239 DecemberJanuary 20121 DRAFT

Formatted: Don't suppress line numbers Formatted: Section start: New column 1

2

3

4

5

6

7 Science Plan

8 International Quiet Ocean Experiment

239 DecemberJanuary 20121 DRAFT

9 Table of Contents 10

11 1. Executive Summary 4 12 13 2. Introduction and overview 5 14 2.1 Purpose of the document, 5 15 2.2 Rationale for an IQOE, 6 16 2.3 Background, 7 17 2.4 Defining the Questions, 11 18 2.5 IQOE Objectives, 12 19 2.6 Anticipated benefits, 13 20 2.7 Activities, 15 21 22 3. Theme 1: Ocean Soundscapes 17 23 3.1 Characterization of Soundscapes, 17 24 3.2 Ambient Sound and the Components of Soundscapes, 21 25 3.3 Modeling Soundscapes (Sources and Propagation), 27 26 27 4. Theme 2: Defining the effects of sound on marine organisms 31 28 4.1 Introduction, 31 29 4.2 Acoustic Environments, 32 30 4.3 Unplanned or Semi-planned Experiments, 34 31 4.4 Research Approach, 39 32 4.5 Summary, 44 33 34 5. Theme 3: Observing sound in the ocean 45 35 5.1 Introduction, 45 36 5.2 Acoustic Observation Networks, 45 37 5.3 Acoustics and Global and Regional Ocean Observing Systems, 46 38 5.4 Integration Within Existing Systems, 47 39 5.5 New Systems Designed for IQOE, 48 40 5.6 Extracting useful scientific information from data collected for regulations, 49 41 5.7 Data collection (including standardization), quality control, analysis, reporting, 42 management, and accessibility, 50 43 5.8 Biological Observing Systems, 51 44 5.9 Synthesis and modeling: physical, biological, and acoustic, 52 45 5.10 Recommendations, 53 46 47 6. Theme 4: Industry and regulation 56 48 6.1 Introduction, 56 49 6.2 Risk Frameworks, 58 50 6.3 Routine Sound Monitoring, 61 51 6.4 Research Priorities for Regulators and Industry, 62 52 6.5 Finding Novel Solutions, 66 53 6.6 Key Recommendations, 66 54 6.7 Engaging Industry and Other Stakeholders, 66 55 56 7. Implementation 69 57 7.1 Project Structure, 69 58 7.2 Operating Approach, 70 59 7.3 Governance, 70 60 7.4 Project Management, 71 61 7.5 Communication, 72

239 DecemberJanuary 20121 DRAFT

62 7.6 Education and Capacity Building, 72 63 7.7 Relationship with other organizations, programs and activities, 73 64 7.8 Work Streams and Workshops, 74 65 7.9 Funding, 75 66 67 References 81 68 Appendix I 86 69 Appendix II 87

239 DecemberJanuary 20121 DRAFT

70 1 71 1 72 Executive Summary 73 74 Executive summary 75 76 The International Quiet Ocean Experiment (IQOE) will create an international global 77 program of research, observation, and modeling to better characterize ocean sound 78 fields and to promote understanding of the effects of sound on marine life. The effect 79 of anthropogenic sound on marine life is a large gap in our current knowledge; the 80 resulting scientific uncertainty is leading to the implementation of precautionary 81 legislation with potentially large economic consequences. 82 83 A central feature of the IQOE will be an International Year of Ocean Acoustics 84 (IYOA), which will focus the participating scientific, industrial, environmental, and 85 naval communities on the goal of an intense period of scientific activity, coordinated 86 across regions to create a global program. The IYOA will raise awareness of the 87 effects of sound in the ocean within the participating communities and in the broader 88 public realm. 89 90 The IQOE will address 5 key questions: 91 92 1. What are the current levels and distribution of anthropogenic sound in the ocean? 93 2. What are the trends in anthropogenic sound levels across the global ocean? 94 3. What are the effects of anthropogenic sound on important marine animal 95 populations? 96 4. What was the global ocean soundscape before humans arrived? 97 5. What are the potential future effects of sound on marine life? 98 99 The IQOE has been motivated by evidence of increasing sound levels from human 100 activities. Considerable evidence exists that the human contribution to ocean sound 101 has increased during the past few decades and that anthropogenic sound has 102 become the dominant component of marine sound in some regions. Anthropogenic 103 sound levels are likely to be directly correlated with the increasing industrialization of 104 the ocean, thus making the measurement of ocean sound fields an important tool for 105 assessing industrial presence in the ocean. 106 107 Sound is an important factor in the lives of many marine organisms and increasing 108 theory and observations suggest that human-generated sound in the ocean could be 109 approaching levels at which negative effects on marine life may be occurring. Certain 110 species already show symptoms of the effects of sound. Although some of these 111 effects are acute and rare, chronic sub-lethal effects may be more prevalent, but are 112 difficult to measure. 113 114 The IQOE will mobilize the participating communities to investigate sound in the 115 ocean in a way that will be useful for mitigation and for management of sound 116 sources. The IQOE will (1) ensure that the measurement of the sound field becomes 117 an integrated part of global ocean observations; (2) develop a global approach to

1 The contributors to the Science Plan and their roles are given in Appendix II.

4 IQOE Science Plan 239 DecemberJanuary 20121 DRAFT

118 investigating ocean sound, including a community-based approach to research; (3) 119 support innovation in ocean observing systems; (4) support data management and 120 the development of data standards; (5) develop acoustic models; (6) support the 121 planning and implementation of regional experiments; and (7) ensure constructive 122 engagement with industry and the public. 123 To achieve these goals, the IQOE will be implemented across four themes: 124 125 1. The Ocean Soundscapes theme will describe ocean soundscapes from regional to 126 global scales. This theme will include the identification of sound sources, the 127 modeling of acoustic propagation, and the validation of these models using ocean 128 observation systems. This theme will be the main focus of efforts to measure trends 129 in ocean sound and to define sound budgets within regions. It will also investigate 130 soundscape diversity and examine the concept that the conservation of soundscapes 131 may be an appropriate objective for the integrated management of the marine 132 environment. 133 134 2. The Effects of sound on marine organisms theme will plan and carry out 135 experiments. This may include experiments to quiet regions and to observe the 136 responses of marine organisms. This theme will include the use of planned 137 experiments as well as opportunistic studies using post-hoc statistical modeling to 138 test for effects. This theme is the main vehicle through which the biological 139 significance of sound will be assessed and, where possible, this will be focused on 140 measuring dose-response relationships so that assessments of the effects of sound 141 can be predictive, with special emphasis on the Population Consequences of 142 Acoustic Disturbance (PCAD) approach. Much of this theme will rely upon the use of 143 a small set of representative species. 144 145 3. The Observing sound in the ocean theme will be the primary focus for addition of 146 sound measurements to existing and future observing systems and will encourage 147 technical innovation. This theme will develop data standards—where these do not 148 exist—and will promote observation of the key biological and physical variables. 149 Much of the synthesis, data management, and acoustic modeling needed by the 150 IQOE will be managed from within this theme. 151 152 4. The Industry and Regulation theme will develop the methodology for “noise” 153 monitoring within regulatory regimes. This theme will examine the operational 154 management of sound in the ocean through risk analysis by, among other 155 approaches, defining appropriate thresholds for disturbance and damage to marine 156 life. It will also help regulators to measure compliance, and industry to maintain its 157 activities, by providing innovative solutions to barriers presented by regulation. The 158 theme will organize and establish the communications strategy for the IQOE. 159 160 Each of these four themes will be important in preparations for the IYOA. 161 162 The IQOE will be implemented under the governance of the Scientific Committee on 163 Oceanic Research (SCOR) and the Partnership for Observation of the Global 164 Oceans (POGO). The project will be managed by a Steering Committee that will be 165 supported by a secretariat. Two standing committees, the Scientific Committee and 166 the Data Management and Communication Committee, will be tasked with ensuring 167 the themes are implemented and, under these, temporary ad hoc working groups will 168 be established to plan and implement particular activities. The Steering Committee 169 will be responsible for international planning and coordination of contributing national 170 activities funded from national sources. The IQOE will plan a series of workshops in

5 IQOE Science Plan 239 DecemberJanuary 20121 DRAFT

171 its first two years, which will lay the foundation for IQOE implementation. Particular 172 emphasis in the early years of the project will be given to characterizing global trends 173 in ocean sound, gaining access to existing long time series of ocean sound, 174 specifying standards for observations and experimentation, designing a data 175 management system, and planning for the IYOA. 176 177

178 We will write this section after the other sections of the report have been reviewed by the 179 OSM participants.

6 IQOE Science Plan 239 DecemberJanuary 20121 DRAFT

180 2 181 182 Introduction and overview 183 184 2.1 Purpose of the document Formatted: Heading 2, Justified 185 Underwater sound is important for many marine organisms. In general, marine species rely 186 more on sound to support life functions than terrestrial organisms because the ocean is 187 relatively opaque to light and transparent to sound. Ocean sound has many components, 188 including naturally occurring sound sources such as storms and breaking waves, the 189 fracturing of sea ice, subsurface volcanoes and seismic activity, along with the calls and 190 other sounds produced by a great variety of marine species. In the mid-1800s, a new 191 source of sound began to fill the ocean, driven by the rapid spread of mechanical propulsion 192 in the shipping industry (Figure 2.1). 193 Formatted: Centered

194 195 196 Figure 2.1. Examples of spectrograms showing different forms of sound in the ocean. (a) Comment [EU1]: We will try to add a color bar 197 Sound from a ship where the vessel had been travelling at full power between intervals of to show the relationship between color and sound 198 power off; and (b) several different natural sounds. The intensity of the color represents intensity. 199 sound intensity with darker color showing high intensity of sound.

7 IQOE Science Plan 239 DecemberJanuary 20121 DRAFT

200 Closely linked to the global economy (Frisk, 2007), the sound environment is changing 201 quickly; additional sources related to naval activities, offshore construction, oil and gas 202 exploration, fisheries and recreational boating now add to the mix. Intense sound can have 203 acute impacts on animal life, but lower levels of continuous sound may lead to chronic 204 effects, many of which will be difficult to detect. To the extent that human activities are 205 steadily increasing ocean sound levels, it can be inferred that chronic effects will increase. 206 207 A scientific analysis of adverse effects from anthropogenic sound in the ocean is difficult 208 because of the wide range of acoustical sources, variations in frequency, intensity and 209 occurrence, and the complexity of acoustic propagation, especially in strongly stratified and 210 shallow water. A realistic approach to understanding underwater sound and its effects on 211 organisms requires new interdisciplinary and international collaboration, generally lacking in 212 the past. 213 214 The International Quiet Ocean Experiment (IQOE) provides a framework for a decade-long 215 project of research, observations, and modeling, aimed at improving our understanding of 216 sound in the ocean and its effects on marine organisms. The project will include carefully 217 designed observations exploiting situations of varying sound inputs in conjunction with 218 detailed model analysis. The project will build toward a period of intensive study of sound in 219 the ocean, a Year of Ocean Acoustics (see Chapter 7). 220

221 2.2 Rationale for an IQOE 222 Does the sound made by humans harm marine life? At present, we can offer only 223 preliminary answers to this important question, for only a few species. We know that the 224 ocean has become more industrialized and that the sound levels associated with human 225 activities have increased (NRC, 2003). For example, in areas where measurements have 226 been made, anthropogenic sound in the ocean has been increasing across much of the 227 frequency spectrum (Andrew et al., 2002; McDonald et al., 2008), and especially at lower 228 frequencies (<500 Hz) (Frisk 2007, Andrew et al. 2011) (Figure 2.2). However, given the 229 spatial and temporal complexity and variability in all sound sources, the relative contribution 230 of anthropogenic sound is not always readily distinguishable. 231 232 The effect of sound on marine life is an important unknown of marine science. Considerable 233 evidence has accumulated that the human contribution to ocean sound has increased during 234 the past few decades; anthropogenic sound has become the dominant component of some 235 marine soundscapes. Sound is an important factor in the lives of many marine organisms 236 (Figure 2.3), and increasing theory and observations suggest that human sound could be 237 approaching levels at which marine life may be experiencing chronic negative effects. 238 239 Some species show symptoms indicative of negative effects of sound. Although some of 240 these effects are acute and rare in occurrence, chronic sub-lethal effects may be more 241 prevalent, but are much more difficult to measure. We need to identify the thresholds of such 242 effects for different species and be in a position to predict how increasing anthropogenic 243 sound will increase the chronic effects. The IQOE is being developed with an objective of 244 coordinating the international research community to both quantify the ocean soundscape 245 and examine the functional relationship between sound and the viability of key marine 246 organisms. This has implications for the stability and resilience of marine ecosystems and 247 future marine productivity. 248 249

8 IQOE Science Plan 239 DecemberJanuary 20121 DRAFT

250 251 Figure 2.2. Historical and contemporary shipping traffic sound levels from Andrew et al. Formatted: Justified, Don't adjust space 252 (2011).Solid lines represent the model fits to the APL/UW data in Andrew et al. (2011), between Latin and Asian text, Don't adjust 253 shown over the temporal span of the actual dataset. Thin dotted lines connect space between Asian text and numbers 254 measurements for the same band for each system. The heavy dashed line indicates the 255 trend suggested by Ross (2005), which was based broadly on data from many systems in 256 both the Atlantic and Pacific oceans, and not specifically on data from the systems used in 257 Andrew et al. (2011). 258 259

260 2.3 Background 261 The potential impact of anthropogenic sound on marine life is a matter of societal concern. 262 Most people are unaware of ocean sound, yet it is a vital part of the ocean environment, 263 important not only to large marine mammals, but also to fish and other animal groups. 264 Environmental non-profit organizations are more aware of the issue and can motivate action of 265 various kinds, such as industrial guidelines and possible regulation designed to reduce the 266 impact of sound on marine life. Such action can be expensive and must therefore be based on 267 robust scientific evidence. Moreover, the results of such understanding need to be effectively 268 communicated to the public so as to foster rational discussion and public support for 269 meaningful and justifiable action. 270 271 272 The basic scientific foundation for management of sound in the ocean falls into two related 273 categories. First, the existing sound field should be described adequately. This description 274 cannot be represented by a single number, but must rather be a quantitative description of the 275 kinds of sound that exist, and their frequencies, intensities, and variations in both space and 276 time. An important example of the value of long time series observations of key 277 environmental parameters is the ”Keeling Curve” which documents the changes in

9 IQOE Science Plan 239 DecemberJanuary 20121 DRAFT

278 atmospheric CO2 concentration (Keeling, 1960) Documenting changes in the ocean sound 279 environment will require long-term measurement with appropriate hydrophone arrays across 280 many regions, together with analysis that identifies trends in different contributions. 281 Technological advancements allow such measurements to span a broad frequency range and 282 to exploit internal data recording, thus avoiding the need for long and expensive transmission 283 cables, although existing systems such as hydrophones used for Comprehensive Test Ban 284 Treaty verification would provide a helpful contribution. Underwater sound needs to be 285 understood and modeled in terms of the different sound sources—both natural and 286 anthropogenic—and the propagation characteristics that contribute to the building of a 287 ‘soundscape’2. This will require development of comprehensive numerical models of ocean 288 sound fields based on knowledge and measurements of the sources and of the propagation 289 environment. Such models can be used to explore the relative significance of different 290 sources, guide design of further measurements, and provide valuable tools for planning 291 mitigation efforts where these are found necessary. 292 293

294 295 Figure 2.3. The hearing ranges of different kinds of fish and mammals together with the 296 overlap in frequency with different sources of man-made sound (from Boyd et al., 2011). 297 298 299 In addition to a basic characterization of ocean soundscapes, the biological impact of the 300 sound must be studied to guide appropriate management of sound in the ocean. For a 301 particular area this will include knowledge of the species that occur and their sensitivity to 302 acoustical interference. Given the general state of our knowledge of biological sensitivity to

2A soundscape is a description of an acoustic environment.

10 IQOE Science Plan 239 DecemberJanuary 20121 DRAFT

303 sound this represents, by far, the most challenging aspect of the scientific and management 304 problem. This challenge is unlikely to be resolved quickly or completely. Nevertheless, a good 305 understanding of the effects of anthropogenic sound on marine life remains essential to 306 rational decision-making and is a central goal of the IQOE. 307 308 The deep ocean environment creates a sound channel by which low-frequency acoustic 309 waves can be transmitted over large distances, sometimes hundreds of kilometers and often 310 much further. The complex pathways taken by this sound affect the final received levels, but 311 if they are averaged through time at the receiver they provide an integrated signal defined by 312 the relative locations of all the sound producers, the architecture of the ocean basin, and the 313 properties of the water through which the sound has passed. It is sometimes possible to 314 distinguish among different sound sources based on sound characteristics. 315 316 Humans introduce sound to the ocean through many different activities. Each source may 317 have different effects, depending upon frequency range, its intensity and whether it is an 318 intermittent, pulsed, or a continuous sound. Some anthropogenic sounds—such as some 319 military sonar, seismic air guns used extensively for oil and gas exploration, and pile 320 driving—are both impulsive and high intensity. These types of sounds can elicit strong 321 negative reactions, or even physical injury, in some marine species. This concern has led to 322 higher levels of scrutiny for many of those sources. Recently, military sonars have been a 323 particular focus of attention because of their association with the stranding of beaked whales 324 (Cox et al., 2006). Nevertheless, the acute effects of sonars upon beaked whales probably 325 occur only rarely because the effects of sonars themselves co-vary with other factors, such 326 as context of the exposure (i.e., bathymetry, presence of surface temperature ducts, 327 behavior, and number of naval vessels). Animal strandings are probably the most easily 328 observed end point of a syndrome of behavioral responses to sound in these species (Boyd 329 et al., 2007), leading through some unknown progression to physical harm and/or mortality. 330 There is a strong suspicion, supported by increasing evidence, that a similar syndrome of 331 reduced capacity to perform normal life functions is present across a wide range of marine 332 fauna, including fish (Slabbekoorn et al., 2010) and marine mammals (Southall et al., 2007; 333 Tyack, 2008). 334 335 A major unanswered question in almost all of these cases is whether there is a significant 336 impact on the fitness of individuals within populations that jeopardizes the viability of those 337 populations. This was addressed by the U.S. National Research Council in its 2005 report on 338 marine mammals and ocean sound (NRC, 2005), but the principles apply equally to all forms 339 of marine life. The NRC report developed an approach known as Population Consequences 340 of Acoustic Disturbance (PCAD) (Figure 2.4), which defines a rationale for developing 341 assessments of the significance of sub-lethal effects and for identifying the most important 342 gaps in our knowledge. The greatest challenge is to define the functional relationships 343 between behavioral and physiological responses to sound and then the population effects— 344 an essential requirement of this assessment process (Figure 2.5). Defining the transfer 345 functions between the different boxes of the PCAD framework (Figures 2.4 and 2.5) and 346 building knowledge of biological significance will be challenging. However, it may be possible 347 to make progress through a combination of modeling transfer functions, leveraging multiple 348 data sets derived from ongoing observational studies, passive acoustic monitoring, direct 349 measurements, and the attachment of tags capable of measuring the received level of sound 350 to marine organisms. 351 352 Shipping is an important anthropogenic sound source (Wenz, 1962). The volume of cargo 353 transported by sea has been doubling approximately every 20 years 354 (http://www.marisec.org/shippingfacts/worldtrade/volume-world-trade-sea.php), resulting in 355 an increase in anthropogenic sound from this source. Although the measurement of sound in 356 relation to these changes has been mostly local and is incomplete, the current estimate is 357 that increased shipping has been accompanied by an increase in anthropogenic sound in

11 IQOE Science Plan 239 DecemberJanuary 20121 DRAFT

358 frequencies below 500 Hz. From 1950 to 2000, the shipping contribution to ambient sound 359 increased by as much as 15 dB, corresponding to a rate of increase of approximately 3 dB 360 per decade (Andrew et al., 2002, 2011; Frisk, 2007; Hildebrand, 2009). We also know that 361 offshore oil and gas exploration and production, as well as renewable energy developments, 362 have expanded during the same period, as has the fishing industry. 363 364 Many animals use sound in the ocean, either passively to listen and orient relative to their 365 surroundings, or they actively produce sound themselves to search for prey or other objects, 366 communicate, or in some cases as a by-product of other activity. The active use of sound is 367 relatively easy to detect, but passive use is not. It is likely that most multi-cellular marine 368 organisms use sound passively as a way of sensing their environment, including listening for 369 prey and predators, and changing behavior in relation to weather and obstacles (including 370 moving ships or stationary propellers such as proposed for tidal turbines). The idea that 371 animals may use something analogous to “acoustic daylight” (Buckingham et al., 1992) to 372 gain an image of their surroundings is gaining momentum, even if it is difficult to 373 demonstrate empirically. The properties of sound in water and the low levels of light 374 penetration below the surface in many circumstances mean that sound has essentially 375 replaced light, for some species, as the principal source of environmental information. 376 Indeed, sound is so important for many species that understanding the acoustic environment 377 amounts to describing the acoustic ecology of many species.

12 IQOE Science Plan 239 DecemberJanuary 20121 DRAFT

378 379 380 Figure 2.4: The Population Consequences of Acoustic Disturbance (PCAD) model 381 developed by a committee of the National Research Council (2005). The arrows define 382 transfer functions leading from the presence of a sound to its effects in ways that are 383 biologically significant. The number of plus signs within each box shows the level of 384 knowledge about each of the processes represented and the number of plus signs between 385 the boxes represents the current ability to infer an effect in this sequence. Note that the 386 ability to infer an effect between Life Function and Vital Rates is currently absent. 387 388 389

13 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

390 391 392 393 Figure 2.5. A diagrammatic view of the issues being investigated by the IQOE (from Boyd et 394 al., 2011). We define 3 major sources of sound in the ocean: physical, biological, and 395 anthropogenic. The sounds involved in marine animal communication and echolocation can 396 be “masked” by physical and other biological sound sources. Animal communication is likely 397 to have evolved to cope with this type of masking. However, overlaid on this soundscape is 398 new sound added by humans, and marine animals may not be able to handle the additional 399 masking to the same extent. The characteristics of the sound received by organisms 400 (“receivers”) will determine responses that could cascade through physiological or behavioral 401 effects that affect an animal’s ability to feed, migrate, and breed and which, in turn, may lead 402 to changes in reproduction and survival of the individual. Relatively few responses at the 403 level of physiology and behavior will have a direct effect on populations, but the effects of 404 increasing levels of sound could accumulate across individuals, thus pushing these effects 405 gradually from physiological and behavioral responses to population-level effects. 406 407 408 2.4 Defining the questions 409 410 Much evidence points to sound in the low frequencies (<10 kHz) being most important, 411 except in the case of some invertebrates (e.g., snapping shrimp, Alpheidae species) and 412 some marine mammals (dolphins, some whales, and seals) that have developed the 413 capacity to both hear and, in some cases, produce complex sounds at much higher 414 frequencies (up to >120 kHz in smaller cetaceans). Our basic knowledge of the way in which 415 the majority of marine organisms sense sound and then respond behaviorally to different 416 sound stimuli is quite rudimentary for most species and groups. Similarly, the extent to which 417 higher background sound levels masks the ability of marine animals to interpret sound 418 signals from their environment is largely unknown, as is their reaction to loud human- 419 produced sounds in their vicinity. 14 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

420 421 For example, we now know that several species of whales have adjusted their 422 communication calls in a manner that suggests they are “raising their voices” or otherwise 423 changing their calls to be heard in the context of potentially masking sounds (e.g., Miller et 424 al., 2000; Foote et al., 2004; Holt et al., 2008; Parks et al., 2010). This is known as the 425 Lombard effect (Lombard, 1911), originally reported for humans, but also seen in terrestrial 426 species such as birds that use sound in social activities (Lengange, 2008; Slabberkoorn et 427 al., 2008). There is evidence that, in the presence of high levels of background sound, some 428 species simply stop vocalizing, either because they are being disturbed or because, like 429 humans trying to talk in the presence of loud background sound, they give up because 430 communication becomes ineffective. Acoustic masking of marine mammals from increased 431 ambient sound is of particular concern in low-frequency specialists, such as the large baleen 432 whales (Clark et al., 2009). Although it is possible that whales could be especially sensitive 433 (and we know that not all whale species share the same sensitivities), the presence of 434 masking and the Lombard effect leads to two questions: (1) are these general effects 435 widespread among marine organisms, and (2) even if they are widespread, are they 436 important to the function and survival of viable populations? 437

438 The IQOE will address the following major questions:

439  What are the levels and distribution of anthropogenic sound in the ocean?

440  What are the long-term trends in anthropogenic sound levels across the global ocean?

441  What is the effect of anthropogenic sound on the viability of important marine animal 442 populations?

443  What were the global ocean sound levels before humans depleted sound-producing 444 organisms and introduced sound from industrial, recreational, and naval activities?

445  What is the future effect of sound on marine life?

446

447 2.5 IQOE objectives 448 The IQOE will last for 10 years. In very broad terms, the IQOE will test the hypothesis that 449 there is an increasing level of man-made sound in the ocean and that this affects marine life 450 in important ways. Because sound travels over ocean basin scales and because almost all 451 human activities in the ocean generate sound in some way, simple calibrated measurement 452 of sound over various time scales could be used to measure industrialization (Frisk, 2007). 453 Consequently, over its duration IQOE will have the following objectives, which are described 454 within the themes of this Science Plan: 455 456 1. To document changes in man-made sound in the ocean, including its distribution in 457 space and time and to place this within the context of natural variations in sound; 458 459 2. To increase our understanding of the role that sound has within the global ocean, 460 especially in terms of how it affects ecological structure and function; 461 3. To measure the effects of anthropogenic sound on key marine organisms, including 462 the level of populations and communities to estimate the biological significance of 463 anthropogenic sound; and 464 4. To develop approaches to observing the current state of sound within the global 465 ocean and, through this, to hind-cast and forecast ocean sound at multiple spatial 466 scales. 467

15 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

468 In parallel with researching these objectives we will build a sustainable, global scientific 469 community with a critical mass of expertise across a broad range of disciplines and with the 470 technologies and tools, including databases, necessary to support research in ocean sound 471 as a core activity within future ocean science. 472 473 The goals will be achieved through the organization of four research themes: Comment [ILBoyd2]: We need to re‐visit these 474 structures to make sure they are aligned with the 475 Theme 1: Ocean soundscapes themes structures once they are finalized. 476  Characterization of soundscapes 477  Identification of sound sources 478  Modeling of soundscapes 479  Trends in ocean sound 480  Sound budgets 481  Soundscape diversity 482 483 Theme 2: Effects of sound on marine organisms 484  Experimental approaches 485  Comparative and baseline studies 486  Biological significance 487  Measuring effects 488  Determining dose-response relationships 489  Model species 490  Large-scale experiments 491 492 Theme 3: Observing the physical and biological sound field 493  Data standards 494  Global and regional observing systems 495  Observing system integration 496  New observing systems 497  Biological observing systems 498  Synthesis and modeling 499 500 Theme 4: Scientific information needed for regulation of sound in the ocean 501  “Noise” monitoring and regulation 502  Risk management 503  Defining thresholds 504  Measuring compliance 505  Communicating results 506 507 The IQOE will develop an understanding of the current state of the ocean sound field, how 508 this is changing because of human activity, and what effects this might have on marine life. It 509 will also develop the capability to understand how ocean sound is evolving and how this may 510 have affected marine life in the distant past and also into the future.

511 2.6 Anticipated benefits 512 The questions addressed by the IQOE are important for two main reasons. The first reason 513 is that industrialization of the ocean is likely to increase in the next few decades. A very large 514 proportion of the manufactured goods and raw materials needed by a growing global 515 economy are being shipped around the globe on the ocean. The demand for hydrocarbons 516 is also pushing exploration and production further offshore into deep waters at continental 517 shelf edges, at depths where sound can more easily enter the “deep sound channel.” 518 Energy extraction from the ocean wind, waves, and tides—although relatively small at 519 present is expected to increase rapidly over the next few decades. In coastal areas, 16 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

520 recreation is also leading to increasing sound levels from pleasure boats. There are serious 521 concerns that this process of increasing industrialization and recreation will lead us in small 522 steps towards an intolerable acoustic environment for many marine organisms. 523 524 It is in the best interests of sound producers to help study the effects of sound on marine 525 organisms, because the precautionary principle is slowly but progressively constraining the 526 ability of sound producers to operate (Gillespie, 2007). Precaution in the face of uncertainty 527 is rational and is an approach that is now deeply embedded in the way that environmental 528 management operates in many countries and internationally. Reducing uncertainty by 529 increasing our knowledge and understanding of the effects of human-generated sound on 530 marine organisms will help avoid excessive regulation. 531 532 The second reason why we should give attention to the issue of sound in the ocean is even 533 more profound and is one of concern to all responsible people. We are slow to learn the 534 lessons from the negative impacts of the past industrialization of the ocean and that the 535 dangers of causing irreversible declines in the quality of the planet’s self-regulating 536 environment are tangible and real. We know that the non-linear, complex nature of the Earth 537 system means that collapses could happen quickly and without much warning. At some 538 point, small changes could lead to very large shifts in the state of the system. Although there 539 is some evidence that many parts of the ocean show remarkable resilience to the direct 540 exploitation of fish, whales, plankton and other forms of biological productivity, there is 541 increasing evidence that there are definite limits. Ecological collapse is an emotive and 542 poorly defined term. However, if we view it from a human perspective, as ecosystems that 543 can no longer support normal goods and services, it has already happened locally as a 544 result of direct exploitation (Bakun and Weeks, 2006; Thurstan and Roberts, 2010). The 545 danger we face is that the uncontrolled increase of sound in the ocean—some of which 546 could be avoided with appropriate design, planning, and technological innovation—could 547 add significant stress to already-stressed oceanic biota. Unless we improve our knowledge 548 of the consequences of sound pollution, we may be cruising blindly towards consequences 549 that, in terms of a simple cost-benefit trade-off, could cost us much more than we will ever 550 gain from ignoring them. 551 552 Therefore, the IQOE recognizes the benefits of its activities accruing across many 553 stakeholder groups: 554 555  Regulators who codify emerging legal frameworks as constraints on the sound 556 radiated by industrial activity; 557  Legislators who define the legal setting for sound and who respond to the public 558 good; 559  Public that has an increasing jaundiced view of the activities and motivation of 560 industry, especially in the ocean, where experience of poor management in fisheries 561 has sensitized the public to issues of marine management; 562  Managers who need relatively simple targets and reference points to establish as 563 objectives for managing anthropogenic sound in the ocean, and that have some 564 objective foundation; 565  Scientists because, while sound is usually overlooked as part of the physical 566 structure of the ocean, it has such an important role within ocean biology that it is 567 likely to have much more widespread importance than is currently appreciated. The 568 importance of sound in the ocean has been appreciated by submariners for many 569 decades and much of our current knowledge of ocean acoustics derives from studies 570 conducted to support defense and submarine warfare. We need to broaden the 571 foundation for our knowledge of sound in the ocean. 572  Industries that produce sound and technologists who are seeking ways to reduce the 573 inputs of sound from commercial activities. 17 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

574 575

576 2.7 Activities

577 2.7.1 Experimental approaches 578 To address the challenging questions posed by the effects of increasing ocean sound we 579 need to ensure that science activities are coordinated across international boundaries and 580 across disciplines. This is why an IQOE has been proposed. Such a project would employ 581 two types of methods to help increase our understanding of sound in the ocean and its 582 effects. One of these methods will be to conduct experiments involving the active 583 manipulation of anthropogenic sound sources, either through directed, temporary reductions 584 of these sound sources at regional scales, or to use planned lulls in sound production (e.g., 585 due to planned shut-down of offshore construction, the diversion of shipping lanes, or the 586 temporary presence and absence of sound sources). The second method will be to use 587 opportunities to make observations in an unplanned or semi-planned experiment design, 588 either through a comparative approach or an a posteriori opportunistic approach. These 589 experimental approaches will require expanded observations and modeling. 590

591 2.7.2 Ocean soundscapes 592 A first step in this direction will be to define what we call ocean soundscapes. An ocean 593 soundscape is a description of the acoustic environment that fully describes its spatial, 594 temporal, and spectral content. Although we have identified at least 30 sites or networks 595 globally that have currently or recently collected data about ocean sound (see Appendix 1), 596 in almost all cases the monitoring stations involved have been established to perform 597 specific functions. This is reflected in the disparity of sensor designs and of data collection 598 and transmission protocols. We need to find ways to use these data in a unified framework, 599 and to establish other measurement systems, to understand the complex global sound field 600 in the ocean. Building a picture of this global sound field, even in a relatively unrefined form, 601 is a high priority as a baseline for other studies. Sound propagation modeling—based on 602 ship position and activity (from Automatic Identification System data), data for wind and 603 rainfall, and data for seismic surveying, sonars, and pile driving—may provide a general view 604 of the sound fields across the global ocean. The biggest “unknown” in estimating the global 605 soundscape will be the contribution of biological sound, which will require better 606 understanding of animal vocal behavior, particularly when species vocalize in large numbers 607 to produce “choruses.” Refinement of this model will be possible with increasing knowledge 608 of the sound production from ships and other human activities, many of which are currently 609 poorly characterized. 610 611 The IQOE will promote the establishment of a Global Ocean Acoustic Observing System 612 (e.g., Dushaw et al., 2009). This system will build on the existing and planned capability of 613 the Global Ocean Observing System, and local and regional systems such as the U.S. 614 Integrated Ocean Observing System and the Australia Integrated Marine Observing System, 615 by helping to define standards and protocols for sensors and for the analysis, storage, and 616 distribution of data across a global research community. 617

618 2.7.3 Predicting Sound Fields and Managing Sound Inventories 619 Establishing the global ocean soundscape, with appropriate statistical characterization of its 620 variation in space, time, and frequency, is a necessary step towards predicting ocean sound 621 fields in particular locations. These predictions can then be challenged with the in situ 622 measurements from existing sites, and a process of tuning sound field models to maximize 623 the fit to the empirical observations will eventually refine the descriptions of ocean 624 soundscapes. 18 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

625 626 Predicting sound fields in this way will also feed directly into the emerging processes for the 627 regulation of offshore human activities and general industrial development. In both the 628 United States and Europe, for example, legislation is moving rapidly to embrace marine 629 spatial planning and setting standards for sound production, principally on a precautionary 630 basis. But existing information is not sufficient to build the rationale for spatial management 631 of industrial activities in order to reduce potential sound impacts on sensitive species or 632 habitats. The development of sound budgets3 on the global scale will enable regional and 633 local managers to refine these budgets to reflect their own needs at regional and local 634 scales, and to help define the kinds of threshold values that managers often need in order to 635 be able to set legally binding conditions on use of the ocean. This nested approach to model 636 development and validation is necessary because sound is a problem that needs to be 637 tackled initially at large scales because of the long-range propagation of low-frequency 638 sound. Even local models need to have the boundary conditions specified in order to build 639 local sound budgets and we intend to provide this capability. 640

641 2.7.4 Exploration in Deep Time 642 What was the global ocean like before humans arrived? Many have explored this question 643 with respect to the removal of marine mammals and fish, in particular, but we also want to 644 know how noisy the ocean was in the past. In other words, can we back-cast the ocean 645 soundscape to a pre-industrial era? Similarly, can we predict the ocean soundscape in the 646 future if current trends continue? What is the cost-benefit trade-off if regulations are set to 647 reduce the sound produced by human activities? These questions, though interesting in their 648 own right, have most relevance if they are accompanied by robust functional relationships 649 between sound and the growth or decline of populations of marine organisms. 650 651 The challenge and opportunity of the IQOE is to coordinate scientific activities on the effects 652 of ocean sound on marine organisms internationally, whether conducted in the academic, 653 governmental, or industrial (e.g., Joint Industry Program) sectors. The development of a 654 body of knowledge that begins to flesh out the types of responses to different levels of sound 655 in the life functions of individual organisms—such as changes in the reproductive rate, 656 growth rate, use of habitat, survival rate and the social structure—is an essential part of the 657 strategy being adopted by the IQOE. The species that need to be included vary across the 658 full range of marine organisms, but perhaps could focus principally on some of the keystone 659 or indicator species within major, or important, ecological systems, as well as species 660 already recognized as endangered. Many of the resulting “effects” studies will be small-scale 661 in situ experiments and some may be possible in controlled conditions in the laboratory. 662 However, all will need to be designed carefully with controls and also with a view to ensuring 663 that the effects observed can be built into larger-scale strategic models of effects at 664 population and ecological levels, such as the PCAD model referred to previously.

3 Sound budgets are defined as the overall distribution of sound energy at a particular location within a defined period of time. 19 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

665 3 666 667 Theme 1: Ocean Soundscapes 668

669 3.1 Characterization of Soundscapes 670 Quantitative description of the ocean sound field, which is how we define soundscapes, is 671 fundamental to any analysis of the trends in sound in the ocean and the effects of these 672 trends. If we wish to understand the consequences of variation in sound within the ocean it is 673 essential to define the mean and variance of the sound field, spatially temporally, and with 674 respect to the frequency. The term soundscape used in the context of the ocean has 675 resonance with a landscape because whereas light is the principal source of environmental 676 information within landscapes, sound takes precedence within the aquatic environment. Our 677 ability to perceive the land and the ocean reflect our ability to sense light and sound, 678 respectively. Indeed, both have the same level of spatial and temporal dynamics. 679 680 The term soundscape has developed recently within the terrestrial environment (Pijanowski 681 et al. 2011). A soundscape is a description of an acoustic environment. However, while an Comment [EU3]: From George: There are 682 acoustic environment can be perceived from the various perspectives of receivers of signals, additional references from the J. Acoust. Soc. Am. addressing urban soundscapes that should be added 683 here we intend to provide a view of the ocean sound field and how it varies that is here. 684 independent of the characteristics of the receiver. For example, the acoustic environment of 685 two co-located organisms that have different capacities for sound perception could be very 686 different, but both experience identical soundscapes. We need to be careful to ensure these 687 are distinct concepts. 688 689 The variation in sound, which is really what the soundscape describes, is central to almost 690 every aspect of the IQOE. The soundscape is the context within which any field experiments 691 (Theme 2) have to operate by observing the response of organisms to intentional or 692 serendipitous changes in the soundscape. It may also be used to document the contribution 693 made to ocean sound by humans and as a measure of the extent of industrialization of the 694 ocean. 695 696 At any place and time, a soundscape may be described quantitatively in terms of different 697 measures such as 698 699  the acoustic waveform as a function of time (which contains all the information about 700 the sound sources in the soundscape); 701  the spectrum (sound intensity versus frequency over appropriate time scales; for 702 example, narrow band spectrum, 1/3 octave band levels); 703  the spectrogram (time history of spectra); and 704  sound energy over specified time scales, often referred to as sound exposure level 705 (particularly to describe high level transients). 706 707 Figure 3.1 illustrates a cross-section through a soundscape (also illustrated in Figure 2.1), 708 where a broad-band acoustic spectrum has been measured across time, in this case for one 709 month. However, presented in this form, which is the normal form for the depiction of the 710 dynamics of ocean sound measured from a particular observation station, the soundscape 711 can be rather difficult to interpret. It is extremely difficult to independently verify the 712 interpretations given in such figures, which rely on the experience of the observer. 713 Spectrograms of this type often contain sounds for which there is no obvious explanation. 714 715 The variation in the soundscape represented by the section in Figure 3.1 is remarkable, and 716 this kind of variability is what would normally be expected within ocean regions where there 20 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

717 are a wide variety of different sound sources, including natural physical and biological sound, 718 as well as human-generated sound. Note that in this case the spectrum ranges from 10 Hz 719 to about 20 kHz and the soundscape could contain high energy below 10 Hz representing 720 the pulsing beat of ship propellers. Many marine mammals also use frequencies >20 kHz, 721 but the propagation of sound to long ranges at these higher frequencies is much less 722 effective than at the lower end of the spectrum of sound in water. This frequency-dependent 723 attenuation means that the spatial scale reflected in a spectrogram of the type provided by 724 Figure 3.1 changes from small scale (a few kilometers) at the top of the diagram to a very 725 large scale of hundreds of kilometers at the bottom. Soundscape analysis aims to present 726 the distribution of acoustic energy at a single spatial scale. 727 728 Characterization of the soundscape can be tackled from two opposing directions, each of 729 which should ideally lead to a convergent result. The directions represent the derivation of 730 soundscapes based upon (1) empirical observation of isolated sections through 731 soundscapes, as illustrated by Figure 3.1, and stitching together many of these across 732 required spatial scales, and (2) observation of the sound sources and the derivation of a 733 sound field by modeling the propagation of sound from these sources. 734 Po w Night-time chorus of snapping shrimp er Sp Fre ec que tru ncy Humpback whales m [Hz] Le Fish ve Distant seismic l survey [d B re Tidal flow Tidal flow 1µ Pa 2 Day of Month / 735 736 Figure 3.1: Spectrogram of a soundscape including snapping shrimp (pink box), evening 737 fish choruses (white ellipses, only drawn for four nights), nearby humpback whales (red 738 ellipses) and distant humpback whales (red box), and a distant seismic survey (yellow box). 739 The thin grey boxes at low frequencies highlight sound from fluid flow around the recorder. 740 This is an artifact of measurement and is not considered part of the soundscape. Recording Comment [ILBoyd4]: This might be right but 741 by JASCO Applied Sciences.Source? we need to be sure. Could it not just as easily be caused by sediment movement? Comment [EU5R4]: From Christine: Sand moving over the hydrophone has much more energy at high frequencies. You can hear the sand swishing. In this recording, the current flow about the hydrophone was not audible to the human ear.

21 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

742 743 744 Both these approaches have strengths and weaknesses that are expanded on later in this 745 Science Plan but, because they are independent of one another, they present an opportunity 746 to provide independent validation. Soundscape descriptions are most often derived from 747 modeling sound sources using propagation models (i.e. the second approach described 748 above). Consequently, to provide validation there would be a need in these circumstances to 749 use this approach to predict the amplitudes of sound within the kind of soundscape sections 750 illustrated by Figure 3.1 and then compare the prediction with the actual soundscape 751 section. Thus the IQOE has an opportunity to test and improve sound propagation models. 752 This approach was adopted during the ATOC (Acoustic Tomography of Ocean Climate), 753 experiment that ended in 2006 (Munk and Wunsch, 1983; Munk et al., 1994). Based upon a 754 known sound source located close to Heard Island in the southern Indian Ocean, this 755 experiment made use of the idea that if one knows the sound source characteristics, and 756 can differentiate the sound from other sounds in the ocean, and one controls the position of 757 the receiver, then the way in which the sound is modified between the source and the 758 receiver is indicative of the properties of the water through which the sound has travelled 759 (Munk et al., 1994). This includes the temperature of the ocean and the ATOC experiment Comment [EU6]: Stocker‐‐ATOC and the Heard 760 was designed to examine long-term changes in temperature at whole-ocean scales. Island Feasibility Test were two different programs. ATOC had a source on the Pioneer Seamount off the 761 However, ATOC was an approach that measured a defined sound signal at a receiver. This coast of California, and a receiver in Hawaii. HIFT 762 design could be generalized to validate sound propagation models or, more likely, to define was an earlier experiment. Munk was the PI on both 763 the general level of uncertainty in these models, both of which are tools for predicting programs 764 soundscapes. 765 766 This kind of validation process would most likely only be appropriate at low frequencies. As 767 described for Figure 3.1, the spatial range represented by the spectrogram of the cross- 768 section of the soundscape is progressively biased towards sources closer to the receiver as 769 frequency increases. Consequently, the frequencies available for validation will depend upon 770 the spatial scales at which the propagation modeling can be carried out. Indeed, the ATOC 771 experiment had a frequency centered on 75 Hz. 772 773 Sound levels are usually determined by measuring pressure and this is largely what has 774 been considered so far, but fish and invertebrates also sense particle velocity representing 775 the actual motions of fluid elements in response to the fluctuating pressure of the sound 776 field. The particle velocity has a has a vector component and a time domain component both 777 dependent on proximity and direction between the source and the receiver. As fish and many 778 invertebrates detect particle motion, they may be especially capable of determining the 779 direction of sources in the horizontal and vertical planes. . Hence, methods of measuring or 780 estimating particle velocity need to be included to characterize sound sources and 781 exposures. directional component depending on the principle direction in which the sound is 782 propagating. Hence, methods of measuring or estimating particle velocity need to be 783 included to characterize soundscapes. Directionality of sound can also be sensed by marine Comment [EU7]: Stocker‐‐ Actually sound 784 animals, so this variable needs to be understood when including marine mammal sources, characterizing particle velocity as a component of a soundscape would be way too 785 vocalizations within soundscapes. As fish and many invertebrates detect particle motion, complicated for the return 786 they may be especially capable of determining the direction of sources in the horizontal and 787 vertical planes. 788 789 The complete characterization of the ocean soundscape using propagation modeling can 790 only be achieved to within certain bounds of accuracy. Ideally, we desire to have the 791 capacity to predict the sound spectrum in any particular part of the three-dimensional ocean 792 at the finest possible scales. However, there are several major classes of constraint. These 793 constraints include the scale at which other relevant data are available, such as bathymetry, 794 and temperature and salinity profiles that dictate sound speed at different depths. Other 795 limitations exist because of fundamental inaccuracies within sound propagation models

22 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

796 themselves, which are necessarily approximations of the physics involved in sound 797 propagation, and many of these limitations will co-vary with the scarcity of data that are used 798 to provide inputs to these models. However, perhaps the greatest constraint comes from 799 information about the distribution, abundance and characteristics of sound sources.

800 Key Questions 801 802 Relevant questions related to the characterization of soundscapes include: 803 804  What physical quantities and metrics are useful to measure components of 805 soundscapes? 806  What are the contributing components of soundscapes? 807  How should the contributing components be combined to characterize and differentiate 808 soundscapes?

809 Research approach 810 811 The current state of knowledge allows us to approximate the sound propagation from 812 particular sources with particular characteristics. The most productive approach to increasing 813 knowledge of soundscapes would be to develop a framework for calculating the 814 soundscape. Ideally, this would be at all scales in three dimensions and it would also 815 illustrate how the soundscape changes through time, but we will need to carefully structure 816 the approach to allow progress to be made within the current constraints of the data. 817 818 We will use a nested approach by first focusing on the large, global or ocean-basin scale 819 with relatively coarse spatial grids so that this can be used to define the boundary conditions 820 for calculations of the soundscape at regional and local scales, both of which may be able to 821 use much smaller grid sizes. Models to describe sound propagation and the oceanography 822 to determine sound speed are both available at the large scale. In some specific places, 823 regional and local characterizations of the soundscape will also be possible because of the 824 quality of the data available. 825 826 The global-scale characterization will use global approximations for oceanic conditions, 827 including seasonal variation and using the kind of grid sizes commonly applied on global 828 oceanographic data sets. Data sets that have similar levels of accuracy are available for the 829 three additional layers required to describe a global soundscape, namely, (1) weather as a 830 physical sound source, using global weather charts; (2) the distribution and density of 831 biological sound producers within broad classes (e.g. toothed whales, baleen whales, 832 snapping shrimp, demersal fish biomass); and (3) anthropogenic sound sources using the 833 best available knowledge of the distribution of human activity. This approach should include 834 the statistical uncertainty connected with those components. 835 836 Initial versions of the implementation of this global soundscape calculation will produce 837 highly uncertain outputs but will create a clearer view of where the greatest uncertainties lie 838 and the most effective way in which uncertainty can be reduced. Early implementation of 839 soundscape characterization is vital to define those variables that lead to the greatest 840 uncertainty because it is important to focus, in the remainder of the IQOE, on assimilating 841 and/or collecting the data that will contribute most to the reduction of variance in soundscape 842 calculations, much of which will be delivered through Theme 3 on Ocean Observation. 843 844 The output from this work will be an ocean sound map showing the intensity of sound across 845 the ocean. The method of representing this intensity will require more work; there are many 846 different ways of representing sound intensity but it is likely that this will consider the

23 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

847 integrated total power (integrated across frequency and depth) at a particular location in the 848 ocean. 849 850 Future iterations of the global soundscape characterization will include variations based 851 upon different scenarios associated with the physical, biological and anthropogenic data 852 layers. Three example scenarios that will be explored are: 853 854 1. The outputs from global circulation models and long-term weather prediction will be used 855 to examine the impact of climate change on ocean ambient sound from weather and the 856 way in which changes in ocean stratification may affect the intensity of ocean sound in 857 particular regions. 858 2. Changing the biological layers to those more representative of a pre-fishing and pre- 859 whaling era will allow us to test the hypothesis that the biological contribution to ocean 860 ambient sound is considerably less now than it has been in the past. 861 3. Prediction of the changes on ocean ambient sound in the Arctic associated with ice 862 retreat (including sound generated by iceberg calving and the progressive break-up of 863 Antarctic ice sheets) and increasing industrialization of the ocean. 864 865 Ambient sound is the term used to describe background sounds that merge in such a way 866 that their individual sources cannot be distinguished. Consequently, the level of ambient 867 sound is a function of several factors, including the number and bandwidth of sound sources 868 and their range from the detector. If different sound sources overlap in terms of frequency, 869 the signal received will be a merged signal from all the sources producing sound in that 870 frequency band and it may be impossible to untangle these different sources. Figure 3.2 871 shows this pattern of complexity involving overlapping sound sources. 872 873 Finally, an aspiration is to make soundscape characterization the center-piece of the 874 assessment of the global trends in ambient sound. This is only aspirational at this stage 875 because of the immense amount of work required before this can be achieved. For the 876 immediate future the development of assessments of global trends in ambient sound need to 877 use time series of data that are, by definition, collected at single points. Since these points 878 are relatively few in number, many have only short time series and, in most cases, it is 879 difficult to filter near-field transient sound sources, it is difficult to compile an authoritative 880 view of whether ambient sound is increasing. Soundscape characterization is one approach 881 that can integrate data from many sources—physical, biological, human—to examine trends 882 in sound based upon a broad range of historical time series and forecasting mechanisms. 883 Where there are individual time series of sparsely distributed observations (see Theme 3), it 884 should then be possible to validate the resulting soundscape predictions against these 885 sparse observations. This approach has the potential to provide a truly global integrated 886 assessment of trends in ambient sound. 887 888 One imaginative approach to establishing long-term observation of trends in ocean sound 889 would be to place an observatory in the sea below the landward end of an Antarctic ice shelf. 890 Although technically challenging to implement, this would allow monitoring of a deep-ocean 891 acoustic environment in the absence of near-field biological and weather sound. Both the 892 Ross and Filchner ice shelves would be appropriate for this purpose and would “look” out 893 into the South Atlantic and South Pacific oceans, respectively. 894

24 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

895 3.2 Ambient sound and the components of soundscapes 896 897 898 Ocean soundscapes are composed of a combination of ambient sound and sounds from 899 sources that can be localized, which are often transient or pulsed sounds, or occasionally 900 they are from sources that are completely characterized in terms of their spectrum and their 901 contribution to ambient sound can be inferred or calculated. However, the term ambient 902 sound is typically used as a surrogate for a spectrogram of a sound field in which there are 903 no specific identifiable sources. 904 905 Ambient sound (often referred to as ambient noise in the published literature) includes sound 906 from the sea surface, sound from distant shipping (defined as traffic sound) and biological 907 choruses when animals are so numerous that the sounds of individuals merge into a 908 continuous background. Localized individual sources include passing ships, other 909 anthropogenic activity (e.g., pile driving, seismic air guns and sonar), as well as sounds of 910 individual animals vocalizing. Soundscapes include both natural and anthropogenic 911 components and an example of the sound spectrum from pile driving is given in Figure 3.3. 912 The natural components can be further sub-divided into the physical sound sources such as 913 from weather or ice and the biological sources such as from snapping shrimp, fish and 914 whales. 915 916 Ambient sound has been studied extensively for about 70 years so that we have substantial 917 knowledge on the characteristics of ambient sound in general, and in many environments in 918 particular. Historically, most of the measurements of ambient sound have been motivated by 919 naval strategic considerations. Much of the research has been in areas near North America 920 and Europe where the highest density of human-induced sound is observed. This leaves 921 significant deficiencies in knowledge for areas of the world where there is lower or very little 922 human activity, although there has been some work around Australia, New Zealand and the 923 Antarctic. 924 925 Ambient sound variation with weather conditions can form a substantial component of the 926 background sound field (Figure 3.2). However, wind speed can be used as a proxy for 927 ambient sound. Figure 3.4 shows a study carried out in the Tongue of the Ocean in the 928 Bahamas, which is almost completely isolated from other background ocean sound and 929 therefore provides a useful environment in which to calibrate the relationship between wind 930 speed and ambient sound. As shown by the figure, the relationship shows a consistent and 931 linear pattern, raising the possibility that surface wind speed data can be easily translated 932 into an ambient sound prediction, although this assumes similarity between the Tongue of 933 the Ocean and other locations. Before using this type of relationship in a general sense, it 934 would be useful to see this result replicated for other environments. 935

25 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

100 Comment [ILBoyd8]: We will need to edit “noise” out of the caption here. Maybe Doug could CA 90s do this?

/Hz) 90 2 Fish Comment [EU9R8]: Also, “biological” is

Pa chorus Evening misspelled and there needs to be a hyphen after

 CA 60s choruses “Wind” 80 TS IO Shrimps 70 inshore RD

60 Shrimps

50 Wind 30 kn Bilogical noise dependent 20 kn Wind dependent noise noise 10 kn 40 Traffic noise NOISE SPECTRUM LEVEL (dB re 1 LEVEL SPECTRUM re NOISE (dB 5 kn 30 10 100 1000 10000

FREQUENCY (Hz) 936 937 Figure 3.2: Summary of ambient sound spectral components shown as averages of 938 sustained background sound levels. These components can be combined to predict 939 soundscapes for particular conditions. Traffic sound is the sound of distant shipping and 940 excludes close ships. “CA 90s” shows levels measured off California in the late 1990s 941 (Andrew et al., 2002) at the same place and using the same methods as the curve “CA 60s” 942 measured in the early 1960s (Wenz, 1969). Lower levels of traffic sound occur off Australia: 943 “TS” Tasman Sea, SW Pacific, “IO” SE Indian Ocean, “RD” remote (from shipping lanes) 944 deep water. Wind-dependent sound is from breaking waves for the wind speeds shown. 945 The biological choruses vary with time of day and season (typical maximum levels shown). 946 “Shrimps” refers to the sustained background sound from snapping shrimps in shallow water 947 (modified from Cato, 1997). 948 949 Temporal variation in ambient sound is substantial, typically around 20 dB re 1 uPa @ 1 m 950 due to variation in weather and biological activity, with extremes of variation in excess of 30 951 dB re 1 uPa @ 1 m. Localized individual sources can add substantially to the sound levels 952 and variability in a soundscape. An example of the spectral and temporal variability of a 953 soundscape off the Queensland coast is shown in Figure 3.5. Various natural and 954 anthropogenic, continuous and transient, near-by and distant sources contributed to this 955 spectrum but, in general, only very specific sound sources, including two pingers, can be 956 clearly distinguished. This type of plot showing the percentiles of variability in the amplitude 957 of the ambient sound spectrum helps to summarize the ambient sound field. 958 959 However, it is also important to understand why ambient sound varies through time and as 960 well as variations in the wavelength of the kind of variation in the kind of ambient sound 961 profile illustrated in Figure 3.3 can provide clues about this source of sound and, 962 consequently, allow a certain amount of dis-entanglement of the ambient sound profile. For Comment [EU10]: Stocker‐‐ However, it is 963 example, in regions where there is unlikely to be a large amount of local pleasure boat also important to understand why ambient sound amplitudes as well as frequency profiles 964 traffic, diel variation is very probably caused by biological sources that are often much more vary over time in both the diel and seasonal 965 active at night than during daylight (see snapping shrimp in Figure 3.1). Similarly strong time scales. 966 seasonal trends in ambient sound in particular frequencies have been observed by the 967 hydrophones deployed by the Comprehensive Test Ban Treaty Organisation (CTBTO) in the 968 Indian Ocean and these are most likely to have been associated with whale migrations. 969 Annual changes with specific timings may also be related to weather. 970

26 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

971 Comment [EU11]: Stocker‐‐ This graph does not illustrate the point made in the preceding paragraph.

972 973 974 975 Figure 3.3. The average 1/3-octave band spectra of the sound exposure level (SEL) at two 976 different distances from the piling, compared with the SEL of the average environmental 977 noise. De Jong & Ainslie (2008). 978 979 The identification, characterization and localization of sources that contribute to a complex 980 soundscape in a given location at a given time is a key factor in understanding soundscapes 981 and their effects on marine life. ‘Source separation’ can be spatial, spectral or statistical (or a 982 combination). It is important to appreciate the physical mobility and statistical non-stationarity 983 of most of the sources of interest (at various scales) that often lead to using non-stationary 984 descriptors such as spectrograms, in spite of their shortcomings. 985 986 The intrinsic characteristics of the sources (in terms of factors such as spectral content, time 987 evolution, radiation pattern), need to be characterized independently from the way in which 988 that sound is modified as it propagates through seawater, especially for physically large 989 sources such as ships.

27 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

990 991 Figure. 3.4. Ambient sound level (NSL) in relation to wind speed. Small data points depict 992 495 NSL values used to establish the correlation with integer values of the log of the wind 993 speed. Diamonds depict mean values of NSL for each value of wind speed. The solid line 994 indicates the linear regression based on mean values (diamonds). Dashed lines indicate 995 regions of extrapolation of the regression. From Reeder et al. (2011). 996 997 998 Nevertheless, in spite of the importance of localized sources importance for the 999 characterization of soundscapes, most practical circumstances represent soundscapes in 1000 terms of a spectrogram (e.g. Figures 3.1 and 3.5) and the motivation is to disaggregate the 1001 contributing components from a complex soundscape measured at a specific time and 1002 location. This has attracted a large amount of attention in the past because of the 1003 importance of localized sound sources in military applications for detecting the signal of a 1004 submerged object from within the background ambient sound. Defining the position and 1005 main characteristics of contributing sources (in particular anthropogenic ones) relies on 1006 ‘accurate’ modeling of sound propagation from the source to the measurement location, 1007 based on ‘representative’ modeling of oceanographic features affecting sound propagation 1008 such as wind speed, wave height, sound velocity profiles, and ocean bathymetry and 1009 sediment type. 1010 1011 It will be challenging to design systems that classify sound sources based upon all available 1012 a priori information that could help in resolving the problem, such as oceanographic 1013 measurements, visual observations, ship tracks from the Automatic Identification System 1014 (AIS), information about the presence of other industrial activities such as pile driving or 1015 seismic surveying. Access to historical data will allow post hoc analysis. The identification of 1016 sound sources will need to rely on access to central libraries of recorded and identified 1017 sounds. 1018 1019 Automatic detectors and classifiers can be used for streamlined analysis of data. 1020 Classification systems will form an important part of the IQOE. PAMGUARD 1021 (http://www.pamguard.org/home.shtml) has been designed as an open-source software tool 1022 that can be downloaded free and that software designers can add to. Its development was 1023 sponsored by the oil and gas industry to promote the development of a satisfactory system 1024 for detecting, localizing and classifying marine mammal sounds during seismic surveys.

28 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

1025 However, its open source architecture means that individuals can design detectors for any 1026 sound source and load them into PAMGUARD, allowing this software tool, and 1027 improvements made to it by any member of the research community, to be available to all 1028 researchers. PAMGUARD can operate in real time or as a post-hoc system for analyzing 1029 detections using WAV files. 1030 1031 1032

1033 1034 1035 Figure 3.5: Temporal variability of sound recorded over a 3-week-period off the Queensland 1036 coast. The nth percentile gives the level that was exceeded n% of the time. The recorder was 1037 deployed in 10m deep water, 500m from a shark net with pingers operating at 3 and 10 KHz 1038 and 10 km from a sand pump (Erbe et al., 2011). 1039

1040 Key Questions 1041 1042 Relevant questions related to understanding variation in ambient sound include the 1043 following: 1044 1045  Can sound sources be classified into broad groupings that will provide sufficient detail to 1046 enable the construction of predicted ambient sound maps? 1047  Are there important sound sources for which we have insufficient information about both 1048 the sounds they produce and their distribution? 1049  Are there patterns in the behavior of sound sources that might be important to help 1050 predict the underlying causes of ambient sound variation? 1051

29 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

1052 1053 1054 Figure 3.6: The IQOE will estimate the contribution made to ambient sound by the breakup 1055 of Arctic ice floes. Photo courtesy of JASCO field staff. 1056

1057 Research approach 1058 1059 The number and type of potential sound sources are almost limitless, making it impractical to 1060 include every individual source in models to produce estimated soundscapes. In such 1061 situations, modelers often include a small number of representative elements in their 1062 models, scaling up individuals to populations based on estimated numbers and/or densities 1063 of individuals. IQOE modeling activities will use such an approach, selecting a small number 1064 of representative examples of the three main classes of sound sources: physical, biological 1065 and human. For example, a model soundscape might include one or two species of baleen 1066 whales, one dolphin species, one fish species, a small number of storms in specific areas, 1067 and a few types of ships, naval sonars, and seismic surveys in specific locations.. 1068 1069 In the case of physical sources of sound, we will explore further examples of the use of wind 1070 data as a proxy for sound produced by weather and examine the extent to which the 1071 relationship between wind speed and ocean sound is consistent among these examples. A 1072 further factor determining sound levels, at in polar waters is that associated with ice movement 1073 and breakup especially in the Arctic, is that associated with ice breakup. However, we will 1074 also test the hypothesis that this sound is related to wind speed. The sound produced by sea 1075 ice can take many forms. For example, during winter thermal cracking is a major source, 1076 especially during clear nights (in this sense the sound is dependent on the weather). Ice 1077 fracture and compression related cracks produce distinctive signals which dependent upon 1078 the larger scale current and wind stresses. The closing of winter leads also produces a 1079 distinctive sound. Interaction of ice-floes in the marginal ice zone is affected by wind and 1080 waves and the source of significant sound. An important source of sound, especially in the 1081 Southern Ocean, is that generated by calving icebergs and the progressive failure of these 1082 as they break up. is the noise generated by icebergs as they calve and break up, and ice 1083 sheets as they are moved by tides and winds. 1084 1085 Databases and libraries, or links to existing databases and libraries, of appropriate sounds 1086 will be created on the central data Web site for the IQOE by building on the Aquatic Acoustic 1087 Archive (http://aquaticacousticarchive.com/) that has already been created. 1088 1089 Furthermore, we will support the continuation of an extant, but ad hoc, Detection- 1090 Classification-Localization (DCL) Working Group, including researchers with a common 1091 interest in the detection and classification of marine biological sound, especially from marine 1092 mammals. In this context, we will also support the further development of PAMGUARD as an 1093 appropriate software system for data collection and off-line analysis.

30 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

1094

1095 1096 1097 Figure 3.7: Examples of biotic contributors to ocean sound. Left and right panels by 1098 Christine Erbe, permission to use photos granted. 1099

1100

1101 3.3 Modeling soundscapes (Sources and Propagation) 1102 1103 Ocean soundscapes are characterized in a variety of different ways. In many cases 1104 biologists want actual time-series data, representing precisely the sounds heard by whales, 1105 fish, and other organisms in the marine environment. This allows careful analysis, for 1106 example, of how shipping sound might mask communication between animals, or interfere 1107 with foraging. On a much broader scale, the sound power can be averaged globally and 1108 annually to describe trends (see section 3.1). The sound sources span a wide range of 1109 frequencies from the low-frequency rumble of earthquakes, the drumbeat of seismic air 1110 guns, and the churning sounds of ship propellers, to mid- or high-frequency sonar pings. 1111 1112 Significant effort has been devoted to developing sound propagation models over the years, 1113 mainly for naval applications. As a result, environments outside of the U.S. Navy range are 1114 sometimes poorly modeled; for exampke, an adequate model for range-dependent acousto- 1115 elastic environments (such as sandstone and limestone seafloors, or tropical reef 1116 environments) is still lacking. HoweverFurthermore, much of the navy-drivenis work has 1117 focused on detecting extremely quiet sources (submarines) by listening to them against the 1118 background of these other masking sources. Thus the focus of modeling efforts essentially 1119 reverses foreground and background. 1120 1121 Modeling soundscapes is carried out using a variety of sound propagation models (Frisk, 1122 1994). This requires information about sound sources (see section 3.2) and about the 1123 structure of the ocean through which the sound is likely to travel. The prediction of 1124 propagation depends upon the specific model used—and there are many from which to 1125 choose—and the characteristics of the environment. The models normally work by 1126 simulating the path of sound through water in 2-dimensional slices radiating away from the 1127 sound source. As sound encounters different discontinuities, such as the surface, the sea 1128 bed or an ocean front, its path is determined by the characteristics of the discontinuity due to 1129 diffraction or reflection. The gradient and position of the thermocline as well as the density of 1130 the seawater, determined by measuring the salinity changes with depth are, therefore, 1131 important input variables within these models. Comment [EU12]: From George: NAVOCEANO 1132 databases should be mentioned here. 1133 Consequently, to predict the propagation of a known sound in the ocean accurately, the 1134 following information is needed: (1) ocean ‘weather’ or oceanography in terms of its 1135 temperature and salinity structure; (2) sea state in terms of surface wave conditions; (3) 1136 water depth or bottom topography; and (4) geoacoustic sub-bottom structure in terms of the 1137 sound speed, shear wave speed, and attenuation of sound in the sediment. Sound speed in 1138 the ocean is also affected by pressure, creating areas of the ocean in which sound can be 1139 trapped in “waveguides” and propagated for long distances. Sound waves refract and reflect, 1140 similar to light waves.

31 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

1141 1142 Global bathymetry is probably the most readily available data, through sources such as the 1143 National Geophysical Data Center. International consortia (HyCOM, Mercator, and FOAM) 1144 provide global forecasts of oceanographic information. Global sources of information for the 1145 sea state need further investigation; however, in many cases satellite data are available that 1146 give information about the roughness of the sea surface. 1147 1148 Probably the biggest challenge in this process is to model the ocean bottom reflectivity 1149 accurately. Some limited global databases are available. Interestingly, it may be possible to 1150 ‘bootstrap’—as we learn more about key sound sources, such as shipping traffic, ships may 1151 be seen as beacons that ‘illuminate’ the environment. Their brightness as they crisscross the 1152 global ocean is itself a measure of the reflectivity of the ocean bottom. Exploiting this 1153 information to provide information about the structure of the ocean, as was the case in the 1154 ATOC experiments (see section 3.1) is an interesting challenge for the IQOE. 1155 1156 Overall, there will be a need to develop approaches to modeling that use approximations for 1157 information required. For example, we are uncertain how critical precise variability in oceanic 1158 conditions will be for the investigations of general questions at large spatial scales or 1159 whether approximations over large space and time scales may suffice. Similarly, it may be 1160 that functional relationships between ocean sound and wind speed may provide an 1161 appropriate surrogate for ocean weather (see section 3.2). Consequently, any modeling will 1162 need to be explicit about the scales over which it is operating and about the uncertainties 1163 involved. 1164 1165 Modeling the propagation of sound for the purpose of characterizing soundscapes is feasible 1166 at large scales within the open ocean. However, the complexity of shallow coastal 1167 environments will make it immensely difficult to develop useful predictive models in such 1168 areas. Instead, sound propagation models have been run on independent bearing lines 1169 assuming that the sound stays in a single 2D vertical slice. The IQOE will encourage the 1170 development of modern advances in 3D sound propagation modeling (Fig. 3.8). 1171 1172 By making several broad assumptions, it should be relatively easy to generate soundscape 1173 models, but validation of these models will be much more difficult. The process of generating 1174 models will probably be most easily accomplished at large spatial and temporal scales but 1175 validation data, most probably generated through ocean observation (see Theme 3), will be 1176 most relevant to smaller spatial and temporal scales. 1177 1178

32 IQOE Science Plan

239 DecemberJanuary 20121 DRAFT

1179 1180 1181 Figure 3.8: Rays of sound travelling over a complicated topography that produces three- Comment [EU13]: From George: We should use 1182 dimensional (out of plane) effects. a better figure(s) than this one. 1183 1184 1185 To reduce the discrepancy between the scales at which soundscape models are likely to be 1186 developed and those associated with data collection, validation may most likely succeed 1187 initially at lower frequencies because the measurement of low frequencies at single points 1188 where observations are made are likely to be most representative of larger spatial and 1189 temporal scales. Validation will become more complex with increasing frequency. Comment [EU14]: From George: This paragraph 1190 should be re‐written. 1191 Model validation may be possible by fitting the model results to observation data. Since the 1192 models themselves can be generated independently of the observation data, it should be 1193 possible to establish Bayesian fitting procedures for soundscape models. Not only will this 1194 allow generation of an expression of the accuracy of models are because the output will be a 1195 statistical distribution, but it will also help to further define the parameters in the soundscape 1196 models that provide most information. 1197 1198 Experimental validation is the ultimate test. To achieve effective validation, it is necessary to 1199 be able to separate out the individual components of the soundscape. This will require a 1200 variety of sensors, including horizontal and vertical line arrays, as well as vector sensors that 1201 respond directly to the particle velocity field sensed by many species of fish. 1202 Site selection is also a key part of the validation. Sites in the Southern Ocean where human 1203 contributions are less important will be useful to assess the prediction of naturally occurring 1204 sound due to storms, lightning, wind, etc. In particular, a site under the Ross or Filchner ice 1205 shelves could provide a unique opportunity to establish conditions that would allow 1206 calibration of the ambient background sound level without contamination by local biotic or 1207 abiotic sounds. This is essentially a 1000 km square cave well isolated from anthropogenic 1208 sources of sound and that has its entrance pointing toward the South Pacific and South 1209 Atlantic oceans.

33

239 DecemberJanuary 20121 DRAFT

1210 Key Questions 1211 1212 Relevant questions related to modeling soundscapes include the following: 1213 1214  To what extent can current models be used to characterize soundscapes and be explicit 1215 about the uncertainties within the models? 1216  What are the limits of modeling in terms of the scales at which it is possible to obtain 1217 reliable results and how does this vary between contrasting locations (e.g. offshore 1218 versus coastal)?

1219 Research approach 1220 1221 Ensemble modeling, the use of multiple models to evaluate specific scenarios, will be used 1222 to examine the influence of model structure and assumptions. We will establish a modeling 1223 working group to exploit situations in which it may be possible to test the validity, as well as 1224 the uncertainty, in different models. This model validation may be carried out in conjunction 1225 with other studies in the Experimental Approaches and the Ocean Observations themes 1226 (Themes 2 and 3 respectively). There may be areas, such as ice shelves in the Antarctic or 1227 isolated regions that are known in considerable detail and are acoustically quiet like the 1228 Tongue of the Ocean, that provide opportunities for model development and validation. 1229 Indeed, the U.S. naval underwater ranges may be ideal for this purpose because of the 1230 hydrophone arrays present and the high level of knowledge of acoustic propagation 1231 conditions in these locations. 1232 1233 The IQOE also wishes to support the development of new models that include 3-D capability 1234 and that are dynamically linked to oceanographic models (Theme 3), as a way of developing 1235 improved real-time model certainty. 1236

34

239 DecemberJanuary 20121 DRAFT

1237 4 1238

1239 Theme 2: Defining the effects of sound on marine 1240 organisms 1241

1242 4.1 Introduction 1243 1244 This theme describes the IQOE approaches to studying the responses of marine organisms 1245 to their acoustic environments, including human-induced changes in ocean soundscapes. 1246 This concerns the interaction between organisms and their environment and is also, 1247 therefore, about the acoustic ecology of marine organisms. There are two distinctive 1248 approaches to examining this problem: (1) using opportunities to make observations in an 1249 unplanned or semi-planned experimental design and (2) the use of planned experiments. 1250 This theme will deal with these two approaches separately, although there is some overlap. 1251 1252 However, before considering these approaches to addressing the question of how sound 1253 affects marine organisms, there is a need to consider how soundscapes—the totality of the 1254 sound field within the close vicinity of an organism—translates into a specific organism’s 1255 acoustic environment. This translation process depends upon what organisms hear and how 1256 they hear. 1257 1258 Relatively little is known about the effects of anthropogenic sound on marine organisms, in 1259 relation to what we need to know. Popper and Hastings (2009) reviewed the literature for fish 1260 and concluded that very little is known about the effects of anthropogenic sounds on fishes, 1261 and that it is not yet possible to extrapolate from the results of one experiment to other 1262 parameters of the same sound, to other types of sounds, to other effects, or to other species. 1263 This is a typical situation. It is debatable as to whether any more is known about effects of 1264 sound on marine mammals (Southall et al. 2008). In this case, some acute effects are known 1265 (e.g., Cox et al. 2006) but, overall, the effects of masking of communication by 1266 anthropogenic sound, disturbance by sound, or the direct physical effects of sound are 1267 poorly understood. In theory, we expect a dose-response relationship like the one illustrated 1268 in Figure 4.1. In this particular case the 50% probability of harassment has been defined as 1269 happening at a sound exposure level of 165 dB re. 1 µPa @ 1 m. The derivation of these Comment [EU15]: From Yvan Simard: Sound 1270 types of relationships is central to examining the effects of anthropogenic sound on marine exposure levels (SEL ) is probably wrongly used here instead of sound pressure levels (SPL), and the units 1271 organisms. For example, recent analysis of beaked whale responses to sonar and other presented (dB re 1 uPa @ 1 m) are wrong (for either 1272 pulsed sound has suggested a response in the region of 140 dB re. 1 µPa @ 1 m. In these SEL , which is dB re 1 uPa^2‐s, or SPL, which is dB re 1273 circumstances, we might expect a dose-response relationship more like the blue curve 1 uPa). 1274 illustrated in Figure 4.1. 1275 1276 In some cases, responses are defined by physiological criteria, such as permanent or 1277 temporary threshold shifts (Southall et al. 2008), but behavioral thresholds, such as the blue 1278 curve for beaked whales, are becoming increasingly important. Exactly which threshold is 1279 most important will vary with the signal being tested and with species and circumstances but, 1280 in general, we need more information to create such response functions. Similarly, although 1281 Figure 4.1 shows “harassment” as the response criterion, many other criteria could be used; 1282 establishing which criterion is most important biologically—as opposed to which is most 1283 easily measured—will make a considerable difference to assessments of the effects of 1284 sound on the life functions of marine organisms (see Figure 2.4). 1285

35

239 DecemberJanuary 20121 DRAFT

1286 Consequently, a priority for the IQOE will be to better define the response functions for key 1287 species, especially keystone species within ecosystems, commercially important species, 1288 and species of conservation concern.

1.0

0.8

0.6

0.4

0.2

Probability of harassment 0.0

120 140 160 180 200 1289 Sound exposure level (dB re. 1Pa @ 1m) 1290 1291 Figure 4.1 The dose-response relationship used by the U.S. Navy to define the threshold for 1292 harassment of odontocete cetaceans (red, excluding harbor porpoises and beaked whales) 1293 and a dose-response function defining a lower threshold that may be more realistic for 1294 beaked whales (blue). 1295 1296

1297 4.2 Acoustic environments 1298 1299 Theme 1 will deliver a comprehensive description of soundscapes and, while this will help to 1300 predict the sound environment that any organism is exposed to at any specific place and 1301 time, it does not define what an organism may hear and, therefore, how it may respond. The 1302 hearing thresholds of marine organisms (Fig. 4.2) are not well described. Although there is 1303 considerable knowledge of the anatomy of auditory systems in marine vertebrates (Fay and 1304 Popper 2000; Ketten 2010) and of the mechanics of hearing, significant uncertainties remain 1305 regarding the diversity of frequency-dependence of hearing capability, as well as hearing 1306 sensitivity, in many species. Fish detect particle motion as well as sound pressure (Popper 1307 and Fay 2011), and there is relatively little understanding of hearing capabilities in 1308 invertebrates. Even in species whose hearing capabilities have been studied in considerable 1309 detail, mainly a few fish species and a small number of marine mammals, little is known 1310 about how hearing varies with age and other life-history features, such as sound exposure 1311 history. Virtually nothing is known about the importance, if any, of sound to the early life 1312 stages of fish and invertebrates, but there is evidence of the capacity of young animals to 1313 use sound (Radford et al. 2011). 1314 1315 Sounds generated by human activities have changed in many marine environments, and 1316 continue to modify the acoustic habitats of many marine species and may affect ecological 1317 interactions. The acute, short-term impacts and chronic long-term influences of these 1318 changes at biologically meaningful spatio-temporal-spectral scales are poorly understood, 1319 but there is increasing evidence that animals are responding to and behaviorally 1320 compensating for influences from anthropogenic sounds (refs.). The relative risks of these 1321 different effects at either individual or population levels can be partially constrained by 1322 relating the dynamics of the physical acoustic field (i.e., the soundscape) to which animals 1323 are exposed with the bioacoustical functions of a species’ acoustic signals (e.g., 1324 communication, navigation, foraging) and its auditory perception capabilities (e.g., sensitivity, 1325 thresholds). In some cases these changes could have a direct and acute impact on an

36

239 DecemberJanuary 20121 DRAFT

1326 individual (e.g., a beaked whale responding to mid-frequency sonar, Tyack et al. 2011) or 1327 more of an indirect and long-term influence on a population (e.g., prolonged and large-scale 1328 reduction in communication space for northern right whales, reduction in foraging efficiency 1329 in resident killer whales, Williams and Ashe 2007). 1330 High frequency marine mammal Medium frequency marine mammal Low frequency whale Pelagic schooling fish Demersal fish 160 Shrimp

140 Pa @ 1m) Pa @  120

100

80

60

40

10 100 1000 10000 100000 1000000 Hearing threshold (dB re re 1 (dB threshold Hearing 1331 Frequency (Hz) 1332 1333 Figure 4.2 Stylized examples of the hearing thresholds (audiograms) of a range of marine 1334 organisms. Comment [EU16]: From Yvan SImard: Y‐axis 1335 units are wrong, it should not be @ 1 m, but simply 1336 dB re 1 uPa 1337 Sound production by human activities becomes biologically significant to an individual animal 1338 when it affects an individual’s survival or ability to reproduce. Such effects on individuals can 1339 then cascade into population-level consequences and impact the function of ecosystems. A 1340 characteristic of the problems concerning the effects of sound in the ocean is that it modifies 1341 the acoustic habitats and acoustic ecosystems of a broad suite of species, thus amplifying 1342 the ecosystem risk and complicating our ability to detect cause-and-effect relationships. This 1343 becomes especially difficult when considering that the largest ocean acoustic habitat 1344 modifications are likely to be occurring in the low-frequency range (<1000 Hz), over large 1345 ocean areas (e.g., 100,000 nm2) and for long periods of time (e.g., annually for many months 1346 to decades). 1347 1348 Consequently, acoustic environments of different organisms, even if they are co-located, 1349 could be very different and this is important when examining the consequences of sound. 1350 Sound at particular frequencies has the potential to elicit widely varying responses from 1351 organisms and to change community dynamics in ways that are much greater than simple 1352 dose-response relationships might suggest. This is because of the potential for sound to 1353 cause uneven effects on organisms at the community level. In some circumstances, 1354 changing the competitive balance among species can result in cascade effects within 1355 communities (Ruttenberg et al. 2011). 1356

37

239 DecemberJanuary 20121 DRAFT

1357 Overall, therefore, there is a need for more research focused on how marine organisms 1358 sense sound and the variation in acoustic sensitivity among species, as well as the range of 1359 sensitivity within individual species and how changes in sound affect interactions among 1360 individuals.

1361 4.3 Unplanned or semi-planned experiments 1362 1363 Ethical and practical constraints often make it difficult to carry out precisely planned 1364 experiments on the effects of sound on marine life. Indeed, any experiment conducted in the 1365 field is likely to be more or less of the semi-planned variety because of the difficulties in 1366 ensuring that there are proper control treatments in place. Even in the laboratory, it may 1367 often be difficult to conduct fully controlled experiments because of limiting factors such as 1368 sample size where, especially for species like marine mammals, very small numbers of 1369 individuals and species are available for study. This can result in biased results because it is 1370 impossible to control for individual variation, inter-species differences, and/or serial 1371 correlation within experiments. 1372 1373 We define two main forms of unplanned or semi-planned experimental approach. One, 1374 termed the comparative approach, has been used traditionally as an empirical method in 1375 animal physiology and anatomy that has formed the basis of much of what we know about 1376 functional anatomy and physiology in non-captive, non-agricultural species. The other is the 1377 a posteriori opportunistic approach in which observations made during some form of 1378 unplanned event allow the development of a functional explanation about a connection 1379 between the event and an organism’s response. 1380

1381 Comparative approach 1382 1383 This approach uses comparisons between the responses of similar organisms in similar 1384 regions when exposed to different sound fields. Comparison provides the means by which 1385 an effect of sound may be measured against some form of experimental control treatment. 1386 The control may be some measure of an organism’s behavior or physiology before sound 1387 exposure or for the same species in a similar, but quieter or noisier, environment. In such 1388 studies the dose of sound is uncontrolled and the response variables being measured are 1389 usually unplanned and detected post hoc from a range of measurements because there is 1390 normally no a priori hypothesis about the exact nature of the response to sound. Often, 1391 some form of multivariate statistical methods (e.g., Bayesian techniques) are used to 1392 separate the signal in the data from noise associated with effects from uncontrolled 1393 variables.4 1394 1395 With the development of offshore industries that are increasingly regulated to limit the sound 1396 they produce, protocols are being developed to measure the impacts of these developments 1397 on some marine organisms. The required monitoring can often amount to studies conducted 1398 over many years and they may include several components: 1399 1400 1. Baseline assessment: documenting the state of the organisms of concern before the 1401 introduction of anthropogenic sound; 1402 2. Impact monitoring: documenting any change in the state of organisms during the 1403 period when anthropogenic sound is produced; and 1404 3. Post-effects monitoring: documenting the return of the organisms to their original 1405 state after the period of anthropogenic sound production.

4 Uncontrolled or confounding variables are those factors that are not of direct interest, but cannot be considered to be the same in the control treatment and all experimental treatments. It is important to measure uncontrolled variables that are likely to have an effect on the variable(s) of interest.

38

239 DecemberJanuary 20121 DRAFT

1406 1407 Variations on this approach include the capacity to compare the state of marine organisms in 1408 similar, and possibly contiguous, undisturbed and disturbed habitats simultaneously. In this 1409 case the “state” of organisms could include changes of behavior, physiology, social structure 1410 or population density and population size. Ultimately, however, population size is likely to 1411 represent the end point of an assessment of effects that result from acoustic stress because 1412 this will reflect changes of fitness in individuals that accumulate at the population level to 1413 affect survival and/or reproduction. This means that measurements may have to be made 1414 over time scales of many years for species that have long generation times. However, more 1415 often, short-term changes in behavior, physiology or social structure can be used as proxies 1416 for potentially significant (in terms of population trajectory) effects of acoustic stress. 1417 1418 Opportunistic approach 1419 1420 This approach to gathering experimental data is the most extreme form of unplanned 1421 experimental design. Often the data have been collected before any research question has 1422 been developed and there is a need in these circumstances to rely on post-hoc analyses to 1423 examine relationships between the responses of organisms and sound levels. In many 1424 circumstances this will involve the use of statistical models to help partition the variance in 1425 the states or responses of organisms to particular causes. Often Bayesian statistics can be 1426 used in this context.

1427 Baseline studies 1428 1429 Baseline studies provide a quantitative assessment of the normal, undisturbed state of 1430 organisms, communities, social structures or populations. One of the greatest challenges to 1431 characterizing baseline conditions is that there may be high natural variability in the 1432 parameter(s) of interest. This natural variability makes the detection of changes resulting 1433 from the introduction of sound much more difficult because the effect must be differentiated 1434 statistically from natural variation. Higher statistical power to detect changes can be obtained 1435 by either extending the duration of the baseline or by measuring a broad range of co-variates 1436 alongside the baseline. These co-variates could include other potential anthropogenic 1437 stressors that may change during the course of the study, but may also include indicators of 1438 the physical and biological system. Including these within a statistical analysis to look for 1439 significant effects from the acoustic exposure may greatly increase the chances of detecting 1440 these effects because the analysis, in effect, controls for the “noise” effects of the co-variates 1441 on the “signal” of interest. Nevertheless, even in circumstances where no “signal” is 1442 detected, there remains the possibility that the signal exists but there is insufficient statistical 1443 power to detect it. Moreover, just because a signal cannot be detected does not mean it is 1444 not biologically significant and, conversely, even if a signal is detected this does not mean it 1445 is biologically significant. 1446 1447 We expect that the IQOE will include many studies of this type. Most will be characterized by 1448 f some form of observation associated with some form of activity that is generating sound. 1449 Observations of this type will be facilitated by the rapid increase in the availability of 1450 relatively inexpensive and mobile observation systems (see Figure 4.3) often collecting and 1451 processing data in real time on multiple channels. These multivariate time series in which 1452 sample sizes can be very large are amenable to the application of statistical modeling to 1453 discriminate among effects caused by changes in specific factors, including levels of 1454 anthropogenic sound.

39

239 DecemberJanuary 20121 DRAFT

1455

1456 1457 1458 1459 Fig. 4.3 PAMBuoyTM is a passive acoustic monitoring system mounted on a moored or 1460 drifting buoy that measures marine sounds and carries out acoustic detection and 1461 classification of different pre-programmed sound sources. It transmits these in near-real time 1462 via cell phone or satellite phone networks to base stations, including cell phones. Source, 1463 SMRU. 1464 1465 1466 Measuring effects using quieting 1467 1468 The origins of the IQOE were based on the idea that, rather than introducing additional 1469 sound and observing the effects of this, there was a need to examine the responses of 1470 organisms to quieting. Comparative and semi-unplanned approaches provide an opportunity 1471 to make progress with this ambition. 1472 1473 By using variation in the background sound levels, including intermittent noisy and quiet 1474 periods, it is possible to examine responses to the full range of sound levels and types (e.g. 1475 chronic or acute conditions). This is as much about responses to quiet conditions as it is 1476 about the opposite. Statistical models applied to these data can then be used to predict the 1477 effects of reduction in noise.

1478 Measuring rare effects 1479 1480 In uncontrolled experiments, significant effects of sound on marine organisms can be difficult 1481 to observe. Moreover, effects can be acute or chronic and, in general, it may be easier to 1482 observe acute rather than chronic effects. However, experience with some marine mammals 1483 (Cox et al. 2006) shows that the occurrence of acute effects of sound exposure can be 1484 classified as rare. Many organisms may have thresholds of response rather than a graded 1485 dose response (Fig. 4.1) and when these thresholds are exceeded an acute response 1486 occurs. However, even in these circumstances, acute responses may be context-specific, so 1487 that the probability of an effect depends on both the probability of sound exceeding a 1488 threshold and the exposure happening in a context in which the animal is susceptible to the 1489 effect. In spite of this low probability, acute effects may still be biologically significant when 1490 they occur in species, such as marine mammals, that have a relatively low population 1491 resilience. 1492

40

239 DecemberJanuary 20121 DRAFT

1493 Measuring stress 1494 1495 Behavioral responses to exposure or quieting are often observable and measurable 1496 responses to anthropogenic sound, but population-level effects of this sound will also be 1497 modulated through physiological stress responses. The main stress response in marine 1498 mammals is similar to the generalized stress response for other mammals, which is defined 1499 by activation of the hypothalamic–pituitary–adrenal (HPA) axis in response to an internal or 1500 external stimulus (or stressor), resulting in elevated levels of glucocorticoid (GC) hormones 1501 (i.e., cortisol and corticosterone). Whether the response is beneficial or deleterious depends 1502 on the magnitude and duration of the response and the condition of the animal exposed to 1503 the stressor. Prolonged exposure to stress may result in immune system suppression, 1504 reproductive failure, accelerated aging, and slowed growth. If GCs are not the primary 1505 mechanism, they and other biomarkers may well be indicators of a cascade of effects 1506 leading from behavioral changes to alterations in reproduction and survival. However, even 1507 among well-studied mammal species, finding individuals exhibiting stress indicators outside 1508 the “normal range” may not be indicative of stress because different individuals have widely 1509 varying baseline levels of these stress indicators. 1510 1511 Stress, measured in terms of similar hormonal indicators, is less well understood in non- 1512 mammals, such as fish and invertebrates. But the concept of stress involving incapacity to 1513 adapt or acclimate to immediate environmental conditions is well understood. Nevertheless, 1514 the ways in which marine species adapt or acclimate to stress are many and varied, and 1515 they reflect an underlying complexity of physiological function that may not yield simple 1516 indicators of stress related to the effects of noise. 1517 1518 Interpreting point measures of endocrine responses to a stressor requires a good 1519 understanding of the natural variation in hormones associated with the generalized stress 1520 response. In free-ranging animals, where blood is difficult or impossible to sample, this 1521 understanding must rely on collecting biological samples that are more amenable to 1522 sampling. Although levels of hormones that potentially indicate stress, such as cortisol in the 1523 bloodstream, provide relevant information about stress, accumulation in non-blood tissues 1524 and excretions such as blubber, skin, hair, feces, and exhaled breath may provide measures 1525 of chronic stress because they are integrated measures of the magnitude and duration of 1526 physiological stress responses. Thus, to use stress hormones from non-blood matrices as 1527 indices of stress, the relationship between the levels and dynamics of hormones in blood 1528 and other matrices must be determined. 1529 1530 It remains to be seen whether it is possible to define “stress” with sufficient rigor and 1531 consistency as to make it a general goal of the IQOE to measure stress in marine organisms 1532 as a general response variable. In its most general form, stress simply measures an aspect 1533 of an organism’s physiology that is outside its normal range. Making and using such 1534 measurements will need to be judged based on individual cases and, in some specific 1535 circumstances, they may prove to be useful. 1536

1537 Focal animal studies 1538 1539 Since different animal taxa may show fundamentally different responses to sound, the IQOE 1540 should focus upon key examples of several different species or taxa. These should be 1541 chosen to capture different lifestyles and population demographic features, while 1542 representing species for which we have good ancillary knowledge in a variety of study areas 1543 and situations, or that are especially amenable to study. Some animals may have a size and 1544 physiology allowing for large data loggers to be deployed on them for extended periods (e.g., 1545 seals), whereas deployment of large tags can be very difficult for other species (e.g., fish,

41

239 DecemberJanuary 20121 DRAFT

1546 small odontocetes). Some species may have a lifestyle allowing for very direct measures of 1547 fitness (e.g., damselfish), whereas it may be virtually impossible to measure this directly in 1548 others (e.g., baleen whales). Some locations may have favorable patterns of disturbance 1549 (long periods of silence, followed by long periods of activity). 1550 1551 The following is a representative list of potential study species: 1552 1553  Baleen whales: blue whale, northern/southern right whales 1554  Toothed whales and dolphins: killer whale, bottlenose dolphin, harbor porpoise 1555  Pinnipeds: harbor seal, ringed seal 1556  Sirenians: dugongs and manatees 1557  Birds: penguins, auks 1558  Fish: damselfish, tuna, plainfin midshipman 1559  Turtles: green turtle, leatherbacks 1560  Invertebrates (molluscs, crustaceans): squid, ghost crab, lobster, krill 1561 1562 Key studies 1563 1564 The types of comparisons that can be made include 1565 1566 a) Comparison across species/populations 1567 b) Comparison across habitats/locations/time 1568 c) Comparison before/after sound exposure 1569 d) Comparison between pristine and noisy environments (and grades in between) 1570 e) Comparison among treatments (e.g., pile driving, seismic, sonar, shipping sound) 1571 f) Quieting as a treatment 1572 1573 The IQOE will use opportunities to conduct studies on the effects of marine noise around the 1574 following: 1575 1576 1. Comparing noisy and quiet environments 1577 1578 Comparing animal behavior, abundance, and productivity between noisy and quiet 1579 environments is fraught with difficulties associated with attempting to ensure that the 1580 differences in the sites being compared are predominantly related to the soundscape. 1581 No two environments are identical. However, if chosen carefully it may be possible to 1582 examine the responses of animals in different circumstances. This could be achieved 1583 using cross-sectional sampling of animals from resident populations within the 1584 locations being compared and longitudinally using the same individuals if they migrate 1585 between the contrasting acoustic habitats. Specifically, for example, comparisons will 1586 be made between the behavior of the same killer whales in fjord habitats with or 1587 without boat sound and there will be comparisons between the key faunal attributes of 1588 fjords and tropical reefs where there are contrasting influences of anthropogenic 1589 sound. 1590 1591 2. Marine pile driving 1592 1593 Offshore developments involving wind farms will result in pile driving on an 1594 unprecedented scale within some coastal regions. The impacts of pile driving on 1595 marine life are poorly understood but, mainly as a result of the needs for industry to 1596 comply with regulation, research will be undertaken to examine the degree to which 1597 construction operations are compliant and to assess any effects of marine species. We 1598 propose that these research efforts would benefit from being included within the IQOE. 1599

42

239 DecemberJanuary 20121 DRAFT

1600 3. Seismic exploration 1601 1602 The oil and gas industry has already done much to advance knowledge about the 1603 effects of sound on marine life through the direct sponsorship of research through its 1604 Joint Industry Program. The IQOE will continue to undertake studies of the effects of 1605 seismic survey sound on marine life. This will be conducted in collaboration with the oil 1606 and gas industry (see Theme 4). However, there are also important opportunities to 1607 undertake research in association with academic scientists undertaking seismic 1608 surveys for geophysical studies. 1609 1610 4. Shipping 1611 1612 The IQOE will also examine the effects of ship sound. Since ship sound is probably 1613 one of the most pervasive sources of anthropogenic sound, this will attract most 1614 attention. The approaches used will capitalize upon the highly constrained nature of 1615 shipping lanes involving pinch points at which ship sound is likely to be greatest. Many 1616 of these are easily defined, but circumstances will exist where habitat is shadowed 1617 from the acoustic effects of the predictable ship tracks . Some of these shadow regions Comment [EU17]: Add reference to the Santa 1618 will be identified as outputs from the soundscape modeling described under Theme 1. Barbara Channel experience 1619 1620 5. Unusual events 1621 1622 The effects of sound on marine life have been brought to public attention mainly 1623 because of the unusual and extreme response of some species of beaked whales to 1624 military sonar. Indeed, much of the current knowledge of the effects of noise on 1625 marine mammals comes from Behavioral Response Studies (or Controlled Exposure 1626 Experiments) conducted on behalf of the U.S. Navy to help resolve this problem. 1627 These types of studies will mainly be dealt with in the following section, but there 1628 remains considerable interest in the occurrence of unusual events, especially the 1629 stranding of cetaceans, in relation to offshore industrial activities. The IQOE will 1630 undertake analyses, where appropriate, of the circumstances under which unusual 1631 events have happened as a way of potentially identifying evidence for causes and 1632 effects in relation to anthropogenic noise. 1633

1634 4.4 Research Approach

1635 Strategic approaches to experiments 1636 1637 A key of the IQOE will be to explore the effects on marine life of both increasing and 1638 decreasing sound using independent controls. Although they will be much more constrained 1639 in their extent and generality, the establishment of planned experiments will complement the 1640 more opportunistic comparative and semi-controlled approach described above. 1641 1642 Although large-scale manipulations (up to ocean basin scale) will be important, such as 1643 moving shipping lanes from north to south of the Aleutian Islands, there will be an important 1644 trade-off between the spatial and temporal scales at which experiments take place and the 1645 feasibility of those experiments (Fig. 4.4). Opportunities where the sound exposure can be 1646 changed over a 5-10 year period need to be identified. At present, the establishment of 1647 many marine reserves presents an opportunity to establish the capability to measure 1648 biological changes associated with changing levels of sound. 1649 1650 In ideal circumstances the most appropriate protocol will be to move sound sources to create 1651 contrasting (increase and decrease) conditions, and a key element to make this work will be 1652 to have enough advance notice (up to 5 years) to establish baseline observations before

43

239 DecemberJanuary 20121 DRAFT

1653 changes that are undertaken for other reasons. In the previous section, we also described a 1654 similar situation with opportunities associated with the introduction of sound because 1655 industry is expanding into new areas. The proposal here is similar, but in this case the 1656 experimenter has control of the sources and the likelihood will be that the effects will be to 1657 shift from a noisy to a quiet environment. 1658 1659 1660

1661 1662 1663 Figure 4.4. Matrix of quieting feasibility. The difficulty and financial cost of a shutdown of Comment [EU18]: This figure was developed at 1664 noise sources increases from left to right in the matrix. The feasible time that a noise the first IQOE exploratory meeting to put into context the feasibility of different levels of quieting. 1665 shutdown could be accomplished decreases from left to right (orange row). Different Should any changes be made to the figure? 1666 experimental activities (blue row) might be possible at different spatial scales (green row). 1667 The goal of the IQOE would be to conduct activities at many different scales. The 1668 relationship of the different temporal and spatial scales means that the most feasible 1669 approaches are likely to be several experiments carried out over long durations at small 1670 scales (i.e., toward the left of the diagram). Two roles that the IQOE will play will be (1) to 1671 help reduce the difficulty of experiments as one moves to the right in this diagram, and (2) to 1672 coordinate experiments of the type defined to the left of the diagram so that they will 1673 combine to deliver some of the benefits that would emerge if we were able to carry out 1674 experiments lying to the right of the diagram (from Boyd et al., 2011). 1675 1676

44

239 DecemberJanuary 20121 DRAFT

1677 The long-term nature of experimental studies is essential; short (weekly to monthly) studies 1678 are unlikely to capture the vital life-history effects except in short-lived species with rapid 1679 turnover, and they are often not the main focus of concern. The observations should include 1680 population counts and survival for short-lived species, and at least reproductive cycles for 1681 longer-lived species. In general, experiments should also include measurements of 1682 ecosystem co-variates. 1683 1684 Experiments should not make the assumption that a directional change in sound will produce 1685 proportional change in the response; non-linear responses should be expected and 1686 sampling designed to capture such responses. 1687

1688 Design of experiments 1689 1690 A large-scale, long-term (lasting from months to years) experiment is proposed. This 1691 experiment will test the hypothesis that changes in chronic sound levels have biologically 1692 significant effects on individual target species (see above). In this case, biological 1693 significance is defined by its meaning described in Figure 2.4 and the related text. 1694 1695 To achieve the experimental controls, chronic sound conditions will be experimentally 1696 changed within defined study areas. Each area should ideally contain 3 sub-sites, one in 1697 which the exposure is increased, one in which it is decreased, and a third in which the level 1698 remains relatively unchanged (Fig. 4.5). Within each site, a targeted set of observations 1699 would be made and models constructed to quantify treatment variables (chronic background 1700 sound levels), response variables (to individual species and ecosystem dynamics), and 1701 potentially confounding variables (physical and biological oceanographic conditions). 1702 Although it is likely to be introduced initially at a regional or local scale because of the 1703 practicalities involved, this design could be used as a blueprint for an ocean-scale 1704 experiment. 1705 1706 Alternative, and simpler, approaches to this experimental design would be to compare loud 1707 versus quiet areas or to measure effects over time in relation to changes in the levels of 1708 sound in small localized areas.

45

239 DecemberJanuary 20121 DRAFT

1709 1710 Comment [EU19]: From George: “Noise” should be “sound”. Comment [EU20R19]: Stocker‐‐Thus a need to define “noise” and “sound’ ‐ because I would use “noise” here as used. Noise in this case would be “unwanted” sound – which would be needed in order to determine if there is a disturbance. “Sound” might just infer a manipulation of otherwise common environmental ambient sounds. The “sounds in this graphic are “wanted” by the investigator, but not by the subject.

1711 1712 1713 1714 Figure 4.5 Schematic showing the experimental design involving three locations. 1715

1716 Predicting outcomes 1717 1718 Although we have identified a general hypothesis to test, careful consideration will need to 1719 be given to developing a framework to predict outcomes of manipulations and the capacity to 1720 do this will have a strong influence upon the location, species, and general situation chosen 1721 for an experiment. The general assumption underlying the design is that reducing sound will 1722 cause improvement in vital rates because sound as an external stressor is expected to have 1723 negative consequences for marine life. However, we should also consider the following 1724 factors: 1725 1726 1. How animals use background sound and may have become acclimated to higher 1727 sound levels or even have experienced selection to sustain high performance under 1728 those conditions. Under these circumstances, it is possible that removing sound will 1729 lead to negative consequences. 1730 2. Sound may alter predator-prey interactions. We raised the possibility above that 1731 changes in sound could have non-linear consequences for community structure 1732 because even small changes in competitive interactions could have large effects upon 1733 the dynamics in marine ecosystems. The same applies to predator-prey interactions. 1734 1735 The outputs from Theme 1 on ocean soundscapes will be important as inputs to this element 1736 of the IQOE. However, biological models of the system being studied will need to be 1737 developed in advance in order to establish the a priori predictions of the effects of the

46

239 DecemberJanuary 20121 DRAFT

1738 experiment. The exact nature of these models will depend on the circumstances, but they 1739 could include population dynamics models or end-to-end ecosystem models. However, 1740 whatever model is chosen will need to be validated in advance of the experiment.

1741 Study species 1742 1743 Ideally, study species should include a range of taxa—including invertebrates, fish, and 1744 mammals—each of which will provide different challenges to study. Organisms could be 1745 divided into categories by the role of sound in their lives, and their ability to hear and 1746 produce sound at frequencies of most interest. For example, if our focus is low-frequency 1747 sound, this defines the types of focus taxa as those with sensitivity to low-frequency sound. 1748 1749 The criteria for selecting species for inclusion in experiments might include the following: 1750 1751 a. High sensitivity of the species to sound; 1752 b. Resident individuals should be preferred over migratory species/individuals; 1753 c. There should ideally be a high level of background knowledge of the species and even 1754 individuals if long-lived species are involved. (some whales are known by marking 1755 patterns and have been monitored over years); 1756 d. The species is important in the ecosystem, and/or is commercially important, or it has 1757 some specific significance to the stakeholder community; and 1758 e. The species needs to be accessible and measuring responses must be feasible. It will 1759 also be important to distinguish among treatment, confounding, and response variables. 1760 1761 Given these constraints, there are relatively few species that will fit all of these criteria. In 1762 particular, pinnipeds (Figure 4.6), because of their size and propensity to return predictably 1763 to breeding colonies, and some long-lived resident fish species within reef habitats, would be 1764 appropriate candidates. 1765

1766 1767 1768 Figure 4.6 Pinnipeds are likely to be appropriate for experimental studies. They are large 1769 enough to carry instruments, have predictable migration routes to and from predictable 1770 feeding locations, and return to specific locations on land, making individuals and 1771 populations easy to monitor. 1772 However, these types of experiment could be carried out at small spatial scales with species 1773 that are short-lived and accessible, perhaps using natural mesocosms in which species 1774 composition and ecosystem structure are well defined and possibly also controlled. This type 1775 of design also has the potential to include multiple species.

47

239 DecemberJanuary 20121 DRAFT

1776 Site selection 1777 1778 The study sites will depend upon the species being selected for study and the specific 1779 outcomes predicted for the experiment. A secondary feature will be the availability of 1780 baseline data from the site, to reduce uncertainty. Ideally, selected locations should have a 1781 long history of data collection on the species concerned. This type of criterion narrows the 1782 possibilities considerably. 1783 1784 Additional considerations include 1785 1786  Is there industrial activity in the area that could be used as the exposure and control? In 1787 some settings, it might be possible to shift the sound in a systematic way, but scientists 1788 need to work with industry to develop a consensus plan. For example, it may be possible 1789 to divert shipping for several years at a time, with enough advance notice and if this did 1790 not entail additional cost. In some cases, industry decides on its own to change shipping 1791 routes for economic or regulatory reasons. 1792  Can the studies be replicated? If there were multiple independent sites (e.g. separate 1793 seal colonies) that could be monitored over a long period of time, this would provide an 1794 opportunity for replication. The choice of sites could lead to the development of different 1795 exposure scenarios for each site, for example, (1) increased sound, (2) decreased sound, 1796 and (3) no change in sound levels. 1797

1798 Variables to be measured 1799 1800 1. Dose or treatment variables 1801 Measurement standards will be developed across different regions and species to make it 1802 possible to extrapolate beyond individual study sites and to determine global implications. 1803 1804 The most important dose or treatment variable is the sound received by the study animals 1805 (Boyd et al. 2008). Ideally, this should be measured directly from an instrument placed on or 1806 near the experimental animals but could be modeled based upon information from the sound 1807 field (Theme 1). 1808 1809 2. Response variables 1810 For each dose, the experiment will have defined response variables. Responses will be 1811 measured at different levels of the PCAD model (Figure 2.4), depending on the targeted 1812 organisms, for example, for short- versus long-lived species. With short-lived species, 1813 research will focus on vital rates as much as possible. For other species, it will be necessary 1814 to measure parameters that will allow estimation of vital rates. Individuals/species that leave 1815 the area and others that re-colonize can also be measured in terms of how they change their 1816 migration routes in relation to the sound sources. A combination of tracking and survey 1817 techniques may be applied, including mark-recapture studies and tracking of a subset of 1818 individuals. Tracking a subset of individuals over the observational period could also be used 1819 to study habituation and sensitivity. 1820 1821 3. Confounding variables 1822 Confounding variables are those that can affect the responses to a specific dose or 1823 treatment in an experiment in a way that makes it hard to understand the results, particularly 1824 if the confounding variables are not measured. Some examples of confounding variables 1825 include ocean currents, temperature, salinity, chlorophyll, depth, bottom types (biological and 1826 acoustic), turbidity, and ice cover. 1827

48

239 DecemberJanuary 20121 DRAFT

1828 4.5 Summary 1829 1830 This theme aims to be the main driver for developing a deeper understanding of the 1831 connection between soundscapes and animal responses to the soundscapes. It is, therefore, 1832 important that considerable effort is committed to the task set out here. 1833 1834 When complete, these tasks should deliver representative dose-response functions to the 1835 effects of anthropogenic sound for key species. Although, ideally, these should be in the 1836 form of classical dose-response curves, it is much more likely that they will amount to a 1837 mixture of these types of precise functional relationships between animal responses and 1838 sound levels and heuristic assessments of the effects of sound at various levels from 1839 behavioral response to effects upon populations. Nevertheless, this will represent a major 1840 step forward and, when combined with the outcomes of Theme 1 on ocean soundscapes, 1841 will form the basis for making broad predictions about the potential effects of future changes 1842 in anthropogenic sound in the ocean. 1843

49

239 DecemberJanuary 20121 DRAFT

1844 5 1845

1846 Theme 3: Observing sound in the ocean 1847

1848 5.1 Introduction 1849 1850 Sound in the ocean is challenging to both detect and visualize. This problem is one that 1851 requires measurements in five dimensions: three spatial dimensions, plus time and acoustic 1852 frequenc characteristics, which might be considered three additional dimensions – amplitude, 1853 frequency, and fine‐pitch time domain variability.y. The process of measuring ocean 1854 soundscapes (see Theme 1) is concerned with characterizing these five dimensions to form 1855 a coherent picture. In this theme, we address the requirements of the instruments and 1856 observing systems needed to provide the raw data to allow the measurement of 1857 soundscapes at a very wide range of spatial and temporal scales. The theme also addresses 1858 the need to observe sound fields from the perspective of the marine life that may be affected 1859 by sound. Being able to observe the sounds that organisms are exposed to, and that they 1860 can hear, has been identified as one of the most critical first steps towards being able to 1861 measure the effects of sound on organisms (Boyd et al. 2008). 1862 1863 Only limited data are currently available on ocean soundscapes. Information on long-term 1864 changes in ocean sound levels, whether anthropogenic or natural in origin, is available at 1865 only a few locations in the world ocean, for a limited period. Measurements of underwater 1866 sound also provide data that can be used to track, count, and study the behavior of 1867 vocalizing marine mammals and fish, which can be used to help determine the effects of 1868 anthropogenic sound on marine life. Finally, active acoustic measurements, using 1869 instruments such as scientific echo sounders, can provide information on aspects of the 1870 ocean environment important to marine life, such as the density and distribution of marine 1871 life, especially within the water column. 1872 1873 In recent years, there has been a strong emphasis on the development of ocean observation 1874 systems (Kite-Powell 2009). System development has been enabled partly by increasingly 1875 technological capability, but also by recognition of the need for new data about the ocean, 1876 that sometimes need to be delivered in real time, such as in the case of tsunami warning 1877 systems. These requirements have driven innovation and it is likely that the need for 1878 observation systems and their capabilities will 1879 1880 increase greatly in the next decade. Traditional ocean observatories using moored systems 1881 of sensors are being augmented by mobile sensors on 1882 1883 floats (Roemmich et al. 2009), AUVs (Nicholls et al. 2008), gliders (Johnson et al. 2009) 1884 and even instruments carried by marine mammals (Grist et al. 2011). Acoustic observation 1885 has not, in general, been a part of many of these systems and, when present, it is usually 1886 recording at very low frequencies that may be of greatest interest in terms of observing 1887 seismic events or other physical changes, such as sea ice breakup, but has less importance 1888 in terms of the frequencies that are important to most marine organisms. 1889

1890 5.2 Acoustic observation networks 1891 1892 Dushaw et al. (2009) presented a vision for a Global Ocean Acoustic Observing Network. 1893 Wherever possible, acoustic observations need to be included in ocean observing systems 1894 designed for other observations. This approach will both contribute to providing the

50

239 DecemberJanuary 20121 DRAFT

1895 information required to characterize the global ocean soundscape and make use of current 1896 and future infrastructure with minimum additional cost. 1897 1898 Hydrophone systems are already deployed for the purpose of recording sound in the ocean 1899 and many of these are listed in Appendix 1. Other observing systems have been deployed 1900 for specific oceanographic, biological, chemical, or other environmental purposes, but have 1901 not included ocean acoustic sensors. One key benefit of integrating acoustic capabilities 1902 into such systems is that they would acquire diverse oceanographic and atmospheric data 1903 concurrent with the acoustical signals. These ancillary environmental data may be essential 1904 in determining the relationship between sound, the ecology of target organisms, and the 1905 environment. Some systems under development are cabled observatories, such as the U.S. 1906 Ocean Observatories Initiative, and offer unique platforms with power, timekeeping, and 1907 communications capabilities thus providing opportunities for acoustical instrumentation. 1908 The IQOE has begun the process of identifying and cataloging existing and planned 1909 acoustical observation systems (see Appendix 1). These systems have been described and 1910 tabulated according to several important criteria. The aims of this effort were to show what 1911 systems are presently available to address specific IQOE questions and to identify new 1912 acoustical capabilities that need to be developed for IQOE studies. 1913 1914 The remainder of this chapter places the IQOE into the context of the larger ocean observing 1915 systems; addresses the existing observation systems that either directly support acoustical 1916 measurements or could be configured to do so; suggests some examples of new 1917 technologies that would augment existing acoustical measurement capabilities; and 1918 recommends investigating the assimilation of acoustic measurements, either archived from 1919 prior monitoring activities, or collected by contemporary or planned regulatory activities. 1920 Finally, the theme addresses the issues of standardization of data in terms of quality, 1921 calibrations, formats, and management in order to enable the comparison of results among 1922 international collaborators. 1923 1924

1925 5.3 Acoustics and Global and Regional Ocean Observing Systems 1926 1927 The IQOE cannot unilaterally deploy ocean observing systems because of the resources, 1928 experience, and effort required for establishing and operating global and regional ocean 1929 observing systems (OOSs). These ocean observing elements include a range of observing 1930 technologies, from satellite observations of a variety of oceanic variables to standard NOAA 1931 weather buoys to acoustic rain gauges deployed on Argo floats. As noted previously, the Comment [EU21]: From George: Add Jeff 1932 ocean is largely transparent to sound, hence oceanographic, biological, engineering, and Nystuen reference. 1933 signal-processing acoustic techniques are primary tools for ocean observation (Dushaw et 1934 al., 2009) and represent essential components of the Ocean Observing Systems. 1935 1936 As the IQOE evolves, it will be important to maintain an inventory of OOS capabilities to 1937 assess how they can assist the science, and to identify important information that is still 1938 missing. Time-synchronized multivariate sensing systems will be increasingly important as 1939 attention focuses on interpreting the potential ecological impacts of sound. Consequently, 1940 the application and integration of OOSs as contributors to IQOE monitoring and 1941 experimental efforts is preferred over solely acoustic measurements. However, the type and 1942 nature of these ancillary data streams is dependent on the specific question and the 1943 environment in which measurements are being made. Indeed, information or data that are 1944 missing from existing systems, but are essential for addressing IQOE science questions, will 1945 require deployment of additional instrumentation. Some of these deployments will probably 1946 transition into operational components of the OOS. 1947

51

239 DecemberJanuary 20121 DRAFT

1948 Experience and technologies to promoted good Data Management and Communication 1949 (DMAC) have grown out of the OOSs. The IQOE will need to take advantage of DMAC 1950 technology, and begin a dialog with observing organizations to help incorporate the 1951 acoustical data streams within the OOS, where appropriate. 1952 1953 The Ocean Observing Systems include within their mandate educational and public outreach 1954 efforts. These efforts would naturally extend to any acoustical activities related to the Ocean 1955 Observing System (see Theme 4). 1956

1957 5.4 Integration within existing systems 1958 1959 Most or all of the envisioned IQOE monitoring and experimental efforts should leverage data 1960 from existing capabilities and/or should promote the integration of acoustics into existing 1961 observing systems. An early objective of the IQOE will be to complete an initial draft of what 1962 is envisioned as a continuously updated survey of the known systems (included in Appendix 1963 ___ and online at http://aquaticacousticarchive.com/). The survey matrix will be available for 1964 public contributions on the IQOE Web site and related to a global map showing locations of 1965 existing acoustic observing systems (Figure 5.1). 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 Figure 5.1. The locations of existing hydrophone stations of the Comprehensive Test Ban 1984 Treaty Organization (CTBTO) (magenta circles) and the regional Australian hydrophone 1985 facilities (black triangles) with ocean bathymetry derived from the Smith-Sandwell atlas 1986 (ref?). Each CTBTO receiver consists of a triplet of hydrophones to make it possible to 1987 determine the direction of acoustic signals. Data are recorded, processed, and transmitted to 1988 shore in real time from these arrays. From Dushaw et al. (2009).

1989 1990 These systems were categorized as cabled arrays (e.g., fiber-optic systems), remotely 1991 deployed archival systems (e.g., bottom-mounted recorders), and mobile systems (e.g., 1992 drifting buoys, gliders, animal-borne instruments). Each system was assessed relative to 1993 1994  geographic location; 1995  whether acoustics is a current operational capability; 1996  various system characteristics; 1997  the inclusion and type of ancillary (non-acoustic) data; 1998  relative accessibility of data from each system; 1999  potential integration with other systems;

52

239 DecemberJanuary 20121 DRAFT

2000  sponsoring entity; 2001  general societal benefit or product of each system; and 2002  installation duration/life expectancy. 2003 2004 The derivation of the system matrix in Appendix 1 according to these criteria was intended to 2005 provide a basis for system selection when the experimental/monitoring areas and objectives 2006 are chosen (e.g., in Theme 2). The anticipated process would be the consideration of 2007 specific experiments, each with its own time-space resolution and objectives.

2008 2009 2010 Figure 5.2 Sound amplitude in relation to frequency through time recorded at one of the 2011 CTBTO sites in Western Australia (see Fig. 5.1). The seasonal peaks at low frequencies are 2012 related to Antarctic ice breakup and at the higher frequencies these relate to seasonal calling 2013 by baleen whales. This is an example of the kind of depiction of the soundscape that could 2014 be achieved across many OOS sites. 2015 2016

2017 5.5 New systems designed for IQOE 2018 2019 It should be possible to use existing acoustic technologies within existing and planned ocean 2020 observation systems. However, it appears unlikely that sufficient monitoring systems with 2021 acoustic capabilities exist in enough areas of the ocean to accomplish the broad and 2022 sustained monitoring objectives of the IQOE. There is a need to conduct a detailed 2023 assessment, probably through one or more workshops, of the observation capacity 2024 that is required to meet the objectives of the IQOE, and to assess the extent to which 2025 modification of current and planned capabilities in ocean observation are likely to fulfill these 2026 objectives. 2027 2028 Although it is preferable to utilize and improve upon the capabilities of existing systems, 2029 there will be instances where integration of currently available sensors into existing systems 2030 will not be possible, or existing system nodes may not be located in the appropriate 2031 geographical area(s) targeted for program experiments. In such cases, the IQOE may need

53

239 DecemberJanuary 20121 DRAFT

2032 to establish dedicated monitoring. The development of technology that would directly apply 2033 and enhance the information available from existing systems and sensors has been 2034 identified in 5 topic areas: 2035 2036 1. Particle motion/vector sensors—Research has shown that a majority of fish species are 2037 more sensitive to the particle motion component of sound compared than to the pressure 2038 component (Popper and Fay 2011). The need to measure this parameter is important for 2039 providing the proper environmental context for fish in response to sound exposure. 2040 2. New portable system designed for IQOE (single hydrophone)— In areas where OOS 2041 networks do not have regional nodes or coverage, it will be necessary to develop small, 2042 inexpensive, portable systems designed to provide required acoustic, and where 2043 necessary, other measurements. These portable systems would provide information 2044 relating to the survey of global ocean soundscapes and short-term experiments. Such 2045 portable and inexpensive devices will encourage wider international participation in the 2046 IQOE. 2047 3. Modular hydrophones to assemble H/V line arrays—The ability to quickly and efficiently 2048 assemble modular arrays will enhance our capability to provide directional acoustic data 2049 to the global soundscape survey effort and short-term experiments 2050 4. Data transmission technology—The limited bandwidth of current satellite transmissions is 2051 often the bottleneck for the transfer of high-volume acoustic data. Developments in this 2052 area on either the recording hardware, processing, or data transfer sides would be 2053 beneficial to the IQOE effort. 2054 5. Acoustic backscatter sensors and echosounders —Existing OOS and satellite networks 2055 provide valuable information on physical ocean properties and primary productivity. 2056 Passive Acoustic Monitoring (PAM) provides information on the presence of vocalizing 2057 animals (mammals and fish). The development and incorporation of scientific 2058 echosounders into OOS networks would provide the capacity to measure zooplankton 2059 and fish distribution and concentration in the water column, and to study the predator/prey 2060 dynamics of an area, which is needed to provide proper context for interpreting the effect 2061 of changing sound levels on marine animals. NEPTUNE and VENUS already have active 2062 acoustics capabilities, but this is not a widespread capability across OOS networks. 2063 2064 5.6 Extracting useful scientific information from data collected for regulations 2065 2066 So far, we have considered only scientific OOS networks, but there will be a need in the 2067 future for observation systems that assess compliance with limits on the additional sound in 2068 the ocean from anthropogenic sources. Some of these observation systems may be in place 2069 for short periods when industrial development is proceeding, but there may also be other 2070 networks operated by coastal nations to demonstrate national compliance with targets for 2071 anthropogenic sound production. 2072 2073 There are two approaches to building on this opportunity that could be adopted by the IQOE. 2074 One would involve the analysis of acoustic data obtained in the course of regulatory 2075 monitoring of industrial developments. Alternatively, as regulation of ocean acoustical 2076 pollution is initiated throughout the world, their associated monitoring systems could be 2077 sources of future datasets and the IQOE has an opportunity to influence the design and 2078 placement of such systems. 2079

2080 5.6.1 A survey of historical data to establish the nature of soundscapes of the past 2081 2082 Time series of acoustical data have been collected at multiple locations in multiple regions 2083 over the past 50 years. Many of these datasets were generated by private industry, military, 2084 research institutions, and regulatory agencies for the purposes of regulatory compliance, 2085 exploration, research, and targeted surveillance. Some of these data are proprietary or have

54

239 DecemberJanuary 20121 DRAFT

2086 national security classifications; whereas other datasets are openly available. At an early 2087 stage, the IQOE will undertake a comprehensive survey of historical data. This could 2088 be accomplished by establishing a database on the IQOE Web site that hosts the working 2089 document and allows for the continual contribution and updating from individual inputs. 2090 2091 Historical data may not be in the format agreed upon by the IQOE, but targeted datasets 2092 could be re-processed for contribution to the IQOE. The information resulting from the 2093 historical data survey and resulting data acquisition would provide information to the IQOE 2094 for providing historical soundscapes in areas of interest for comparison to the present and 2095 for validating contemporary acoustic models. 2096

2097 5.6.2 Sources of future data for IQOE 2098 2099 Government-mandated regulation of either radiated sound from individual sources or 2100 cumulative anthropogenic sound contribution in a targeted region will require monitoring 2101 instrumentation that may be a source of acoustic data for the IQOE. As an example, the 2102 Marine Strategy Framework Directive (see Tasker et al., 2010) specifies that all EU member 2103 states monitor their marine environment in order to regulate the contribution of 2104 anthropogenic sound energy. This directive will require new monitoring systems throughout 2105 European waters. While the actual legally binding monitoring requirements are likely to be 2106 very narrow, the instruments being used to provide this information will have the capacity to 2107 collect considerable additional data about sound. Consequently, the IQOE should establish 2108 data sharing agreements that permit continuous, on-going collection of these data to 2109 an IQOE data assembly center. 2110 2111 Since these kinds of data will be formatted primarily to meet the needs of the regulatory 2112 agencies, it will be critical for the IQOE to coordinate with regulatory agencies as early as 2113 possible in order to influence the data formats, and subsequently to devise any necessary 2114 reformatting procedures to transform the available data products into the IQOE formats. 2115 Technical contacts representing the IQOE will need to be appointed to interface with these 2116 regulatory agencies, and the details of the data interface and any subsequent data 2117 reformatting may profit from attention as a subtopic at an IQOE technical workshop (if the 2118 issue occurs early in the IQOE). We propose that a standing committee on data 2119 management should emerge from this workshop. Although this example is specific to 2120 Europe, the IQOE should investigate whether similar opportunities or initiatives exist in other 2121 regions of the world and ensure that there is effective liaison between those initiatives and 2122 the IQOE. An important activity of the IQOE will be to work with navies and all offshore 2123 engineering industries, including those involved in oil and gas exploration and production, 2124 wind-farm deployment and operation, bridge and tunnel construction, offshore mining, etc., 2125 in relation to access to proprietary and classified data in a way that will advance the science 2126 without compromising the interests of those providing the data. 2127 2128 5.7 Data collection (including standardization), quality control, analysis, reporting, 2129 management, and accessibility 2130 2131 Of similar importance to synoptic measurements is the use of standards and the application 2132 of a systematic and standardized data management structure. Information and potentially 2133 important trends and observations are likely to be lost or unutilized unless an explicit 2134 strategy is implemented for data archiving, analysis, and sharing. Data acquisition and 2135 reporting standards are an important part of data management, as is the development of 2136 data‐sharing agreements that ensure the rights of individual data originators. As with 2137 ancillary data measurements, data management strategies require considerable deliberation 2138 and planning, and will vary depending on the systems employed and questions asked. 2139

55

239 DecemberJanuary 20121 DRAFT

2140 A specific recommendation is to convene a technology workshop to define standards 2141 that will lead to a proposal for the global soundscape project. This workshop would 2142 include representatives from the major observing systems, but would be necessarily 2143 preceded by specification by IQOE acousticians regarding experimental design. These 2144 specifications would strive to provide data acquisition and management standards and 2145 protocols for (at least) the following variables: 2146 2147  Bandwidth 2148  Bits, resolution 2149  Sensitivity 2150  Units 2151  Sample rate 2152  Data format 2153  Analysis methods 2154  Calibration 2155  Metadata and Data Accessibility 2156 2157 There are two parts to this issue, to (1) enable agreement among the acousticians on the 2158 standards for data and metadata, and (2) coordinate the acoustical data with OOS or the 2159 IQOE DMAC. Appointing a DMAC before the IQOE experimental plan is sorted out is not 2160 likely to be helpful. 2161

2162 5.8 Biological Observing Systems 2163 2164 The development of biological components of the OOSs have lagged the physical 2165 components, but the biological components were recently highlighted during OceanObs'09 2166 (e.g., Gunn et al., 2010). An important class of acoustical systems are those that employ 2167 passive acoustics for monitoring marine life, including distribution, abundance and behavior 2168 (Dushaw et al 2009; André et al. 2011). Satellite observations of some biological variables 2169 also are available. 2170 2171 Biological observations collected for the purposes of the IQOE may demonstrate their long- 2172 term importance and consequently transition to observational status and become elements 2173 of ocean observing systems. Combining biological observation with observations of sound 2174 will have two specific advantages. First, it will make it possible to develop experiments that 2175 relate the general bioscape (i.e., the acoustically determined distribution and abundance of 2176 components of marine communities, most probably in the pelagic context) to other ocean 2177 sound variations. This will enable some options for developing effects studies as described 2178 under Theme 2. 2179

56

239 DecemberJanuary 20121 DRAFT

2180 2181 Figure 5.3. Spectrogram of a typical dolphin whistle, showing the effects of the various 2182 processing stages of whistle contour extraction. This is typical of the processing carried out 2183 within the PAMGUARD software to enable identification of species. Courtesy of SMRU Ltd.

2184 2185 Second, combining biological and acoustic measurements will enable identification, 2186 classification and possible estimation of the abundance of organisms that are the sources of 2187 sound. Software to achieve such goals is in a fairly advanced state of development (Figure 2188 5.3), but the IQOE will stimulate the development of open-source software for the automated 2189 identification, localization and classification of biological sound sources from ocean 2190 observation platforms (Figure 4.3). 2191 2192 5.9 Synthesis and modeling: physical, biological, and acoustic 2193 2194 Modeling will be an essential component of the IQOE for predicting ocean sound levels 2195 across the globe, estimating acoustic propagation of sound over space and time, and 2196 assessing impacts of changing sound on animal populations. Theme 1 concerning ocean

57

239 DecemberJanuary 20121 DRAFT

2197 soundscapes and Theme 2 concerning the response of marine life to sound both require the 2198 application and further development of models of how sound travels physically in the ocean. 2199 2200 A three-input modeling approach will be needed to integrate the acoustic, biological, and 2201 oceanographic data necessary to relate sound to biological dynamics because the three 2202 separate datasets are inter-related when assessing the impacts of acoustic change at the 2203 population level of animal groups. There are currently no models that predict the effects of 2204 chronic sound on marine animals, and much will be learned by an ongoing review of models 2205 that are presently being used to predict impacts of acute acoustic exposures. The 2206 Population Consequences of Acoustic Disturbance (PCAD) model will be a major conceptual 2207 tool for the IQOE (see Figure 2.4), but other modeling approaches will also be used. 2208 2209 Models used within the IQOE will ultimately allow point measurements made locally, 2210 regionally, and globally to be used to gain an understanding of how soundscapes change 2211 over space and time. In order to develop the most appropriate models, more accurate 2212 characterization/measurement of sound sources (biological and anthropogenic) is needed. 2213 The utilization of available historical data (e.g., decades of SOSUS/Navy data, data from 2214 industry, data resulting from regulatory requirements) will be valuable for validating models 2215 and testing model predictions. However, there will also be a need for high-quality 2216 bathymetric data and integration with regional oceanographic models to enable accurate 2217 predictions of the sound field in particular locations. Consequently, there is a need for the 2218 IQOE to take an active and leading role in the development and implementation of 2219 new acoustic models that better integrate fine-scale details derived from new data and 2220 oceanographic modeling. See Theme 1 for further discussion of modeling and model 2221 validation.

2222 5.10 Recommendations

2223 Monitoring/Experiments for IQOE 2224 2225 Four specific, though not mutually exclusive, types of monitoring and/or experimental efforts 2226 are recommended: 2227 2228 1) “Year of Ocean Sound” or “Ocean Soundscape Monitoring Week/Year” 2229 2230 This lies within the broader concepts of the IQOE, but there are opportunities to use a 2231 focused period of activity to make important progress. What is envisioned is a high-visibility 2232 international effort with coordinated observations around the globe over a short period to 2233 compare with modeling results. This would occur within a short period of time, but would 2234 most logically represent just the beginning of such coordinated observations and modeling. 2235 The intention is that this approach would hopefully produce a global map of soundscapes, 2236 and that these point measurements would inform subsequent models (see Theme 1). To 2237 establish a baseline of the soundscapes of the world's ocean basins, international 2238 coordination would be required to obtain comparable data in different locations. This is 2239 envisioned as including two years of organizing, a third year of data collection, and a final 2240 year for data analysis and integration. See Theme 4 for further development of this concept. 2241 2242 2) Long-term measurements of chronic sound 2243 2244 A high-priority effort for IQOE should be the initiation of long-term monitoring of sound, 2245 particularly at low frequencies, over large/basin scales. The intent would be to focus in a 2246 sustained way on characterizing variability in overall sound, anthropogenic contributions, and 2247 biological use of areas and possible impacts of anthropogenic sound. Low frequencies 2248 would be a particular focus because of the propagation of low frequencies over ocean basins 2249 and the likelihood that many of the animals that might respond to sound are those that use

58

239 DecemberJanuary 20121 DRAFT

2250 low-frequency sound (e.g., whales, fish). Observation systems are available in many likely 2251 study areas, but will probably need to be augmented for more complete coverage, 2252 particularly in abyssal areas of an ocean basin. 2253 2254 3) Regional “experiments” or monitoring efforts 2255 2256 A geographically focused study with potentially short-term changes in the sound field is 2257 envisaged. This could include the comparison of two similar habitats in an area of somewhat 2258 rapid change and/or contrasting anthropogenic activity, such as comparing the Gulf of 2259 Mexico with the Gulf of California. Such an experiment would occur over a regional scale 2260 (e.g., tens to hundreds of kilometers), weekly to decadal time scales, and would necessarily 2261 consider a broader frequency range than long-term measurements and would consider a 2262 larger number of individuals of the target species and possibly also examine community- 2263 level effects. 2264 2265 This could include opportunistic studies (see Theme 2), and could focus on areas of planned 2266 changes in shipping regulations, the no-boat zone in San Juan Islands (before/during/after), 2267 changes in shipping lanes around the California Channel Islands, designation of Particularly 2268 Sensitive Areas (PSA) (subcommittee within IMO) involving re-routing of ships, and the 2269 presence of transient industrial noise, such as pile driving for the construction of wind farms. 2270 The re-routing of shipping from Unimak Pass to south of the Aleutian Islands to Japan is a 2271 further opportunity. Other opportunities could be available in relation to new gas platforms 2272 installed off Russia, in the Barents Sea involved with new leases, and to changes in 2273 maritime traffic in the Canary Islands area due to logistic and overload issues at other large 2274 container-ship harbors in the Mediterranean Sea and Suez Canal area. 2275 2276 An important action for the IQOE will be to establish an appropriate mechanism for 2277 interacting with the organizations and agencies involved in observations. To this end, 2278 the IQOE will appoint and fund a representative to attend meetings and make the case for 2279 participation of the IQOE in observation activities. 2280 2281 One typical purely logistical example is that the Suez Canal authority is developing a project 2282 to increase the depth of western channels of the Suez Canal from 48 ft. to 52 ft. It is 2283 anticipated that this project will have an impact on the traffic during its implementation, some 2284 of which would have to be redirected temporarily along the west coast of Africa. The Las 2285 Palmas harbor in the Canary Islands has the capacity to handle large container ship and 2286 thus it is expected that during these project periods, which may be repeated in the course of 2287 the IQOE project, a notable increase in traffic may occur and would be monitored by the 2288 PLOCAN observatory station and at the ESTOC site 2289 (seehttp://www.palmasport.es/00000/paginas/asp/noticias_ficha_user.asp?Id_noticia=19292 2290 &codigo_N1=47&codigo_N2=&codigo_N3=). 2291 2292 4) Arctic study comparison 2293 2294 The increasing retreat of the Arctic ice cover is opening up that region to increases in human 2295 activity, which is expected to bring profound changes to the natural (but not quiet) 2296 soundscape. The expected climate changes suggest there is a unique opportunity to 2297 observe the effects of introducing higher levels of anthropogenic noise into this region. A 2298 challenge for the IQOE will be to design experiments that can distinguish between the 2299 effects of changes in sound levels from other environmental change, such as ice cover. 2300 2301 Numerous researchers have deployed and will continue to deploy autonomous and cabled 2302 acoustical recording and monitoring systems in the Arctic, and it would be highly 2303 advantageous to incorporate these capabilities into the IQOE. Such systems include 2304 nascent implementation of an observing system for the Arctic (Figure 5.4), for which

59

239 DecemberJanuary 20121 DRAFT

2305 acoustical applications were highlighted during the OceanObs’09 conference (Dushaw et al. 2306 2009). Acoustical applications provide a unique approach to sense or operate under sea 2307 ice. 2308 2309 The ecological changes in response to a changing soundscape are not expected to occur 2310 instantaneously, but rather are expected to occur over at least the duration of the IQOE and 2311 likely decades into the future; therefore, in addition to programs of short-term autonomous 2312 measurements, this suggests that the IQOE press for a long-term monitoring effort. 2313 2314 Interpreting the results of Arctic studies will be confounded by the influence on the ecology of 2315 the natural climatic change in this environment. It might be possible to utilize the conditions 2316 in the Antarctic for comparison studies to separate the contribution of climate stressors from 2317 sound stressors. 2318 2319 5) Antarctic study comparison 2320 Numerous observational efforts are also underway in the Antarctic using autonomous 2321 systems, and the IQOE should coordinate with these efforts as well. However, if we wish to 2322 develop a system for making long-term measurements of the background ocean ambient 2323 sound, then placing sound observatories under Antarctic ice sheets would enable the 2324 collection of data that is free of near-field interferences. Although technically challenging, 2325 both the Ross and Filchner ice shelves would be appropriate for this purpose and would 2326 “look” out into the South Atlantic and South Pacific oceans, respectively. 2327 2328 2329

2330 2331 Figure 5.4. An envisioned basin-wide AURORA mooring grid in the Arctic Ocean (left panel) 2332 and the existing moored observatory in Fram Strait (right panels). In addition to moored 2333 instrumentation, an array of drifting Acoustic Ice Tethered Platforms (AITP) with acoustic 2334 modems for communications will be deployed in the Arctic Ocean, while in Fram Strait 2335 profiling gliders, capable of under-ice acoustic navigation, will be employed. In the right lower 2336 panel, positions of moorings are overlaid on the temperature distribution in Fram Strait, an 2337 inflow of the warm Atlantic water in the West Spitsbergen Current depicted by red color (XR: 2338 [43]). (From Dushaw et al. (2009).

60

239 DecemberJanuary 20121 DRAFT

2339 6 2340

2341 Theme 4: Industry and regulation

2342 6.1 Introduction 2343 The societal response to concerns about the effects of underwater sound generated from 2344 human activities has been to introduce legislation and regulation of sound-generating 2345 activities. Although still at an early stage in development, mainly because of the limited 2346 evidence for effects of sound on marine life, the legislative frameworks currently in existence 2347 tend to give government policymakers the option to introduce highly precautionary 2348 regulations. This tendency toward caution is mainly the result of high scientific uncertainty. 2349 As examples, both Europe and the United States have these types of legislative frameworks. 2350 2351 The legislative basis for most U.S. regulation focuses on the protection and recovery of 2352 particular species (Hatch and Fristrup 2009), whereas European Union legislation is focused 2353 mainly on reducing the introduction of sound energy into the water (Tasker et al. 2010). EU 2354 regulation also includes aspects of species protection, both directed at listed species and at 2355 the protection of critical habitats, within the Habitats Directive. Under the U.S. Endangered 2356 Species Act (ESA), acoustic injury or disturbance of any listed marine species or population 2357 is considered when determining if an activity will ‘jeopardize’ the existence of the species or 2358 population. The ESA also may consider whether human-generated sound will destroy or 2359 adversely modify habitats that are critical to the listed species. The U.S. Marine Mammal 2360 Protection Act (MMPA) requires that human activities that could harass marine mammals, 2361 including harassment by sound, are subject to a permitting process. Exposure thresholds 2362 relevant to the MMPA have been established by the National Marine Fisheries Service of the 2363 U.S. National Oceanic and Atmospheric Administration (NOAA) to regulate potential impacts 2364 of sound on marine mammals. In the case of cetaceans exposed to sequences of pulsed 2365 sounds, the threshold at which harassment begins, as defined by regulators is 160 dB re 1 2366 μPa @ 1 m. For continuous sounds, the threshold is lower: 120 dB re 1 μPa @ 1 m. In the 2367 case of pinnipeds, the thresholds are 180 dB re 1 μPa at 1 m and 160 dB re 1μPa @ 1 m, Comment [EU22]: From Yvan Simard: again the 2368 respectively (NOAA 2005, 50 CFR 216). These thresholds correspond closely with the dose- units are wrong, they should not be @ 1 m (which is only used for source levels, SL’ not for received 2369 response thresholds defined by the U.S. Navy (see Theme 2). The effectiveness of these levels, RL), but simply dB re 1 uPa. 2370 types of regulatory approaches relative to the costs to human activities that produce sound 2371 in the ocean have been debated extensively and inconclusively. 2372 2373 The effectiveness of these types of regulatory approaches and their impacts (potentially 2374 needless) on human activities that produce sound in the ocean have been debated 2375 extensively. 2376 2377 The global commercial shipping fleet expanded from about 30,000 vessels (of about 2378 85,000,000 gross metric tons) in 1950 to more than 85,000 vessels (about 525,000,000 2379 gross metric tons) in 1998 (NRC 2003) because ships are the most economical means of 2380 bulk transport (Round Table of International Shipping Associations 2011). Perhaps more 2381 than 90% of world trade (in gross tonnage) depends on ship transport and, apart from 2382 declines during global economic downturns, the gross tonnage of goods transported by sea 2383 has steadily increased since the early 1970s (Figure 6.1).

61

239 DecemberJanuary 20121 DRAFT

2384 2385 2386 Figure 6.1 Changes in the metric ton-miles of different types of cargos moved by shipping 2387 from 1970 to 2010. Source: Fearnley's Review. 2388 2389 2390 A continuing increase in shipping traffic is not certain because there is likely to be an upper 2391 limit in the growth of ship transportation of goods brought about by (1) periods of slow or 2392 stagnant economic growth; (2) increased efficiency of the movement of goods; (3) reduced 2393 availability of raw materials or more efficient local sourcing of raw materials; and in the very 2394 long term (4) slower increases in demand because of leveling out of the global human 2395 population and especially the population of richer, Northern Hemisphere nations. Comment [ILBoyd23]: I think most of the 2396 Furthermore, the sound produced by ships is unlikely to increase in proportion to either the projections show it leveling off. 2397 number of ships or the tonnage of goods moved because of increasingly efficient and quiet Comment [E24]: I don’t know about this one. 2398 ship designs. This may be borne out in part by the difference between the rapid increase in The BRIC countries are likely to more than replace 2399 shipping since 2000 shown in Figure 6.1 and the decline in apparent shipping sound in the any decreases in developed countries. 2400 ocean shown in Figure 2.2 (Andrew et al. 2011). An objective of Theme 1, on ocean 2401 soundscapes, is to resolve whether there is a relationship between low-frequency 2402 omnidirectional ambient sound and shipping. Figure 6.2 shows the radiated sound spectrum 2403 of a passing ship. 2404 2405 It is much more difficult to compile evidence of changes in ocean sound from other key 2406 human activities, such as oil and gas exploration and other offshore engineering. The IQOE 2407 is aimed principally at resolving some of the critical scientific uncertainties associated with 2408 our understanding of how sound travels in the ocean from source to organisms and how 2409 organisms react, both individually and as populations. However, only partial progress will be 2410 made during the IQOE and considerable uncertainties will remain. It is important that the

62

239 DecemberJanuary 20121 DRAFT

2411 IQOE has the capacity to maximize its effectiveness within the arenas of policy and 2412 regulation to render current approaches to regulating marine sound more effective. This 2413 theme will develop the applied axis of the IQOE research activities to complement the more 2414 fundamental research of the other themes. Some of this theme’s approach will concern 2415 undertaking specific research that could provide the basis for more informed approaches to 2416 regulation, such as that used by NOAA for regulating effects upon marine mammals (NOAA 2417 2005), but some will involve weaving an app*lied thread through the activities defined in 2418 Themes 1, 2 and 3, and ensuring that the knowledge gained is used in future regulatory 2419 activity. Consequently, some of the activities specified in this theme refer to those in other 2420 themes. 2421

2422

2423 Figure 6.2 The radiated sound spectrum from a passing ship in which the brighter colors 2424 show high sound intensity and where the distance is measured in meters from the ship. 2425 From Kozaczka et al. (2010).

2426

2427 6.2 Risk frameworks

2428 6.2.1 General description of risk frameworks 2429 2430 Exploitation of the ocean and its resources is a necessary part of human economic and 2431 social development. Continued expansion of the global human population for the 2432 foreseeable future, together with declines in the availability of basic raw materials, including 2433 energy resources, creates a strong imperative to continue to exploit ocean resources at least 2434 as much as in the present day and probably much more in to the future. 2435 2436 Therefore, industrial development will continue even in the face of increasing regulatory 2437 constraints. In these circumstances, it can be costlyis impossible to wait until scientific 2438 knowledge catches up to provide appropriate levels of certainty so that industry can move 2439 forward with a high degree of certainty that the options chosen for future development will 2440 not significantly impact the sustainability of the ocean environment. Consequently, we need 2441 a framework within which progress can be made in a measured manner, while 2442 simultaneously minimizing risks to the environment and the costs to industry in terms of both 2443 direct financial costs and those related to lost opportunities. Such a framework will explicitly 2444 assess risk and incorporate adaptive management of the industrial process and 2445 development (Boyd et al. 2008; Figure 6.3). 2446 2447

63

239 DecemberJanuary 20121 DRAFT

2448 2449 2450 Figure 6.3. An illustration of the information flow and decision pathway for a risk assessment 2451 process. This shows a feedback process involving mitigation when the risk exceeds the 2452 trigger level for management action. This is an adaptive approach to managing risk. From 2453 Boyd et al. (2008). 2454 2455 2456 2457 The advantage of a science-based framework for regulation of sound in the ocean is that it 2458 allows industrial activity to proceed in a precautionary manner and establishes procedures 2459 for collecting information about its effects as activity proceeds. Effects are then assessed 2460 against pre-determined objectives. If those objectives are not met, mitigation is introduced, 2461 the mitigated activity is allowed to continue, and once again assessed against specific 2462 objectives. This procedure is continued until a satisfactory operating procedure or design is 2463 found for the industrial activity. Many circumstances lend themselves to this approach, but 2464 some activities will always be found to be too harmful to continue. 2465 2466 • Which sound sources need additional characterization? 2467 • What can we do to develop acceptable (by industry, regulators and stakeholders) 2468 standards and methods for measurement? 2469 • Can we develop alternative sound sources for seismic airguns for example, marine 2470 vibrator? 2471 • What environmental impacts do alternative sound sources have on the environment? 2472 • What can be done to existing sound sources to reduce unwanted sound? 2473 • How does industry measure its contribution to the ocean sound budget? 2474 • What do we need to do to better understand the global background sound status? 2475 2476 b. Dose-response assessment - Quantitative relation between dose (i.e., received 2477 sound characteristics) and response (behavioral response, masking, TTS, PTS, 2478 injury). Information needs in this category include the following. 2479

64

239 DecemberJanuary 20121 DRAFT

2480 Physical and physiological effects 2481 2482 • Does background sound impact masking and, if so, to what extent? 2483 • How do we measure or assess how the received sound is perceived by the animal? Does 2484 this change with environmental context? 2485 • Which animals are more susceptible to exploration and production (E&P) sounds? For 2486 example, are large whales more susceptible than small whales or fish? 2487 • Can the hearing range of animals be modeled? 2488 • Can hearing damage from sound be modeled? 2489 • What frequencies can be heard by various species of marine life? 2490 • What impact does long-term sound exposure have on hearing ability of marine life? Is 2491 there an effective recovery period? 2492 • Could seismic sources cause animals to become frightened, resulting in vertical flight and 2493 the “bends”? 2494 • What are the impacts of industrial sounds on fish and turtles? 2495 • What are the impacts of industrial sounds on prey species? 2496 • What is the relationship between equal loudness curves and audiograms? 2497 2498 Behavioral effects 2499 2500 • What aspects of the sound source are responsible for behavioral response: exposure 2501 level, peak pressure, frequency content, harmonic context, texture, etc.? 2502 • What behavioral responses occur when animals are exposed to exploration and 2503 production (E&P) sound sources? 2504 • How does E&P sound impact animal displacement? Is it species-specific or do other 2505 environmental context factors prevail? 2506 • Do long-term industrial operations have impacts on animal residency? If so, which 2507 species are most affected and to what extent? 2508 • What is the impact of masking on animal behavior? 2509 • Do animals become sensitized or habituated to repeated particular sound exposures? 2510 2511 c. Exposure Assessment – Specifying the population that might be exposed to the 2512 hazard, identification of routes through which exposure can occur and estimating 2513 the magnitude, duration, and timing of the dose that marine mammals might 2514 receive as a results of exposure. Information needs in this category include 2515 2516  Distribution and abundance of cetacean/fish stocks over long time periods to identify 2517 overlap between sources and receivers 2518  Quantification of industrial activity in the areas under question 2519  How does industrial activity translate into sound budgets? 2520  What is the baseline stressor situation (e.g., what other stressors than sound act on the 2521 population under question?) 2522 2523 d. Risk Characterization – Overall assessment of risk by integration of information 2524 from the first three steps to develop a qualitative or quantitative estimate of the 2525 likelihood that any of the hazards associated with the sound source will be 2526 realized in marine life. Information needs in this category include 2527 • How can we measure biological significance? 2528 • At what level of effect do we need to be able to detect biological significance? 2529 • With respect to biological significance, are all behavioral reactions equal in importance? 2530 • Are there special biological “hotspots” for animal production that should be avoided at 2531 times? 2532 • What is the population growth in areas where sound is prevalent?

65

239 DecemberJanuary 20121 DRAFT

2533 • Can models (population, energetic, etc.) be used to predict changes in vital rates that 2534 could cause biological significant impacts to animal populations? 2535 • How do we determine Population Consequences of Acoustic Disturbance (PCAD) 2536 transfer functions? 2537 • How do we use model output in risk assessments? 2538 • Cumulative Impacts: Sound in the context of other pressures to gather a more holistic 2539 view of baseline situations. 2540

2541 6.3 Routine sound monitoring 2542 2543 Measurement of the characteristics of the sources of human-generated sound is required so 2544 that these can be used within sound propagation models (see Theme 1). Figure 3.3 shows 2545 the amplitude of different frequencies with distance from a pile driver but, not only is it 2546 necessary to examine amplitude as in this case, other characteristics of the sound need to 2547 be determined, such as directionality, bandwidth, kurtosis, particle motion or, in the case of 2548 pulsed sound, pulse width, height and rise time. The sound radiated from a source (Figure 2549 6.2) can vary with orientation and, in the case of ships, their speed and whether they are 2550 loaded. In addition, a fuller understanding of how the characteristics of these sounds may 2551 change with propagation over larger ranges is required. Even though sound from shipping is 2552 probably the greatest single source of human-generated sound in the ocean, this will not 2553 always be the case on local or regional scales. Pile driving is a fast-growing activity in some 2554 coastal regions, mainly because of construction of offshore wind farms. Seismic surveys 2555 using airguns are very widespread. Other sources of marine sound include—but are not 2556 limited to—construction sound, dredging, acoustic communications, supersonic aircraft, 2557 ammunition, explosives, and naval training and surveillance sonars. Although sound spectral 2558 characteristics for many of these sources are available as examples that may (or may not) 2559 typify those types of sources, there is a need to compile information about the characteristics 2560 of these different sources and the extent to which they can be described by typical 2561 examples. The compilation then needs to be made available and the IQOE will promote 2562 this by developing a Web-based repository for spectral information about sound 2563 sources. 2564 2565 The IQOE will also adopt and promote standards for measuring these values, such as 2566 developed by ISO on coastal shipping sound (Ainslie 2011; Carey and Evans 2011). A 2567 working group is currently being established by ISO to develop standards for a variety of 2568 sound sources, including natural, biological, and anthropogenic sounds, and the IQOE will 2569 work closely with this group to adopt appropriate standards. The IQOE will promote these 2570 standards through the provision of Web tutorials about their application. 2571 2572 In many regions, statutory monitoring of sound levels is being developed around offshore 2573 developments or is being implemented by regional management authorities in response to 2574 wide-ranging legislation concerning management of the marine environment. This includes 2575 the Marine Strategy Framework Directive in European waters, where there is a commitment 2576 to comply with limits to the level of human-generated sound in the ocean. 2577 2578 Methodological standards need to be developed for ambient sound measurements over long 2579 periods, to allow legislative standards to be established and enforced. The outcomes of both 2580 Themes 1 and 3 will support this requirement and the presence of observatories established 2581 for the statutory monitoring of marine noise also presents opportunities for low-cost data 2582 collection in support of research with broader objectives. 2583 2584 Sound recording systems of this type are likely to become routine in future and the IQOE 2585 has the opportunity to influence the design of monitoring protocols, the hardware systems for 2586 carrying out monitoring and the data storage and analysis systems, and also to benefit from

66

239 DecemberJanuary 20121 DRAFT

2587 these systems. However, in order to achieve these goals, the IQOE will need to engage 2588 closely with those who are responsible for establishing sound monitoring to maximize these 2589 benefits. Figure 6.4 illustrates a system of sound monitoring on a local scale that has been 2590 developed to provide both regulators and ship operators with a solution to a conflict 2591 between, in this case, ship strikes on northern right whales. This system uses acoustic 2592 detection of the right whales to alert the ships to their presence. 2593 2594

2595 2596 2597 Figure 6.4 Diagram showing a real-time auto buoy system that is operational off the 2598 northeast coast of the United States. The system alerts mariners to the presence of right 2599 whales to reduce the probability of ship strikes. From van Parijs et al. (2009).

2600

2601 6.4 Research priorities for regulators and industry

2602 6.4.1 Defining the biological significance of sound for marine organisms 2603 2604 The effects of specific doses of sound on protected species—as well as other species that 2605 may have ecological or economic value—is important for regulators and, therefore, also for 2606 industry. In the Introduction to this Science Plan, the Population Consequences of Acoustic 2607 Disturbance (PCAD) model developed by the NRC (2005) was described. This approach 2608 leads to the definition of significant biological effects of sound as well as the knowledge 2609 needed to define the significance of an effect. From an industry and regulatory point of view, 2610 while it would be ideal to understand the mechanisms underlying effects, there is a greater 2611 need to move quickly to developing precautionary indicators of significant effects. This may 2612 mean that simple, but robust, empirical relationships between sound and responses in 2613 marine organisms need to be discovered and used. The behavioral responses of beaked 2614 whales to sonars (Tyack et aI. 2011) are an indicator of potential harm. Translating the 2615 probability of behavioral disturbance into a probability of a significant population effect can 2616 probably be achieved relatively easily, but with low levels of confidence. Additional work will 2617 then be necessary to reduce the confidence intervals around the estimated level of effect. 2618 2619 Studies concerning individual impacts can deal with behavioral response, masking, 2620 temporary threshold shift (TTS), and auditory or non-auditory injury. Although it has been 2621 shown that, in some cases, sound can injure or even kill marine life, population-level 2622 consequences are unlikely, as only a very limited proportion of the population are generally

67

239 DecemberJanuary 20121 DRAFT

2623 affected. In contrast, behavioral and masking effects can occur at lower sound levels and 2624 over vastly larger areas, and therefore may affect a larger and potentially significant portion 2625 of the population. 2626 2627 The effects of multiple sound exposures (both sequentially and simultaneously) may 2628 accumulate and add to the effects of other stressors, such as the case of ship collisions and 2629 northern right whales (Figure 6.4); the IQOE needs to consider such cumulative effects. 2630 Therefore, the program will undertake modeling to examine how the measured, usually 2631 adaptive, responses of animals to underwater sound can be re-scaled to develop realistic 2632 representations of risk to populations. 2633 2634 The research needed to define biological significance of sound for marine organisms can be 2635 summarized as follows: 2636 2637  To make it possible to scale up the problem through the accumulation of effects on 2638 individuals to populations, it will be important to better identify the effects of underwater 2639 sound on individual marine animals. This could involve investigations of 2640 o TTS: More studies are needed to assess temporary threshold shifts for risk 2641 assessments; 2642 o Behavioral effects: There is very little understanding on the behavioral effects of sound 2643 on marine life; 2644 o Masking: experimental evidence that an impact occurs, not only forecasts of effects 2645 based on modeling. 2646  Knowing the spatial distribution of sensitive species and how this changes through time 2647 as a way of defining critical habitats for industry to avoid. 2648  Whether the PCAD approach can provide practical solutions to the problem of regulating 2649 sound production by industrial activity. Both industry and regulators may need solutions 2650 that are less precise, but more tractable. 2651  Providing key information as defined by a risk-based approach to adaptive management 2652 of industrial activity when faced with high uncertainty about the effects of sound.

2653 6.4.2 Key topics 2654 2655 Information needs with respect to mitigation measures for industry and regulators include the 2656 following: 2657 2658 • What is the effectiveness of existing monitoring techniques and tools? 2659 • What can be done to improve current monitoring techniques and tools? 2660 • What additional monitoring tools (short- and long-term) can be developed to assist in 2661 marine mammal observation? 2662 • Can International Maritime Organization (IMO) data be used as an analysis tool on a 2663 local and regional basis? 2664 • Can Active Acoustic Monitoring become a viable monitoring tool? Will it be 2665 acceptable to regulators and stakeholders? 2666 • Are “soft-start” or “ramp-up” effective mitigation techniques? 2667 • Are there other ways to mitigate unwanted sound, for example, air bubble curtains? 2668 2669 For all industrial activities, a fuller understanding of current and projected trends in activity 2670 levels and sounds produced is important. 2671 2672 Shipping sound and its potential impact on marine life is an area of increasing concern, and 2673 has received relatively limited study. Shipping sound can mask communication signals of 2674 marine mammals and fish, and both taxa have been shown to change behavior in reaction to 2675 these sounds. However, there are uncertainties in assessing impacts of shipping sound. 2676 Predictions based on theory indicate that communication ranges can be decreased as a

68

239 DecemberJanuary 20121 DRAFT

2677 result of increased sound levels; however, understanding the ecologically relevant functional 2678 communication range is critically important. There are also large differences in potential 2679 effects between deep and shallow waters and among the taxonomic groups affected. 2680 2681 Cavitation sound by propellers is a major source of shipping sound, so mitigation attempts 2682 should be targeted there. The Marine Environmental Protection Committee (MEPC) within 2683 IMO has formed a correspondence group on sound from commercial shipping that deals with 2684 mitigation measures such as ship-quieting technologies. Propellers are likely to be re- 2685 designed as a response to requirements to make more efficient propulsion. Sound radiation 2686 should be considered as part of the design process from the start. 2687 2688 Impact pile driving, for installation of offshore wind farm turbines, has been shown to lead to 2689 wide-ranging behavioral impacts on small odontocetes, such as harbor porpoises, and can 2690 injure marine life very close to the source. The issue, however, has been addressed in 2691 some regulations such as the MSFD (also see OSPAR, 2009). Although marine pile-driving 2692 is generally well regulated in many regions and Environmental Impact Assessments (EIA) 2693 are often required, cumulative effects of multiple piling activity, or cumulative effects of piling 2694 with other stressors (acoustic and non-acoustic) are not well understood and are beyond the 2695 scope of individual EIAs. Furthermore, particle acceleration is a primary concern for fish, 2696 especially in the near field, and should be measured. 2697 2698 In examining the effects of seismic airguns on marine life, the OGP has set a commendable 2699 example of the engagement of industry in researching acoustic impacts through the Joint 2700 Industry Program (OGP-JIP). While there are currently ongoing studies characterizing the 2701 source characteristics of seismic airgun arrays, among a range of other studies (e.g., 2702 behavioral response study of Australian Humpback whales to seismic airgun exposure), 2703 there remains a need for further research in the areas of behavioral effects, masking, and 2704 efficacy of mitigation measures. 2705 2706 Additional sound sources to be considered include naval sonars, wave- and tide-energy 2707 generation, new exploration technology, dredging, echo sounders, active positioning 2708 systems, and fish farming, among a range of others (overview provided by the World Ocean 2709 Council). Development of new technologies, such as marine vibroseis, that replaces, where 2710 possible, existing sources is strongly encouraged. 2711 2712 Research needed to determine how best to conduct routine sound monitoring can be 2713 summarized as follows: 2714 2715  Define what is meant by ‘routine sound monitoring’. For example, it is not always clear if 2716 this should include measurements of specific activities or monitoring of ambient sound. 2717  In the case of ambient sound monitoring, it is necessary to identify the objective in terms 2718 of meaningful and valued outcomes. Monitoring ambient sound can produce sound maps 2719 and sound budgets, and can identify trends of ambient sound over predetermined time 2720 scales in specific areas. This last objective is required, for example, by the EU Marine 2721 Strategy Framework Directive. Each of these objectives could require different 2722 approaches. 2723  Modeling sound propagation will be an important approach in designing monitoring 2724 networks and also for analyzing data. 2725  The development of compliance monitoring for sound levels at regional scales will require 2726 several steps, including 2727 2728 o Identification of existing ambient sound measurement data within the study area. 2729 o Identification of suitable measurement systems (there is a database on suitable 2730 devices).

69

239 DecemberJanuary 20121 DRAFT

2731 o Identification of existing sea observatories in the study area. 2732 o Assessment of the feasibility of using existing observatories for ambient sound 2733 monitoring. 2734 o Identification of representative sites (e.g., pressure areas or areas of high sensitivity); 2735 and the possibility of establishing reference areas where there is very little human- 2736 generated ambient sound to describe natural fluctuations. 2737 o Development of a work plan, including a maintenance schedule and data analysis 2738 reporting cycle. 2739  If ambient sound monitoring is attempted in sensitive areas, the distribution and 2740 abundance of sensitive receivers (marine mammals and fish sensitive to the source in 2741 question) needs to be documented to a higher standard than elsewhere so that there is 2742 sufficient statistical power to detect important changes in population status, and to be 2743 able to report these in sufficient time for management action to be taken. 2744  Sound frequencies that are most biologically important should be monitored. 2745 Consequently, defining these frequencies is important. 2746  An investigation of the costs and benefits of monitoring is needed to ensure that the 2747 outcomes lead to a net benefit for society. 2748

2749 6.4.3 Modeling the relationship between industrial activities and sound levels 2750 2751 An important first step in modeling the relationship between industrial activity and sound 2752 levels is to define the permissible sound budgets for a region and to compile inventories of 2753 human-generated sound in a region to establish whether regulatory targets have been met. 2754 This is a particular need of the regulators, but industry also requires simple and cost- 2755 effective tools for demonstrating compliance as well as real-time feedback to allow optimal 2756 decision-making during marine operations. For example, if the captain of a container ship 2757 knows the radiated, speed-dependent sound profile of the ship and also knows the 2758 contribution that the ship is allowed to make to the sound budget of the region, the captain 2759 can make a judgment about the speed at which the ship can travel, while also knowing that if 2760 the ship exceeds this speed, its excess contribution to the sound budget may be detected. 2761 Regulation is moving towards developing sound budgets, but we do not currently have the 2762 sophisticated mechanisms in place to optimize human behavior in a way that matches 2763 aspirations. 2764 2765 Some progress may be made through the compilation of sound inventories that 2766 quantitatively assess the contribution and characteristics of different sources to the overall 2767 sound field in a given region. Such compilations can be used to identify research and 2768 regulation priorities. Furthermore, sound inventories might help in determining whether, and 2769 to what degree, human activities contribute to the ambient sound field. Thus, they are tools 2770 that can be used in marine management. 2771 2772 Robust, validated models of ambient sound can provide an essential tool for marine spatial 2773 planning and informed regulation. The need extends from regional areas of heavy traffic to 2774 ocean-basin scales. The goal should be to characterize the present ambient sound 2775 conditions and how ambient sound is changing, and to explain the sound field in terms of 2776 known sources (natural and anthropogenic) and known propagation conditions using 2777 appropriate models. The requirements for model development are 2778 2779  establishing a program of long-term sound measurements (see Theme 3), 2780  cataloguing sound sources (see 6.2), 2781  integrating these into propagation models (see 3.3), and 2782  validating these models with direct observations (see 5.1). 2783

70

239 DecemberJanuary 20121 DRAFT

2784 Robust validated models can be used to evaluate results of sound-constraining regulation, 2785 estimating historical sound levels (changing whale populations, economic activity, ship 2786 propulsion efficiency) and estimating sound radiated from potential traffic patterns.

2787 6.5 Finding novel solutions 2788 2789 Finally, the IQOE is concerned with developing novel solutions to the suspected problem of 2790 the effects of sound on marine life but, in so doing, this drives innovation that will have spin- 2791 offs for the overall efficiency with which society operates. If by simply questioning whether 2792 we really need to transport >90% of the world’s goods by sea, decision makers can consider 2793 the costs involved, and this can drive the development of fundamental innovations that may 2794 transform the efficiency of global transport networks. It will be an objective of the IQOE to 2795 work alongside others by modeling transport systems to identify savings that could 2796 reduce the human-generated component of ocean sound and, in so doing, find more 2797 efficient solutions overall. The IQOE will work with the World Ocean Council and others to 2798 provide appropriate information to maximize the effectiveness of these models. 2799

2800 6.6 Key recommendation 2801 2802 The IQOE encourages the IMO to further address the issue of shipping sound as one of the 2803 main contributors to underwater sound in the sea. The IQOE will work in collaboration with 2804 the WOC by developing a joint working group on the issue of underwater sound to 2805 strengthen the links between industry and research. This should be supported by a clear 2806 communication and outreach plan to convey the results of IQOE to global stakeholders, 2807 policymakers, and the public. 2808 2809 The IQOE will, as far as possible, promote the use of scientific results to harmonize national 2810 regulations. There is a danger of different standards being applied by different countries, 2811 resulting in confusion and added costs for industry. Although the regulation of human 2812 activities is a policy decision that is based on more than scientific facts, the same results 2813 could be used very differently across national boundaries. For example, many EU member 2814 states prioritize underwater sound in very different ways. Researchers operating within the 2815 context of the IQOE have an opportunity to harmonize the foundation of scientific information 2816 behind the decisions of environmental management. Wherever possible, environmental 2817 management should be based on empirical studies and standardization of measurement 2818 techniques globally. 2819

2820 6.7 Engaging Industry and Other Stakeholders 2821 2822 A comprehensive plan for the engagement of stakeholders and the public will be a priority for 2823 the IQOE. This plan will include targeted activities that reach policymakers, industry 2824 representatives, the media, and other stakeholders. Strategic activities will build awareness 2825 of the IQOE’s research portfolio and encourage the participation of important industries in 2826 research activities. The following potential activities would support these goals. 2827

2828 6.7.1 Program “Launch” 2829 To accomplish IQOE goals within an international framework, it is essential that the program 2830 become widely known as a credible and trustworthy source of authoritative information and a 2831 basis for new measurements and new understanding of the effects of sound on marine life. 2832 An initial outreach and communications effort should reach professionals and scientists 2833 working on related research, print, and broadcast news media; the scientific community at 2834 large; stakeholders (e.g. policy and decision-makers, fishers, oil and gas industry

71

239 DecemberJanuary 20121 DRAFT

2835 representatives, and the environmental community); and the public to increase awareness of 2836 the IQOE, its mission, and potential projects. Public trust and confidence in the IQOE is 2837 critical and will be advanced as these parties become convinced of the scientific integrity of 2838 the research being conducted by the program’s participants and discover the utility and 2839 timeliness of the end products being created by this significant international research effort. 2840 2841 It is recognized that the long-term viability of the IQOE depends on broadly based funding 2842 from government, industry, and private sources. This funding will only come if the program 2843 achieves broadly based public support. A highly focused and vigorous outreach program to 2844 develop credibility of IQOE activities and eventually inform stakeholders and the public of the 2845 program’s achievements will play a fundamental role in building this support and ensuring 2846 the program’s long-term viability. To help achieve this, a common message will be 2847 developed to present a consistent image. 2848 2849 It is recommended that the program be “launched” with a suite of activities that include a 2850 scientific symposium, a public lecture, and evening reception with important international 2851 government officials and dignitaries. To attract the international press, these activities should 2852 be held in a high-profile venue such as the Natural History Museum in London, the 2853 Smithsonian Institution in Washington, DC, the Museum of Oceanography in Monaco, or 2854 perhaps several such venues simultaneously. 2855

2856 6.7.2 Central Web Portal 2857 A Web portal will provide a convenient “entry point” for the internal project participants and 2858 external community. It should reflect and highlight the state-of-the-art research being carried 2859 out by IQOE scientists. It should provide overviews and links to and from each field project. 2860 2861 The portal should include a password-protected IQOE section for project participants. It 2862 should also allow for regulator and industry access to appropriate resources. 2863 2864 Integrated social media should be considered in developing the site. Also provisions will be 2865 necessary for cataloguing image, audio, and video files. 2866

2867 6.7.3 Interactive On-line Database 2868 An on-line database must be developed to allow for the cataloging, sharing, and archiving of 2869 program data. It should provide data access to industry and regulatory representatives. 2870 Periodic on-line training on how to use the data tools and contribute to the database will be 2871 important.

2872 6.7.4 Media Relations 2873 The ongoing work of the IQOE will be brought to the attention of the international news 2874 media and relationships between media and project representatives must be established. 2875 Resources such as “backgrounders” on important issues and new findings should be 2876 available on the central Web portal for the media, as well as audio and video material. 2877

2878 6.7.5 Highlights Reports 2879 Stakeholder and public interest can be developed through the sharing of new discoveries 2880 and research findings in well-publicized annual “highlights” reports. The release of these 2881 reports should be timed to coincide with an annual international media campaign. 2882

72

239 DecemberJanuary 20121 DRAFT

2883 6.7.6 Partner Resources 2884 It is imperative that the IQOE gain the attention and support of public officials and other 2885 stakeholders worldwide. To help facilitate this, a suite of informational materials should be 2886 developed, which can be easily replicated, translated, and distributed by program 2887 participants. Common resources will assist project partners in communicating about the 2888 program and to engage stakeholders in their regions. These resources could include fact 2889 sheets, maps highlighting on-going research, reproducible graphics, etc. It is also important 2890 that the program have a consistent “graphic identity” that is print and Web friendly. Other 2891 resources could include PowerPoint presentations, archived webinars, print materials, 2892 applications, graphic visualizations, etc. 2893 2894 Engagement of stakeholders is critical in all these processes. In fact, many data are already 2895 available, for example as recorded in various environmental assessments, and these data 2896 should be made more widely accessible. Initiatives engaging the industry such as the JIP 2897 and the recently formed working group on underwater sound within the Central Dredging 2898 Agency (CEDA) are very important first steps. In general, important roles will be played by 2899 the industry associations such as the International Oil and Gas Producers (OGP) and crucial 2900 international bodies such as the IMO to help us fill the knowledge gaps.

73

239 DecemberJanuary 20121 DRAFT

2901 7 2902 2903 Implementation 2904

2905 7.1 Project structure 2906 2907 The IQOE is planned to last 10 years in total. The approximate timeline is presented in 2908 Figure 7.1. A major feature will be a focus on the work around an International Year of 2909 Ocean Acoustics (IYOA). An objective will be to conduct as much activity as possible during 2910 this year. This coordination will make it possible to focus intensive research, observations, 2911 and modeling worldwide and at large scales simultaneously. The IYOA will be used as a 2912 focus of public attention on sound in the ocean. However, considerable research will take 2913 place in the years before and after the IYOA. 2914 2915 The decadal time scale is needed because the field is beginning from a very low base of 2916 organization and working up to an IYOA that will require considerable planning. Similarly, 2917 reporting upon studies carried out will require several years to complete partly because of 2918 the potential complexity of the data that will emerge and the time that will be required to 2919 analyze it. The community of scientists involved at the IQOE Open Science Meeting was 2920 enthusiastic about a decadal project. 2921 2922 2923

2924 2925 Figure 7.1 The proposed timeline for the IQOE, showing the implementation schedule for 2926 the five major elements of the program. Additional arrows could be added to show the timing 2927 of specific events; those that are labeled in this diagram are for illustration only. More 2928 workshops are scheduled during the planning stage than later in the program, but we expect 2929 specific meetings to take place to develop funding proposals and to guide the 2930 implementation of practical research. There is a plan to hold a conference around Year 9/10 2931 in order to report the results. 2932

74

239 DecemberJanuary 20121 DRAFT

2933 7.2 Operating approach 2934 The IQOE will be structured to be a coordinating mechanism for all researchers with an 2935 interest in underwater sound and its effects on marine organisms. Research will be funded 2936 through traditional national and regional sources. However, the IQOE will not be a simple 2937 badging exercise. Any research that will be part of the IQOE will need to be approved by an 2938 IQOE Steering Committee (see 7.4). The Steering Committee will apply standards 2939 associated with the objective, planning, data collection and reporting and will monitor 2940 progress of the collected set of IQOE-endorsed scientific activities. In return for complying 2941 with IQOE standards, researchers organizing specific studies will benefit from being a part of 2942 the process that defines the data standards, organizes outputs into common formats and 2943 provides central coordination of data management and modeling. In addition, they will be a 2944 part of a community that achieves the critical mass sufficient to sustain a high profile within 2945 the stakeholder community and that provides representation for their scientific outputs to 2946 policy-makers, industry and the public. 2947

2948 7.3 Governance 2949 The IQOE derives its authority from its organizational co-sponsors, the Scientific Committee 2950 on Oceanic Research (SCOR) and the Partnership for Observation of the Global Oceans 2951 (POGO). These organizations also provide the international context within which the IQOE 2952 will operate. However, it is possible that national or regional subsections of the IQOE may 2953 develop, especially in association with conducting regional experiments. As an example, the 2954 activities in the United Kingdom of the Underwater Sound Forum, which meets under the 2955 auspices of the UK Marine Science Coordinating Committee5, has the capacity to link 2956 directly to the IQOE as a way of building its capacity to develop the evidence base for 2957 policies concerning marine sound. 2958

2959 2960 2961 Figure 7.2 The governance structure of the IQOE. 2962 2963 2964 Operational governance of the IQOE will be the responsibility of the Steering Committee 2965 (Figure 7.2). This committee and its chair(s) will be appointed by the sponsors. The Chair(s)

5 The MSCC is and inter-departmental committee of the UK government that considers the organisation of marine science because marine issues cut across many different government interests.

75

239 DecemberJanuary 20121 DRAFT

2966 of this Steering Committee will report annually to SCOR and POGO about all aspects of the 2967 IQOE, including progress in implementing the Science Plan and with outreach and 2968 stakeholder engagement. The Steering Committee may have joint chairs, particularly 2969 representing biological and acoustical expertise. The Steering Committee will be formed of 2970 its chairman, the chairs of the permanent working groups, and as many as 10 additional 2971 members. The current plan is to have two standing working groups: 2972 2973 (1) The Working Group on Data Management and Communication will have responsibility 2974 for defining the standards of data collection and managing the various systems used by 2975 IQOE for communication, including the Web site and associated data portal. 2976 (2) The Science Working Group will have responsibility for determining whether projects 2977 proposed by the research community should be included within the IQOE. 2978 2979 Other standing working groups may be set up as the project develops to handle specific 2980 tasks. 2981 2982 Appointments to working groups will be the responsibility of the Steering Committee. The 2983 IQOE will be managed on a day-to-day basis by the IQOE Executive. 2984 2985 Much of the international coordination and planning of the IQOE will be conducted by ad hoc 2986 working groups, each of which will be established to perform a specific task related to the 2987 IQOE objectives. Such tasks could be to conduct a specific experiment or study, but could 2988 equally be associated with coordination of public outreach or stakeholder engagement. 2989

2990 7.4 Project management 2991 Following review and approval of the Science Plan by the sponsors, the Steering Committee 2992 (SC) will be selected by project sponsors. With the assistance of the IQOE IPO, the SC will: 2993 2994  Manage implementation of the IQOE Science Plan and coordinate IQOE activities 2995 among different nations; 2996  Oversee the budget of the project; 2997  Establish appropriate policies for data management and sharing, and for standards 2998 and inter-calibration to ensure that IQOE data collected by different investigators are 2999 comparable and to promote sharing and preservation of IQOE-related data; 3000  Collaborate, as appropriate, with other related programs; 3001  Create and implement the communication strategy (see 7.5); and 3002  Report annually to SCOR, POGO and any subsequent sponsors, on the state of 3003 planning and accomplishments of IQOE. 3004 3005 SCOR will provide the primary administrative support for the IQOE, at least initially. The 3006 project will require at least one staff person whose time is devoted to helping implement and 3007 represent IQOE. Duties that will need to be handled by the IQOE IPO include: 3008 3009  Helping the SC with logistics for meetings and publications 3010  Representing the project at various meetings 3011  Fund raising for project activities, working with the SC and sponsors 3012  Communications and outreach, including Web site, newsletter, etc. 3013  Management of metadata 3014 3015 Many international projects benefit from the creation of national committees, which can lead 3016 national efforts for planning science activities, raising funding for these activities, promoting 3017 national data management activities, promoting capacity building, etc. The IQOE will 3018 investigate the formation of national and/or regional project committees. These could be

76

239 DecemberJanuary 20121 DRAFT

3019 particularly relevant in relation to basin-scale observations and experiments. The chairs of 3020 these national/regional committees will be ex-officio members of the IQOE Steering 3021 Committee. 3022

3023 7.5 Communication 3024 The IQOE will develop a communication strategy at an early stage. This strategy will identify 3025 all the major stakeholder groups and define the methods that will be used to communicate 3026 the purpose and results of the IQOE, as well as to encourage the involvement of 3027 stakeholders, when appropriate. The strategy will lead to the provision of literature and other 3028 materials that will allow those associated with the IQOE to provide a consistent and clear 3029 message. Figure 7.3 illustrates this kind of approach when explaining the derivation of the 3030 name of the program. Within 4 words, it is possible to transmit in very simple terms who will 3031 be involved, what the program is about, where it will take place (including the frame of 3032 reference) and how the program will be conducted. 3033

3034

3035

3036

3037 Figure 7.3 A diagrammatic representation of how the program name coveys a complex 3038 message in simple terms.

3039

3040 7.6 Education and capacity building 3041 3042 An important aspect of large-scale international research projects is to help build capacity for 3043 science related to the project. In part, this is accomplished through involvement of graduate 3044 students in the work of their advisors. However, the sponsors also expect their projects to 3045 encourage the development of science capacity in developing countries. Research on 3046 sound in the ocean and its effects on marine organisms can be carried out anywhere that 3047 hydrophones could be deployed, but it will be important to provide opportunities for scientists 3048 from developing countries to participant in the IQOE and to receive training as a result of the 3049 project. SCOR and POGO conduct a great deal of capacity building through their various

77

239 DecemberJanuary 20121 DRAFT

3050 programs. Some specific opportunities that could specifically relate to IQOE topics include 3051 the following: 3052 3053  POGO Visiting Professorships—POGO offers the opportunity for institutions in 3054 developing countries to host scientists from other countries (developed or 3055 developing) for periods of about 6 months, to serve as teachers and mentors, and 3056 potentially to conduct joint research (see http://ocean-partners.org/index.php/training- 3057 and-education/pogo-visiting-professorship). 3058  POGO/SCOR Visiting Fellowships for Ocean Observations—POGO and SCOR co- 3059 fund a program that provides an opportunity for students, technicians, post-doctoral 3060 fellows, and other early-career scientists to visit an institution in another country to 3061 learn how to deploy and operate, and analyze data from, observing systems (see 3062 http://www.ocean-partners.org/training-and-education/pogo-scor-fellowship). 3063  SCOR Visiting Scholars—SCOR operates a program similar to the POGO Visiting 3064 Professorship program, with the major difference being that the terms of SCOR 3065 Visiting Scholars is shorter (2-8 weeks) and SCOR does not require the host and 3066 Scholar to be pre-matched (see http://www.scor- 3067 int.org/SCOR_Visiting_Scholars.pdf). 3068 3069 The broader objectives of the IQOE include improved public appreciation of the ocean, the 3070 life within it and pressures upon it. Opportunities exist to stimulate the development of 3071 “citizen science” projects by streaming spectrograms in real time from observatories and by 3072 providing the possibility for the public to record sounds of marine organisms (e.g., through an 3073 iPhone application) and to download these to a Web site with an accompanying photograph. 3074 Links with Google and social networking will be important to encourage broad uptake of 3075 IQOE activities and information. 3076 3077 As an example of what should be possible on a larger scale, Ecological Acoustic Recorders 3078 (EARs) were developed by the Census of Coral Reefs project of the Census of Marine Life 3079 for monitoring of coral reefs, including the appraisal of coral reef biodiversity, activity of 3080 sound-producing organisms, and human activities in reef areas (Lammers et al., 2008). 3081 Sounds recorded by EARs can be heard at http://www.pifsc.noaa.gov/cred/ear_sounds.php. 3082 There are many other examples of sound from marine organisms available through the 3083 Aquatic Acoustic Archive (http://aquaticacousticarchive.com/) and DOSITS 3084 (http://www.dosits.org/) 3085 3086 Promotion of the use of acoustics within the wider marine science community should be a 3087 further spin-off from the IQOE. Researchers often do not use acoustics where it could 3088 actually benefit their research (and conversely their research can benefit the overarching 3089 questions being addressed by this project). These missed opportunities are largely because 3090 of the nature of the specialist technologies, including their complexity, cost, and accessibility. 3091 However, the means now exist to make acoustic recording and data logging devices that can 3092 be readily used by other researchers without recourse to specialized technical knowledge. 3093 3094 7.7 Relationship with other organizations, programs and activities 3095 3096 The IQOE is being established in the context of organizations, other projects, and observing 3097 system activities concerned with sound in the ocean and its effects on marine organisms. 3098 The IQOE will make connections with these entities through inviting individuals involved in 3099 them to serve on the SSC or other IQOE groups, through workshops of the project, and 3100 through regular communication. Some particularly important relationships will include: 3101 3102  Observing Systems—This report catalogs ocean observing systems that the 3103 community believes could be important for implementation of IQOE (see Appendix

78

239 DecemberJanuary 20121 DRAFT

3104 1). These include the Global Ocean Observing System (GOOS)—which is a global 3105 consortium of specific observing systems—as well as specific systems that may or 3106 may not consider themselves to be part of GOOS, but which either currently include 3107 hydrophones or to which hydrophones could be added. 3108  Industry and Regulators—The major sources of human-generated sound are from 3109 industries such as shipping, oil and gas exploration and production, energy 3110 production, etc. It will be important through the life of the project to keep contact with 3111 the relevant ocean industries to ensure that industry needs for information are being 3112 met and to gain access to sound data from industry that could be useful for IQOE 3113 implementation, The project will work with the World Ocean Council, as well as 3114 individual industry groups, as appropriate. Regulators of industrial activity and 3115 environmental quality are also major potential users of IQOE results and the project 3116 will ensure communication with such organizations. 3117  Scientific Community—Members of the SSC will provide the primary linkage to the 3118 relevant portions of the scientific community, including acousticians, marine 3119 biologists, physical oceanographers, and others. The IQOE will endeavor to present 3120 information about the IQOE and its progress at international ocean science and 3121 acoustic meetings. 3122  Related Projects—The IQOE is being developed in the context of other research 3123 projects that are seeking to understand the global ocean and how it is changing over 3124 time. This is the first international project on sound in the ocean and there are no 3125 closely related research projects. However, other SCOR-sponsored projects are 3126 focused on ocean biology and projects sponsored by other organizations are 3127 interested in aspects of physical oceanography and climate. These projects will be 3128 kept informed about the IQOE. 3129

3130 7.8 Work streams and workshops 3131 IQOE will build its activities around work streams. These will be designed to capture as 3132 many of the scientific requirements defined in this Science Plan as possible. Establishing 3133 where synergies and dependencies lie within the program structure will be an early task for 3134 the Steering Committee but many of these will only emerge as the IQOE develops. There 3135 are, for example, strong dependencies between progress on defining soundscapes and 3136 progress on observing and modeling ocean sound. 3137 3138 A preliminary set of work streams defined by the Open Science Meeting included: 3139 3140 • Mapping the global soundscape 3141 • Conservation of soundscapes 3142 • Mini-quiet ocean experiment (e.g. at AUTEC) 3143 • Quantifying the increase in anthropogenic ocean sound 3144 • Hydrophones on deep-sea communication cables 3145 • Observing ocean sound field from under Antarctic ice shelves 3146 • Changing soundscapes in the Arctic 3147 • Citizen science: hydrophone on an iPhone (iHydrophone or iDrophone) 3148 3149 In practice, most work streams will be developed through the mechanism of focused, 3150 organized workshops, some of which will help the IQOE make early progress on some high- 3151 profile topics. The IQOE will, therefore, plan early workshops that will help lay a foundation 3152 for the project, providing information needed for future research, observations, and 3153 modeling, but which will also provide early visibility for the project through high-impact 3154 papers. A list of potential meetings and their sequencing is given in Table 7.1 3155

79

239 DecemberJanuary 20121 DRAFT

3156 Four examples of emerging proposals follow, to illustrate the development of the 3157 implementation process. The IQOE will develop a process for soliciting and reviewing ideas 3158 like these, developing those with sufficient appeal to the community of IQOE scientists. 3159

3160 7.9 Funding 3161 Funding for the IQOE Open Science Meeting was provided by the Alfred P. Sloan 3162 Foundation. Two kinds of financial support will be necessary to implement the IQOE. First, 3163 support for planning and coordination will provide a foundation for science activities. Support 3164 for planning and coordination will be sought from the Sloan Foundation, but also from 3165 national science agencies and other national sources, industry, and from other foundations 3166 and non-governmental organizations. This support will be used for workshops to create more 3167 detailed implementation plans related to IQOE activities described in this Science Plan; to 3168 support meetings to establish a data management system, observational standards and 3169 protocols, and other infrastructure needed to ensure comparability and access to data 3170 gathered worldwide; and to maintain a small centralized staff function. The second kind of 3171 financial support will be national funding to conduct the research, observations, and 3172 modeling described in this Science Plan. This kind of support will be sought and obtained 3173 from traditional sources and will leverage the much smaller planning and coordination 3174 support. For the Census of Marine Life, the funding provided by nations through traditional 3175 channels for science activities was approximately 10 times the amount of funding for 3176 planning and coordination provided by the Sloan Foundation. To implement the IQOE 3177 successfully, it will be necessary for individuals to use this Science Plan and subsequent 3178 documents as a basis for proposals to their usual national and multinational (in some 3179 regions) funding agencies. Support from the Sloan Foundation would be sought particularly 3180 to prepare for the intensive International Year of Ocean Acoustics.

3181 . 3182

80

239 DecemberJanuary 20121 DRAFT

3183 Table 7.1 List and sequence of potential meetings and workshops to implement the IQOE 3184 Science Plan 3185 1. Meeting with industry representatives regarding their participation in the IQOE. This meeting would be planned in conjunction with the World Ocean Council and could result in an industry advisory group to the IQOE. The goal of this meeting would be to produce one or more memoranda of agreement between IQOE and industry groups. 2. Workshop on access to proprietary and classified information, past and future. Success of the IQOE will depend on access to past data collected by navies, commercial oil and gas exploration companies, and the Comprehensive Test Ban Treaty Organization, as well as the data collected as part of the IQOE will need to take into account sensitivities of stakeholders about collection and public access to acoustic data. The goal of this workshop would be to develop a written agreements between IQOE and potential providers of information about what information they would make available, under what circumstances, and would describe a mechanism for future discussions. 3. Workshop on design of IQOE data collection, management and access. Issues related to standardization of data collection and storage will need to be addresses early in the project. The goal of this workshop would be to agree to standards and procedures that would guide IQOE observations and research to make observations and research conducted in different locations to be comparable. 4. Workshop of status of modeling relevant to IQOE and needs for the project. This workshop would include discussion of information needed for Population Consequences of Acoustic Disturbance (PCAD)0 models. The goal of this workshop would be to provide a roadmap for development of the modeling component of the IQOE. A report of Tthe meeting document will be published in a suitable journal. 5. Workshop on the Global Ocean Acoustical Observing System. A proposal for such a system was made at OceanObs’09 (Dushaw et al. 2009). The workshop would revisit that plan and would seek commitments from observation systems to install and support hydrophones, as well as to identify areas of the global ocean where new hydrophone systems should be deployed. The workshop would produce a white paper thatn might be published in the peer-reviewed literature. 6. Workshop on opportunistic observations. It will be necessary early in the project to establish mechanisms to identify opportunities related to changes in shipping lanes, planned large-scale pile driving activities, and other opportunities to study before and after noise levels and effects on marine organisms. This workshop would conduct detailed planning for opportunities that have already been identified, as well as look forward to detect approaches to identify future approaches. The goal of this workshop would be to produce detailed plans for IQOE observations and experiments related to known opportunities, as well as a document that would specify how future opportunities would be approached. 7. Workshop to develop a world map of anthropogenic ocean sound (see Example 1 below). This workshop would work as far back in time as possible to determine whether it is possible to develop time series for ocean sound similar to the time series for atmospheric CO2 concentrations known as Keeling Ccurves. This workshop would produce a paper for a high-profile journal. 8. Workshop to plan an Arctic Ocean acoustic survey (see Example 2 below). This workshop would need to involve scientists involved in acoustics of the Arctic Ocean, as well as organizations involved in Arctic Ocean science and observations. This workshop would produce a plan for an Arctic Ocean experiment that would be part of IQOE.

81

239 DecemberJanuary 20121 DRAFT

3186

3187 Example 1 World Map of Anthropogenic Ocean Sound

Objectives:  Produce a map of the Earth’s oceans showing total acoustic energy (over a meaningful duration, e.g. 1 year) from human sources  Create layers for the different sources: shipping, seismics, pile driving  The map will highlight regions characterized by high levels of sound (hot spots) and other regions that are still relatively unimpacted by human-generated sound  The map will be made publicly available (in print and in GIS data)

Background and context: Two studies indicate that ambient sound in the ocean has increased significantly over the past 50 years (Andrew et al., 2011; Chapman and Price, 2011) at frequencies below 100 Hz. Sound at such low frequencies travels across ocean basins with little loss of energy. Both studies attributed the increases to distant shipping. A world noise map will begin to expand our understanding of the spatial and temporal variation in sound within important ecological regions that are changing rapidly, such as the Arctic Ocean.

Approach: 1. Identify key people who will collaboratively be able to produce this map. 2. Identify the types of anthropogenic sources that can and should be included in this map. 3. Identify data needs and sources. 4. Devise modeling approach. 5. Arrange peer review of both the approach and the input data (e.g., through a workshop). 6. Write (in software) and run the model. 7. Arrange peer review of the results. 8. Write and publish report and map in a high‐impact journal (Nature or Science).

A model will be used to impose a grid on the Earth’s oceans. The total number of hours that each source operated within each cell needs to be extracted from the underlying databases. Source levels need to be assigned to each source. Transmission loss models will be applied (varying by acoustic zone) to populate cells with received energy. Maps of cumulative energy will be produced for each source type (shipping, seismics, pile driving) and as a cumulative total. Validation of the map with field measurements will use long‐term recordings and error analysis that will consider the effects on the modeled received level as a function of source depth and receiver depth, uncertainty in source level, variability in sound speed profile, uncertainty in seafloor geoacoustics, and other relevant factors.

The layers of the map will look similar to the shipping density map by Halpern et al. (2008), illustrated in Figure 7.4. The map will be in units of energy. The transform from a shipping density map to a cumulative energy map, however, will not be linear, that is, the energy map will not exhibit the same range and distribution of ‘colors’.

Figure 7.4 Shipping density map (Halpern et al. 2008) 82

239 DecemberJanuary 20121 DRAFT

3188

3189 Example 2 3190 Acoustic ecology of the Arctic Ocean: A survey along the ice edge 3191 Objectives: 3192  Make recordings of ambient sound conditions differentiating, when possible, anthropogenic, 3193 physical, and biological sound sources.  Identify sounds from specific target study organisms, which may play an important role in ocean 3194 food webs. The goal here is to survey across the entire ecosystem (zooplankton-nekton), but 3195 also to target highly valued species like whales. 3196  Examine how the changing level of activities will change sound levels and where there is likely 3197 to be an intersection between the biology and sounds. 3198 Background and context:

3199 The Arctic Ocean is likely to be one of the most changed biomes as a result of climate warming. Loss of 3200 sea ice during summer and reduced ice cover in winter present opportunities for economic development 3201 that have hitherto been technologically impossible and/or not economically viable. In particular, the expansion of shipping traffic and the extension of oil and gas explorations and production into the Arctic Ocean are likely to be the main anthropogenic stressors. This is exemplified by the Shtokman gas field that exploits gas condensates and is estimated to be one of the world's largest natural gas deposits. These activities cause significant sound pollution (OSPAR 2009). Considerable uncertainty exists around how these activities should be regulated in order to reduce their environmental effects.

Approach: 1. Conduct observations starting after the period of maximum ice extent in March, but before the minimum in Sept. (see Figure 7.5) using a vessel transit of the NE passage back and forth close to the retreating ice edge; 2. Use dipping hydrophone arrays, towed hydrophones, and bottom-mounted and drifting buoys; 3. Data from buoys and dipping stations will be used to quantify ambient sound levels; 4. Conduct targeted acoustic recordings of specific target individual species; 5. Assemble a team of acoustic and biological experts as a project steering group; and 6. The most practical platform to conduct the study is likely to be a yacht.

Figure 7.5 Map of the Arctic Ocean ice extent together with the median ice edge showing the maximum (March) and minimum extent (September) during 2010.

83

239 DecemberJanuary 20121 DRAFT

3202 Example 3 3203 Formatted: Centered, Space After: 0 pt, Line Average ocean noise as environmental status indicator spacing: single

Objectives:  Monitor globally averaged low-frequency sound (“ocean noise”)  Identify key parameters to which the ocean noise is sensitive  Detect changes in the key parameters from measured ocean noise

Background and context: Several studies (e.g., Andrew et al, 2011) indicate that the level of ambient sound in the ocean has increased significantly since the 1960s at frequencies below 300 Hz. The level of ocean noise is sensitive to the number and strength of the sources (mainly ships; also whales and air guns) and to propagation conditions. Monitoring the globally averaged low-frequency sound (“ocean noise”) provides a useful indicator in its own right because of its possible impact on communication ranges of baleen whales (Parks et al., 2007; Stafford et al., 2007; Clark et al., 2009). It also offers the prospect of monitoring climate change (through the propagation conditions, e.g., average temperature or wind speed) and changes in sources of sound (e.g., total sound power radiated by all ocean-going ships). A two-step approach is proposed, starting with a feasibility study at the Atlantic Undersea Test and Evaluation Center (AUTEC).

Approach: 9. Identify team able to collaboratively monitor and interpret ocean noise 10. Identify and gather key input data for feasibility study (AUTEC) 11. Model sound sources and propagation at AUTEC 12. Measure sound field at AUTEC and relate to key model input parameters 13. Peer review (e.g., through a workshop) before proceeding to ocean scale 14. Identify suitable ocean (suitable basin; suitable monitoring network) 15. Identify and gather key input data for selected ocean basin 16. Model sound sources and propagation in ocean basin (see also “World Map of Anthropogenic Ocean Sound”) 17. Identify and fill critical gaps in hydrophone network 18. Measure sound field in ocean basin and relate to key input parameters 19. Assess feasibility for detecting changes in key parameters (e.g., global thermometer) 20. Write and publish results in a high-impact journal (Nature or Science).

AUTEC feasibility study: The purpose of the feasibility study is to mitigate the risk of an ocean-scale experiment by identifying and testing the methods in advance. AUTEC’s Tongue of the Ocean (TOTO) provides a natural quiet and deep basin, an existing monitoring network, and known shipping activity. TOTO is surrounded by shallow water that for a model-scale experiment would represent the continental shelves surrounding a deep ocean. An important aspect of the feasibility study is understanding how the geometry and frequency scale to a larger (ocean-scale) experiment.

Ocean basin study: After a progress review, monitor sound on an ocean scale. The choice of a suitable ocean depends on having a hydrophone network with wide coverage, for which best use will be made of existing CTBTO, LIDO and MPEL stations, as well as newly installed stations (EU Member States are required to monitor trends in low-frequency underwater noise from 2014). New hydrophone stations might be needed. In order to provide an indication of any trend, the ocean noise will be monitored for a period of several years. The first result is the ocean noise itself, a possible proxy for global economic activity (Frisk 2007). Because propagation conditions also depend on parameters related to climate change (e.g., sea surface temperature), the possibility also exists of monitoring climate change through changes in the ocean noise Figure 7.6).

84 Figure 7.6 Does ocean noise reflect trends in sea surface temperature?

239 DecemberJanuary 20121 DRAFT

3204 3205 Example 4 Marsdiep powerboat race: Effect on soundscape and marine life

Objectives: a. Produce a three-dimensional map of the sound field in the Marsdiep (see Figure 7.7) before, during and after a powerboat race b. Assess the impact of the powerboat race on marine life in the Marsdiep c. Identify key inputs needed to model ambient noise on a larger scale

Background and context: The Marsdiep is a natural basin between den Helder on the Dutch mainland and the island of Texel. It is also a foraging area for harbor porpoise and harbor seal populations. A unique opportunity to study the effect of changing levels of anthropogenic sound on marine life in the Marsdiep will be available during a powerboat race (“the Event”) scheduled for August 2012, as well as before and after the Event. This race may become an annual event.

Approach: 1. Measure sound field before, during, and after the Event 2. Measure sound sources both during Event and separately under controlled conditions 3. Identify suitable sound propagation model and identify key model input parameters 4. Model the sound field before, during, and after the Event 5. Measure distribution of marine life before, during, and after the Event 6. Identify possible cause-effect relationships between animal behavior and the sound field 7. Write and publish report in a peer-reviewed journal.

The noise background to the Event comprises contributions from both natural (wind and biological) and anthropogenic (boat traffic) sources. The total noise during the Event will be the sum of this background (possibly influenced by the Event), the contributions from the powerboats and any supporting activities. The spatial distribution of the sound will be characterized, as well as the source levels of individual boats. Measurements are envisaged also on a single powerboat under controlled conditions with varying speed.

The effect of the Event on harbor porpoises and harbor seals will be studied. We hypothesize that the sound emissions from the powerboats will have an effect on the animals’ behavior and habitat use. The presence and behavior of harbor porpoises will be monitored by means of aerial and land-based visual observations and acoustic click detectors; harbor seals will be tagged with telemetry units and satellite transmitters. Understanding the behavior of the animals will help in managing the impact of similar anthropogenic events in the future.

It is hypothesized that the Marsdiep’s unique shape, with relatively deep water (ca. 50 m) surrounded by shallow water (ca. 10 m) can be used as a scale model of a larger basin, and that these observations will therefore provide valuable insight to future research on larger scale experiments. An important aspect of the study is understanding how the geometry and frequency scale to a larger basin (see “Average ocean noise as environmental status indicator”).

Figure 7.7 Bathymetry of the Marsdiep (Dastgheib et al. 2008) 85

239 DecemberJanuary 20121 DRAFT

3206 Example 5

Sound Propagation Over Rocky Seafloors

Objectives:

Develop a sound propagation model for underwater sources and receivers in regions of “rocky”, i.e. acoustically elastic, seabeds with variable (range-dependent) characteristics.

Background and context:

Underwater sound propagation models can be classified by dimension: 1D models consider the ocean as vertically stratified; acoustic properties vary only with depth (range-independent models, RI). In 2D models properties vary as a function of depth and range (range-dependent models, RD). Azimuth is the dimension added to 3D models. Modelling over long ranges requires range-dependent models, because the bathymetry as well as the hydro- and geo-acoustic parameters are usually not constant over long distances in real ocean environments. The most commonly used range-dependent models are ray and Gaussian beam models for high frequencies and parabolic equation models for low frequencies (Table 1). Both approaches are not adequate for range-dependent acousto-elastic environments, where a substantial amount of water-borne acoustic energy travels through the seafloor, and losses occur due to conversion of compressional waves to shear waves. Although numerical propagation models have been developed for range-independent acousto-elastic environments, these models are not appropriate for range-dependent environments.

Table 1: Underwater sound propagation models and their applicability (Etter, 2003).

Model Type Shallow Water Deep Water

< 500 Hz > 500 Hz < 500 Hz > 500 Hz

RI RD RI RD RI RD RI RD

Ray Theory no no limited yes limited limited yes yes

Normal Mode yes limited yes limited yes limited limited no

Multipath Expansion no no limited no limited no yes no

Fast Field yes no yes no yes no limited no

Parabolic Equation limited yes no no limited yes limited limited

Acousto-elastic environments include thin unconsolidated sediment layers and rocky seafloors (blue colours in Fig. 1). On the continental shelves, these environments are common in tropical coral-reef regions, and where the seafloor geology is dominated by limestones and sandstones.

86

239 DecemberJanuary 20121 DRAFT

Fig. 1: Total sediment thickness of the world’s oceans (http://www.ngdc.no aa.gov/mgg/image/s edthick9.jpg).

Currently, inappropriate sound propagation models are used for these areas, often leading to several 10s of dB of errors in modelled received levels. Range-dependent acousto-elastic sound propagation models are a necessity for noise prediction, bioacoustic impact assessments of anthropogenic marine operations, and noise mapping (e.g., for marine spatial planning).

Approach:

A workshop should be convened bringing together experts in sound propagation model development. Prior to the workshop, participants will be given ground truth data (including geo- and hydro-acoustic measurements) from a well-characterised acousto-elastic environment. Participants will be asked to model this environment prior to the workshop, and to present their approach and results at the workshop. Discussions will focus on benchmarking, model development & validation. The workshop will result in a written research & development plan that can be presented to offshore industries, governments and navies for support.

Etter, P. C. (2003). Underwater acoustic modeling and simulation (3rd ed.). New York: Spon Press.

3207

87

239 DecemberJanuary 20121 DRAFT

3208 REFERENCES 3209 3210 Ainslie, M.A. 2011. Standard for measurement and monitoring of underwater noise, Part I: 3211 physical quantities and their units. TNO report, TNO-DV 2011 C235, 67pp. 3212 Allen, S., and D.A. Demer. 2003. Detection and characterization of yellowfin and bluefin tuna 3213 using passive-acoustical techniques. Fisheries Research 63(3):393-403. 3214 André, M., M. van der Schaar, S. Zaugg, L. Houégnigan, A.M. Sánchez, and J.V. Castell. 3215 2011. Listening to the deep: live monitoring of ocean noise and cetacean acoustic 3216 signals. Marine Pollution Bulletin 63:18-26. 3217 Andrew, R.K., B.M. Howe, and J.A. Mercer. 2011. Long-time trends in ship traffic noise for 3218 four sites off the North American West Coast. Journal of the Acoustical Society of 3219 America 129(2):642-651. 3220 Andrew, R.K., B.M. Howe, J.A. Mercer, and M.A. Dzieciuch. 2002. Ocean ambient sound: 3221 Comparing the 1960s with the 1990s for a receiver off the California coast. Acoustic 3222 Research Letters Online 3:65-70. 3223 Ausubel, J. 2009. Rethinking sound and light. Seed: November 23, 2009. 3224 http://seedmagazine.com/content/article/rethinking_light_and_sound/. 3225 Bakun, A., and S.J. Weeks. 2006. Adverse feedback sequences in exploited marine 3226 systems: Are deliberate interruptive actions warranted? Fish and Fisheries 7:316- 3227 333. 3228 Boyd, I.L., R. Brownell, D. Cato, C. Clarke, D. Costa, P. Evans, J. Gedamke, R. Gentry, R. 3229 Gisiner, J. Gordon, P. Jepson, P. Miller, L. Rendell, M. Tasker, P. Tyack, E. Vos, H. 3230 Whitehead, D. Wartzok, and W. Zimmer. 2008. The effects of anthropogenic sound 3231 on marine mammals -- A draft research strategy. European Science Foundation 3232 Marine Board Position Paper 13, June 2008, www.esf.org/marineboard. 3233 Boyd, I.L., D.E. Claridge, C.W. Clark, B.L. Southall, and P.L. Tyack. 2007. Behavioral 3234 Response Study Cruise Report (BRS-2007), http://www.sea- 3235 inc.net/resources/brs_08_finalcruisereport.pdf. 3236 Boyd, I., G. Frisk , E. Urban, P. Tyack, J. Ausubel, S. Seeyave, D. Cato, B. Southall, M. 3237 Weise, R. Andrew, T. Akamatsu, R. Dekeling, C. Erbe, D. Farmer, R. Gentry, T. 3238 Gross , A. Hawkins , F. Li , K. Metcalf, J.H. Miller, D. Moretti, C. Rodrigo, and T. 3239 Shinke. 2011. An International Quiet Ocean Experiment. Oceanography 24(2):174– 3240 181, doi:10.5670/oceanog.2011.37. 3241 Buckingham, M.J., B.V. Berkhout, and S.A.L. Glegg. 1992. Imaging the ocean with ambient 3242 sound. Nature 356:327-329. 3243 Carey, W.M. and Evans, R.B. 2011. Ocean Ambient Noise. The Underwater Acoustics 3244 Series, Springer 263 pp. 3245 Cato, D.H. 1997. Ambient sea sound in Australian waters. Pp. 2813-2818 in. Proc. Fifth 3246 International Congress on Sound and Vibration, Adelaide, Casual Productions, 3247 Adelaide, Vol. 5, 3248 Cato, D.H. 2008. Ocean ambient sound: its measurement and its significance to marine 3249 animals. Proc. Institute of Acoustics, 30 (5). 3250 Clark, C.W., W.T. Ellison, B.L. Southall, L. Hatch, S.M. Van Parijs, A. Frankel, and D. 3251 Ponirakis. 2009. Acoustic masking in marine ecosystems: intuitions, analysis, and 3252 implication. Marine Ecology Progress Series 395:201-222. 3253 Chapman, N.R., and A. Price. 2011. Low frequency deep ocean ambient noise trend in the 3254 Northeast Pacific Ocean. Journal of the Acoustical Society of America 129(5):EL161- 3255 EL165. 3256 Cox, T.M., T.J. Ragen, A.J. Read, E. Vos, R.W. Baird, K. Balcomb, J. Barlow, J. Caldwell, T. 3257 Cranford, L. Crum, A. D'Amico, G. D'Spain, A. Fernández, J. Finneran, R. Gentry, W. 3258 Gerth, F. Gulland, J. Hildebrand, D. Houser, T. Hullar, P.D. Jepson, D. Ketten, C.D. 3259 Macleod, P. Miller, S. Moore, D.C. Mountain, D. Palka, P. Ponganis, S. Rommel, T. 3260 Rowles, B. Taylor, P. Tyack, D. Wartzok, R. Gisiner, J. Mead, and L. Benner. 2006. 3261 Understanding the impacts of anthropogenic sound on beaked whales. Journal of 3262 Cetacean Research and Management 7:177-187.

88

239 DecemberJanuary 20121 DRAFT

3263 Dastgheib, A., J.A. Roelvink, and Z.B. Wang. 2008. Long-term process-based morphological 3264 modeling of the Marsdiep Tidal Basin. Marine Geology 256:90-100. 3265 De Jong C.A.F., and M.A. Ainslie. 2008. Underwater radiated noise due to the piling for the 3266 offshore wind park. Acoustic 08, June 29-July 4 2008, Paris. 3267 Dushaw, B. & Co-Authors (2010). "A Global Ocean Acoustic Observing Network" in 3268 Proceedings of OceanObs’09: Sustained Ocean Observations and Information for 3269 Society (Vol. 2), Venice, Italy, 21-25 September 2009, Hall, J., Harrison, D.E. & 3270 Stammer, D., Eds., ESA Publication WPP-306, doi:10.5270/OceanObs09.cwp.25 3271 Erbe, C., C. McPherson, and A. Craven. 2011. Acoustic Investigation of Bycatch Mitigation 3272 Pingers. JASCO Applied Sciences report to the Australian Marine Mammal Centre, 3273 Kingston, Australia. 89 pp. 3274 Evans, G.W., M. Bullinger, and S. Hygge. 1998. Chronic sound exposure and physiological 3275 response: a prospective study of children living under environmental stress. 3276 Psychological Science 9(1):75-77. 3277 Grist, J.P., S.A. Josey, L. Boehme, M.P. Meredith, F.J.M. Davidson, G.B. Stenson, and 3278 M.O. Hammill. 2011. Temperature signature of high latitude Atlantic boundary 3279 currents revealed by marine mammal-borne sensor and Argo data. Geophysical 3280 Research Letters 38:L15601 DOI: 10.1029/2011GL048204. 3281 Fay, R.R., and A.N. Popper. 2000. Evolution of hearing in vertebrates: the inner ears and 3282 processing. Hearing Research 149:1-10. 3283 Foote, A.D., R.W. Asborne, and A.R. Hoelzel. 2004. Whale-call response to masking boat 3284 sound. Nature 428:910. 3285 Frisk, G.V. 1994. Ocean and Seabed Acoustics: A Theory of Wave Propagation. Prentice- 3286 Hall 299 pp. 3287 Frisk, G.V. 2007. Noiseonomics:: The Relationship between Ambient Noise Levels and 3288 Global Economic Trends – Invited paper presented at Pacific Rim Underwater 3289 Acoustics Conference 2007, Vancouver, BC, Canada (3-5 October 2007). 3290 Gillespie, A. 2007. The precautionary principle in the 21st century: a case study of sound 3291 pollution in the oceans. The International Journal of Marine and Coastal Law 22:61- 3292 87. 3293 Grist et al. 3294 Gunn, J., A. Rogers. and E. Urban. 2010. Observation of Ocean Biology on a Global Scale: 3295 Implementing Bio-GOOS in Proceedings of OceanObs’09: Sustained Ocean 3296 Observations and Information for Society (Vol. 1), Venice, Italy, 21-25 September 3297 2009, Hall, J., Harrison, D.E. & Stammer, D., Eds., ESA Publication WPP-306, 3298 doi:10.5270/OceanObs09.pp.20 3299 Halpern, B.S., S. Walbridge, K.A. Selkoe, C.V. Kappel, F. Micheli, C. D’Agrosa J.F. Bruno, 3300 K.S. Casey, C. Ebert, H.E. Fox, R. Fujita, D. Heinemann, H.S. Lenihan, E.M.P. 3301 Madin, M.T. Perry, E.R. Selig, M. Spalding, R. Steneck, and R. Watson. 2008. A 3302 Global Map of Human Impact on Marine Ecosystems. Science 319:948-952. 3303 Hatch, L.T., and K.M. Fristrup. 2009. No barrier at the boundaries: implementing regional 3304 frameworks for noise management in protected natural areas. Marine Ecology 3305 Progress Series 395:223-244. 3306 Hildebrand, J.A. 2009. Anthropogenic and natural sources of ambient sound in the ocean. 3307 Marine Ecology Progress Series 395:5-20. 3308 Holt, M.M., D.P. Noren, V. Veirs, C.K. Emmonds, and S. Veirs. 2008. Speaking up: Killer 3309 whales (Orcinus orca) increase their call amplitude in response to vessel sound. 3310 Journal of the Acoustical Society of America Express Letters 125:EL27-EL32. 3311 Jakobsson, M., R. Macnab, L. Mayer, R. Anderson, M. Edwards, J. Hatzky, H.-W. Schenke, 3312 and P. Johnson. 2008, An improved bathymetric portrayal of the Arctic Ocean: 3313 Implications for ocean modeling and geological, geophysical and oceanographic 3314 analyses, Geophysical Research Letters,35:L07602doi:10.1029/2008GL033520. 3315 Johnson, K.S., W.M. Berelson, E.S. Boss, Z. Chase, H. Claustre, S.R. Emerson,, N. Gruber, 3316 A. Kortzinger, M.J. Perry, and S.C. Riser. 2009. Observing biogeochemical cycles at

89

239 DecemberJanuary 20121 DRAFT

3317 global scales with profiling floats and gliders: prospects for a global array. 3318 Oceanography 22:216-225. 3319 Keeling, C.D.1960 The Concentration and Isotopic Abundances of Carbon Dioxide in the 3320 Atmosphere, Tellus 12: 200-203. 3321 Ketten, D.R. 2010. Underwater ears and potential impacts: how what whales can hear. Israel 3322 Journal of Ecology and Evolution 56:87. 3323 Kite-Powell, H.L. 2009. Economic Considerations in the Design of Ocean Observing 3324 Systems. Oceanography 22:44-49. 3325 Kozaczka, E., J. Domagalski, and I. Gloza. 2010. Investigation of the underwater noise 3326 produced by ships by means of intensity method. Polish Maritime Research 17:26- 3327 36. 3328 Lammers, M.O., R.E. Brainard, W.W.L. Au, T.A. Mooney, and K.B. Wong. 2008. An 3329 ecological acoustic recorder (EAR) for long-term monitoring of biological and 3330 anthropogenic sounds on coral reefs and other habitats. Journal of the Acoustic 3331 Society of America 123(3). DOI: 10.1121/1.2836780. 3332 Lengagne, T. 2008. Traffic sound affects communication behaviour in a breeding anuran, 3333 Hyla arborea. Biological Conservation 141:2023-2031 DOI: 3334 10.1016/j.biocon.2008.05.017. 3335 Lombard, E. .1911. Le signe de le elevation de la voix. Ann. Malad.I'Orielle. Larynx. Nez. 3336 Pharynx 37:101-119. 3337 Marine Strategy Framework Directive; Task Group 11 Report: Underwater sound and other 3338 forms of energy, April 2010, M.L. Tasker, M. Amundin, M. Andre, A. Hawkins, W. 3339 Lang, T. Merck, A. Scholik-Schlomer, J. Teilmann, F. Thomsen, S. Werner & M. 3340 Zakharia 3341 McDonald, M.A., J.A. Hildebrand, S.M. Wiggins, D. and Ross. 2008. A 50 Year comparison 3342 of ambient ocean sound near San Clemente Island: A bathymetrically complex 3343 coastal region off Southern California Journal of the Acoustical Society of America 3344 124:1985-1992. 3345 Matsumoto, H., J.H. Haxel, R.P. Dziak, D.R. Bohnenstiehl, and R.W. Embley, 2011. 3346 Mapping the sound field of an erupting submarine volcano using an acoustic glider. 3347 Journal of the Acoustical Society of America 129: EL94-EL99. 3348 Miller, P.J.O., N. Biassoni, A. Samuels, and P.L. Tyack. 2000. Whale songs lengthen in 3349 response to sonar. Nature 405:903. 3350 Munk, W.H., R.C. Spindel, A. Baggeroer, and T.G. Birdsall. 1994. The Heard Island 3351 Feasibility test. Journal of the Acoustical Society of America 94:2330-2342. 3352 Munk, W., and C. Wunsch. 1983. Ocean acoustic tomography – rays and modes. Reviews 3353 of Geophysics 21:777-793. 3354 National Research Council. 2003. Ocean Sound and Marine Mammals. National Academy 3355 Press, Washington, D.C. 3356 National Research Council. 2005. Marine Mammal Populations and Ocean Sound: 3357 Determining When Sound Causes Biologically Significant Effects. National Academy 3358 Press, Washington, D.C. 3359 Nicholls, K.W., E.P. Abrahamsen, K.J. Heywood, K. Stansfield, and S. Osterhus. 2008. High- 3360 latitude oceanography using the Autosub autonomous underwater vehicle. Limnology 3361 and Oceanography 53:2309-2320. 3362 NOAA (US National Oceanic and Atmospheric Administration) 2005. Regulations governing 3363 the taking and importing of marine mammals. Subchapter C: Marine Mammal 3364 Protection Act regulations (50 CFR 216). 3365 OSPAR, 2009. Overview of the impacts of anthropogenic underwater sound in the marine 3366 environment. OSPAR Commission, 2009. Publication number 441/2009.

90

239 DecemberJanuary 20121 DRAFT

3367 Parks, S.E., M. Johnson, D. Nowacek, and P.L. Tyack. 2010. Individual right whales call 3368 louder in increased environmental sound. Biology Letters 7:33-35, 3369 doi:10.1098/rsbl.2010.0451. 3370 Pijanowski, B.C., L.J. Villanueva-Rivera, S.L. Dumyahn, A. Farina, B.L. Krause, B.M. 3371 Napoletano, S.H. Gage, and N. Pieretti. 2011. Soundscape Ecology: The Science of 3372 Sound in the Landscape. BioScience 61:203-216. 3373 Popper, A.N., and R.R. Fay. 2011. Rethinking sound detection by fishes. Hearing Research 3374 273:25-36. 3375 Popper A.N, and M.C. Hastings. 2009. The effect of anthropogenic sound on fishes. Journal 3376 of Fish Biology 75:455-489. 3377 Radford, C.A., J.A. Stanley, S.D. Simpson, and A.G. Jeffs. 2011. Juvenile coral reef fish use 3378 sound to locate habitats. Coral Reefs 30:295-305. 3379 Reeder, D.B., E. Sheffield, and S.M. Mack. 2011. Wind-generated ambient noise in a 3380 topographically isolated basin: A pre-industrial era proxy. Journal of the Acoustical 3381 Society of America 129:64-73. 3382 Roemmich, D., G.C. Johnson, S. Riser, R. Davis, J. Gilson, W.B. Owens, S.L. Garzoli, C. 3383 Schmid, and M. Ignaszewski. 2009. The Argo Program Observing the Global Ocean 3384 with Profiling Floats. Oceanography 22:34-43. 3385 Ross, D.G. 2005. Ship sources of ambient noise, Journal of Oceanic Engineering 30, 257– 3386 261. 3387 Round Table of International Shipping Associations 2011. Shipping and world trade. 3388 http://www.marisec.org/shippingfacts/worldtrade/volume-world-trade-sea.php 3389 (accessed 12 November 2011) 3390 Ruttenberg, B.I., S.L. Hamilton, S.M. Walsh, M.K. Donovan, A. Fiedlander, E. DeMartini, E. 3391 Sala, and S.A. Sandin. 2011. Predator-Induced Demographic Shifts in Coral Reef 3392 Fish Assemblages PLOS ONE 6: e21062 DOI: 10.1371/journal.pone.0021062. 3393 Slabbekoorn, H., N. Bouton, I. van Opzeeland, A. Coers, C. ten Cate, and A.N. Popper. 3394 2010. A noisy spring: the impact of globally rising underwater sound levels on fish. 3395 Trends in Ecology & Evolution 25:419e427, doi:10.1016/ j.tree.2010.04.005. 3396 Smith, W.H.F., and D.T. Sandwell, Global seafloor topography from satellite altimetry and 3397 ship depth soundings. Science 277:1957-1962. 3398 Southall, B.L. 2005. Shipping noise and marine mammals: a forum for science, 3399 management, and technology. Final Report of the National Oceanic and Atmospheric 3400 Administration (NOAA) International Symposium. US NOAA Fisheries, Arlington, VA.. 3401 Southall, B.L., and A. Scholik-Schlomer. 2009. Potential application of vessel-quieting 3402 technology on large commercial vessels. Final Report of the National Oceanic and 3403 Atmospheric Administration (NOAA) International Conference. US NOAA Fisheries, 3404 Silver Spring, MD 3405 Southall, B.L., A.E. Bowles, W.T. Ellison, J.J. Finneran, R.L. Gentry, et al. 2007 Marine 3406 mammal sound exposure criteria: initial scientific recommendations. Aquatic 3407 Mammals 33:411-522. 3408 Tasker, M.L., M. Amundin, M. Andre, A. Hawkins, W. Lang, T. Merck, A. Scholik-Schlomer, 3409 J. Teilmann, F. Thomsen, S. Werner, and M. Zakharia. 2010. Marine Strategy 3410 Framework Directive Task Group 11 Report: Underwater noise and other forms of 3411 energy. European Commission and ICES, EUR24341 EN – 2010. 3412 Thurstan, R.H., and C.M. Roberts. 2010. Ecological Meltdown in the Firth of Clyde, Scotland: 3413 Two Centuries of Change in a Coastal Marine Ecosystem. PLoS ONE 5:e11767. 3414 doi:10.1371/journal.pone.0011767. 3415 Tyack, P.L., W.X.M. Zimmer, D. Moretti, B.L. Southall, D.E. Claridge, J. Durban, C.W. Clark, 3416 A. D'Amico, N. DiMarzio, S. Jarvis, E.M. McCarthy, R. Morrissey, J. Ward, and I.L. 3417 Boyd. 2011. Beaked Whales Respond to Simulated and Actual Navy Sonar. PLoS 3418 ONE 6(3):e17009. doi:10.1371/journal.pone.0017009. 3419 Tyack (2008) 3420 Tyack (2010)

91

239 DecemberJanuary 20121 DRAFT

3421 Van Parijs, S.M., C.W. Clark, R.S. Sousa-Lima, S.E. Parks, S. Rankin, D. Risch,, and I.C. 3422 Van Opzeeland. 2009. Mesoscale applications of near real-time and archival passive 3423 acoustic arrays. Marine Ecology Progress Series 395:21–36. 3424 Vitousek, P.M., H.A. Mooney, J. Lubchenco, and J.M. Melillo. 1997. Human domination of 3425 earth’s ecosystems. Science 277:494–499. 3426 Wartzok D., and D.R. Ketten. 1999. Marine mammal sensory systems. Pp. 117–175 in 3427 Reynolds JE II, Rommel SA (eds.) Biology of marine mammals. Smithsonian 3428 Institution Press, Washington, D.C.. 3429 Weilgart, L. 2006. Managing noise through marine protected areas around global hot spots. 3430 International Whaling Commission SC/58/E25. 3431 Wenz 1967 3432 Williams, R., and E. Ashe. 2007. Killer whale evasive tactics vary with boat number. Journal 3433 of Zoology 272:390-397. 3434

92

239 DecemberJanuary 20121 DRAFT

3435 Appendix I 3436 3437 Matrix of Acoustic Capabilities of Existing Observing 3438 Systems 3439

93

239 DecemberJanuary 20121 DRAFT

3440 Appendix II

3441 Contributors to Science Plan

3442 3443 Editors 3444 These individuals transformed the outputs from the IQOE Open Science Meeting into a 3445 Science Plan, based on experience with previous projects: Ian Boyd, George Frisk, David 3446 Farmer, Ed Urban, and Sophie Seeyave. 3447 3448 Discussion Session Participants 3449 Most of the time at the IQOE Open Science Meeting was devoted to discussion sessions. 3450 Each group was led by two co-chairs and two co-rapporteurs. These groups are responsible 3451 for the good ideas and much of the text in this Science Plan. 3452 3453 Theme 1: Doug Cato and Manell Zakharia (chairs) and Christine Erbe and Tony Hawkins 3454 (rapporteurs) 3455 Other contributors: Michael Ainslie, Caroline Carter, Ross Chapman, Thomas Folegot, Lars 3456 Kindermann, David Mann, Jeffrey Nystuen, Michael Porter, Mark Prior, and George 3457 Shillinger. 3458 3459 Theme 2 (this theme was created from the outputs from two different discussion groups): 3460 Christopher Clark, Robert Gisiner, Vincent Janik and Jakob Tougaard (chairs) and Sophie 3461 Brasseur, Peter Evans, Roger Gentry, and Patrick Miller (rapporteurs) 3462 Other contributors: Tomonari Akamatsu, Clara Amorim, Susannah Buchan, Dan Costa 3463 Yong-Min Jiang, Darlene Ketten, Megan McKenna, Andy Radford, Steve Simpson, Peter 3464 Tyack, Michael Weise, and Rob Williams, 3465 3466 Theme 3: Brian Dushaw and Brandon Southall (chairs) and Rex Andrew and Jennifer 3467 Miksis-Olds (rapporteurs). Other contributors: Michel Andre, Olaf Boebel, Del Dakin, Eric 3468 Delory, Richard Dewey, Albert Fischer, Lee Freitag, George Frisk, Tom Gross, Zygmunt 3469 Klusek, Fenghua Li, Ellen Livingston, David Moreti, Sophie Seeyave, Yvan Simard, 3470 Alexander Vedenev, Ed Urban, and Peter Worcester. 3471 3472 Theme 4: Frank Thomsen and John Young (chairs) and René Dekeling and Jason 3473 Gedamke (rapporteurs) 3474 Other contributors: Hussein Alidina, Aurélien Carbonnière, David Farmer, Paul Holthus, Gail 3475 Scowcroft, Michael Stocker, Anne-Isabelle Vichot, and Dietrich Wittekind. 3476 3477 Reviewers

3478

94