The Impact of Anthropogenic Noise on Fish and ...randd.defra.gov.uk/Document.aspx?Document=13847_ME... · impact of anthropogenic noise, as evidenced by the publication of five peer-reviewed

The Impact of

Anthropogenic Noise on

Fish and Invertebrates

(ME5207)

Final Report for Defra

Andrew N. Radford1, Julia Purser1, Cato ten

Hallers-Tjabbes2 & Stephen D. Simpson1,3

1School of Biological Sciences, University of Bristol, Woodland Road,

Bristol, UK

2CaTO Marine Ecosystems, Oosterweg 1, 9995 VJ Kantens, The

Netherlands

3Biosciences, College of Life and Environmental Sciences, University of

Exeter, UK

Contents

1. Executive Summary ................................................................................................................ 1

2. Aims and Objectives ............................................................................................................... 3

2.1 Background ................................................................................................................... 3

2.2 Objectives ..................................................................................................................... 3

3. Methods and Results ............................................................................................................. 5

3.1 The potential impact of anthropogenic noise on development, physiology and

behaviour ............................................................................................................................ 6

3.2 Interspecific variation in responses ............................................................................ 11

3.3 Inter-individual variation in responses ....................................................................... 12

3.4 Intra-individual variation in responses ....................................................................... 12

3.5 Repeated exposure and prior experience .................................................................. 13

3.6 Variation in noise characteristics ............................................................................... 14

3.7 Recovery from noise exposure ................................................................................... 14

3.8 Results summary ........................................................................................................ 15

4. Discussion of Results ............................................................................................................ 15

5. Future Work ......................................................................................................................... 16

6. Actions Resulting from the Work ......................................................................................... 17

1

1. Executive Summary Since the Industrial Revolution, human activities have substantially changed the acoustic environment around the globe. Consequently, anthropogenic (man-made) noise presents a very real, and often novel, challenge to animals and is now recognised as a pollutant of international concern. While there has been a recent burgeoning research effort to investigate the potential impacts of anthropogenic noise on non-human animals, initial studies of aquatic animals tended to focus on marine mammals; very little work considered the potential impacts on fish or invertebrates. However, given their abundance and diversity and their value to community structure and overall biodiversity, as a food source and commercially, detrimental effects on these taxonomic groups are of clear biological, ecological and societal importance. Thus, the overall aim of this Defra-commissioned research programme (ME5207) was to further our understanding of the impacts of anthropogenic noise on fish and marine invertebrates, and specifically to address five objectives. Objective 1: Identify key species which need to be considered. A summary document relating to marine mammals (‘The impact of anthropogenic noise on UK marine mammals: a review and future suggestions’) was prepared. Fish and invertebrate species targeted for experimental study were those requested by the Steering Group (European eels, Atlantic salmon), those that provide model systems for developing and validating methodologies (three-spined sticklebacks, European minnows, coral reef fish, sea hares), and those of potential commercial importance and that forge fundamental links in marine food webs in UK waters (Atlantic cod, European sea bass, shore crabs). Objective 2: Identify key marine noise generation activities which need to be considered. Noise sources studied were vessels (ships and boats) and pile-driving. Over 50,000 merchant ships carry 90% of world trade around the globe, while small boats with outboard motors are ubiquitous where humans inhabit coastal areas. Pile-driving is a critical element in any construction project, including the development of Marine Renewable Energy resources such as windfarms. Objective 3: Consider how anthropogenic noise potentially impacts the behaviour, physiology and development of key species. A ‘tool kit’ of working paradigms and methods was developed. This included the setting-up of the only Home-Office licenced, mobile electrophysiology unit for the assessment of hearing thresholds in fish. In addition, a range of carefully controlled experiments were designed to consider developmental, physiological and behavioural impacts of noise. Initial work was conducted using playbacks of sounds in aquatic tanks before validation in open-water settings, first using playbacks and finally with real noise sources. Further experiments explored possible interspecific, inter-individual (e.g. size- and condition-dependent) and intra-individual (e.g. context-dependent) variation in responses. Objective 4: Consider the potential impact of different noise exposure types and levels. Using the tool kit developed in Objective 3, experiments were designed and conducted to look beyond the immediate effect of acute exposure to anthropogenic noise. Issues such as repeated and prior exposure, variation in noise characteristics, and recovery time were investigated to explore the possibility of different and changing responses to the same noise source. Objective 5: Provide a greater understanding of the impacts of anthropogenic noise on fish and invertebrates to be used for future policy decisions. Throughout the project, a contemporary knowledge exchange approach was adopted by: (i) co-developing research questions and experiments in collaboration with members of the end-user community; (ii) co-producing the science in collaboration with maritime management organisations; and (iii) co-delivering the science through professional organisations and institutions. Much of the co-development was undertaken at a workshop, attended by 40 academics and 30 industry, policy and funding representatives, which identified key issues before exploring potential ways to solve them. This workshop has led to several ongoing and further engagements with end-users.

2

Results from the extensive array of experiments conducted indicate that both vessel noise (generated by ships and boats) and pile-driving noise have the potential to impact a range of crucial life-history processes, including development (e.g. survival, growth, condition), physiology (e.g. respiratory rate, metabolic rate) and behaviour (e.g. anti-predator, foraging), in a variety of fish and invertebrate species. Qualitatively similar results were found in open-water tests (using both playbacks of sounds and real noise sources) as in those utilising playbacks of sounds in aquatic tanks, validating the latter approach as an exploratory process. However, while tank-based experiments offer the opportunity for tightly controlled conditions and detailed data collection, care must be taken when extrapolating results to real-world situations; experiments in natural conditions with real sound sources are required to generate absolute threshold values for use by regulators. Additional experiments demonstrated that the magnitude of the impact from anthropogenic noise is likely to differ between species and between individual members of the same species, depending on such characteristics as size and condition. Moreover, the response of a given individual to a given type of noise is likely to differ depending on, for instance, the context, prior noise exposure and the particular characteristics of the noise source. Finally, it is possible that individuals will quickly recover from short-term exposure to noise, although whether that occurs following longer exposures and whether there are lasting consequences remain to be tested. In summary, responses to anthropogenic noise are variable and can change, which must be taken into account when assessing ultimate consequences for individual fitness (survival and reproductive success), population viability and community structure. The project has generated many novel and scientifically robust findings relating to the potential impact of anthropogenic noise, as evidenced by the publication of five peer-reviewed papers, with another 14 in review or in preparation. Moreover, the results have been presented at over 20 conferences, workshops and symposia to both academic and end-user communities. A 70-delegate workshop attended by academics and research end-users, exploring issues and potential solutions around biological impacts of marine noise, has resulted in two NERC Business Internships and a Knowledge Transfer Partnership, as well as contracts with Marine Scotland and industrial partners to investigate further the impacts of anthropogenic noise. The policy process within the International Maritime Organization (IMO), where underwater noise from ships now appears on the agenda of the Marine Environment Protection Committee (MEPC), has been monitored throughout and information provided to the relevant UK representative to aid in the drafting of guidelines. An additional major output of this project is an integrated and validated tool kit of experimental approaches and an understanding of which biologically meaningful response parameters to focus on when considering the impacts of anthropogenic noise. There is thus the opportunity to conduct systematic assessments of focal study species of fish and invertebrates near to real-world disturbances (e.g. shipping lanes, harbours, windfarm sites), and thus directly measure impacts of anthropogenic noise on marine life in ways of direct relevance to managers, policy-makers and regulators. Moreover, the tool kit can be used to assess mitigation measures as they are developed. As such, this project has laid the groundwork to deliver on environmental impact assessments, reduce the uncertainty surrounding major planned Marine Renewable Energy programmes, and aid in the delivery of Good Environmental Status as per Descriptor 11 of the European Commission Marine Directive.

3

2. Aims and Objectives

2.1 Background

Since the Industrial Revolution, anthropogenic (man-made) noise has substantially changed the acoustic environment around the globe, both on land and underwater. In aquatic environments (the focus of this project), noise-generating human activities such as commercial shipping, recreational boating, pile-driving, seismic exploration and energy production are more prevalent, widespread and frequent today than in the past (Jasny 1999; McDonald et al. 2006). In addition to the greater scale of impact, the nature of the sound generated by human activities is often very different from that arising from natural sources; anthropogenic noises may differ substantially from abiotic or biotic sounds in such acoustic characteristics as constancy, rise time, duty cycle and impulsiveness (Hildebrand 2009; Popper & Hastings 2009). Consequently, anthropogenic noise presents a very real, and often novel, challenge to animals and is now recognised as a pollutant of international concern (e.g. inclusion in the European Commission Marine Strategy Framework Directive and as a permanent item on the International Maritime Organisation Marine Environmental Protection Committee agenda). In the last decade, there has been a burgeoning research effort to investigate the potential impacts of anthropogenic noise on non-human animals; a recent survey of the peer-reviewed literature for Defra Project NO0235 indicated that over 30% of studies published by the end of 2011 appeared within that last year alone (Radford et al. 2012). Effects have been demonstrated in a variety of taxonomic groups across a range of scales, from the physiology and behaviour of individuals to changes at the population and community level (see Tyack 2008; Barber et al. 2009; Slabbekoorn et al. 2010; Kight and Swaddle 2011 for reviews). In the aquatic environment, the major focus of early research was marine mammals, and particularly movement patterns and acoustic communication (NRC 2003, 2005; Nowacek et al. 2007; Weilgart 2007; Tyack 2008; Clark et al. 2009; Appendix A). Only a few studies initially considered the potential impacts of anthropogenic noise on fish (Popper & Hastings 2009; Slabbekoorn et al. 2010), and virtually no work has investigated invertebrates (Morley et al. in review). Fish provide the primary source of protein for >1 billion people and the principal livelihoods for 100s of millions (FAO 2012). They represent more than half of all vertebrate species, possess a broad range of hearing and sound-production mechanisms, and exhibit a diverse array of vocal, reproductive and social traits (Popper & Fay 1999; Ladich 2004; Bone & Moore 2008). Since the majority of fish species live in coastal or freshwater environments, they are exposed to many forms of anthropogenic noise. The relative lack of research into how they are affected by this pollutant therefore represents a critical knowledge gap. The paucity of attention on invertebrates is also not commensurate with their abundance and diversity (they make up 60% of marine species), their importance ecologically (as essential components of food webs) and economically (especially in light of changing fisheries), or their value in terms of new natural products (Cesar et al. 2003; Ausubel et al. 2010; Leal et al. 2012). Since marine invertebrates are capable of hearing (Salmon 1971; Goodall et al. 1990) and use sound for a variety of reasons (e.g. Jeffs et al. 2003; Stanley et al. 2010; Simpson et al. 2011), they are also likely to be affected by anthropogenic noise. Assessing the impacts of anthropogenic noise on fish and marine invertebrates is thus of clear biological, ecological and societal importance, and was the overall aim of this Defra-commissioned research programme (ME5207).

2.2 Objectives

The following represent the project objectives following annual input from the Steering Group:

4

Objective 1: Identify key species which need to be considered. Consider species from a range of taxa, using both reviews of the literature (mammals) and experimental data collection (fish and invertebrates), to provide crucial information across different trophic levels.

This objective was achieved. A summary document relating to marine mammals (‘The impact of anthropogenic noise on UK marine mammals: a review and future suggestions’; Appendix A) was prepared and submitted to Defra in 2011. Fish and invertebrate species targeted for experimental study were those specifically requested by the Steering Group (European eels Anguilla anguilla, Atlantic salmon Salmo salar), those that provide model systems for developing and validating methodologies (three-spined sticklebacks Gasterosteus aculeatus, European minnows Phoxinus phoxinus, coral reef fish, sea hares Stylocheilus striatus), and those of potential commercial importance and that forge fundamental links in marine food webs in UK waters (Atlantic cod Gadus morhua, European sea bass Dicentrarchus labrax, shore crabs Carcinus maenas). Objective 2: Identify key marine noise generation activities which need to be considered. Consider common noise sources that present different challenges for species in terms of their intensity, duration and predictability.

This objective was achieved. Noise sources studied were vessels (ships and boats) and pile-driving. Over 50,000 merchant ships carry 90% of world trade around the globe (Slabbekoorn et al. 2010), while small boats with outboard motors are ubiquitous where humans inhabit coastal areas. Vessels are therefore the most common and widespread source of aquatic anthropogenic noise and ships alone have caused up to a 100-fold increase of low-frequency sound levels during the last few decades (Tyack 2008). Previous studies have focussed on the chronic effects of background shipping noise (e.g. Richardson et al. 1998; Codarin et al. 2009); less research has considered the potential impact of the sporadic intense noises generated by the passing of single vessels. Pile-driving is the mechanical process used to drive piles into the substrate to form the foundation support for buildings and other structures. It therefore forms a critical stage in any construction project, including the development of Marine Renewable Energy resources such as windfarms. The Crown Estate (the UK marine space leasing authority) recently warned that 75% of planned UK windfarm developments are at risk due to uncertainties about impacts of noise. Experiments were conducted to examine the potential impact of these two noise sources on various fish and invertebrate species identified in Objective 1 (see Objective 3) and to explore how variations in intensity, duration and predictability affect the response (see Objective 4). Objective 3: Consider how anthropogenic noise potentially impacts the behaviour, physiology and development of key species. Use carefully designed and controlled experiments to examine the impacts of different sound sources and noise regimes on various aspects of fish and invertebrate life; work will focus on life-history parameters that have biologically meaningful consequences for individual fitness and population viability.

This objective was achieved. A ‘tool kit’ of working paradigms and methods was developed and validated. This included the setting-up of the only Home-Office licenced, mobile electrophysiology unit for the assessment of hearing thresholds in fish; this set-up was tested successfully in Bristol on a range of species and then taken to the Marine Scotland Laboratory, Aberdeen to use on salmon (Appendix B). In addition, a range of developmental (e.g. survival, growth, condition), physiological (e.g. respiratory rate, metabolic rate) and behavioural (e.g. anti-predator, foraging) experimental methods were designed to investigate the potential impact of anthropogenic noise (Section 3.1). All of these utilised controlled comparisons with matched ambient sounds. Initial work was developed and conducted using playbacks of sounds in aquatic tanks before validation in open-water settings, first using playbacks and finally with real noise sources. Having established that anthropogenic noise does have the potential to impact a range of crucial life-history processes, additional experiments explored possible interspecific (Section 3.2), inter-individual (e.g. size- and condition-dependent; Section 3.3) and intra-individual (e.g. context-dependent; Section 3.4) variation in responses.

5

Objective 4: Consider the potential impact of different noise exposure types and levels. As well as considering the immediate and subsequent effects of both acute and chronic exposure, we will investigate the possibility of habituation and sensitisation to noise following repeated exposures of differing predictability.

This objective was achieved. Using the tool kit developed in Objective 3, experiments were designed and conducted to look beyond a simple effect of anthropogenic noise and consider such aspects as repeated and prior exposure (Section 3.5), variation in noise characteristics (Section 3.6), and recovery time (Section 3.7). Objective 5: Provide a greater understanding of the impacts of anthropogenic noise on fish and invertebrates to be used for future policy decisions. We will ensure that our research programme is relevant to end-users and that they are made aware of our scientific findings.

This objective was achieved. Throughout the project, a contemporary knowledge exchange approach was adopted. Specifically, we: (i) co-developed our research questions and experiments in collaboration with members of the end-user community (e.g. IUCN delegation to IMO and Marine Scotland, as well as Defra); (ii) co-produced the science in collaboration with maritime management organisations (e.g. Port of Bristol Authority, Bristol Harbour Authority); and (iii) co-delivered the science through professional organisations and institutions (e.g. presentations to Underwater Sound Forum, further workshops with Institute of Marine Engineering, Science & Technology). Much of the co-development was undertaken at a workshop we organised in Bristol in February 2012, funded by the Natural Environment Research Council through the Strategic Oceans Funding Initiative and Marine Renewable Energy Knowledge Exchange Programme. This workshop was attended by 40 academics and 30 industry, policy and funding representatives, and identified key issues before exploring potential ways to solve them. This workshop has led to several important engagements (Section 6). Our key findings are summarised in this concise report that is accessible to the end-user community; fuller details are available in peer-reviewed publications that lend weight to the quality of the underpinning science.

3. Methods and Results

All data collected were from experiments in which noise was manipulated as an experimental factor (using either playbacks or real noise sources); stronger conclusions can therefore be drawn than if correlative datasets were used (Radford et al. 2012). Experiments all included an appropriate control condition, usually ambient noise from the same area as the anthropogenic noise source, and thus responses could be compared against a baseline level (with holding conditions of fish kept close to these control baseline acoustic levels wherever possible). In most cases, an independent-measures design was adopted, where different cohorts of individuals (randomly assigned) received each of the different noise treatments in a given experiment; this eliminates any possibility of carry-over effects between treatments. In some cases, the change in response from a baseline level during ambient noise to either another period of ambient noise or one of anthropogenic noise was considered; this additionally controlled for inter-individual variation. Wherever possible, experiments or scoring of data were conducted ‘blind’ to the acoustic treatment, to minimise any observer bias. All work was approved by the Ethical Review Committees of the University of Bristol (University Investigator Number: UB/10/034) and the University of Exeter (2013/247), and by the Home Office (Project Licence PPL 30/2860) where appropriate. In all experiments, the received sound pressure level was measured using a calibrated omnidirectional hydrophone (HiTech HTI-96-MIN with inbuilt preamplifier; sensitivity -165 dB re 1 V/μPa; frequency range 2 Hz – 30 kHz; High Tech Inc., Gulfport, MS, U.S.A.) and an Edirol R09-HR 24-Bit recorder (44.1 kHz sampling rate, Roland Systems Group, Bellingham, WA, U.S.A). The recording level was calibrated for the R09-HR using pure sine wave signals, measured in line with an oscilloscope, produced by a function generator. In open-water experiments, the received particle

6

velocity level was measured using an M30 particle velocity sensor (sensitivity 0–3 kHz, GeoSpectrum Technologies, Dartmouth, Canada) recorded on a laptop via a calibrated USB soundcard (MAYA44, ESI Audiotechnik GmbH, Leonberg, Germany; sampling rate 44.1 kHz); logistical constraints currently prevent the accurate assessment of particle velocity levels in tanks (see Section 4). All received levels fell within the range of the relevant naturally occurring sound source (ambient noise, ship noise, boat noise, pile-driving noise). We do not present values here because we did not attempt to establish absolute sensitivity thresholds for policy and regulatory decision-making (see Sections 4 and 5), but rather provide an indication of organismal responses to the addition of noise to the environment (relative to appropriate reference, baseline control, conditions). Sounds used in playback experiments were recorded in relevant harbour and coastal areas using the hydrophone and recorder set-up described above. Playback tracks were prepared in Audacity 1.3.13 (http://audacity.sourceforge.net/), including fade-in and fade-out sections as appropriate. Experimental tracks were played back as WAV or mp3 files (no detectable compression artefacts) using: WAV/MP3 Player (Ultradisk DVR2 560 h / TrekStor GmbH & Co / Sony PCM-M10 or similar; frequency range 20–20 000 Hz); amplifier (Kemo Electronic GmbH; 18 W or 40W; frequency response: 40–20 000 Hz); potentiometer (set to minimum resistance; Omeg Ltd; 10 K logarithmic); and Aqua 30 or UW30 underwater speaker (DNH or Lubell Labs Inc.; effective frequency range 80–20,000 Hz). Full details of all aspects of experimental procedures are provided in published papers (Purser & Radford 2011; Bruintjes & Radford 2013; Holles et al. 2013; Wale et al. 2013a, b). The Defra project (ME5207) provided for one PDRA. We were able to expand the remit with five independently funded postgraduate students (two PhD, three Masters), who conducted complementary work utilising the methods developed in this project and the overall infrastructure. These extra researchers allowed us to provide additional data and valuable validation. In the following sections, published results are summarised with reference to the relevant citation (papers included as Appendices); major findings from the work that are yet to be published are supported with illustrative statistical analyses and figures (there is insufficient space to provide comprehensive reports of all results). 3.1 The potential impact of anthropogenic noise on development, physiology and behaviour

Work to investigate this issue involved three stages. First, methods were developed using a tank-based approach, which also allowed the tightest control of experimental conditions for exploring potential impacts; tank-based work utilised playback of noise sources. Second, a subset of the experiments were re-run in open-water conditions, utilising playback of noise sources, to validate results and ensure that findings were not the result of acoustic constraints imposed by tank conditions. Finally, open-water experiments utilising real noise sources were conducted to validate findings further, ensure results are not the consequence of utilising playbacks and assess the feasibility of this approach for future work (see Section 5). 3.1.1 Development

Survival through development influences population dynamics and underpins population resilience (Gosselin & Qian 1997; Gagliano et al. 2007). While early-life stages are well adapted to tolerate normal environmental fluctuations and challenges (Gilbert 2001; Hamdoun & Epel 2007), accelerating anthropogenic change with its associated environmental perturbation can push fluctuations outside of the bounds of normal variability. However, virtually no empirical work has investigated the potential impact of anthropogenic noise on development (Kight & Swaddle 2011). Tank-based experiments at Ardtoe Marine Laboratories, Viking Fish Farms Ltd., Scotland examined the impact of exposure to playbacks of vessel (ship and boat) noise on the early development of Atlantic cod (Holles et al. in prep. a). Batches of eggs and subsequent larvae experienced one of three noise treatments: ambient noise (control), regular vessel noise and irregular vessel noise (see also Section 3.6). Newly hatched larvae used significantly more of their yolk sac in the first two days

7

post-hatching when exposed to vessel-noise playbacks compared to control playbacks (GLMM, interaction term between noise treatment and day post-hatching: χ2

2 = 31.43, p < 0.001; Figure 1a). There was no significant effect of noise treatment on growth (body length) at 16 days post-hatching (interaction term between noise treatment and day post-hatching: χ2

4 = 10.56, p = 0.032, result driven by difference at day 2; Figure 1b). However, larvae exposed to playback of vessel noise were in significantly worse body condition (myotome length/body length) than those that had experienced ambient-noise playback (interaction term between noise treatment and day post-hatching: χ2

4 = 5.89, p = 0.015; Figure 1c).

Figure 1. Playback of vessel noise affects early-life development of Atlantic cod. Shown are mean ± SE (a) yolk sac centroid size, (b) body length and (c) body condition at different ages (days post-hatching; dph) for larvae exposed to three different repeated-noise regimes.

Complementary work, based on the methods developed for this project and utilising playbacks of boat noise and ambient noise in open-water conditions, found that anthropogenic noise has the potential to affect early-life stages of both marine invertebrates (Holles et al. submitted) and coral reef fish (Holles et al. in prep. b). For instance, compared to ambient-noise playback, boat-noise playback resulted in a significant increase in the number of sea hare (a marine mollusc) eggs that failed to develop at the cleavage stage (paired t-test: t12 = 2.99, p = 0.011; Figure 2a) and the number of individuals that died shortly after hatching (Wilcoxon signed ranks test: V = 10, N = 11, p = 0.045; Figure 2b), although there was no significant impact on developmental rate (paired t-test: t12 = 0.63, p = 0.538). Repeated playback of boat noise, compared to ambient-noise playback, did not significantly affect the standard length (GLMM: χ2

1 = 0.01, p = 0.998), wet weight (χ21 = 0.50, p =

0.478) or body condition (percentage weight of liver to body weight; χ21 = 0.92, p = 0.339) of

threespot dascyllus Dascyllus trimaculatus (coral reef fish species) larvae, but those individuals exposed to boat-noise playback had significantly lower levels of blood cortisol than those exposed to ambient-noise playback (χ2

1 = 8.80, p = 0.003; Figure 2c).

Figure 2. Playback of boat noise affects sea hares and coral reef fish early in life. Shown are the mean ± SE (a) proportion of sea hare eggs that failed to develop, (b) proportion of recently hatched sea hare veligers that died, and (c) blood plasma cortisol concentrations (pg/ml) of threespot dascyllus larvae following repeated exposure to two noise treatments.

8

3.1.2 Physiology

Assessing how noise affects physiology, in addition to behaviour, is vital for a full understanding of both proximate and ultimate impacts on individual fitness (Kight & Swaddle 2011). In response to noise, animals might be expected to mobilise energetic reserves and alter resource allocation in preparation for action (potentially as part of a more generalised stress response); possible indicators of this response would be enhanced metabolic rate, and thus increased oxygen consumption, and a greater respiratory rate (Wendelaar-Bonga 1997).

Metabolic rate was assessed using closed respirometry, in which the amount of oxygen consumed was determined. In a tank-based experiment, exposure to playback of the noise generated by a single ship pass (hereafter ship noise) led to a significant increase in oxygen consumption in shore crabs (Wale et al. 2013a; Appendix C). Since there was no concomitant increase in movement, the increased oxygen consumption is likely to reflect an increase in metabolic rate (see Wale et al. 2013a). Similarly, single, acute exposure to playback of anthropogenic noise affected the physiology of both European eels (Simpson et al. submitted) and European sea bass (Everley et al. in prep.). European eels in a tank-based experiment significantly increased their oxygen consumption in response to ship-noise playback (ANOVA: F1,43 = 13.77, P = 0.001; Figure 3a), a result that was confirmed by a playback experiment in open-water conditions (F1,43 = 47.44, P < 0.001; Figure 3b). Respiratory rate was assessed by comparing the change when there was a switch from one ambient-noise playback track (baseline level) to either another ambient-noise track or an anthropogenic noise track; all tracks lasted 2 min. Switch to playback of ship-noise significantly increased the respiratory rate of European eels (ANOVA: F1,21 = 21.80, P < 0.001; Figure 3c), while playback of pile-driving noise had a similar effect on sea bass (paired t-test: t17 = 3.42, P = 0.003; Figure 3d). The latter result was confirmed in open-water conditions (t17 = 6.20, P < 0.001; Figure 3e).

Figure 3. Playback of anthropogenic noise affects metabolic rate and respiratory rate. Shown are the mean ± SE levels of oxygen consumption by European eels in (a) a tank-based experiment and (b) an open-water test and the mean ± SE respiratory rate of (c) European eels in tank conditions, (d) European sea bass in tank conditions, and (e) European sea bass in an open-water test.

9

3.1.3 Behaviour

The majority of the previous work examining the potential impacts of anthropogenic noise has focused on behavioural responses and, in particular, movement patterns and acoustic communication (Radford et al. 2012). Ultimately what is needed are assessments of the effect on individual fitness and population viability; if these cannot be measured, behaviours (e.g. anti-predator behaviour and foraging) that translate more directly to such fitness impacts are potentially of most value. In tank-based experiments, both the foraging and anti-predator behaviour of shore crabs was affected by playback of ship noise (Wale et al. 2013b; Appendix D). Ship-noise playback was significantly more likely than ambient-noise playback to disrupt feeding once it had started, although crabs experiencing the two sound treatments did not differ in their likelihood of, or speed at, finding a food source in the first place. While crabs exposed to ship-noise playback were just as likely as ambient-noise controls to detect and respond to a simulated predatory attack, they were significantly slower to retreat to shelter. Ship noise playback also resulted in crabs that had been turned on their backs righting themselves significantly faster than those experiencing ambient-noise playback; remaining immobile may reduce the likelihood of further predatory attention. The anti-predator behaviour of European eels, in response to both simulated pursuit and ambush predators, was also negatively affected by playback of ship noise in comparison to ambient-noise playback (Simpson et al. submitted). European eels experiencing ship-noise playback were caught significantly faster (ANOVA: F1,54 = 10.78, P = 0.002; Figure 4a), were significantly less likely to startle to a looming stimulus (chi-squared test: χ2

1 = 8.57, n = 48, P = 0.003; Figure 4b), and were significantly slower to startle when they did so (ANOVA: F1,23 = 10.26, P = 0.004; Figure 4c). European sea bass were also significantly less likely to startle to a looming stimulus when exposed to playback of pile-driving noise compared to ambient harbour noise (chi-squared test: χ2

1 = 5.46, n = 36, p = 0.019; Figure 4d; Everley et al. in prep.). Utilising the tool kit developed in this project, complementary work on cooperatively breeding Neolamprologus pulcher (a cichlid fish species) found that, in comparison to ambient-noise playback, playback of boat-pass noise: (i) reduced digging behaviour, which is vital to maintain hiding and breeding shelters; (ii) decreased defence against predators of eggs and fry, with direct consequences for fitness; and (iii) increased the amount of aggression received and submission shown by subordinates (Bruintjes & Radford 2013; Appendix E). Using a combination of playbacks and choice chambers in open-water conditions, boat noise was found to have a disruptive effect on the response of fish larvae to natural reef sounds, with implications for settlement and population dynamics (Holles et al. 2013; Appendix F). Moreover, work at Lizard Island, Australia on coral reef fish has found that the balance between predators (Pseudochromis fuscus) and prey (Pomacentrus amboinensis) is affected by anthropogenic noise (Simpson et al. in prep.). Significantly more prey were consumed in the presence of boat noise compared to ambient noise conditions in both tank-based experiments utilising playbacks (two-sample t-test: t34 = 4.40, p < 0.001; Figure 5a) and open-water experiments with real boats (t42 = 3.96, p < 0.001; Figure 5b). Part of the reason for this increase in predation was that significantly fewer attempts were needed by the predator for each successful capture during exposure to boat noise compared to ambient conditions (t28 = 4.34, p < 0.001; Figure 5c).

10

Figure 4. Playback of anthropogenic noise impacts anti-predator behaviour. Shown are (a) the mean ± SE time to capture (in a simulated pursuit paradigm), (b) number of individuals exhibiting a startle response to a looming stimulus, and (c) the mean ± SE time to startle to the looming stimulus by European eels, and (d) the number of European sea bass exhibiting a startle response to a looming stimulus.

Figure 5. Boat noise alters the balance between predators and prey. Shown are the mean ± SE number of prey captured in (a) tank-based and (b) open-water experiments, and (c) the mean ± SE number of attempts made by predators per successful capture in two noise treatments.

3.1.4 Combined approach

There is typically a difficulty in marrying the understanding of environmental disturbance effects on individuals and understanding possible population-level effects (Angeloni et al. 2008). A resource-allocation approach has the potential to bridge these different scales of focus (Congdon et al. 2001), since the trade-off between allocation of energetic resources to basic body maintenance and current activity with growth, storage and reproduction will directly constrain each individual’s contribution

11

to population dynamics. An integrated approach, examining disturbance effects through multiple mechanisms (both physiological and behavioural), is required to obtain the best appreciation of likely impacts (Beale 2007; Cooke et al. In press). This approach was explored using playback experiments with juvenile European eels (Purser et al. in review; Appendix G). Eels significantly increased their metabolic rate (oxygen usage) in response to ship-sound playbacks compared to playbacks of location-matched ambient sounds in a closed respirometry test, but did not show any significant alterations to their activity budgets in a familiar tank test. In a modified ‘open field test’, a novel non-enriched arena designed to act as a mildly adverse scenario, eels showed significant increases in classic anxiety- (erratic swimming, tank-diving) and fear-related (freezing) responses during exposure to ship sounds compared to ambient control playbacks. These results suggest that the addition of acoustic noise to an otherwise quiet environment has the potential to act as a costly stressor: increased metabolic costs and responses indicative of a classic stress-response, associated with allocation of energy resources away from investment in growth and reproduction, towards a state of readiness or reaction to the stressor. 3.2 Interspecific variation in responses

It is likely that there will be stable interspecific differences in susceptibility and responses to elevated noise levels depending on variation in, for example, hearing ability (Fay et al. 2008) and mechanisms of physiological stress response (Hofer & East 1998). Any such differences could affect the relative success of, for instance, competitors or predators and prey, and so potentially impact community composition and structure. However, direct comparisons of species in response to the same anthropogenic noise source in the same contexts are rare (but see Halvorsen et al. 2012). Utilising the tool kit developed and validated during this project (see Section 3.1), complementary work has demonstrated that both the foraging (Voellmy et al. submitted) and anti-predator behaviour (Voellmy et al. in prep. a) of sympatric fish species are affected differently by exposure to the same noise source. For example, using an established foraging paradigm (see Purser & Radford 2011), it was found that both three-spined sticklebacks and European minnows consumed significantly less food during playback of ship noise compared to control playbacks (GLM: F1,53 = 7.09, p = 0.010; Figure 6a). However, whereas the reduced food intake by sticklebacks resulted from an increased number of foraging errors (interaction between species and noise treatment: LRT1,53 =3.49, p = 0.062; Figure 6b), that in minnows was the consequence of a greater level of inactivity (interaction between species and noise treatment: LRT1,53 = 3.91, p=0.048; Figure 6c) at the expense of foraging time.

Figure 6. Playback of ship noise affects foraging behaviour of sticklebacks and minnows differently. Shown are (a) the mean ± SE proportion of Daphnia consumed, (b) the mean ± SE number of foraging errors made, and (c) the number of individuals that exhibited inactive behaviour in two noise conditions.

12

3.3 Inter-individual variation in responses

Factors such as sex, dominance status, age, size and condition may all influence how members of the same population are affected by a given stimulus, including environmental change arising from human activities (Kiffney & Clements 1996; Huntingford et al. 2006; Xu et al. 2010). Consistent inter-individual differences in response could have impacts on population dynamics and for harvests of commercially important species. Previous empirical work on the impacts of anthropogenic noise has tended to focus on the overall response of cohorts of individuals; as part of this project, we considered potential intra-population differences in response. The increased oxygen usage of shore crabs in response to playback of ship noise compared to ambient noise (Section 3.1.2) was size-dependent: individuals of larger size were found to respond disproportionately more strongly than their smaller counterparts (Wale et al. 2013a; Appendix C). In complementary work using the methods designed in this project, male and female individuals of the cooperatively breeding cichlid fish species Neolamprogus pulcher were found to exhibit different behavioural responses to the same playback of boat noise (Bruintjes & Radford 2013; Appendix E). The behavioural and physiological responses of European eels to playback of ship noise (see Section 3.1) differed significantly depending on individual condition (relative weight) (Purser et al. in prep.). Individuals in poorer condition were more strongly affected than their better-condition counterparts by ship-noise playback, being less likely to detect a predatory threat (interaction between condition and noise treatment: LRT1,84 = 5.34, p=0.021; Figure 7a) and exhibiting a greater increase in respiratory rate (F1,148 = 4.54, p=0.034; Figure 7b).

Figure 7. Condition-dependent responses of European eels to playback of ship noise. Shown are (a) the proportion of individuals that startled to a looming stimulus and (b) the mean ± SE change in respiratory rate when an ambient-noise playback was switched to another ambient-noise playback or playback of ship noise.

3.4 Intra-individual variation in responses

For a variety of reasons, it is unlikely that the same individual will always respond in the same fashion to the same noise source. The response of an animal can, for instance, be dependent on its current situation (e.g. Bell et al. 2009, 2010), with increasing evidence that context can influence the harmful effects of human activities on animal welfare (see Huntingford et al. 2006). In complementary work utilising the tool kit developed for this project, it was found that the impact of anthropogenic noise can be context-dependent: playback of boat noise resulted in a reduction in anti-predator defence by Neolamprogus pulcher (cichlid fish) group members if no eggs were present in a nest, but not if eggs were present (Bruintjes & Radford 2013; Appendix E). Moreover,

13

social interactions between dominants and subordinates were affected differently by the same noise playbacks depending on whether group members were engaged in defence behaviour or nest digging (Bruintjes & Radford 2013). The implication is that responses to anthropogenic noise are not fixed, but rather show some element of flexibility, which may or may not be under the control of the individual. 3.5 Repeated exposure and prior experience

Responses to pollutants may change across time (Piola & Johnson 2009; Whitehead et al., 2010) as a result of such processes as habituation, tolerance and sensitization (Bejder et al. 2009). Although noise-related experiments have, from an understandable logistical perspective, often involved a single presentation of the relevant stimulus (see also Section 3.1), organisms in most natural situations are likely to experience chronic or repeated noise exposure. Exploring how responses can change and are dependent on prior experience is thus important (Simpson et al. 2010). While shore crabs repeatedly exposed to ambient-noise playback increased their oxygen consumption (perhaps due to handling stress), those individuals repeatedly exposed to playback of ship noise did not exhibit a similar change (Wale et al. 2013a; Appendix C). It is possible that they already show a maximum response on first exposure to ship-noise playback, but they might also be habituating or becoming tolerant over time. While acute exposure to playback of ship noise results in a greater number of startle responses from larval Atlantic cod, individuals that had been previously exposed to ship-noise playback during rearing startled less often than those reared in ambient noise conditions (GLMM, interaction between rearing condition and acute noise exposure treatment: χ2

2 = 6.69, p = 0.035; Figure 8a; Holles et al. in prep. a). In complementary work in open-water conditions with playback of boat noise, responses of marine fish to anthropogenic noise were also found to change with repeated exposure during early life stages (Holles et al. in prep. b). For instance, acute exposure to boat-noise playback resulted in an increase in respiratory rate by threespot dascyllus (see also Section 3.1.2), but the effect was significantly less strong in individuals that had previously been exposed to repeated boat-noise playback compared to those exposed to ambient-noise playback (GLMM, interaction between repeated-exposure noise treatment and acute noise exposure treatment: χ2

3 = 81.80, p < 0.001; Figure 8b).

Figure 8. Repeated playback of vessel noise alters subsequent responses to anthropogenic noise playback. Shown are (a) the mean ± SE number of startle responses by Atlantic cod larvae reared in three noise regimes and (b) the mean ± SE change in respiratory rate of threespot dascyllus larvae when playback was switched from one ambient track to another or one of boat noise.

By experimentally manipulating the noise in holding tanks, the importance of prior exposure was demonstrated: European minnows kept in louder and quieter conditions exhibited different baseline

14

behaviour and also responded differently to the same acute playbacks of ship noise (Voellmy et al. in prep. b). For instance, minnows from loud holding tanks responded significantly more often than those from quiet holding tanks to the same visual predatory stimulus (GLMM: χ2

1 = 4.67, p = 0.031; Figure 9a). However, whereas minnows from loud holding tanks did not differ significantly in their latency to respond to the predatory stimulus depending on whether they were currently experiencing ship-noise or ambient-noise playback (GLMM: χ2

1 = 1.27, p = 0.260), those from quiet holding tanks took significantly longer to respond during acute playback of ship noise compared to ambient noise (χ2

1 = 4.63, p = 0.031; Figure 9b).

Figure 9. Prior noise exposure conditions influence subsequent behaviour and responses to ship-noise playback in European minnows. Shown are (a) the number of individuals startling to a visual predatory stimulus and (b) the mean ± SE time to respond to the visual predatory threat.

3.6 Variation in noise characteristics

As distance from the source increases, sound intensity will naturally decrease, but dose-dependent relationships for the impact of anthropogenic noise are currently lacking (Popper & Hastings 2009). Moreover, variation in characteristics beyond just absolute amplitude should be considered; differences in, for instance, periodicity, frequency and amplitude fluctuations, and temporal patterns may well result in different impacts (De Boer et al. 1989; Popper & Hastings 2009). Tank-based experiments were used to explore the shape of the dose-response curve in European eels (Kerridge 2012); absolute values would need to be established in natural conditions (see Section 4). Early indications are that there is an S-shaped (polynomial) relationship, rather than a sudden threshold. The importance of noise regularity on responses was examined as part of the developmental work on Atlantic cod (Holles et al. in prep. a). Larvae appeared to be more strongly affected by regular vessel noise than by irregular vessel noise: when exposed to regular compared to irregular noise, larvae used up more of their yolk sac (see Section 3.1.1; Figure 1a) and were in worse body condition 16 days post-hatching (Section 3.1.1; Figure 1c); larvae exposed to irregular noise had a reduced startle response to acute ship-noise playback compared to those that had been repeatedly exposed to irregular noise (Section 3.5; Figure 8a). 3.7 Recovery from noise exposure

Many anthropogenic noise events are transient in nature (Hildebrand 2009; Popper & Hastings 2009), and short-term impacts of noise may not necessarily translate into long-term consequences (see Bejder et al. 2006). While it is clear that behaviour and physiology can be detrimentally impacted during the period of elevated noise (see Section 3.1), effects on survival and reproductive success will be dependent on whether, and how quickly, the affected individuals recover to baseline performance levels and if they can compensate. Species will differ in their ability to recover and

15

compensate (Voellmy 2013), and compensation itself may carry a variety of inherent costs (see Purser & Radford 2011). In tank-based experiments, European eels quickly returned to baseline levels following acute exposure to ship-noise playback (Bruintjes et al. in prep.). The significant reduction in startling to a looming stimulus observed during ship-noise playback (Section 3.1.1) was not sustained when the noise ceased (chi-squared test: χ2

2 = 8.29, P = 0.016; Figure 10a). Similarly, the increased respiratory rate of eels during exposure to ship-noise playback compared to ambient-noise playback (Section 3.1.2; two-sample t-test: t154 = 3.97, P < 0.001; Figure 10b) was not sustained after the anthropogenic noise ceased (t154 = 0.91, P = 0.363; Figure 10b). Likewise, the increased respiratory rate of European sea bass in response to playback of pile-driving noise compared to ambient noise (Section 3.1.2; paired t-test: t17 = 6.20, P < 0.001; Figure 10c) quickly disappeared once the noise ceased (t17 = 1.46, P = 0.162; Figure 10c; Everley et al. in prep.).

Figure 10. Fast recovery to baseline levels following acute playback of anthropogenic noise. Shown are (a) the number of European eels startling to a looming predatory stimulus during the second of two playback tracks (ambient then ambient, ambient then ship noise, ship noise then ambient noise), (b) are mean ± SE change in respiratory rate of European eels in response to ship-noise playback, and (c) are mean ± SE change in respiratory rate of European sea bass in response to playback of pile-driving noise.

3.8 Results summary

Our experiments indicate that both vessel noise (generated by ships and boats) and pile-driving noise have the potential to impact the development, physiology and behaviour of a range of fish and invertebrate species. However, the magnitude of the impact is likely to differ between species and between individual members of the same species, depending on such characteristics as size and condition. Moreover, the response of a given individual to a given type of noise is likely to differ depending on, for instance, the context, prior noise exposure and the particular characteristics of the noise source. Finally, it is possible that individuals will quickly recover from short-term exposure to noise, although whether that occurs following longer exposures and whether there are lasting consequences remain to be tested.

4. Discussion of Results

Ultimately required are assessments of how anthropogenic noise impacts individual fitness (survival and reproductive success), population viability and community structure. Correlative studies do not allow the effect of noise to be isolated and therefore do not allow strong conclusions (Radford et al. 2012); hence our focus on experimental manipulations. Animals might potentially compensate for many demonstrated responses to anthropogenic noise (e.g. changes in movement patterns or acoustic communication), and thus there might not be a direct link to long-term impacts (Bejder et

16

al. 2006); hence our focus on biological parameters that have a meaningful link to long-term fitness and contribution to population parameters (e.g. anti-predator behaviour, mortality). Responses to a given noise stimulus are unlikely to be fixed – differences may exist between species, individuals and across time – and such differences will affect ultimate consequences (Radford et al. 2014); hence our inclusion of experiments beyond an initial consideration of whether noise has an impact. We chose to conduct our initial experiments in tanks to control carefully the conditions and contexts of the study animals, and to establish what developmental, physiological and behavioural responses are potentially impacted by anthropogenic noise and thus worth considering further. Such an approach also allows more detailed data collection than is often possible in natural conditions, thus making more feasible the investigation of questions beyond whether noise has an impact (see Sections 3.2 – 3.7). Care must be taken when extrapolating results from tank-based experiments to meaningful implications for free-ranging animals in open water for a variety of reasons. From a biological perspective, captive animals are usually more constrained than in the wild and individuals are receiving husbandry regimes that may not replicate natural conditions of resource availability and other environmental stressors. From an acoustics perspective, playbacks cannot perfectly replicate natural sound sources and the sound field in a tank is complex (Parvulescu 1964, 1967; Okumura et al. 2002). Moreover, although some species of fishes can detect sound pressure, all teleost fishes and marine invertebrates are likely to detect the particle motion component of sound (Bleckmann 2004; Mooney et al. 2010); measurements of particle velocity are not currently possible in tanks. However, our open-water experiments, using both playbacks and real noise sources, generated qualitatively similar results to those conducted in tanks, thus validating those initial findings. While tank-based experiments can never generate absolute threshold values for use by regulators – experiments in natural conditions with real sound sources are required for this – they can provide a valuable first step in that process (in determining how best to measure responses and what responses to target), as well as offering an opportunity to answer related, important questions that would be logistically challenging in the wild.

5. Future Work

In addition to the generation of many novel and scientifically robust findings relating to the potential impact of anthropogenic noise (see Section 3), a major output of this project is an integrated and validated tool kit of experimental approaches and an understanding of which biologically meaningful response parameters to focus on. This puts us in a unique position to conduct systematic assessments of focal study species of fish and invertebrates near to real-world disturbances (e.g. shipping lanes, harbours, windfarm sites), and thus directly measure impacts of anthropogenic noise on marine life in ways of direct relevance to managers, policy-makers and regulators. Moreover, we can utilise that tool kit to assess mitigation measures as they are developed; currently the acoustic output of such measures are determined, but not any change in impact on marine life. The strong links we have developed with, for instance, bioacousticians who are measuring both sound-pressure and particle-velocity components of sound, as well as developing soundscape models (e.g. Geospectrum, Pete Theobold at the National Physics Laboratory, Paul Lepper at Loughborough University), and with theoreticians who are developing agent-based models to predict the impact of noise sources (e.g. HR Wallingford) further enhance our capabilities moving forward. We are therefore keen to explore further funding opportunities that will allow us to make valuable headway in the next 5 years in measuring and managing biologically-important impacts of marine anthropogenic noise, and thus help deliver on environmental impact assessments, reduce the uncertainty surrounding major planned Marine Renewable Energy programmes, and aid in the delivery of Good Environmental Status as per Descriptor 11 of the European Commission Marine Directive.

17

6. Actions Resulting from the Work

We have so far published five peer-reviewed papers (Purser & Radford 2011; Bruintjes & Radford 2013; Holles et al. 2013; Wale et al. 2013a, b) and have another 14 either in review or soon to be submitted; five conference proceedings (from The Effects of Noise on Aquatic Life Conference, Budapest) are also in press:

Bruintjes, R., Purser, J., Radford, A.N. & Simpson, S.D. (In prep.) Fast recovery from playback of anthropogenic noise.

Bruintjes, R. & Radford, A.N. (2013) Context-dependent impacts of anthropogenic noise on individual and social behaviour in a cooperatively breeding fish. Animal Behaviour 85: 1343–1349.

Everley, K.A., Radford, A.N. & Simpson, S.D. (2014) Pile-driving noise impairs anti-predator behavior of the European sea bass Dicentrarchus labrax. Conference Proceedings.

Everley, K.A., Radford, A.N., Bruintjes, R. & Simpson, S.D. (In prep.) Playback of pile-driving noise impacts the behaviour and physiology of European sea bass.

Holles, S., Lecchini, D., Simpson, S.D. & Radford, A.N. (In prep. b). Early-life impacts of anthropogenic noise in coral reef fish.

Holles, S., Radford, A.N., Simpson, S.D., Lecchini, D. & Mills, S.C. (Submitted) Anthropogenic noise impairs embryonic development and increases mortality in sea hares. Current Biology.

Holles, S., Simpson, S.D., Morley, E. & Radford, A.N. (In prep. a). Anthropogenic noise impacts early-life development and subsequent behaviour in Atlantic cod.

Holles, S., Simpson, S.D., Radford, A.N., Berten, L. & Lecchini, D. (2013) Boat noise disrupts orientation behaviour in a coral reef fish. Marine Ecology Progress Series 485: 295–300.

Holles S, Simpson, S.D., Lecchini, D. & Radford, A.N. (2014) Playback experiments for repeated boat noise exposure. Conference Proceedings.

Kerridge, E., Purser, J., Simpson, S.D. & Radford, A.N. (In prep.) Dose-dependent responses of European eels to playback of anthropogenic noise.

Purser, J. & Radford, A.N. (2011) Acoustic noise induces attention shifts and reduces foraging performance in three-spined sticklebacks (Gasterosteus aculeatus). PLoS ONE 6: e17478.

Purser, J., Bruintjes, R., Simpson, S.D. & Radford, A.N. (In prep.) Condition-dependent responses of European eels to anthropogenic noise.

Purser, J., Simpson, S.D. & Radford, A.N. (In review) The cost of an acoustic stressor: a resource-allocation approach. Functional Ecology.

Purser, J. & Radford, A.N. (In prep.) Flexible responses to acoustic noise: managing the costs during prolonged and repeated exposures to an acoustic stressor.

Radford, A.N., Purser, J., Bruintjes, R., Voellmy, I.K., Everley, K.A., Wale, M.A., Holles, S. & Simpson, S.D. Beyond a simple effect: variable and changing responses to anthropogenic noise. Conference Proceedings.

Simpson, S.D., Purser, J. & Radford, A.N. (Submitted) Ship noise has the potential to compromise fish anti-predator behaviour, physiology and cognitive functioning. Nature Communications.

Simpson, S.D., Radford, A.N., Holles, S., Ferrari, M.C.O., Chivers, D.P., McCormick, M.I. & Meekan, M.G. (2014) Small boat noise impacts natural settlement behaviour of coral reef fish larvae. Conference Proceedings.

Simpson, S.D., Radford, A.N., Holles, S., Ferrari, M.C.O., Chivers, D.P., McCormick, M.I. & Meekan, M.G. (In prep.) Anthropogenic noise fundamentally impacts the ecology of reef fishes.

Voellmy, I.K., Purser, J., Simpson, S.D. & Radford, A.N. (2014) Effects of previous acoustic experience on behavioral responses to experimental sound stimuli and implications for research. Conference Proceedings.

Voellmy, I.K., Purser, J., Simpson, S.D. & Radford, A.N. (Submitted) Acoustic noise reduces foraging success in two sympatric fish species via different mechanisms. Animal Behaviour.

18

Voellmy, I.K., Purser, J., Simpson, S.D. & Radford, A.N. (In prep. a) Acoustic noise affects anti-predator behaviour of two sympatric species differently.

Voellmy, I.K., Simpson, S.D., Purser, J. & Radford, A.N. (In prep. b) Prior acoustic exposure affects fish responses to anthropogenic noise.

Wale, M.A., Simpson, S.D. & Radford, A.N. (2013a) Size-dependent physiological responses of shore crabs to single and repeated playback of ship noise. Biology Letters 9: 20121103.

Wale, M.A., Simpson, S.D. & Radford, A.N. (2013) Noise negatively affects foraging and antipredator behaviour in shore crabs. Animal Behaviour 86: 111–118.

We have also presented our work at:

NERC Marine Renewable Energy Programme Advisory Board, London, UK

Underwater Sound Forum, Southampton, UK

Mathematics and Theoretical Ecology, Colchester, UK

International Bioacoustics Council, la Rochelle, France

Fisheries Society of the British Isles, Norwich, UK

International Coral Reef Symposium, Cairns, Australia

Student Conservation Conference, Cambridge, UK

Association for the Study of Animal Behaviour, Aberystwyth, UK

Invertebrate Sound and Vibration, Strathclyde, UK

Impact of Noise and Vibration on Fish, Sheffield, UK

European Conference on Underwater Acoustics, Corfu

Bureau of Ocean and Energy Management, San Diego, USA

Society of Biology and The Centre for Science and Policy workshop, London, UK

NERC Marine Renewable Energy Knowledge Exchange Programme Advisory Group, Edinburgh

International Ethological Conference and Association for the Study of Animal Behaviour, Newcastle, UK

Effects of Noise on Aquatic Life, Budapest, Hungary

Invited seminars at the Universities of Auckland, Bristol, Exeter, Glasgow, Liverpool and St Andrews, Universite de Perpignan, and James Cook University

In February 2012, we organised a 70-delegate workshop attended by academics and research end-users, exploring issues and potential solutions around biological impacts of marine noise. Resulting from this workshop and the ongoing programme of research through this Defra project, we have:

Secured a 4-month NERC Business Internship for one of our post-docs to work with HR Wallingford on developing a predictive model of responses of animals to anthropogenic noise.

Secured a 4-month NERC Business Internship for one of our PhD students to work with Ultra Electronics on integrating passive acoustic monitoring with active sonar to monitor cetacean behaviour around tidal turbines.

Secured a 3-year Knowledge Transfer Partnership for one of our post-docs to work with HR Wallingford on further developing the HAMMER model to predict animal responses to pile-driving noise.

Secured a small contract with Marine Scotland to measure salmon hearing. We anticipate this leading to wider collaboration with studies of salmon behaviour and physiology, and field-based experimentation.

Secured a PhD studentship with CGG (global geoscience company) to explore impacts of traditional and novel seismic survey approaches on marine life

Secured a postdoc contract with Ecocean (French biodiversity restoration company) to write a review on the impacts of noise

Throughout the project, we have kept a close eye on the policy process within the International Maritime Organization (IMO), where underwater noise from ships had been flagged as an

19

environmental issue of concern, now appearing on the agenda of the Marine Environment Protection Committee (MEPC). Project collaborator Cato ten Hallers-Tjabbes has been briefing the relevant UK government representative in MEPC (Mr Jonathan Simpson, Head of Environmental Policy, Marine Coast Guard Agency) on the findings of this Defra project and its relevance to the present developments in IMO. Those findings have fed into the drafting of guidelines for the reduction of underwater noise from commercial shipping that have been prepared for consideration at the next MEPC (66th meeting, 31 March – 4 April 2014). The project results fit into several relevant policy issues, such as: early warning, monitoring, spatio-temporal measures, species- or population-specific measures. As Dr ten Hallers-Tjabbes represents IUCN (International Union for the Conservation of Nature) in MEPC and other IMO meetings, she has also discussed with Mr Simpson the option to have a submission to MEPC 66 alongside the one from the UK Government.