692.2002.tb01020.x/abstract Clark, C. W. (1995). Application of US Navy underwater hydrophone arrays for scientific research on whales Reports Int. Whal. Comm., International Whaling Commission, Vol. 45, 210–212 pages. Clark, C. W., Charif, R., Mitchell, S., and Colby, J. (1996). “Distribution and behavior of the bowhead whale, Balaena mysticetus, based on analysis of acoustic data collected during the 1993 spring migration off Point Barrow, Alaska,” Forty-sixth Rep. Int. Whal. Comm.,. Clark, C. W., Ellison, W. T., Southall, B. L., Hatch, L., Van Parijs, S. M., Frankel, A., and Ponirakis, D. (2009). “Acoustic masking in marine ecosystems: intuitions, analysis, and implication,” Mar. Ecol. Prog. Ser., 395, 201–222. doi:10.3354/meps08402 Clausen, K. T., Wahlberg, M., Beedholm, K., DeRuiter, S., and Madsen, P. T. (2010). “Click communication in harbour porpoises Phocoena phocoena,” Bioacoustics, 20, 1–28. doi:10.1080/09524622.2011.9753630 Codarin, A., and Picciulin, M. (2015). “Underwater noise assessment in the Gulf of Trieste (Northern Adriatic Sea, Italy) using an MSFD approach,” Mar. Pollut. Bull., 101, 694– 700. doi:10.1016/j.marpolbul.2015.10.028 Cummings, W. C., and Thompson, P. O. (1971). “Underwater sounds from the blue whale, Balaenoptera musculus,” J. Acoust. Soc. Am., 50, 1193–1198. Curé, C., Isojunno, S., Visser, F., Wensveen, P., Sivle, L., Kvadsheim, P., Lam, F., et al. (2016). “Biological significance of sperm whale responses to sonar: comparison with anti-predator responses,” Endanger. Species Res., 31, 89–102. doi:10.3354/esr00748 CWSS (2010). “Wadden Sea Plan 2010,” 11th Trilateral Gov. Conf. Prot. Wadden Sea,. CWSS (2017). “Introduction,” Wadden Sea Qual. Status Rep. 2017,. Dahlheim, M. E., Fisher, D. H., and Schempp, J. D. (1984). “Sound production by the gray whale and ambient noise levels in Laguna San Ignacio, Baja California Sur, Mexico,” In M. Lou Jones, S. L. Swartz, and S. Leatherwood (Eds.), gray whale Eschrichtius robustus, Academic Press Inc., Orlando, Florida, pp. 511–541. Dähne, M., Gilles, A., Lucke, K., Peschko, V., Adler, S., Krügel, K., Sundermeyer, J., et al. (2013). “Effects of pile-driving on harbour porpoises (Phocoena phocoena) at the first offshore wind farm in Germany,” Environ. Res. Lett., 8, 025002. doi:10.1088/1748- 9326/8/2/025002 Dähne, M., Tougaard, J., Carstensen, J., Rose, A., and Nabe-Nielsen, J. (2017). “Bubble curtains attenuate noise from offshore wind farm construction and reduce temporary habitat loss for harbour porpoises,” Mar. Ecol. Prog. Ser., 580, 221–237. doi:10.3354/meps12257 DeRuiter, S. L., Bahr, A., Blanchet, M.-A., Hansen, S. F., Kristensen, J. H., Madsen, P. T., Tyack, P. L., et al. (2009). “Acoustic behaviour of echolocating porpoises during prey capture.,” J. Exp. Biol., 212, 3100–3107. doi:10.1242/jeb.030825 DeRuiter, S. L., Tyack, P. L., Lin, Y.-T., Newhall, A. E., Lynch, J. F., and Miller, P. J. O. (2006). “Modeling acoustic propagation of airgun array pulses recorded on tagged sperm

102

Bibliography

whales (Physeter macrocephalus),” J. Acoust. Soc. Am., 120, 4100–4114. doi:10.1121/1.2359705 Duncan, A. J., Weilgart, L. S., Leaper, R., Jasny, M., and Livermore, S. (2017). “A modelling comparison between received sound levels produced by a marine Vibroseis array and those from an airgun array for some typical seismic survey scenarios,” Mar. Pollut. Bull., 119, 277–288. doi:10.1016/J.MARPOLBUL.2017.04.001 Dunlop, R. A., Cato, D. H., and Noad, M. J. (2014). “Evidence of a Lombard response in migrating humpback whales (Megaptera novaeangliae),” J. Acoust. Soc. Am., 136, 430– 437. doi:10.1121/1.4883598 Dyndo, M., Wiśniewska, D. M., Rojano-Doñate, L., and Madsen, P. T. (2015). “Harbour porpoises react to low levels of high frequency vessel noise,” Sci. Rep., 5, 11083. doi:10.1038/srep11083 Elliott, M., Whitfield, A. K., Potter, I. C., Blaber, S. J. M., Cyrus, D. P., Nordlie, F. G., and Harrison, T. D. (2007). “The guild approach to categorizing estuarine fish assemblages: a global review,” Fish Fish., 8, 241–268. Erbe, C., Dunlop, R., and Dolman, S. (2018). “Effects of noise on marine mammals,” In H. Slabbekoorn, A. N. Popper, R. J. Dooling, and R. R. Fay (Eds.), Eff. Anthropog. Noise Anim., Springer, New York, pp. 277–309. doi:10.1007/978-1-4939-8574-6 Erbe, C., Marley, S. A., Schoeman, R. P., Smith, J. N., Trigg, L. E., and Embling, C. B. (2019). “The Effects of Ship Noise on Marine Mammals—A Review,” Front. Mar. Sci., , doi: 10.3389/fmars.2019.00606. doi:10.3389/fmars.2019.00606 Erbe, C., Reichmuth, C., Cunningham, K., Lucke, K., and Dooling, R. (2016). “Communication masking in marine mammals: A review and research strategy,” Mar. Pollut. Bull., , doi: 10.1016/j.marpolbul.2015.12.007. doi:10.1016/j.marpolbul.2015.12.007 European Union (2008). “Directive 2008/56/EC of the European Parliament and of the Council of 17 June 2008, establishing a framework for community action in the field of marine environmental policy (Marine Strategy Framework Directive),” Off. J. Eur. Union,. Faulkner, R. C., Farcas, A., and Merchant, N. D. (2018). “Guiding principles for assessing the impact of underwater noise,” (M. González-Suárez, Ed.) J. Appl. Ecol., , doi: 10.1111/1365-2664.13161. doi:10.1111/1365-2664.13161 Filadelfo, R., Mintz, J., Michlovich, E., D’Amico, A., Tyack, P. L., and Ketten, D. R. (2009). “Correlating Military Sonar Use with Beaked Whale Mass Strandings: What Do the Historical Data Show?,” Aquat. Mamm., 35, 435–444. doi:10.1578/AM.35.4.2009.435 Foote, A. D., Osborne, R. W., and Hoelzel, A. R. (2004). “Whale-call response to masking boat noise,” Nature, 428, 910. François, R. E., and Garrison, G. R. (1982). “Sound absorption based on ocean measurements. Part II: Boric acid contribution and equation for total absorption,” J. Acoust. Soc. Am., 72, 1879–1890. doi:10.1121/1.388673 Frantzis, A. (1998). “Does acoustic testing strand whales?,” Nature, 392, 29–29. doi:10.1038/32068 Gillespie, D. (2004). “Detection and classification of right whale calls using an ‘edge’ detector

103

Bibliography

operating on a smoothed spectrogram,” Can. Acoust. - Acoust. Can., 32, 39–47. Gillespie, D., and Chappell, O. (2002). “An automatic system for detecting and classifying the vocalisations of harbour porpoises,” Bioacoustics, 13, 37–61. doi:10.1080/09524622.2002.9753485 Gillespie, D., Gordon, J., Mchugh, R., Mclaren, D., Mellinger, D. K., Redmond, P., Thode, A., et al. (2008). “PAMGUARD: Semiautomated, open source software for real-time acoustic detection and localisation of cetaceans,” Proc. Inst. Acoust.,. Gordon, J. C. D., Gillespie, D., Potter, J., Frantzis, A., Simmonds, M. P., and Swift, R. (2003). “Investigating the effects of man made noise on marine mammals,” Mar. Technol. Soc. J., 37, 16–34. doi:http://dx.doi.org/10.4031/002533203787536998 Gordon, J., and Tyack, P. L. (2002). “Acoustic techniques for studying cetaceans,” In P. G. H. Evans and J. A. Raga (Eds.), Mar. Mamm. Biol. Conserv., Kluwer Academic Publishers, New York, pp. 293–324. Gospić, N. R., and Picciulin, M. (2016). “Changes in whistle structure of resident bottlenose dolphins in relation to underwater noise and boat traffic,” Mar. Pollut. Bull., 105, 193– 198. doi:10.1016/j.marpolbul.2016.02.030 Govoni, J. J., West, M. A., Settle, L. R., Lynch, R. T., and Greene, M. D. (2008). “Effects of Underwater Explosions on Larval Fish: Implications for a Coastal Engineering Project,” J. Coast. Res., 24, 228–233. doi:10.2112/05-0518.1 van der Graaf, A. J., Ainslie, M. A., André, M., Brensing, K., Dalen, J., Dekeling, R. P. A., Robinson, S. P., et al. (2012). “European Marine Strategy Framework Directive Good Environmental Status (MSFD-GES),” Rep. Tech. Subgr. Underw. noise other forms energy, , doi: 10.1007/s11065-009-9119-9. doi:10.1007/s11065-009-9119-9 Graham, I. M., Merchant, N. D., Farcas, A., Barton, T. R., Cheney, B., Bono, S., and Thompson, P. M. (2019). “Harbour porpoise responses to pile-driving diminish over time,” R. Soc. Open Sci., 6, 190335. doi:10.1098/rsos.190335 Guazzo, R. A., Helble, T. A., Alongi, G. C., Durbach, I. N., Martin, C. R., Martin, S. W., Henderson, E. E., et al. (2020). “ocean noise levels increase The Lombard effect in singing humpback whales : Source levels increase as ambient ocean noise levels increase,” J. Acoust. Soc. Am., 148, 542–555. doi:10.1121/10.0001669 Hafner, S. D., and Katz, J. (2018). “monitoR: Acoustic template detection in R.,” Retrieved from http://www.uvm.edu/rsenr/vtcfwru/R/?Page=monitoR/monitoR.htm Halpern, B. S., Frazier, M., Potapenko, J., Casey, K. S., Koenig, K., Longo, C., Lowndes, J. S., et al. (2015). “Spatial and temporal changes in cumulative human impacts on the world’s ocean,” Nat. Commun., 6, 7615. doi:10.1038/ncomms8615 Halvorsen, M. B., Casper, B. M., Matthews, F., Carlson, T. J., and Popper, A. N. (2012a). “Effects of exposure to pile-driving sounds on the lake sturgeon, Nile tilapia and hogchoker,” Proceeding R. Soc. B Biol. Sci., 279, 4705–4714. doi:10.1098/rspb.2012.1544 Halvorsen, M. B., Casper, B. M., Woodley, C. M., Carlson, T. J., and Popper, A. N. (2012b). “Threshold for Onset of Injury in Chinook Salmon from Exposure to Impulsive Pile

104

Bibliography

Driving Sounds,” PLoS One, 7, e38968. doi:10.1371/journal.pone.0038968 Hardy, T., Williams, R., Caslake, R., and Tregenza, N. (2012). “An investigation of acoustic deterrent devices to reduce cetacean bycatch in an inshore set net fishery,” J. Cetacean Res. Manag., 12, 85–90. Hatch, L. T., Clark, C. W., van Parijs, S. M., Frankel, A. S., and Ponirakis, D. W. (2012). “Quantifying loss of acoustic communication space for right whales in and around a U.S. National Marine Sanctuary.,” Conserv. Biol., 26, 983–994. doi:10.1111/j.1523- 1739.2012.01908.x Hawkins, A. D., and Rasmussen, K. J. (1978). “The calls of gadoid fish,” J. Mar. Biol. Assoc. United Kingdom, 58, 891–911. doi:10.1017/S0025315400056848 Hawkins, A. D., Roberts, L., and Cheesman, S. (2014). “Responses of free-living coastal pelagic fish to impulsive sounds,” J. Acoust. Soc. Am., 135, 3101–3116. doi:10.1121/1.4870697 Heiler, J., Elwen, S. H., Kriesell, H. J., and Gridley, T. (2016). “Changes in bottlenose dolphin whistle parameters related to vessel presence, surface behaviour and group composition,” Anim. Behav., 117, 167–177. doi:10.1016/j.anbehav.2016.04.014 Heinis, F. (2013). Offshore windpark Luchterduinen: effecten van aanleg op zeezoogdieren. unpublished report by HWE. Helble, T. A., Guazzo, R. A., Martin, C. R., Durbach, I. N., Alongi, G. C., Martin, S. W., Boyle, J. K., et al. (2020). “ocean noise Lombard effect : Minke whale boing call source levels vary with natural variations in ocean noise a ),” J. Acoust. Soc. Am., , doi: 10.1121/10.0000596. doi:10.1121/10.0000596 Helble, T. a, Ierley, G. R., D’Spain, G. L., and Martin, S. W. (2015). “Automated acoustic localization and call association for vocalizing humpback whales on the Navy’s Pacific Missile Range Facility,” J. Acoust. Soc. Am., 137, 11–21. doi:10.1121/1.4904505 Herbert-Read, J. E., Kremer, L., Bruintjes, R., Radford, A. N., and Ioannou, C. C. (2017). “Anthropogenic noise pollution from pile-driving disrupts the structure and dynamics of fish shoals,” Proc. R. Soc. B Biol. Sci., 284, 20171627. doi:10.1098/rspb.2017.1627 Hijmans, R. J. (2019). “geosphere: Spherical Trigonometry.,” Retrieved from https://cran.r- project.org/package=geosphere Hild, A. (1999). “Study Area: The Backbarrier Tidal Flats of Spiekeroog,” In S. Dittmann (Ed.), Wadden Sea Ecosyst., Springer Berlin Heidelberg, Heidelberg, 1st ed. doi:10.1007/978- 3-642-60097-5 Hildebrand, J. A. (2009). “Anthropogenic and natural sources of ambient noise in the ocean,” Mar. Ecol. Prog. Ser., 395, 5–20. doi:10.3354/meps08353 Holt, D. E., and Johnston, C. E. (2014). “Evidence of the Lombard effect in fishes,” Behav. Ecol., 25, 819–826. doi:10.1093/beheco/aru028 Holt, M. M., Noren, D. P., Dunkin, R. C., and Williams, T. M. (2015). “Vocal performance affects metabolic rate in dolphins: implications for animals communicating in noisy environments,” J. Exp. Biol., 2, 1647–1654. doi:10.1242/jeb.122424

105

Bibliography

Holt, M. M., Noren, D. P., and Emmons, C. K. (2011). “Effects of noise levels and call types on the source levels of killer whale calls.,” J. Acoust. Soc. Am., 130, 3100–3106. doi:10.1121/1.3641446 Holt, M. M., Noren, D. P., Veirs, V., Emmons, C. K., and Veirs, S. (2009). “Speaking up: Killer whales (Orcinus orca) increase their call amplitude in response to vessel noise,” J. Acoust. Soc. Am., 125, 27–32. doi:10.1121/1.3040028 International Organization for Standardization (2016). Acoustics — Description, measurement and assessment of environmental noise (ISO 1996-1:2016). Retrieved from https://www.iso.org/obp/ui/#iso:std:iso:1996:-1:ed-3:v1:en Di Iorio, L., and Clark, C. W. (2010). “Exposure to seismic survey alters blue whale acoustic communication,” Biol. Lett., 6, 51–54. doi:10.1098/rsbl.2009.0651 ISO 18406 (2017). Underwater acoustics — Measurement of radiated underwater sound from percussive pile driving,. Retrieved from https://www.iso.org/obp/ui/#iso:std:iso:18406:ed-1:v1:en Isojunno, S., Curé, C., Kvadsheim, P. H., Lam, F. P. A., Tyack, P. L., Wensveen, P. J., and Miller, P. J. O. M. (2016). “Sperm whales reduce foraging effort during exposure to 1-2 kH z sonar and killer whale sounds,” Ecol. Appl., 26, 77–93. doi:10.1890/15- 0040.1/suppinfo Iversen, M., Rasmussen, M. H., Magnúsdóttir, E. E., and Lammers, M. O. (2011). “Recordings of blue whale sound with single hydrophone and ecological acoustical recorders (EARs),” 19th Bienn. Conf. by Soc. Mar. Mammology, Tampa, Florida, US. Iversen, M., Rasmussen, M. H., Sigurðardóttur, V., and Örnólfsdóttir, E. B. (2010). “Recordings of blue whales (B. musculus) in Skjálfandi Bay, Iceland,” 24th Annu. Conf. Eur. Cetacean Soc. 06th - 7th March 2010, Stralsund, Germany. Janik, V. M. (2000). “Source levels and the estimated active space of bottlenose dolphin (Tursiops truncatus) whistles in the Moray Firth, Scotland.,” J. Comp. Physiol. A, 186, 673–680. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11016783 Jensen, F. H., Bejder, L., Wahlberg, M., Aguilar Soto, N., Johnson, M., and Madsen, P. T. (2009). “Vessel noise effects on delphinid communication,” Mar. Ecol. Prog. Ser., 395, 161–175. doi:10.3354/meps08204 Jensen, L. F., Teilmann, J., Galatius, A., Pund, R., Czeck, R., Jess, A., Siebert, U., et al. (2017). “Marine mammals,” In S. et al. Kloepper (Ed.), Wadden Sea Qual. Status Rep. 2017, Common Wadden Sea Secretariat (CWSS), Wilhelmshaven, Germany. Retrieved from qsr.waddensea-worldheritage.org/reports/marine-mammals Jochens, A., Biggs, D., Benoit-Bird, K. J., Engelhaupt, D., Gordon, J., Hu, C., Jaquet, N., et al. (2008). Sperm whale seismic study in the Gulf of Mexico: Synthesis report. Johnson, M. P., and Tyack, P. L. (2003). “A digital acoustic recording tag for measuring the response of wild marine mammals to sound,” IEEE J. Ocean. Eng.,. doi:10.1109/JOE.2002.808212 de Jong, K., Amorim, M. C. P., Fonseca, P. J., and Heubel, K. U. (2018). “Noise Affects Multimodal Communication During Courtship in a Marine Fish,” Front. Ecol. Evol., 6, 1–

106

Bibliography

8. doi:10.3389/fevo.2018.00113 de Jong, K., Amorim, M. C. P., Fonseca, P. J., Klein, A., and Heubel, K. U. (2016). “Noise affects acoustic courtship behavior similarly in two species of gobies,” Proc. Meet. Acoust., Fourth International Conference on the Effects of Noise on Aquatic Life, Dublin, 1–10. doi:10.1121/2.0000272 Kastak, D., Reichmuth, C., Holt, M. M., Mulsow, J., Southall, B. L., and Schusterman, R. J. (2007). “Onset , growth , and recovery of in-air temporary threshold shift in a California sea lion (Zalophus californianus),” J. Acoust. Soc. Am., 122, 2916–2924. doi:10.1121/1.2783111 Kastak, D., Southall, B. L., Schusterman, R. J., and Reichmuth Kastak, C. (2005). “Underwater temporary threshold shift in pinnipeds: Effects of noise level and duration,” J. Acoust. Soc. Am., 118, 3154–3163. doi:10.1121/1.2047128 Kastelein, R. A., van Heerden, D., Gransier, R., and Hoek, L. (2013a). “Behavioral responses of a harbor porpoise (Phocoena phocoena) to playbacks of broadband pile driving sounds.,” Mar. Environ. Res., 92, 206–214. doi:10.1016/j.marenvres.2013.09.020 Kastelein, R. A., Helder-Hoek, L., Covi, J., and Gransier, R. (2016). “Pile driving playback sounds and temporary threshold shift in harbor porpoises (Phocoena phocoena): Effect of exposure duration,” J. Acoust. Soc. Am., 139, 2842–2851. doi:10.1121/1.4948571 Kastelein, R. A., Heul, S. van der, Verboom, W. C., Jennings, N., Veen, J. van der, and Haan, D. de (2008). “Startle response of captive North Sea fish species to underwater tones between 0.1 and 64 kHz,” Mar. Environ. Res., 65, 369–377. doi:10.1016/j.marenvres.2008.01.001 Kastelein, R. A., Hoek, L., Gransier, R., and Jennings, N. (2013b). “Hearing thresholds of two harbor seals (Phoca vitulina) for playbacks of multiple pile driving strike sounds.,” J. Acoust. Soc. Am., 134, 2307–2312. doi:10.1121/1.4817889 Kastelein, R. A., Jennings, N., Kommeren, A., Helder-Hoek, L., and Schop, J. (2017). “Acoustic dose-behavioral response relationship in sea bass (Dicentrarchus labrax) exposed to playbacks of pile driving sounds,” Mar. Environ. Res., 130, 315–324. doi:10.1016/j.marenvres.2017.08.010 Kastelein, R. A., Steen, N., Gransier, R., and de Jong, C. A. F. (2013c). “Brief behavioral response threshold level of a harbor porpoise (Phocoena phocoena) to an impulsive sound,” Aquat. Mamm., 39, 315–323. doi:10.1578/AM.39.4.2013.315 Kindt-Larsen, L., Berg, C. W., Northridge, S., and Larsen, F. (2019). “Harbor porpoise (Phocoena phocoena) reactions to pingers,” Mar. Mammal Sci., 35, 552–573. doi:10.1111/mms.12552 Kipple, B., and Gabriele, C. (2003). Glacier Bay watercraft noise Technical Report NSWCCDE-71-TR-2003/522, prepared for Glacier Bay National Park and Preserve, Naval Surface Warfare Center Bremerton, WA, 1–54 pages. Kipple, B., and Gabriele, C. (2004). Glacier Bay watercraft noise — noise characterization for tour, charter, private, and government vessels. Technical Report NSWCCDE-71-TR- 2004/545, prepared for Glacier Bay National Park and Preserve, Naval Surface Warfare Center Bremerton, WA, 1–47 pages.

107

Bibliography

Kirsebom, O. S., Frazao, F., Simard, Y., Roy, N., Matwin, S., and Giard, S. (2020). “Performance of a deep neural network at detecting North Atlantic right whale upcalls,” J. Acoust. Soc. Am., 147, 2636–2646. doi:10.1121/10.0001132 Kragh, I. M., McHugh, K., Wells, R. S., Sayigh, L. S., Janik, V. M., Tyack, P. L., and Jensen, F. H. (2019). “Signal-specific amplitude adjustment to noise in common bottlenose dolphins (Tursiops truncatus),” J. Exp. Biol., , doi: 10.1242/jeb.216606. doi:10.1242/jeb.216606 Kunc, H. P., McLaughlin, K. E., and Schmidt, R. (2016). “Aquatic noise pollution: Implications for individuals, populations, and ecosystems,” Proc. R. Soc. B Biol. Sci., 283, 20160839. doi:10.1098/rspb.2016.0839 Küsel, E. T., Mellinger, D., Thomas, L., Marques, T. A., Moretti, D. J., and Ward, J. (2009). “Beaked whale density estimation from single hydrophones by means of propagation modeling.,” J. Acoust. Soc. Am., 125, 2589–2589. doi:10.1121/1.4783840 Kyhn, L. A., Tougaard, J., Thomas, L., Duve, R. L., Stenback, J., Amundin, M., Desportes, G., et al. (2012). “From echolocation clicks to animal density — Acoustic sampling of harbor porpoises with static dataloggers,” J. Acoust. Soc. Am., 131, 550–560. doi:10.1121/1.3662070 Kyhn, L., Jørgensen, P., Carstensen, J., Bech, N., Tougaard, J., Dabelsteen, T., and Teilmann, J. (2015). “Pingers cause temporary habitat displacement in the harbour porpoise Phocoena phocoena,” Mar. Ecol. Prog. Ser., 526, 253–265. doi:10.3354/meps11181 Ladich, F. (2019). “Ecology of sound communication in fishes,” Fish Fish., 20, 552–563. doi:10.1111/faf.12368 Ladich, F., and Fay, R. R. (2013). Auditory evoked potential audiometry in fish, Rev. Fish Biol. Fish., Vol. 23, 317–364 pages. doi:10.1007/s11160-012-9297-z Lagardère, J. P. (1982). “Effects of noise on growth and reproduction of Crangon crangon in rearing tanks,” Mar. Biol., 71, 177–185. Retrieved from http://link.springer.com/article/10.1007/BF00394627 Latusek-Nabholz, J. N., Whitt, A. D., Ferti, D., Gallien, D. R., Khan, A. A., and Sidorovskaia, N. A. (2020). Literature Synthesis: Passive Acoustic Monitoring Projects and Sound Sources in the Gulf of Mexico OCS Study, 99 pages. Lesage, V., Barrette, C., Kingsley, M. C. S., and Sjare, B. (1999). “The effect of vessel noise on the vocal behavior of belugas in the St. Lawrence River estuary, Canada,” Mar. Mammal Sci., 15, 65–84. doi:10.1111/j.1748-7692.1999.tb00782.x Leunissen, E. M., and Dawson, S. M. (2018). “Underwater noise levels of pile-driving in a New Zealand harbour, and the potential impacts on endangered Hector’s dolphins,” Mar. Pollut. Bull., 135, 195–204. doi:10.1016/j.marpolbul.2018.07.024 Ligges, U., Krey, S., Mersmann, O., and Schnackenberg, S. (2016). “tuneR: Analysis of music.,” Ligges, U., Krey, S., Mersmann, O., and Schnackenberg, S. (2018). “tuneR: Analysis of music and speech.,” Retrieved from https://cran.r-project.org/package=tuneR Lindström, K., and Lugli, M. (2000). “A quantitative analysis of the courtship acoustic

108

Bibliography

behaviour and sound patterning in male sand goby, Pomatoschistus minutus,” Environ. Biol. Fishes, 58, 411–424. doi:10.1023/A:1007695526177 Lombard, É. (1911). “Le signe de l’élévation de la voix.,” Ann. des Mal. l’oreille, du larynx, 37th ed., pp. 101–119. Lopatka, M., Adam, O., Laplanche, C., Zarzycki, J., and Motsch, J.-F. (2005). “An Attractive Alternative for Sperm Whale Click Detection Using the Wavelet Transform in Comparison to the Fourier Spectrogram,” Aquat. Mamm., 31, 463–467. doi:10.1578/am.31.4.2005.463 Lucke, K., Siebert, U., Lepper, P. A., and Blanchet, M.-A. (2009). “Temporary shift in masked hearing thresholds in a harbor porpoise (Phocoena phocoena) after exposure to seismic airgun stimuli.,” J. Acoust. Soc. Am., 125, 4060–4070. doi:10.1121/1.3117443 Luczkovich, J. J., Krahforst, C. S., Kelly, K. E., and Sprague, M. W. (2016). “The Lombard effect in fishes: How boat noise impacts oyster toadfish vocalization amplitudes in natural experiments,” Proc. Meet. Acoust., 010035. doi:10.1121/2.0000340 Lurton, X. (2010). An introduction to underwater acoustics: principles and applications, Springer, New York, 680 p pages. doi:10.3397/1.3525563 Madsen, P. T. (2005). “Marine mammals and noise: Problems with root mean square sound pressure levels for transients,” J. Acoust. Soc. Am., 117, 3952. doi:10.1121/1.1921508 Madsen, P. T., and Wahlberg, M. (2007). “Recording and quantification of ultrasonic echolocation clicks from free-ranging toothed whales,” Deep Sea Res. Part I Oceanogr. Res. Pap., 54, 1421–1444. doi:10.1016/j.dsr.2007.04.020 Marley, S. A., Salgado Kent, C. P., Erbe, C., and Parnum, I. M. (2017). “Effects of vessel traffic and underwater noise on the movement, behaviour and vocalisations of bottlenose dolphins in an urbanised estuary,” Sci. Rep., 7, 1–14. doi:10.1038/s41598-017-13252-z McDonald, M. A. (2004). “Difar hydrophone usage in whale research,” Can. Acoust. - Acoust. Can., 32, 155–160. McDonald, M. A., Calambokidis, J., Teranishi, A. M., and Hildebrand, J. A. (2001). “The acoustic calls of blue whales off California with gender data,” J. Acoust. Soc. Am., 109, 1728–1735. doi:10.1121/1.1353593 McDonald, M. A., Hildebrand, J. A., and Mesnick, S. (2009). “Worldwide decline in tonal frequencies of blue whale songs,” Endanger. Species Res., 9, 13–21. doi:10.3354/esr00217 McDonald, M. A., Mesnick, S. L., and Hildebrand, J. A. (2006). “Biogeographic characterization of blue whale song worldwide: Using song to identify populations,” J. Cetacean Res. Manag., 8, 55–66. Melcón, M. L., Cummins, A. J., Kerosky, S. M., Roche, L. K., Wiggins, S. M., and Hildebrand, J. A. (2012). “Blue whales respond to anthropogenic noise,” PLoS One, 7, 1–6. doi:10.1371/journal.pone.0032681 Mellinger, D., and Barlow, J. (2003). Future Directions for Acoustic Marine Mammal Surveys: Stock Assessment and Habitat Use: Report of a Workshop Held in La Jolla, California, 20-22 Nov 2002 Fish. Sci., Contribution 2557 from NOAA/Pacific Marine Environmental Laboratory, La Jolla, CA, USA, 45 pages. Retrieved from

109

Bibliography

http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:Future+Directions+for +Acoustic+Marine+Mammal+Surveys+:+Stock+Assessment+and+Habitat+Use+Report +of+a+Workshop+Held+in+La+Jolla+,+California+,#1 Mellinger, D. K., and Clark, C. W. (1997). “Methods for automatic detection of mysticete sounds,” Mar. Freshw. Behav. Physiol., 29, 163–181. doi:10.1080/10236249709379005 Mellinger, D. K., and Clark, C. W. (2000). “Recognizing transient low-frequency whale sounds by spectrogram correlation,” J. Acoust. Soc. Am., 107, 3518–3529. doi:10.1121/1.429434 Mellinger, D. K., and Clark, C. W. (2003). “Blue whale (Balaenoptera musculus) sounds from the North Atlantic,” J. Acoust. Soc. Am., 114, 1108. doi:10.1121/1.1593066 Mellinger, D. K., Stafford, K. M., and Fox, C. G. (2004a). “Seasonal occurence of sperm whale (Physeter macrocephalus) sounds in the Gulf of Alaska, 1999-2001,” Mar. Mammal Sci., 20, 48–62. Mellinger, D. K., Stafford, K. M., Moore, S. E., Munger, L., and Fox, C. G. (2004b). “Detection of North Pacific right whale (Eubalaena japonica) calls in the Gulf of Alaska,” Mar. Mammal Sci., 20, 872–879. doi:10.1111/j.1748-7692.2004.tb01198.x Mellinger, D., Stafford, K., Moore, S., Dziak, R., and Matsumoto, H. (2007). “An Overview of Fixed Passive Acoustic Observation Methods for Cetaceans,” Oceanography, 20, 36–45. doi:10.5670/oceanog.2007.03 Møhl, B., and Andersen, S. (1973). “Echolocation: high‐frequency component in the click of the Harbour Porpoise ( Phocoena ph. L. ),” J. Acoust. Soc. Am., 54, 1368–1372. doi:10.1121/1.1914435 Møhl, B., Surlykke, A., and Miller, L. A. (1990). “High intensity narwhal clicks,” In J. A. Thomas and R. A. Kastelein (Eds.), Sens. Abil. Cetaceans - Lab. F. Evid., pp. 295–303. Møhl, B., Wahlberg, M., and Heerfordt, A. (2001). “A large-aperture array of nonlinked receivers for acoustic positioning of biological sound sources.,” J. Acoust. Soc. Am., 109, 434–437. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11206176 Møhl, B., Wahlberg, M., Madsen, P. T., Heerfordt, A., and Lund, A. (2003). “The monopulsed nature of sperm whale clicks,” J. Acoust. Soc. Am., 114, 1143. doi:10.1121/1.1586258 Møhl, B., Wahlberg, M., Madsen, P. T., Miller, L. A., and Surlykke, A. (2000). “Sperm whale clicks: Directionality and source level revisited,” J. Acoust. Soc. Am., 107, 638–648. Mooney, T. A., Nachtigall, P. E., and Vlachos, S. (2009). “Sonar-induced temporary hearing loss in dolphins,” Biol. Lett., 5, 565–567. doi:10.1098/rsbl.2009.0099 Moore, K. A., Miller, J. H., Potty, G. R., Lynch, J. F., and Newhall, A. (2007). “Investigation of three‐dimensional propagation effects at the New Jersey shelf break front,” J. Acoust. Soc. Am., 121, 3126–3126. doi:10.1121/1.4808487 Moore, P. W. B., and Pawloski, D. A. (1990). “Investigations on the control of echolocation pulses in the dolphin Tursiops truncatus,” In J. A. Thomas and R. A. Kastelein (Eds.), Sens. Abil. Cetaceans - Lab. F. Evid., Plenum Press, New York, pp. 305–319. Moore, S. E., Stafford, K. M., Mellinger, D. K., and Hildebrand, J. A. (2006). “Listening for Large Whales in the Offshore Waters of Alaska,” Bioscience, 56, 49. doi:10.1641/0006-

110

Bibliography

3568(2006)056[0049:LFLWIT]2.0.CO;2 Morisaka, T., Shinohara, M., Nakahara, F., and Akamatsu, T. (2005). “Effects of ambient noise on the whistles of Indo-Pacific bottlenose dolphin populations,” J. Mammal., 86, 541–546. doi:10.1644/1545-1542(2005)86%255B541%253AEOANO Mueller-Blenkle, C., McGregor, P. K., Gill, A. B., Andersson, M. H., Metcalfe, J., Bendall, V., Sigray, P., et al. (2010). Effects of Pile-Driving Noise on the Behaviour of Marine Fish COWRIE Ref Fish 06-08, Tech. Rep. 31st March 2010, 57 pages. Retrieved from http://dspace.lib.cranfield.ac.uk/handle/1826/8235 Munger, L. M., Mellinger, D. K., Wiggins, S. M., Moore, S. E., and Hildebrand, J. A. (2005). “Performance of spectrogram cross-correlation in detecting right whale calls in long-term recordings from the Bering Sea,” Can. Acoust., 33, 25–34. Nabe-Nielsen, J., van Beest, F. M., Grimm, V., Sibly, R. M., Teilmann, J., and Thompson, P. M. (2018). “Predicting the impacts of anthropogenic disturbances on marine populations,” Conserv. Lett., , doi: 10.1111/conl.12563. doi:10.1111/conl.12563 Neo, Y. Y., Seitz, J., Kastelein, R. A., Winter, H. V., ten Cate, C., and Slabbekoorn, H. (2014). “Temporal structure of sound affects behavioural recovery from noise impact in European seabass,” Biol. Conserv., 178, 65–73. doi:10.1016/j.biocon.2014.07.012 New, L., Clark, J., Costa, D., Fleishman, E., Hindell, M., Klanjšček, T., Lusseau, D., et al. (2014). “Using short-term measures of behaviour to estimate long-term fitness of southern elephant seals,” Mar. Ecol. Prog. Ser., 496, 99–108. doi:10.3354/meps10547 Nieukirk, S. L., Stafford, K. M., Mellinger, D. K., Dziak, R. P., and Fox, C. G. (2004). “Low- frequency whale and seismic airgun sounds recorded in the mid-Atlantic Ocean,” J. Acoust. Soc. Am., 115, 1832. doi:10.1121/1.1675816 Noren, D. P., Holt, M. M., Dunkin, R. C., and Williams, T. M. (2013). “The metabolic cost of communicative sound production in Bottlenose dolphins (Tursiops truncatus),” J. Exp. Biol., 216, 1624–1629. doi:10.1242/jeb.083212 Nowacek, D., Thorne, L. H., Johnston, D. W., and Tyack, P. L. (2007). “Responses of cetaceans to anthropogenic noise,” Mamm. Rev., 37, 81–115. doi:10.1111/j.1365- 2907.2007.00104.x Oleson, E. M., Calambodikis, J., Burgess, W. C., McDonald, M. A., LeDuc, C. A., and Hildebrand, J. A. (2007a). “Behavioral context of call production by eastern North Pacific blue whales,” Mar. Ecol. Prog. Ser., 330, 269–284. Oleson, E. M., Wiggins, S. M., and Hildebrand, J. A. (2007b). “Temporal separation of blue whale call types on a southern California feeding ground,” Anim. Behav., 74, 881–894. doi:10.1016/j.anbehav.2007.01.022 Van Oosterom, L., Montgomery, J. C., Jeffs, A. G., and Radford, C. A. (2016). “Evidence for contact calls in fish: Conspecific vocalisations and ambient soundscape influence group cohesion in a nocturnal species,” Sci. Rep., 6, 1–8. doi:10.1038/srep19098 Oswald, J. N., Rankin, S., and Barlow, J. (2004). “The effect of recording and analysis bandwidth on acoustic identification of delphinid species,” J. Acoust. Soc. Am., 116, 3178. doi:10.1121/1.1804635

111

Bibliography

Papale, E., Gamba, M., Perez-Gil, M., Martin, V. M., and Giacoma, C. (2015). “Dolphins adjust species-specific frequency parameters to compensate for increasing background noise,” PLoS One, 10, e0121711. doi:10.1371/journal.pone.0121711 Parks, S. E., Clark, C. W., and Tyack, P. L. (2007). “Short- and long-term changes in right whale calling behavior: the potential effects of noise on acoustic communication.,” J. Acoust. Soc. Am., 122, 3725–3731. doi:10.1121/1.2799904 Parks, S. E., Johnson, M., Nowacek, D., and Tyack, P. L. (2011a). “Individual right whales call louder in increased environmental noise,” Biol. Lett., 7, 33–35. doi:10.1098/rsbl.2010.0451 Parks, S. E., Searby, A., Célérier, A., Johnson, M. P., Nowacek, D. P., and Tyack, P. L. (2011b). “Sound production behavior of individual North Atlantic right whales: implications for passive acoustic monitoring,” Endanger. Species Res., 15, 63–76. doi:10.3354/esr00368 Parks, S. E., Urazghildiiev, I., and Clark, C. W. (2009). “Variability in ambient noise levels and call parameters of North Atlantic right whales in three habitat areas.,” J. Acoust. Soc. Am., 125, 1230–1239. doi:10.1121/1.3050282 Parsons, E. C. M., Dolman, S. J., Wright, A. J., Rose, N. A., and Burns, W. C. G. (2008). “Navy sonar and cetaceans: just how much does the gun need to smoke before we act?,” Mar. Pollut. Bull., 56, 1248–1257. doi:10.1016/j.marpolbul.2008.04.025 Pramik, B. (2013). “Marine Vibroseis: shaking up the industry,” First Break, 31, 67–72. Purser, J., and Radford, A. N. (2011). “Acoustic noise induces attention shifts and reduces foraging performance in three-spined sticklebacks (Gasterosteus aculeatus).,” PLoS One, 6, e17478. doi:10.1371/journal.pone.0017478 Putland, R. L., Merchant, N. D., Farcas, A., and Radford, C. A. (2017). “Vessel noise cuts down communication space for vocalising fish and marine mammals,” Glob. Chang. Biol., 12, 3218–3221. doi:10.1111/gcb.13996 R Core Team (2017). “R: A language and environment for statistical computing.,” Retrieved from https://www.r-project.org/ Rako-Gospić, N., and Picciulin, M. (2019). “Underwater noise: Sources and effects on marine life,” In C. Sheppard (Ed.), World Seas An Environ. Eval. Vol. III Ecol. Issues Environ. Impacts, Academic Press, 2nd ed., pp. 367–389. doi:10.1016/B978-0-12-805052-1.00023- 1 Richardson, W. J., Greene Jr, C. R., Malme, C. I., and Thomson, D. I. (1995). Marine mammals and noise, Academic Press, San Diego, 576 pages. Romagosa, M., Boisseau, O., Cucknell, A.-C., Moscrop, A., and McLanaghan, R. (2015). “Source level estimates for sei whale ( Balaenoptera borealis ) vocalizations off the Azores,” J. Acoust. Soc. Am., 138, 2367–2372. doi:10.1121/1.4930900 Rowe, S., and Hutchings, J. A. (2005). “The function of sound production by Atlantic cod as inferred from patterns of variation in drumming muscle mass,” Can. J. Zool., 82, 1391– 1398. doi:10.1139/z04-119 Roy, N., Simard, Y., and Gervaise, C. (2010). “3D tracking of foraging belugas from their clicks: Experiment from a coastal hydrophone array,” Appl. Acoust., 71, 1050–1056.

112

Bibliography

doi:10.1016/j.apacoust.2010.05.008 Russell, D. J. F., Hastie, G. D., Thompson, D., Janik, V. M., Hammond, P. S., Scott-Hayward, L. A. S., Matthiopoulos, J., et al. (2016). “Avoidance of wind farms by harbour seals is limited to pile driving activities,” (S. Votier, Ed.) J. Appl. Ecol., , doi: 10.1111/1365- 2664.12678. doi:10.1111/1365-2664.12678 Samaran, F., Guinet, C., Adam, O., Motsch, J.-F., and Cansi, Y. (2010). “Source level estimation of two blue whale subspecies in southwestern Indian Ocean.,” J. Acoust. Soc. Am., 127, 3800–8. doi:10.1121/1.3409479 Samson, J. E., Mooney, T. A., Gussekloo, S. W. S., and Hanlon, R. T. (2016). “A Brief Review of Cephalopod Behavioral Responses to Sound,” In A. N. Popper and A. Hawkins (Eds.), Eff. Noise Aquat. Life II, Advances in Experimental Medicine and Biology, Springer New York, New York, NY, pp. 969–975. doi:10.1007/978-1-4939-2981-8_120 Sarnocińska, J., Teilmann, J., Balle, J. D., van Beest, F. M., Delefosse, M., and Tougaard, J. (2020). “Harbor Porpoise (Phocoena phocoena) Reaction to a 3D Seismic Airgun Survey in the North Sea,” Front. Mar. Sci., 6, 1–10. doi:10.3389/fmars.2019.00824 Scheidat, M., Tougaard, J., Brasseur, S., Carstensen, J., Petel, T. van P., Teilmann, J., Reijnders, P., et al. (2011). “Harbour porpoises (Phocoena phocoena) and wind farms: a case study in the Dutch North Sea,” Environ. Res. Lett., 6, 1–10. doi:10.1088/1748-9326/6/2/025102 Scheifele, P. M., Andrew, S., Cooper, R. A., Darre, M., Musiek, F. E., and Max, L. (2005). “Indication of a Lombard vocal response in the St. Lawrence River beluga,” J. Acoust. Soc. Am., 117, 1486. doi:10.1121/1.1835508 Sears, R., and Perrin, W. F. (2017). “Blue whale,” In B. Würsig, J. G. M. Thewissen, and K. M. Kovacs (Eds.), Encycl. Mar. Mamm., 3rd ed., pp. 110–114. Shapiro, A. D., Tougaard, J., Jørgensen, P. B., Kyhn, L. A., Balle, J. D., Bernardez, C., Fjälling, A., et al. (2009). “Transmission loss patterns from acoustic harassment and deterrent devices do not always follow geometrical spreading predictions,” Mar. Mammal Sci., 25, 53–67. doi:10.1111/j.1748-7692.2008.00243.x Shiu, Y., Palmer, K. J., Roch, M. A., Fleishman, E., Liu, X., Nosal, E. M., Helble, T., et al. (2020). “Deep neural networks for automated detection of marine mammal species,” Sci. Rep., 10, 1–12. doi:10.1038/s41598-020-57549-y Sidorovskaia, N. A. (2005). “Acoustic Propagation Studies For Sperm Whale Phonation Analysis During LADC Experiments,” 288–295. doi:10.1063/1.1843023 Simard, Y., Bahoura, M., and Roy, N. (2004). “Acoustic detection and localization of whales in bay of fundy and St. Lawrence estuary critical habitats,” Can. Acoust. - Acoust. Can., 32, 107–116. Simard, Y., and Roy, N. (2008). “Detection and localization of blue and fin whales from large- aperture autonomous hydrophone arrays: A case study from the St. Lawrence estuary,” Can. Acoust. - Acoust. Can., 36, 104–110. Simpson, S. D., Radford, A. N., Nedelec, S. L., Ferrari, M. C. O., Chivers, D. P., McCormick, M. I., and Meekan, M. G. (2016). “Anthropogenic noise increases fish mortality by predation,” Nat. Commun., , doi: 10.1038/ncomms10544. doi:10.1038/ncomms10544

113

Bibliography

Širović, A., Hildebrand, J. A., and Wiggins, S. M. (2007). “Blue and fin whale call source levels and propagation range in the Southern Ocean.,” J. Acoust. Soc. Am., 122, 1208–1215. doi:10.1121/1.2749452 Širović, A., Hildebrand, J. A., Wiggins, S. M., McDonald, M. A., Moore, S. E., and Thiele, D. (2004). “Seasonality of blue and fin whale calls and the influence of sea ice in the Western Antarctic Peninsula,” Deep Sea Res. Part II Top. Stud. Oceanogr., 51, 2327–2344. doi:10.1016/j.dsr2.2004.08.005 Slabbekoorn, H., Bouton, N., van Opzeeland, I., Coers, A., ten Cate, C., and Popper, A. N. (2010). “A noisy spring: the impact of globally rising underwater sound levels on fish,” Trends Ecol. Evol., 25, 419–427. doi:10.1016/j.tree.2010.04.005 Smith, M. E. (2004). “Noise-induced stress response and hearing loss in goldfish (Carassius auratus),” J. Exp. Biol., 207, 427–435. doi:10.1242/jeb.00755 Sørensen, P. M., Wisniewska, D. M., Jensen, F. H., Johnson, M., Teilmann, J., and Madsen, P. T. (2018). “Click communication in wild harbour porpoises (Phocoena phocoena),” Sci. Rep., 8, 9702. doi:10.1038/s41598-018-28022-8 Southall, B. L., Bowles, A. E., Ellison, W. T., Finneran, J. J., Gentry, R. L., Greene Jr, C. R., Kastak, D., et al. (2007). “Marine mammal noise exposure criteria: Initial scientific recommendations,” Aquat. Mamm., 33, 411–522. doi:10.1080/09524622.2008.9753846 Spiesberger, J. (2006). “Probability distributions for locations of calling animals, receivers, sound speeds, winds, and data from travel time differences,” J. Acoust. Soc. Am., 119, 3434–3434. doi:10.1121/1.4786902 Spiesberger, J. L., and Fristrup, K. M. (1990). “Passive localization of calling animals and sensing of their acoustic environment using acoustic tomography,” Am. Nat., 135, 107– 153. Retrieved from http://www.jstor.org/stable/10.2307/2462141 Spiga, I., Aldred, N., and Caldwell, G. S. (2017). “Anthropogenic noise compromises the anti- predator behaviour of the European seabass, Dicentrarchus labrax (L.),” Mar. Pollut. Bull., 122, 297–305. doi:10.1016/j.marpolbul.2017.06.067 Stafford, K. M., Fox, C. G., and Clark, D. S. (1998). “Long-range acoustic detection and localization of blue whale calls in the northeast Pacific Ocean.,” J. Acoust. Soc. Am., 104, 3616–3625. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9857519 Stafford, K. M., Mellinger, D. K., Moore, S. E., and Fox, C. G. (2007). “Seasonal variability and detection range modeling of baleen whale calls in the Gulf of Alaska, 1999-2002.,” J. Acoust. Soc. Am., 122, 3378–3390. doi:10.1121/1.2799905 Stafford, K. M., Nieukirk, S. L., and Fox, C. G. (1999). “Low-frequency whale sounds recorded on hydrophones moored in the eastern tropical Pacific,” J. Acoust. Soc. Am., 106, 3687– 3698. doi:10.1121/1.428220 Stafford, K. M., Nieukirk, S. L., and Fox, C. G. (2001). “Geographic and seasonal variation of blue whale calls in the North Pacific,” J. Cetacean Res. Manag., 3, 65–76. Stanley, J. A., Parijs, S. M. Van, and Hatch, L. T. (2017). “Underwater sound from vessel traffic reduces the effective communication range in Atlantic cod and haddock,” Sci. Rep., , doi: 10.1038/s41598-017-14743-9. doi:10.1038/s41598-017-14743-9

114

Bibliography

Stimpert, A. K., DeRuiter, S. L., Falcone, E. A., Joseph, J., Douglas, A. B., Moretti, D. J., Friedlaender, A. S., et al. (2015). “Sound production and associated behavior of tagged fin whales (Balaenoptera physalus) in the Southern California Bight,” Anim. Biotelemetry, 3, 1–12. doi:10.1186/s40317-015-0058-3 Strahan, M. G., Finneran, J. J., Mulsow, J., and Houser, D. S. (2020). “Effects of dolphin hearing bandwidth on biosonar click emissions,” J. Acoust. Soc. Am., 148, 243–252. doi:10.1121/10.0001497 Sueur, J., Aubin, T., and Simonis, C. (2008). “Seewave: a free modular tool for sound analysis and synthesis,” Bioacoustics, 18, 213–226. doi:10.1080/09524622.2008.9753600 Teilmann, J., and Carstensen, J. (2012). “Negative long term effects on harbour porpoises from a large scale offshore wind farm in the Baltic—evidence of slow recovery,” Environ. Res. Lett., 7, 45101. doi:10.1088/1748-9326/7/4/045101 Teilmann, J., Wisniewska, D. M., Johnson, M., Miller, L. A., Siebert, U., Dietz, R., Sveegaard, S., et al. (2015). Acoustic tags on wild harbour porpoises reveal context-specific reactions. Final project report for underwater Cluster 7 to BfN. Tenghamn, R. (2006). “An electrical marine vibrator with a flextensional shell,” Explor. Geophys., 37, 286–291. doi:10.1071/EG06286 Tennessen, J. B., and Parks, S. E. (2016). “Acoustic propagation modeling indicates vocal compensation in noise improves communication range for North Atlantic right whales,” Endanger. Species Res., 30, 225–237. doi:10.3354/esr00738 Tervo, O. M., Christoffersen, M. F., Simon, M., Miller, L. A., Jensen, F. H., Parks, S. E., and Madsen, P. T. (2012). “High Source Levels and Small Active Space of High-Pitched Song in Bowhead Whales (Balaena mysticetus),” PLoS One, , doi: 10.1371/journal.pone.0052072. doi:10.1371/journal.pone.0052072 Thode, A. M., Blackwell, S. B., Conrad, A. S., Kim, K. H., Marques, T., Thomas, L., Oedekoven, C. S., et al. (2020). “Roaring and repetition: How bowhead whales adjust their call density and source level (Lombard effect) in the presence of natural and seismic airgun survey noise,” J. Acoust. Soc. Am., 147, 2061–2080. doi:10.1121/10.0000935 Thode, A. M., D’Spain, G. L., and Kuperman, W. A. (2000). “Matched-field processing, geoacoustic inversion, and source signature recovery of blue whale vocalizations.,” J. Acoust. Soc. Am., 107, 1286–1300. doi:10.1121/1.428417 Thompson, P. O., Findley, L. T., Vidal, O., and Cummings, W. C. (1996). “Underwater sounds of blue whales, Balaenoptera musculus, in the Gulf of California, Mexico,” Mar. Mammal Sci., 12, 288–293. doi:10.1111/j.1748-7692.1996.tb00578.x Thompson, P. O., and Friedl, W. A. (1982). “A long term study of low frequency sounds from several species of whales off Oahu, Hawaii,” Cetology, 45, 1–19. Tiemann, C. O., Thode, A. M., Straley, J., O’Connell, V., Folkert, K., and Connell, V. O. (2006). “Three-dimensional localization of sperm whales using a single hydrophone,” J. Acoust. Soc. Am., 120, 2355–2365. doi:10.1121/1.2335577 Tougaard, J., Carstensen, J., Teilmann, J., Skov, H., and Rasmussen, P. (2009). “Pile driving zone of responsiveness extends beyond 20 km for harbor porpoises (Phocoena phocoena

115

Bibliography

(L.)).,” J. Acoust. Soc. Am., 126, 11–14. doi:10.1121/1.3132523 Tulp, I., Bolle, L. J., Dänhardt, A., de Vries, P., Haslob, H., Jepsen, N., Scholle, J., et al. (2017). “Fish,” Wadden Sea Qual. Status Rep. 2017,. Tyack, P. (2009). “Acoustic playback experiments to study behavioral responses of free- ranging marine animals to anthropogenic sound,” Mar. Ecol. Prog. Ser., 395, 187–200. doi:10.3354/meps08363 Tyack, P., Gordon, J., and Thompson, D. (2003). “Controlled Exposure Experiments to Determine the Effects of Noise on Marine Mammals,” Mar. Technol. Soc. J., 37, 41–53. doi:10.4031/002533203787537087 Tyack, P. L. (2000). “Functional aspects of cetacean communication,” In J. Mann, R. C. Connor, P. L. Tyack, and H. Whitehead (Eds.), Cetacean Soc. F. Stud. dolphins whales, The University of Chicago Press, pp. 270–307. Urazghildiiev, I. R., and Clark, C. W. (2006). “Acoustic detection of North Atlantic right whale contact calls using the generalized likelihood ratio test,” J. Acoust. Soc. Am., 120, 1956– 1963. doi:10.1121/1.2257385 Urick, R. J. (1983). Principles of underwater sound, McGraw-Hill Inc., New York, 3rd ed., 423 pages. Van der Veer, H. W., Berghahn, R., Miller, J. M., and Rijnsdorp, A. D. (2000). “Recruitment in flatfish, with special emphasis on North Atlantic species: progress made by the Flatfish Symposia,” ICES J. Mar. Sci., 57, 202–215. Verfuß, U. K., Miller, L. A., Pilz, P. K. D., and Schnitzler, H.-U. (2009). “Echolocation by two foraging harbour porpoises (Phocoena phocoena),” J. Exp. Biol., 212, 823–834. doi:10.1242/jeb.022137 Voellmy, I. K., Purser, J., Simpson, S. D., and Radford, A. N. (2014). “Increased noise levels have different impacts on the anti-predator behaviour of two sympatric fish species,” PLoS One, 9, 1–8. doi:10.1371/journal.pone.0102946 Wahlberg, M., and Larsen, O. N. (2017). “Propagation of sound,” In C. Brown and T. Riede (Eds.), Comp. Bioacoustics An Overv., Bentham Science Publishers, Oak Park, Illinois, USA, pp. 62–119. Wahlberg, M., Møhl, B., and Madsen, P. T. (2001). “Estimating source position accuracy of a large-aperture hydrophone array for bioacoustics,” J. Acoust. Soc. Am., 109, 397–406. doi:10.1121/1.1329619 Wahlberg, M., and Westerberg, H. (2003). “Sounds produced by herring (Clupea harengus) bubble release,” Aquat. Living Resour., 16, 271–275. doi:10.1016/S0990-7440(03)00017- 2 Wang, D., Wuersig, B., and Evans, W. E. (1995). “Whistles of bottlenose dolphins: comparisons among populations,” Aquat. Mamm., 21, 65–77. Retrieved from http://www.mendeley.com/research/whistles-bottlenose-dolphins-comparisons- populations-4/ Wartzok, D., Altmann, J., Au, W., Ralls, K., Starfield, A., Tyack, P. L., Bintz, J., et al. (2005). Marine Mammal Populations and Ocean Noise - Determining when noise causes

116

Bibliography

biologically significant effects, National Academy of Sciences, Washington DC, USA, 126 pages. Wartzok, D., and Ketten, D. R. (1999). “Marine mammal sensory systems,” In J. E. Reynolds III and S. A. Rommel (Eds.), Biol. Mar. Mamm., Smithsonian Institution Press, Washington, D.C., pp. 117–175. Watkins, W. A., and Schevill, W. E. (1972). “Sound source location by arrival-times on a non- rigid three-dimensional hydrophone array,” Deep. Res. Oceanogr. Abstr., , doi: 10.1016/0011-7471(72)90061-7. doi:10.1016/0011-7471(72)90061-7 Weilgart, L. S. (2007). “The impacts of anthropogenic ocean noise on cetaceans and implications for management,” Can. J. Zool., 85, 1091–1116. doi:10.1139/Z07-101 Wiggins, S. M., and Hildebrand, J. A. (2007). “High-frequency Acoustic Recording Package (HARP) for broad-band, long-term marine mammal monitoring,” Int. Symp. Underw. Technol. UT 2007 - Int. Work. Sci. Use Submar. Cables Relat. Technol. 2007, , doi: 10.1109/UT.2007.370760. doi:10.1109/UT.2007.370760 Wisniewska, D. M., Johnson, M., Teilmann, J., Siebert, U., Galatius, A., Dietz, R., and Madsen, P. T. (2018). “High rates of vessel noise disrupt foraging in wild harbour porpoises (Phocoena phocoena),” Proc. R. Soc. B Biol. Sci., 285, 20172314. doi:10.1098/rspb.2017.2314 Woelfing, B., Rasmussen, M. H., Siebert, U., Stilz, P., Gavrilov, A., Erbe, C., McCauley, R., et al. (2020). Assessment of communication masking in Antarctic marine mammals by underwater sound from airguns 197 pages. Retrieved from http://www.umweltbundesamt.de/publikationen World Health Organization (2017). “Burden of disease from environmental noise: Quantification of healthy life years lost in Europe,” Geneva, Switzerland: World Health Organization.

117

List of figures

Fig. 1: Map of the North Sea coastline including the entire Wadden Sea area with an enlarged section of the area of investigation (square therein)...... 20 Fig. 2: Spectrogram of the first detected anchor pipe vibration embedment noise with a duration of 52 s recorded at around 1 km distance to the pile...... 22 Fig. 3: Median third octave sound exposure level (SEL) for the first detected anchor pipe vibration embedment noise event ...... 23 Fig. 4: Calculated transmission loss ...... 24 Fig. 5: Effect radius to the vibration embedment location for different fish and marine mammal species living in the German Wadden Sea ...... 28 Fig. 6: Area of investigation, Skjálfandi Bay in Northeast Iceland...... 40 Fig. 7: Spectrogram of a detected blue whale down sweep at 4.6 km from the receiver...... 47 Fig. 8: Propagation loss (left) and resulting radiated noise levels (RNL; right) for blue whale down sweeps after model 1 assuming spherical spreading...... 49 Fig. 9: Propagation loss (left) and resulting source levels (SL; right) for blue whale down sweeps after model 2 assuming cylindrical spreading...... 50 Fig. 10: Tag data (vertical speed and depth) of two behavioural categories...... 51 Fig. 11: Vertical speed and dive profile of one tagged blue whale (bm18_180a) with behavioural phases and vocalisation...... 53 Fig. 12: Segment of Fig. 11 between 12:15 and 12:50 (top) and sonogram of 12:18 (bottom)...... 54 Fig. 13: Sound source ‘Argotec SS-2’ for emiting marine vibrator sounds...... 68 Fig. 14: Upsweep signal imitating the marine vibrator like sound...... 69 Fig. 15: The experimental design ...... 70 Fig. 16: Comparison of the ambient noise level between the three experimental phases...... 75 Fig. 17: Blue whale D call (arch sound) among marine vibrator like upsweep signals played back via the SS-2...... 76 Fig. 18: Comparison of the call duration (including 90% of the signal’s energy) between the three experimental phases...... 78 Fig. 19: Comparison of the centroid frequency between the three experimental phases...... 79 Fig. 20: Example of frequency change within the postexposure phase...... 84

118

List of tables

Tab. 1: Acoustic signals of the Wadden Sea fish...... 34 Tab. 2: Call characteristics of blue whale down sweeps in Skjálfandi Bay, Northeast Iceland derived from the array data...... 48 Tab. 3: Characteristics of blue whale down sweeps in Skjálfandi Bay, Northeast Iceland derived from tag data...... 52 Tab. 4: Comparison of estimated source levels of blue whale vocalisation from our study and those recorded in other regions of the world from different blue whale populations...... 61 Tab. 5: Call characteristics during the three experimental phases...... 74

119

Acknowledgement

First of all, I would like to thank my supervisors who gave me the opportunity to write a dissertation on the biggest animal on earth. I am very thankful to Ursula Siebert for giving me the opportunity to achieve my doctor’s degree on a great topic and for her trust and support in every way. Thanks goes to Magnus Wahlberg, who was the first introducing me to the fascinating world of whale acoustics.

I am especially grateful to my bioacoustics group the greatest team I could wish for. You guys always supported me and were always willing to help, especially when I was struggling with scientific problems for a while and didn’t say a word (as always). Tobias Schaffeld, there is no one better to share an office with than you. I really appreciate having you around, especially when there are things at work, I don’t know how to deal with in the beginning. I pretty much like our bullshit talking and also ‘knüselling’ on programming issues. You are a great colleague and friend. Special thanks goes out to Jeff Schnitzler, who pushed me hard, when it was necessary and always believed in me. I really appreciate that you were there all the time. Without you, I would never have got done with the thesis. You are a great mentor and a man of action. THANK YOU!!!

I am greatly thankful for the support of my ITAW-family in Büsum. Thanks to Bianca, that I could always come by to talk when it was needed. You were always there when I was pretty much in despair (oh it rhymes) and calmed me down. I thank you, Abbo for your mental support and for motivating me again and again, when I was a hopeless case in my thoughts. The beer breaks and evenings with finest metal really helped. I would like to thank my doctoral group (Dominik, Britta and Caro) for your support and for being there when I needed to talk. I also like to thank Susanne Carstens-Michaelis especially for reminding me when I lose track of things and that you take care of us all.

Thanks goes also to my Team Iceland for a great experience and very nice fieldwork seasons.

I also thank my buddies Flo and Flo. We started to study together and since then became very close friends. I can always count on you guys when I needed help or needed to talk.

Last but not least, I would like to thank my entire family: Mom, Dad, Josi and Steffi for your constant moral support. You were always there for me and supported me when I was desperate and thought I cannot do it.

120